BOOK OF ABSTRACTS 9 th European Medical and Biological Engineering Conference Portorož, Slovenia 9−13 June, 2024 Book of Abstracts Organised by: IFMBE - International Federation for Medical and Biological Engineering, EAMBES - European Alliance for Medical and Biological Engineering & Science, DMBTS - Društvo za biomedicinsko tehniko Slovenije, and University of Ljubljana Edited by: Tomaž Jarm, Rok Šmerc, and Samo Mahnič-Kalamiza BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB .DWDORåQL]DSLVRSXEOLNDFLML&,3SULSUDYLOLY1DURGQLLQXQLYHU]LWHWQLNQMLåQLFLY/MXEOMDQL &2%,666,,' ,6%13') BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB 85/KWWSVZZZHPEHFRUJZSFRQWHQWXSORDGV (0%(&B%RRNBRIB$EVWUDFWVSGI &RS\ULJKW�=DORåED)($OOULJKWVUHVHUYHG5D]PQRåHYDQMHWXGLIRWRNRSLUDQMHGHODY FHORWLDOLSRGHOLKEUH]SUHGKRGQHJDGRYROMHQMD=DORåEH)(SUHSRYHGDQR =DORåQLN=DORåED)(/MXEOMDQD ,]GDMDWHOM)DNXOHWD]DHOHNWURWHKQLNR/MXEOMDQD 8UHGQLNSURIGU6DãR7RPDåLþ .UDMLQOHWRL]LGD/MXEOMDQD HOHNWURQVNDL]GDMD P R O G R A M M E Plenary Lectures Page Monday morning Plenary, Monday, Jun 10 2024, 9:15-10:00 Location: Lecture room A+B Session: Plenary talks session 33 9:15 New ablation method based on irreversible electroporation – a game 33 PL-1 changer in cardiac ablation Atul Verma Monday afternoon Plenary, Monday, Jun 10 2024, 13:30-14:15 Location: Lecture room A+B Session: Plenary talks session 33 13:30 Cardiac regeneration: where we stand 33 PL-2 Mauro Giacca Tuesday morning Plenary, Tuesday, Jun 11 2024, 8:30-10:00 Location: Lecture room A+B Session: Plenary talks session 33 8:30 Applications of Artificial Intelligence in Clinical Diagnosis, Therapy and 33 PL-3 Operations Shankar Krishnan 9:15 Cybersecurity for connected medical devices 34 PL-4 Dubravka Maljević Tuesday afternoon Plenary, Tuesday, Jun 11 2024, 13:30-14:15 Location: Lecture room A+B Session: Plenary talks session 34 13:30 The Joy of Point-Based Visualizations 34 PL-5 Blaž Zupan Wednesday morning Plenary, Wednesday, Jun 12 2024, 8:30-10:00 Location: Lecture room A+B Session: Plenary talks session 34 8:30 History of irreversible electroporation and future clinical applications 34 PL-6 Rafael V. Davalos 9:15 Deep Brain Stimulation - Data Analysis for Clinical Support 34 PL-7 Karin Wårdell, Teresa Nordin, Dorian Vogel, Simone Hemm 5 Thursday morning Plenary, Thursday, Jun 13 2024, 8:30-10:00 Location: Lecture room A+B Session: Plenary talks session 35 8:30 Transferring scientific findings into medical devices: the power of simpli- 35 PL-8 city Zlatko Matjačić 9:15 Biomedical Engineering for Sustainable Global Health 36 PL-9 Leandro Pecchia 6 Oral Presentations Page Monday morning Track A, Monday, Jun 10 2024, 10:30-12:15 Location: Lecture room A+B Session: Technologies for assistive and preventive healthcare 39 Chairs: Zhivko Bliznakov and Lenka Lhotska 10:30 Digitalization of elderly care delivery 39 OR-001 Lenka Lhotska, Petr Novak, Jaroslav Cibulka, Vitezslav Mergl 10:45 Design and development of an optical prototype system: Preliminary res- 39 OR-002 ults Teodor Minev, Kristina Bliznakova, Nikolay Dukov, Zhivko Bliznakov 11:00 Screening prediabetes using 1 hour glucose. A simulation model to estim- 40 OR-003 ate the lifetime health and economic outcomes Martina Andellini, Melania Manco, Matteo Ritrovato 11:15 Quantifying static and dynamic balance for the sit to stand movement 40 OR-004 assessment Athia Haron, Glen Cooper, Helen Dawes, Maedeh Mansoubi, Andrew Weightman 11:30 Design of a Surgical Device for Performing Percutaneous Dilatational 40 OR-005 Tracheostomy Yuan Tang, Glen Cooper, Bruno Adorno, Brendan McGrath, Andrew Weightman 11:45 Telemedicine system for prehabilitation in urogynecology 41 OR-006 Petra Sládková, Marie Tichá, Radim Kliment, Jan Mužík, Marek Doksanský, Karel Hána, Pavel Smrčka 12:00 Current Challenges and Future Outlook for Extended Reality as Cutting- 41 OR-007 edge Assistive Technology shaping Caring Personnel Fabiano Bini, Michela Franzò, Alessia Finti, Franco Marinozzi Monday morning Track B, Monday, Jun 10 2024, 10:30-12:00 Location: Lecture room C Session: S21 - The role of Large-Language Models and Generative AI in Health Technology Management 42 Chairs: Alessio Luschi and Ernesto Iadanza Organisers/Conveners: Alessio Luschi and Ernesto Iadanza 10:30 Proposal of an XML standard protocol for Evidence-Based Medical Equip- 42 OR-008 ment Maintenance Alessio Luschi, Fabio Crapanzano, Francesca Satta, Lorenzo Sani, Ernesto Iadanza 10:45 Advancing Clinical Decision Support with Large Language Models: A 42 OR-009 Framework for Guideline-Compatible Hepatitis C Management Simone Kresevic, Mauro Giuffrè, Milos Ajcevic, Lory Croce, Dennis Shung, Agostino Accardo 7 11:00 Unveiling a New Resource for Crossing Linguistic Frontiers in Biomedicine: 43 OR-010 A Preliminary Study on English-Italian Translation Luca Bacco, Mario Merone, Leandro Pecchia 11:15 Prompt Engineering for Consistent Healthcare Responses in LLMs 43 OR-011 Ali Salman 11:30 Natural language processing in dermatology: an overview 44 OR-012 Alessandra Cartocci, Alessio Luschi, Linda Tognetti, Pietro Rubegni, Gabriele Cevenini, Ernesto Iadanza Monday morning Track C, Monday, Jun 10 2024, 10:30-12:00 Location: Lecture room D Session: S02 - Advances in BME in the Asia Pacific Region 44 Chairs: James Goh and Ichiro Sakuma Organisers/Conveners: James Goh, Ichiro Sakuma 10:30 Smart Wearable Haptics for AI Healthcare and the Health Metaverse 44 OR-013 Chwee Teck Lim 10:45 Silk-based Enthesis Regeneration in Anterior Cruciate Ligament Recon- 45 OR-014 struction James Goh 11:00 Advances in BME in Japan: JSMBE leads bio-resilience research to create 45 OR-015 new scientific principles Yasuyuki Shiraishi, Yumie Ono, Takeo Matsumoto, Ichiro Sakuma, Ryo Haraguchi, Tomohiro Kuroda 11:15 Automatic ROI Delineation for Semi�quantitative Analysis of Brain SPECT 45 OR-016 Imaging at Striatum Kang-Ping Lin 11:30 Soft Robotics for Healthcare 46 OR-017 Raye Chen-Hua Yeow 11:45 A Novel Direct Left-Ventricular Puncture Cannulation Extracorporeal 46 OR-018 Continuous Flow Blood Pump for Emergency Cardiac Support Francis Chikweto, Yasuyuki Shiraishi, Hanako Suzuki, Aoi Fukaya, Kazushi Ishiyama, Tashihiko Kijima, Tomoyuki Yambe Monday afternoon Track A, Monday, Jun 10 2024, 14:15-15:30 Location: Lecture room A+B Session: S07 - Electrochemical Effects in Bioelectrical Therapies 47 Chairs: Antoni Ivorra and Samo Mahnič-Kalamiza Organisers/Conveners: Antoni Ivorra and Samo Mahnič-Kalamiza 14:15 Unwired bipolar electrodes in neural systems. Main aspects of wireless 47 OR-019 electrochemistry in biological systems Nieves Casan-Pastor 14:30 Electrochemistry of neural stimulation 47 OR-020 Jiří Ehlich 14:45 Electrolytic Electroporation for Tissue Ablation 48 OR-021 Mary Ho 8 15:00 Leveraging the Electrochemical Reactions from Electroporation to Pro- 48 OR-022 mote Anti-Tumor Phenotypes in Immune Cells Zaid Salameh, Manali Patwardhan, Alana Hay, Joanne Tuohy, Rafael Davalos 15:15 Beyond the rainbow: A validated mechanistic model of electrochemically 48 OR-023 produced pH changes in tissue Rok Šmerc, Damijan Miklavčič, Samo Mahnič-Kalamiza Monday afternoon Track B, Monday, Jun 10 2024, 14:15-15:30 Location: Lecture room C Session: S26 - IFMBE Education in BME: Transformation of the BME Education in Asia Pacific Region 49 Chairs: Martha Zequera and James Goh Organisers/Conveners: Martha Zequera Diaz and James Goh 14:15 A new framework of BME education with medical focus at CYCU, NYCU 49 OR-024 and NCKU in Asia-pacific region Kang-Ping Lin, Tse-Ying Liu, Ying-Hui Lai, Wen-Tai Chiu, Ming Chen Wang 14:30 Creating an Authentic Learning Environment in the NUS-BME program 50 OR-025 Raye Yeow, James Goh 14:45 Current status and features of BME education in China 50 OR-026 Hairong Zheng 15:00 Bioengineering education through the promotion of research collabora- 51 OR-027 tion of school of medicine and school of engineering at The University of Tokyo Ichiro Sakuma Monday afternoon Track C, Monday, Jun 10 2024, 14:15-15:30 Location: Lecture room D Session: S03 - Applications of Entropy in Health Care 52 Chair: Alexander Martynenko Organisers/Conveners: Alexander Martynenko 14:15 Entropy-Based Analysis of DNA Sequences and IGHV Mutational Status in 52 OR-028 Chronic Lymphocytic Leukemia: Predicting Patient Survival Alexander Martynenko, Xavier Pastor 14:30 Evaluating Mental Workload Through Cross-Entropy Analysis of Two Pre- 52 OR-029 frontal EEG Channels Matin Beiramvand, Mohammad Shahbakhti, Tarmo Lipping 14:45 Bayesian Shannon Entropy for assessing patient’s data interrelation in 52 OR-030 medical applications Alexander Martynenko, Xavier Pastor 15:00 Relationship of the Correlation between EEG and Heart Rate Variability 53 OR-031 with Cardiovascular Indicators Merilin Vihmaru, Laura Päeske, Hiie Hinrikus, Jaanus Lass, Toomas Põld, Maie Bachmann 9 15:15 Detecting Heart Failure Relations: A Preliminary Study Integrating HRV, 53 OR-032 LVEF, and GLS in Patients with Ischemic Heart Disease and Dilated Car- diomyopathy Katerina Iscra, Laura Munaretto, Aleksandar Miladinovic, Jacopo Giulio Rizzi, Marco Merlo, Agostino Accardo, Milos Ajcevic Monday late afternoon Track A, Monday, Jun 10 2024, 16:00-18:15 Location: Lecture room A+B Session: S08 - Electroporation meets cardiac treatment 53 Chairs: Lea Rems and Bojan Vrtovec Organisers/Conveners: Lea Rems 16:00 What do we (think we) know about electroporation 53 OR-033 Damijan Miklavčič 16:15 Initial single centre experience with pulsed field ablation for treatment of 54 OR-034 cardiac arrhythmias Matevž Jan, Tine Prolič Kalinšek, Jernej Iršič, Damijan Miklavčič, Jernej Štublar 16:30 Numerical modelling in pulsed field ablation 55 OR-035 Bor Kos 16:45 Machine Learning Models for Predicting Electroporation-Dependent Tis- 55 OR-036 sue Properties and Temperature Distributions from Pulsed Electric Fields Edward Jacobs, Pedro Santos, Rafael Davalos 17:00 In Vivo Cardioporation Enhanced Myocardial Gene Delivery 56 OR-037 Richard Heller, Mark Jaroszeski, Carly Boye, Michael Francis, Jacob Hensley, Alex Otten, Sezgi Arpag-McIntosh, Anna Bulysheva 17:30 Gene electrotransfer – the role of electrical and biological parameters in 56 OR-038 in vitro experiments Alenka Maček Lebar 17:45 Precision medicine approach to cell therapy in heart failure 57 OR-039 Bojan Vrtovec 18:00 The LVAD-CD34+ Stem Cell Combination Therapy in Non-ischemic Dilated 57 OR-040 Cardiomyopathy Patients Gregor Poglajen, Sabina Frljak, Gregor Zemljič, Andraž Cerar, Bojan Vrtovec Monday late afternoon Track B, Monday, Jun 10 2024, 16:00-18:00 Location: Lecture room C Session: S12 - Intelligent Health Monitoring for Precision Medicine: A Global Perspective 58 Chairs: Li Li and Hairong Zheng Organisers/Conveners: Li Li 16:00 Multimodal Biomedical Imaging for the Advance Precision Medicine: Past, 58 OR-041 Present and Future Hairong Zheng 16:15 United States Perspective on Intelligent Health Monitoring in Precision 58 OR-042 Medicine Shankar Krishnan 10 16:30 Precision medicine and AI for pesonalised estimation of glucose levels via 59 OR-043 ECG Leandro Pecchia 16:45 Precision medicine in Latin America: Universalizing the promise of innov- 59 OR-044 ation Virgina Laura Ballarin 17:00 Comparative study of machine learning methods for the early prediction 60 OR-045 of adherence to treatment Miguel Rujas, Beatriz Merino, Peña Arroyo, Jim Ingebretsen, Jaime Barrio Cortes, Ana Isabel Villimar Rodriguez, Andrés Castillo, Ana Roca-Umbert, Francisco Lupiañez, Maria Fernanda Cabrera, Maria Teresa Arredondo, Giuseppe Fico 17:15 Bone Metastasis Measurement based on CNN Model for Auto-ROIs 60 OR-046 Kang-Ping Lin, Jimmy Hasugian 17:30 Trusted quality infrastructure federated data space for networks in bio- 60 OR-047 medical engineering and medical diagnostics Tomasz Sołtysiński Monday late afternoon Track C, Monday, Jun 10 2024, 16:00-18:15 Location: Lecture room D Session: Biomedical instrumentation, signal processing and imaging 61 Chairs: Josias Wacker and Kazimierz Pęczalski 16:00 A System for Multimodal Ultrasound Vascular Assessment 61 OR-048 Marcin Lewandowski, Piotr Karwat 16:15 Cardiovascular control alterations associated with genetic polymorph- 62 OR-049 isms of proteins involved in sympathetic pathway signal transduction Michal Javorka, Lenka Matušková, Barbora Czippelova, Jana Čerňanová Krohová, Zuzana Turianiková, Zuzana Kolková, Zora Lasabová 16:30 Control Unit for Phototherapy Sources 62 OR-050 Jan Havlík, Lenka Maierová 16:45 Transfer Learning from the Domain of Diabetic Retinopathy to Aid in the 63 OR-051 Detection of Age-Related Macular Degeneration Roberto Romero-Oraá, María Herrero-Tudela, Roberto Hornero, María I. López, María García 17:00 Simple electrochemical sensor for measuring oxygen tension in blood or 63 OR-052 respiratory gases Tadeusz Pałko, Kazimierz Pęczalski 17:15 Comparison of forehead and finger photoplethysmographic waveform 63 OR-053 index for stress assessment Kristjan Pilt, Ivo Fridolin 17:30 Non-Invasive Continuous Measurement of the Intra-Abdominal Pressure 64 OR-054 Josias Wacker, Srdjan Djordjevic, Blaž Trotovšek, Simon Krašna, Jan Žumer, Etienne Haenni, Grégoire Banderet, Patrick Richard, Gürkan Yilmaz 17:45 Estimation of middle ear characteristics by an innovative Pressure-Less 64 OR-055 Acoustic Immittance (PLAI™) device Francesco Bassi, Agostino Accardo 11 18:00 Synchronization of magnetocardiographic and electrocardiographic re- 65 OR-056 cordings obtained with two separate systems Kazimierz Pęczalski, Teodor Buchner, Judyta Sobiech, Gerard Cybulski, Tadeusz Pałko Tuesday morning Track A, Tuesday, Jun 11 2024, 10:30-12:15 Location: Lecture room A+B Session: S15 - Optimizing Healthcare: Patient-Centered Technology Innovations 65 Chairs: Paulo Carvalho, Fabiola Martinez-Licona and Murilo Conto Organisers/Conveners: Paulo de Carvalho, Fabiola Martinez, and Murilo Conto 10:30 Patient-centered Innovation in Hospital Patient Monitoring 65 OR-057 Jens Muehlsteff 11:00 Shaping the healthcare of tomorrow through patient-centered digital 66 OR-058 technologies Josias Wacker 11:30 Consensus on best practices for the management of Large-Scale Pilot 66 OR-059 projects: results from a Delphi study Davide Piaggio, Jordi De Batlle, Alba Gallego, Gloria Cea, Pedro Checa Rifa, Alessia Maccaro, Giuseppe Fico, Leandro Pecchia 11:45 Let’s Play a Cognitive Game 66 OR-060 Anton Nydal, Ankica Babic 12:00 Comparison of automatic and semiautomatic approach for the posterior 67 OR-061 nipple line calculation Francesca Angelone, Alfonso Maria Ponsiglione, Roberto Grassi, Francesco Amato, Mario Sansone Tuesday morning Track B, Tuesday, Jun 11 2024, 10:30-12:00 Location: Lecture room C Session: Biomedical engineering in developing countries 67 Chairs: Ernesto Iadanza and Wanzi Su 10:30 Smartphone Technology: Manual Validation of Two Eye-tracking Al- 67 OR-062 gorithms Wanzi Su, Damon Hoad, Davide Piaggio, Leandro Pecchia 10:45 A Novel University Course on Medical Devices Design and Certification at 68 OR-063 the University of Siena Ernesto Iadanza 11:00 Demand for Biomedical Engineers in Tanzania – A View from the Outside 68 OR-064 Mirjam Bodenstorfer 11:15 The systematic literature review of automated pupillometry system on 69 OR-065 five common ophthalmological diseases Wanzi Su, Katia Polsi, Jacopo Vitale, Emmanuel Nathaniel, Alvin Vyapooree, Daniel Fox, Damon Hoad, Leandro Pecchia, Davide Piaggio 11:30 A frugal smart tool for screening for diabetic neuropathies 69 OR-066 Davide Piaggio, Rossana Castaldo, Gianluca Garibizzo, Ernesto Iadanza, Leandro Pecchia 12 11:45 Establishment of Femoral Bone Defect Model in Sprague-Dawley Rat for 70 OR-067 Engineered Scaffold Implantation: A Pilot Study Amira Raudhah Abdullah, Intan Maslina Musa Tuesday morning Track C, Tuesday, Jun 11 2024, 10:30-12:00 Location: Lecture room D Session: S17 - Regulatory Learning 70 Chairs: Laura Lopez-Perez and Beatriz Merino Organisers/Conveners: Giuseppe Fico 10:30 Living Labs as Regulatory Learning Tools for Biomedical Engineering and 70 OR-068 Digital Transformation in Healthcare Giuseppe Fico 10:45 Integrating Living Labs for Harmonized Data Collection in Transitional 71 OR-069 Care Beatriz Merino, Gloria Cea, Ivana Lombroni, Irene Mallo, Diego Carvajal, Alejandro Medrano, Maria Fernanda Cabrera, Maria Teresa Arredondo, Giuseppe Fico 11:00 Bridging Digital Transformation through EU-LAC Cooperation 71 OR-070 Maria Fernanda Cabrera 11:15 Regulatory Frameworks and Validation Strategies for Advancing Artificial 72 OR-071 Intelligence in Healthcare Laura Lopez-Perez, Beatriz Merino, Miguel Rujas, Alessia Maccaro, Leandro Pecchia, Sergio Guillen, Maria Fernanda Cabrera, Maria Teresa Arredondo, Giuseppe Fico 11:30 The BEAMER Lab: conceptualizing a living lab frame-work to develop 72 OR-072 predictive models, tools, and support programs to improve adherence to treatment Beatriz Merino, Miguel Rujas, Peña Arroyo, Rodrigo Martín Gómez del Moral Herran, Jim Ingebretsen, Francisco Lupiañez, Giuseppe Fico Tuesday afternoon Track A, Tuesday, Jun 11 2024, 14:15-15:30 Location: Lecture room A+B Session: S08 - Electroporation meets cardiac treatment 73 Chairs: Lea Rems and Chiara Collesi Organisers/Conveners: Lea Rems 14:15 Modelling Dilated Cardiomyopathy through Induced Pluripotent Stem 73 OR-073 Cell-Derived Cardiomyocytes Rebecca Artioli, Rossella Cannilla, Cecilia Thairi, Matteo Dal Ferro, Elisa Di Pasquale, Gianfranco Sinagra, Mauro Giacca, Chiara Collesi 14:30 Exploring the Effects of Electroporation on Primary Rat Ventricular Car- 73 OR-074 diomyocytes: Insights into Action Potential, Calcium Release, and Con- traction Dynamics Vid Jan, Marko Stručić, Tina Turk, Jernej Jurič, Monika Kos, Matej Reberšek, Martina Perše, Lea Rems, Damijan Miklavčič 14:45 Excitation and electroporation in excitable S-HEK cells exposed to elec- 74 OR-075 tric pulses of different durations Tina Batista Napotnik, Tina Cimperman, Lea Rems 13 15:00 Localized electroporation on a chip for effective DNA electrotransfer 74 OR-076 Pouyan Boukany 15:15 Gene electrotransfer into cells in hydrogels as extracellular mimicry 75 OR-077 Saša Haberl Meglič Tuesday afternoon Track B, Tuesday, Jun 11 2024, 14:15-15:30 Location: Lecture room C Session: S27 - Perspectives of BME Education 75 Chairs: Ratko Magjarevic, Nicolas Pallikarakis and Martha Zequera Organisers/Conveners: Ratko Magjarević, Nicolas Pallikarakis and Martha Zequera 14:15 IFMBE project on Survey of BME Programs worldwide by IFMBE regions 75 OR-078 Ratko Magjarević 14:30 How the BME study programs survey is planned to be organized: A pro- 75 OR-079 posal Nicolas Pallikarakis 14:45 Emerging Technologies to Support Health Care and Independent Living 76 OR-080 Summer School: An Innovation Education Model in Latin America Martha L. Zequera, Ratko Magjarević, Virginia Ballarin, Luis Zamudio, Shankar Krishnan, Juan Pastore, Eduardo Pinos, Rosa Flores-Luna, Alvaro Poveda, Piotr Ładyżyński, Nubia Arroyo, Mauricio Cubides 15:00 What does harmonization bring to students, teachers, and researchers in 77 OR-081 BME Shankar Krishnan 15:15 BME educational programs - Between historical foundations and the 77 OR-082 need for interdisciplinarity Damijan Miklavčič Tuesday afternoon Track C, Tuesday, Jun 11 2024, 14:15-15:30 Location: Lecture room D Session: S16 - Out-of-the-lab dry EEG 78 Chairs: Jens Haueisen and Patrique Fiedler Organisers/Conveners: Jens Haueisen and Patrique Fiedler 14:15 Dry electrode technologies for comfortable and reliable out-of-the-lab 78 OR-083 EEG Patrique Fiedler, Abdumumin Olimzoda, Indhika Warsito, Silvia Comani, Jens Haueisen 14:30 A flexible soft cap for neonatal EEG-NIRS recording produced by additive 78 OR-084 manufacturing Mohammed-Reda Mejbar, Ralf Hauffe, Jens Haueisen, Patrique Fiedler 14:45 Comparison of methods for bad channel detection in dry EEG 79 OR-085 Milana Komosar, Patrique Fiedler, Jens Haueisen 15:00 Technology and study protocol for the multimodal and ecological invest- 80 OR-086 igation of joint action in table tennis Silvia Comani, Gabriella Tamburro, Antonio De Fano, Khadijeh Raeisi Nafchi, Mohammad Khazaei, Filippo Zappasodi, Ricardo Bruña, Edson Filho, Hannes Oppermann, Patrique Fiedler 14 15:15 Analysis of hyperbrain data underpinning naturalistic joint action 80 OR-087 Gabriella Tamburro, Ricardo Bruña, Patrique Fiedler, Antonio De Fano, Khadijeh Raeisi Nafchi, Mohammad Khazaei, Filippo Zappasodi, Silvia Comani Tuesday late afternoon Track A, Tuesday, Jun 11 2024, 16:00-18:00 Location: Lecture room A+B Session: IFMBE Young Investigators Competition 81 Chairs: Helena Cindrič and Zhivko Bliznakov 16:00 Online Uric Acid Concentration Estimation in Blood from Spent Dialysate 81 OR-088 Measurements Using an Optical Sensor Joosep Paats, Jürgen Arund, Kristjan Pilt, Annika Adoberg, Liisi Leis, Merike Luman, Jana Holmar, Risto Tanner, Ivo Fridolin 16:15 Comparison of Different Configurations for the Implantable Capacitive 81 OR-089 Intrabody Communication on a Two-Layer Phantom Matija Roglić, Luka Klaić, Ziliang Wei, Yueming Gao, Željka Lučev Vasić 16:30 Computational hemodynamic Evaluation of Different Surgical Designs of 82 OR-090 Systemic to Pulmonary Arterial Shunt under Controlled Pulmonary Flow Jiwen Xiong, Qi Sun, Jinfen Liu, Jinlong Liu 16:45 Accurate and interpretable deep learning model for sleep staging in chil- 82 OR-091 dren with sleep apnea from pulse oximetry Fernando Vaquerizo-Villar, Daniel Álvarez, Gonzalo C. Gutiérrez-Tobal, Adrián Martín-Montero, David Gozal, Eduardo Tamayo, Roberto Hornero 17:00 Combined measurement of brain activation during a motor task using 83 OR-092 fNIRS and OPM-MEG Urban Marhl, Stanislaw Wojtkiewicz, Piotr Sawosz, Vojko Jazbinšek, Zvonko Jagličić, Adam Liebert, Tilmann H. Sander 17:15 Frequency Domain Cluster Analysis of Human Activity Using Triaxial Ac- 83 OR-093 celerometer Data Krunoslav Jurcic, Goran Šeketa, Ratko Magjarević 17:30 Electrochemical detection of desorbed calcium ions from bovine serum 83 OR-094 albumin binding Tamara Boscarino, Antonio D’Ambrosio, Andrea Palermo, Vincenzo Piemonte, Leandro Pecchia 17:45 Computational analysis of large bone defect healing using bone tissue 84 OR-095 scaffolds, degradation, and growth factor delivery: a mechanobiological model of bone tissue formation Adel Alshammari, Fahad Alabdah, Lipeng Song, Glen Cooper Tuesday late afternoon Track C, Tuesday, Jun 11 2024, 16:00-18:00 Location: Lecture room D Session: S24 - Wireless Wearable Networks for Physical Function Rehabilitation Monitoring 84 Chairs: Željka Lučev Vasić, Ivana Culjak and Yueming Gao Organisers/Conveners: Željka Lučev Vasić, Ivana Čuljak, and Yueming Gao 15 16:00 Revealing Statistical Patterns in Shoulder Rehabilitation Exercises Char- 84 OR-096 acteristics Martina Sassi, Margherita Anna Grazia Matarrese, Umile Giuseppe Longo, Leandro Pecchia 16:15 Body Movement Analysis during Sleep Based on Ultra-Wideband Com- 85 OR-097 munication Channel Impulse Response Measurement Ivana Culjak, Željka Lučev Vasić, Domagoj Vidović, Mario Cifrek 16:30 Comparison of Time and Frequency Domain Features For Cluster Analysis 85 OR-098 of Human Activity Using Triaxial Accelerometer Data Krunoslav Jurcic 16:45 Dynamic Equivalent Circuit Models for Intracardiac Communication in 85 OR-099 Leadless Pacemakers Dongming Li, Jiamei Wang, Pedro Antonio Mou, Yadong Yin, Hung Chun Li, Yueming Gao, Peng Un Mak, Sio Hang Pun, Mang I. Vai 17:00 Cardiac Phantoms for Signal Transmission Characteristics Analysis of 86 OR-100 Leadless Pacemakers Han Wang, Jiamei Wang, Yang Shuang, Ziliang Wei, Željka Lučev Vasić, Jiejie Yang, Zhimeng Xu, Mario Cifrek, Yueming Gao 17:15 Impact of the Tightness of a Sensorized Top in the Quality of the ECG it 86 OR-101 Records Marta Guzmán-Alarcón, Natalia Guzmán Rodríguez, Pablo Pérez-Tirador, Constantino A. García, Abraham Otero 17:30 Impedance Cardiography signals in applications for cardiovascular in- 87 OR-102 formation content analysis Gerard Cybulski, Tadeusz Pałko, Kazimierz Pęczalski 17:45 An Everyday Hat for Detection of Eye Blinks and Forehead Clenching 87 OR-103 S. M. Musfequr Rahman, Henna Mattila, Asif Shaikh, Pasi Raumonen, Johanna Virkki 16 Wednesday morning Track A, Wednesday, Jun 12 2024, 10:30-12:00 Location: Lecture room A+B Session: S09 - Health Technology Assessment of Medical Devices - Advances and Challenges 88 Chairs: Murilo Conto and Ernesto Iadanza Organisers/Conveners: Murilo Contó and Ernesto Iadanza 10:30 Exploration and practice of quality control index system of medical 88 OR-104 device management Yueqi Yang, Yingxin Xu, Xin Li, Xiangnan Geng, Xiang Xu, Hong Gao, Yuzhi Yang, Wei Jin, Hui Zhong, Ying Qian 10:45 Cost-effectiveness of Continuous Glucose Monitoring For Paediatric Pa- 88 OR-105 tients With Type 1 Diabetes Compared With Self-Monitoring Of Blood Glucose Martina Andellini, Riccardo Schiaffini, Matteo Ritrovato, Leandro Pecchia 11:00 Optimizing Liver Stiffness Assessment in HCV Patients: A Machine Learn- 89 OR-106 ing Approach to Identify Confounding Factors in Fibrosis Estimation Simone Kresevic, Mauro Giuffrè, Milos Ajcevic, Lory Croce, Agostino Accardo 11:15 Advances in Health Technology Assessment of Wearable Devices: A Clin- 89 OR-107 ical Engineering Perspective Fabiola Martinez-Licona 11:30 Alarms Early Detection in Dialytic Therapies via Machine Learning Models 90 OR-108 Alessia Nicosia, Nunzio Cancilla, Marco Siino, Michele Passerini, Francesca Sau, Ilenia Tinnirello, Andrea Cipollina 11:45 Exploring the Potential of Health Data: EHDS, Secondary Utilization, and 90 OR-109 Stakeholder Perspectives in Czech Healthcare Petra Hospodková, Martin Budil Wednesday morning Track B, Wednesday, Jun 12 2024, 10:30-12:00 Location: Lecture room C Session: S22 - Tools, technologies and computing for point-of-care person-centered health and care delivery 90 Chair: Eleni Kaldoudi Organisers/Conveners: Eleni Kaldoudi 10:30 Pointwise reliability of machine learning models: application to cardi- 91 OR-110 ovascular risk assessment Jorge Henriques, Teresa Rocha, Simão Paredes, Paulo Gil, João Loureiro, Lorena Petrella 10:45 Towards Wearable Continuous Point-of-Care Monitoring for Deep Vein 91 OR-111 Thrombosis of the Lower Limb Eleni Kaldoudi, Vaidotas Marozas, Pavlos Moustakidis, Dmitry Novikov, Jurkonis Rytis, Andrius Sakalauskas, Nicolas Pousset, Mathieu Legros, Marco Kircher, Babajide Ayinde, Lara Alessia Moltani, Susann Balling, Antti Vehkaoja, Niku Oksala, Andrius Macas, Neringa Balciuniene, Maria Bigaki, Michail Potoupnis, Stella-Lida Papadopoulou, Elvira Grandone, Maxime Gautier, Sabrina Bouda, Cord Schloetelburg, Thorsten Prinz, Pietro Dionisio, Spyridon Anagnostopoulos, Ioanna Drougka, Frans Folkvord, George Drosatos, Stylianos Didaskalou 17 11:00 Attention Theory Based 12 Lead Visualization of Ventricular Function Dur- 92 OR-112 ing Ventricular Pacing Lead Implants Neal Duong, Jessica Lee, Paul A. Iaizzo 11:15 Integration of in-vitro technologies, in-silico biophysical and data driven 92 OR-113 models towards better stratification and care for hypertrophic cardi- omyopathy patients Jari Hyttinen 11:30 Development of an Explainable Deep Learning-Based Decision Support 93 OR-114 System for Blood Glucose Levels Forecasting in Type 1 Diabetes Using Edge Computing Isabel Longo, Federico D’Antoni, Lorenzo Petrosino, Vincenzo Piemonte, Mario Merone, Leandro Pecchia 11:45 Enhancing Balance Rehabilitation through use of Augmented Reality 94 OR-115 Aleksandra Vulović, Đorđe Ilić, Nenad Filipovic Wednesday morning Track C, Wednesday, Jun 12 2024, 10:30-12:00 Location: Lecture room D Session: S19 - IFMBE Industry Committee: Medical Technologies - From Concept to Commercialization 94 Chairs: Piotr Ładyżyński and Martha Zequera Organisers/Conveners: Piotr Ladyzynski and Martha Zequera Diaz 10:30 Research collaboration between academia, healthcare and industry. Ex- 94 OR-116 perience from the research profile Embedded Sensor Systems for Health and an adjunct industrial graduate school Maria Lindén, Mats Björkman 10:45 Innovative ultrasound devices — from research to commercial product 95 OR-117 Marcin Lewandowski 11:00 Development and commercialization of a novel stimulation device for 95 OR-118 denervated muscles – challenging under MDD, most likely impossible un- der current MDR Winfried Mayr 11:15 Navigating Innovational Valleys: Advancing Health Technology Innova- 96 OR-119 tion for Sustainable and Affordable Care Sudesh Sivarasu 11:30 Understanding Regulatory Requirements: A Postmortem Analysis of 97 OR-120 Tremitas’ Bankruptcy in the Medtech Sector Tibor Zajki-Zechmeister Wednesday after lunch Track B, Wednesday, Jun 12 2024, 13:30-14:15 Location: Lecture room C Session: IFMBE Scientific Challenge 97 Chairs: Bor Kos and Paulo Carvalho Info: Additional talks in this session: 13:30 Jorge Henriques ”The challenge: past and future” 13:40 Bor Kos ”This year’s challenge topic” 18 13:30 A Machine Learning Approach for Predicting Electrophysiological Re- 97 OR-121 sponses in Genetically Modified HEK Cells Jacopo Vitale, Martina Sassi, Leandro Pecchia 13:45 Optimizing Electroporation Responses in Genetically Engineered HEK 97 OR-122 Cells: An Ensemble Learning Approach Francesco Bassi, Simone Kresevic, Alessandro Biscontin, Aleksandar Miladinovic, Milos Ajcevic, Agostino Accardo Wednesday after lunch Track C, Wednesday, Jun 12 2024, 13:30-14:15 Location: Lecture room D Session: Cellular and tissue engineering 98 Chairs: Adriana Lungu and Dongan Wang 13:30 Decellularized Tissue Engineering Hyaline Cartilage Graft for Articular 98 OR-123 Cartilage Repair and Its Forward-Looking Test for Space Medicine Dongan Wang 13:45 Multicomponent printing inks for the designing of 3D extracellular matrix- 98 OR-124 mimetic biomaterials Adriana Lungu, Raluca Dobrisan, Valentina Ciobanu, Minodora Marin, Sorina Dinescu, Izabela-Cristina Stancu 14:00 Artificial Bone Extracellular Matrices based on 3D-printed nanostruc- 99 OR-125 tured hydrogels Adriana Lungu, Izabela-Cristina Stancu, Elena Olăreț, Filis Curti, Carmen-Valentina Nicolae, Sorina Dinescu, Alexandra-Elena Dobranici, Bogdan Stefan Vasile Wednesday afternoon Track A, Wednesday, Jun 12 2024, 14:15-15:30 Location: Lecture room A+B Session: S20 - System design for biomagnetic applications using optically pumped magnetometers 99 Chairs: Tilmann H. Sander, Daniel Baumgarten and Urban Marhl Organisers/Conveners: Tilmann Sander, Daniel Baumgarten, and Urban Marhl 14:15 Active compensation for OPM-MEG inside a two-layer magnetically 99 OR-126 shielded room Michal Władziński, Anna Jodko-Władzińska, Tilmann H. Sander 14:30 Quantitative magnetic nanoparticle imaging with magnetorelaxometry 100 OR-127 in unshielded environments Aaron Jaufenthaler, Daniel Baumgarten 14:45 Improved degaussing procedure for a magnetically shielded room 100 OR-128 Peter Koss, Jens Voigt, Allard Schnabel, Ronja Rasser 15:00 Full MEG system based on helium OPMs for medical imaging 101 OR-129 Rudy Romain, Jaroslaw Rutkowski, Sergey Mitryukovskiy, Kevin Arth, Matthieu Le Prado, Agustin Palacios-Laloy, Etienne Labyt 15:15 Full OPM-MEG system with limited sensor counts 101 OR-130 Urban Marhl, Rok Hren, Tilmann H. Sander, Vojko Jazbinšek 19 Wednesday afternoon Track B, Wednesday, Jun 12 2024, 14:15-15:30 Location: Lecture room C Session: S28 - BME Education in Europe 102 Chairs: Nicolas Pallikarakis and Ratko Magjarević Organisers/Conveners: Ratko Magjarević and Nicolas Pallikarakis 14:15 Overview of the BME study programs – previous European wide survey 102 OR-131 results Nicolas Pallikarakis 14:30 EU students, teachers and researchers exchange programs 102 OR-132 Ratko Magjarević 14:45 The CRH-BME Tempus project – results and experience a decade after 102 OR-133 Tomaž Jarm 15:00 BME in EC research projects 103 OR-134 Leandro Pecchia 15:15 Ethical Engineer: Enhancing AI Education in Engineering 103 OR-135 Lenka Lhotská, David Macku Wednesday afternoon Track C, Wednesday, Jun 12 2024, 14:15-15:30 Location: Lecture room D Session: BME Miscellaneous topics 104 Chairs: Milica Nikolic and Nikolay Dukov 14:15 M-Health in Prostate Cancer: Professional and Patient Perspectives 104 OR-136 Petra Hospodková, Irina Klubarská, Matyáš Mašek, Martin Budil 14:30 Research on dental materials for their suitability in building anthropo- 104 OR-137 morphic phantoms Nikolay Dukov, Minko Milev, Todor Todorov, Zhivko Bliznakov, Kristina Bliznakova 14:45 Analysis of redox processes from in silico perspective in colorectal cancer 104 OR-138 cell line Milica Nikolic, Nevena Milivojevic Dimitrijevic, Ana Miric, Marko Zivanovic, Nenad Filipovic 15:00 Modeling the conductive characteristics of the human lower back using 105 OR-139 medical imaging and FEM modeling to optimize tSCS therapy Jón Andri Árnason, Ragnhildur Guðmundsdóttir-Korchai, Þórður Helgason 15:15 Assessing Ergonomic Compliance in Industrial Environments with Mark- 106 OR-140 erless 3D Camera-Based Systems Jindřich Adolf, Ilona Kačerová, Kateřina Jurčová, Vladimíra Lipšová, Jaromír Doležal, Lenka Lhotská Wednesday late afternoon Track A, Wednesday, Jun 12 2024, 16:00-18:00 Location: Lecture room A+B Session: S14 - Non-invasive methods for monitoring electrophysiological and hemodynamic brain activity 106 Chairs: Vojko Jazbinšek, Tilmann H. Sander and Stanislaw Wojtkiewicz Organisers/Conveners: Vojko Jazbinšek, Tillman Sander, and Stanislaw Wojtkiewicz 20 16:00 Difference in Mentalizing Function between Face-to-Face and Online 106 OR-141 Communications Joohyeong Kim, Hayato Watanabe, Koichi Yokosawa 16:15 Track-EP: A Semi-automated Tool for Enhancing Interictal EEG Interpret- 106 OR-142 ation Margherita Anna Grazia Matarrese, Simonetta Filippi, Leandro Pecchia, Christos Papadelis 16:30 The Neuromodulation Evaluation with Functional Near-Infrared Spectro- 107 OR-143 scopy for Transcutaneous Auricular Vagus Nerve Stimulation on Upper Ex- tremity Rehabilitation Chien-An Chen, Jia-Jin Chen, Wei-Cheng Sun, Chao Chen Lo 16:45 Can fNIRS be used as a treatment monitoring tool for Alzheimer’s dis- 108 OR-144 ease? Samaneh Azarbarzin, Zahra Moussavi 17:00 Multiparametric measurement of cerebral blood flow in neurointensive 108 OR-145 care Karin Wårdell, Sofie Tapper, Stina Mauritzon, Fredrik Ginstman, Peter Zsigmond 17:15 Measuring Anxiety Quantitatively in a Depression population 109 OR-146 Zahra Moussavi, Brian Lithgow 17:30 Data processing and visualization for multimodal imaging of neuro- 109 OR-147 vascular coupling Stanislaw Wojtkiewicz, Urban Marhl, Piotr Sawosz, Vojko Jazbinšek, Tilmann H. Sander, Adam Liebert 17:45 Applying near infrared spectroscopy for assessment of anxiety 110 OR-148 Chien-An Chen, BO-TSEN Lin, Yi-Jing Huang, Yu-Jui Huang, Pao-Huan Chen, Jia-Jin Chen Wednesday late afternoon Track B, Wednesday, Jun 12 2024, 16:00-18:00 Location: Lecture room C Session: S25 - IFMBE Education in BME: BME Education in Latin America and BME in Africa 110 Chairs: Martha Zequera, Virginia Ballarin and Sudesh Sivarasu Organisers/Conveners: Martha Zequera Diaz and Virginia Laura Ballarín 16:00 ARCUSUR SYSTEM: supporting the way to accreditation programs 110 OR-149 Virgina Laura Ballarin 16:15 Academic Programs in Mexico: The Current State of Biomedical Engin- 111 OR-150 eering 50 Years Later Fabiola Martinez-Licona 16:30 Integration of health technology management in healthcare service de- 111 OR-151 livery systems in Kenya Salome Mwaura 16:45 Managed equipment service in Kenya 112 OR-152 Symon Mbakah 17:00 Needs assessment in Kenyan healthcare facilities for context-sensitive 113 OR-153 design of medical devices Florinda Coro, Bazil Mlamba, Augustine Waswa, Carmelo De Maria, June Madete, Valentina Mangano, Margaret Keraka, Arti Ahluwalia 21 17:15 Standardisation of Medical Devices: A Contended Practice with Potential 113 OR-154 Benefits for South African Healthcare Sudesh Sivarasu 17:30 Digitizing Malaria Case Management Protocol in Ghana: A pilot Study 114 OR-155 George Boadu 17:45 From Emergency to Enhancement: Streamlining Ventilator Procurement 114 OR-156 in Africa for Improved Healthcare Outcome Sudesh Sivarasu Wednesday late afternoon Track C, Wednesday, Jun 12 2024, 16:00-18:15 Location: Lecture room D Session: Artificial intelligence in healthcare 115 Chairs: Jacopo Vitale and Marilena Tarousi 16:00 Insights in Data Generation: A Synthetic Data Approach for Enabling 115 OR-157 Small Datasets in Atrial Fibrillation Research Ali Salman, Francesco Goretti, Alessandra Cartocci, Ernesto Iadanza 16:15 An explainable deep-learning model to aid in the diagnosis of age related 115 OR-158 macular degeneration María Herrero-Tudela, Roberto Romero-Oraá, Roberto Hornero, Gonzalo C. Gutiérrez-Tobal, María I. López, María García 16:30 Wearable sensors and artificial intelligence for blood pressure estimation 115 OR-159 during daily life and sleep, a pilot study Katy Stokes, Salman Haleem, Rossana Castaldo, Francesco Cappuccio, Leandro Pecchia 16:45 Unlocking robotic potential through modern organ segmentation 116 OR-160 Ansh Chaudhary, Robail Yasrab 17:00 Rare Eye Diseases Automatic Classification: A Deep Learning Approach 116 OR-161 Jacopo Vitale, Maria Elisabetta Pagnano, Margherita Anna Grazia Matarrese, Rosa Boccia, Paolo Melillo, Francesco Testa, Francesca Simonelli, Leandro Pecchia 17:15 Empowering Colorectal Cancer Research through Advanced Data Integ- 117 OR-162 ration and Analysis: A case study of the DIOPTRA project Marilena Tarousi, Stavros-Theofanis Miloulis, Maria Haritou, Konstantinos Bromis, Ioannis Kouris, George Botis, Ioannis Kakkos, George Matsopoulos 17:30 Predicting Depression Status After Transcranial Direct Current Stimula- 117 OR-163 tion Treatment Using Machine Learning Sayna Rotbei, Giordano D’Urso, Alessio Botta 17:45 Evaluation of Hydrogel Flow into Osteoporotic Trabecular Bone: A Com- 117 OR-164 putational Fluid Dynamics Study Fahad Alabdah, Adel Alshammari, Araida Hidalgo-Bastida, Glen Cooper 18:00 Enabling COVID-19 detection from multiple audio recordings: a prelimin- 118 OR-165 ary comparison between cough, breath, and speech signals Alfonso Maria Ponsiglione, Francesca Angelone, Rossella Sparaco, Salvatore Piccolo, Amy Parrish, Andrea Calcagno, Guillaume Fournier, Ayana de Brito Martins, Fulvio Cordella, Arianna Arienzo, Lorenzo Castella, Vincenzo Norman Vitale, Francesco Amato, Maria Romano 22 Thursday morning Track B, Thursday, Jun 13 2024, 10:30-12:00 Location: Lecture room C Session: S11 – IFMBE Education in BME: Education in Biomedical Engineering for Students 118 Chairs: Martha Zequera, Ratko Magjarevic, Shankar Krishnan and Piotr Ładyżyński Organisers/Conveners: Martha Zequera Diaz, Ratko Magjarevic and Shankar M. Krishnan 10:30 An introduction to Ventra; a programmable abdominal phantom for 118 OR-166 training, educational, research, and development purposes Salar Tayebi, Ashkan Zarghami, Manu Malbrain, Johan Stiens 10:45 Ethics as part of biomedical engineering and informatics education 119 OR-167 Lenka Lhotska 11:00 Implementation of a Pattern Classifier on Thermograms from Plantar Re- 119 OR-168 gion Santiago Humberto Ramirez Martinez, Martha L. Zequera, Francisco Carlos Calderón Bocanegra 11:15 Using 24-hour Heart Rate Variability indices for prognosis monitoring in 120 OR-169 heart failure patients Shi-Yi Wu, Shao-Hung Lu, Mei-Fen Chen, Wen-Chen Lin, Wen-Chi Lin, Cheng-Lun Tsai, Kang-Ping Lin 11:30 Proposal for the Application of Blockchain in Predictive Management in 120 OR-170 Medical Devices Mariana Brandão, Renato Garcia 11:45 EMG Rehab: An interactive platform controlled by EMG signals to improve 121 OR-171 adherence to rehabilitation therapy in patients who have suffered lower body muscle injuries Leslie Yessenia Cieza Huané, Ana Cristina Aldana Palomino, Sergio Enrique Moreno Elescano, Andre Jesus Cruces Chancahuaña, Angel Eduardo Dianderas Jorge, Pablo Cardenas Caceres Thursday morning Track C, Thursday, Jun 13 2024, 10:30-12:00 Location: Lecture room D Session: Clinical engineering 121 Chairs: Alessio Luschi and Rok Hren 10:30 Assessing Perfusion Changes in Clinical Oncology Applications using Hy- 121 OR-172 perspectral Imaging Rok Hren, Jošt Stergar, Urban Simončič, Gregor Serša, Matija Milanič 10:45 OHIO: Integrating IoT Technologies for Enhanced Clinical Engineering and 121 OR-173 Dynamic Tracking of Medical Equipment Alessio Luschi, Gianpaolo Ghisalberti, Giovanni Luca Daino, Vincenzo Mezzatesta, Ernesto Iadanza 11:00 An Innovative Solution for Efficient Workflow Management in Healthcare 122 OR-174 Alessio Luschi, Ernesto Iadanza 11:15 Towards the Creation of a National Medical Equipment Inventory 122 OR-175 Aris Dermitzakis, Spilios Zisimopoulos, Nicolas Pallikarakis 23 11:30 Implementing a Centralized Medical Equipment Management System on 123 OR-176 a National level Aris Dermitzakis, Spilios Zisimopoulos, Nicolas Pallikarakis 11:45 Calibration of medical equipment - why is it critical? 123 OR-177 Symon Mbakah 24 Poster Presentations Page Tuesday morning poster session, Tuesday, Jun 11 2024, 10:00-10:30 Location: Coffee break and poster display area Session: Poster session 127 Info: The same posters will be presented also during the afternoon coffee break, i.e. this Poster session will continue during the Tuesday afternoon coffee break. Odd-numbered posters (PO-01, PO-03, ...) comprise the Tuesday poster session and presenters are expected to be present at their posters during the two coffee breaks. PO-01 Neural network based fetal ECG extraction from abdominal signals 127 Dragoș-Daniel Țarălungă, Radu Botezatu, Alina-Elena Sultana, Titus Mihai Vasile, Georgeta-Mihaela Neagu PO-03 Fundamental study of Readiness Potential elicited by foot movement 127 Reon Takahashi, Puwadej Leelasiri, Tomoya Oi, Tatsuhiro Kimura, Hiroshi Ohshima, Kiyoyuki Yamazaki, Fumitaka Aki PO-05 Automatic detection of sympathovagal response using HRV analysis. 127 Case study: resident surgeons during training and their first laparoscopic surgery Maria Elisabetta Pagnano, Jacopo Vitale, Margherita Anna Grazia Matarrese, Gianluca Mascianà, Marco Caricato, Leandro Pecchia PO-07 Preliminary development and evaluation of a low-cost digital stetho- 128 scope Muyanga Kampekete, Francis Chikweto, Foster Munsanje, Boyd Mubanga, Conix Sunday, Jimmy Sikwese PO-09 Trade-Off Between Real-Time and Classification Performance in Motor 128 Imagery BCI Aleksandar Miladinovic, Milos Ajcevic, Katerina Iscra, Francesco Bassi, Alessandra Raffini, Joanna Jarmolowska, Uros Marusic, Agostino Accardo PO-11 Evaluation of Precursors to Momentary Sleepiness in Automobile Driving 129 Using Cerebral Blood Flow Variability and Thermography Tomoya Oi, Reon Takahashi, Puwadej Leelasiri, Kiyoyuki Yamazaki, Hiroshi Ohshima, Fumitaka Aki PO-13 Use of ECG gating approach for high frequency ultrasound vector flow 129 imaging Chi-Hung Yang, Chih-Chung Huang PO-15 Intraoral scanner classification and accuracy evaluation using UV map- 130 ping of scan bodies Mykolas Akulauskas, Darius Jegelevičius, Vygandas Rutkūnas, Liudas Auškalnis, Justinas Pletkus PO-17 Multi-input CNN based Classification of EEG and NIRS signal during vol- 130 untary hand movement Puwadej Leelasiri, Reon Takahashi, Fumitaka Aki, Hiroshi Ohshima, Kiyoyuki Yamazaki 25 PO-19 Combining quantitative MRI, fluorescence, and neuropathology in 131 frameless brain tumor needle biopsies Elisabeth Klint, Anders Tisell, Ida Blystad, Martin Hallbeck, Teresa Nordin, Johan Richter, Karin Wårdell PO-21 Indoor Natural Ventilation Assessment in Healthcare Facilities in Low- 132 Resource Setting Nahimiya Husen Ibrahim, Vincenzo Piemonte, Leandro Pecchia PO-23 Towards Equity in Healthcare: Designing an Affordable Spirometer for 132 Low-Resource Settings Pedro Checa Rifa, James Wallace, Davide Piaggio PO-25 Efficacy of a virtual reality–based video game intervention for individuals 133 with schizophrenia spectrum disorder Jen-Suh Chern, Yu Yang, Yi-Ching Tsai PO-27 Range of Motion and Morphological Characteristics of the Sacroiliac 133 Joint Seonjin Shin, Dai-Soon Kwak PO-29 Effect of friction coefficient on subsidence of cemented polished stem in 134 THA Noriyuki Takano, Yuuta Nakajyou, Ayumi Kaneuji PO-31 Effect of Tissue on Stress-Strain Characteristics of Pulmonary Pleura 134 Hitomi Sakai, Hirosane Hayashi, Sumiko Maeda, Noriyuki Takano PO-33 Evaluation of the biocompatibility of interface-free polymer-ceramic fu- 135 sion technology using laser-driven penetration synthesis Minseong Chae, Kang-Sik Lee, Yu-Chan Kim, Hojeong Jeon, Seung-Hoon Um, Hwachul Jung, Dongkyu Koo, Dong-Ho Lee PO-35 Dual growth-factors delivery (VEGF and BMP-2) with alendronate im- 135 prove cell adhesion in osteoinductive bone repair Amira Raudhah Abdullah, Mohd Riduan Mohamad, Ka Liong Tan, Nadia Mohd Effendy, Nur Nabilah Abu Bakar, Intan Maslina Musa PO-37 Cardiac Hypertrophy Simulations Using Echocardiography-based LV 136 Model Bogdan Milicevic, Miljan Milosevic, Milos Kojic, Nenad Filipovic PO-39 Evaluation of brain activity using NIRS to examine the antecedents of mi- 136 crosleep Fumitaka Aki, Tatuhiro Kimura, Hiroshi Ohshima, Kiyoyuki Yamazaki PO-41 An IoT-based Medication Calendar for Home-Visiting Nursing 137 Yasushi Yamauchi, Junto Hosono PO-43 Investigation of the use of 3D modeling and predictive analysis for optim- 137 izing the management of the Miljacka River Madžida Hundur, Merima Smajlhodžić-Deljo, Faruk Bećirović, Naida Babić Jordamović, Lejla Gurbeta Pokvić PO-45 The Development of Medical Imaging in China and Its Impact on the Clin- 138 ical Practice Kui Cai, Guobin Li, Junan Zheng, Lei Chen, Qing Mao, Ruixin Jiang, Shuheng Zhang, Taomin Zhang, Wenjing Ma, Xiaoyue Gu, Ying Qian 26 PO-47 An explainable XGBoost model to predict pediatric sleep apnea resolution 138 after treatment from new phenotypic information Gonzalo C. Gutiérrez-Tobal, Javier Gomez-Pilar, Daniela Ferreira-Santos, Pedro Pereira-Rodrigues, Daniel Álvarez, Félix del Campo, David Gozal, Roberto Hornero PO-49 Analysis of age-related variations in photoplethysmography: a Machine 139 Learning approach Sara Lombardi, Federica Tavernise, Piergiorgio Francia, Leonardo Bocchi PO-51 Influence of fatigue in swimmers suffering from swimmer shoulder pain 139 Alessandra Raffini, Alex Buoite Stella, Miriam Martini, Laura Mazzari, Agostino Accardo PO-53 Semi-Automated Approach to Analyze the Tapering of Coronary Ves- 140 sels from Contrast-Enhanced Micro-CT Scans of Human Coronary Artery Anatomy Amanda DeVos, Paul A. Iaizzo PO-55 Reliability of the SPIMON, a portable 3-D spinal range of movement 141 measurement prototype device Aris Dermitzakis, Athanasios Lampropoulos, Ioannis Kalamaras, Emil Valchinov, Vasileios Syrimpeis, Dimitra-Tania Papanikolaou, Evdokia Billis PO-57 Surface electromyographic (sEMG) biofeedback device as an outcome 141 measure for an individualized pelvic floor muscle (PFM) training program for urinary incontinent women: preliminary results Dimitra-Tania Papanikolaou, Aris Dermitzakis, Emil Valchinov, Konstantinos Giannitsas, Evangelos Dermatas, Epaminondas Mitronikas, Evdokia Billis PO-59 Tempo Certo Project - Adapting EEG systems to perform Multifocal Visual 142 Evoked Potentials for clinical care and research Igor Telles, Arthur Conceição, Paulo Galhanone, Mauricio Cagy, Dimitri Abramov, Antonio Miranda de Sa PO-61 Microstates analysis for dry and gel-based multichannel electroenceph- 142 alography Hannes Oppermann, Patrique Fiedler, Jens Haueisen PO-63 The potential use of electroporation in combination with chemothera- 143 peutics or calcium chloride as a therapeutic option for uveal melanoma patients: first in vitro and in vivo results Miltiadis Fiorentzis PO-65 Acoustic pressure analysis due to high-frequency electric field on the bio- 143 logical cell during reversible electroporation Mayank Kumar, Rochish Thaokar PO-67 Texture Analysis of H-scan Ultrasound Images for the Characterization of 144 Breast Tumors Zhanjie Zhang, Sio Hang Pun, Peng Un Mak, Hung Chun Li, Kung Jui Hou, Mang I. Vai PO-69 Prognosis and Uncertainty Management for Filament Degradation of X- 144 ray CT Tubes with IoMT time series data Jie Zhong, Heng Zhang, Qiang Miao, Qilin Liu, Jin Huang 27 Wednesday morning poster session, Wednesday, Jun 12 2024, 10:00-10:30 Location: Coffee break and poster display area Session: Poster session 144 Info: The same posters will be presented also during the afternoon coffee break, i.e. this Poster session will continue during the Wednesday afternoon coffee break. Even-numbered posters (PO-02, PO-04, ...) comprise the Wednesday poster session and presenters are expected to be present at their posters during the two coffee breaks. PO-02 BIM operations and clinical engineering in the training of the biomedical 144 engineer Luis Vilcahuaman PO-04 Baroreflex function in young patients with obesity: analysis of the cardiac 145 chronotropic and vascular resistance baroreflex arms Jana Čerňanová Krohová, Barbora Czippelova, Zuzana Turianiková, Miriam Kuricova, Dana Cernochova, Luca Faes, Michal Javorka PO-06 Artificial Intelligence Advancements in Fetal Monitoring: Enhancing 145 Prenatal Care Dragoș-Daniel Țarălungă, Ionut Manea, Rares-Marin Preoteasa, Bogdan Cristian Florea, Georgeta-Mihaela Neagu PO-08 Using localized electroporation for transfecting cardiac cells 146 Jernej Jurič, Vid Jan, Tina Turk, Lea Rems PO-10 The Study of Non-Invasive Blood Information Measurement and Monitor- 146 ing Method via Wearable AWPPG Device Chung ChiFu, Lu PoWen, Cheng Chun Chang PO-12 Development of Neurorehabilitation Bicycle with Posture Control Detec- 146 tion Ya-Ju Chang, Hsiao-Lung Chan, Jr-kai Yu, Miao-Ju Hsu, Rou-Shayn Chen PO-14 Mapping side effects for deep brain stimulation in essential tremor 147 Teresa Nordin, Erik Österlund, Dorian Vogel, Anders Fytagoridis, Simone Hemm, Karin Wårdell PO-16 Exploring the impact of data and statistical methods for defining sweet 148 spots in Deep Brain Stimulation Vittoria Bucciarelli, Dorian Vogel, Teresa Nordin, Karin Wårdell, Jérôme Coste, Jean-Jacques Lemaire, Raphael Guzman, Simone Hemm PO-18 REDCap and SQLite: a powerful combination for streamlining metadata 148 capture in Deep Brain Stimulation research Marc Stawiski, Vittoria Bucciarelli, Dorian Vogel, Jérôme Coste, Jean-Jacques Lemaire, Siegward Elsas, Ute Gschwandtner, Peter Fuhr, Ethan Taub, Simone Hemm PO-20 Numerical Study of Hemodynamic Characteristics on the Surgery of Pul- 149 monary Artery Banding Jinlong Liu, Weiru Luo, Jiwen Xiong, Yanjun Sun PO-22 Monte Carlo Simulated Photoplethysmography Signals for the Validation 150 of an In Vitro Wrist Phantom Raghda Al-Halawani, Meha Qassem, James May, Panicos A. Kyriacou 28 PO-24 Identification of Barriers in the Adoption and Promotion of Advanced En- 150 doscopic Systems and Minimally Invasive Surgical Equipment in China: Evidence from a Survey of 320 Healthcare Facilities Zhichen Wang, Fei Sun, Qian Zhang, Jing Sun, Jingyi Feng PO-26 Co-transfection of anti-fibrotic microRNAs as a treatment for oral sub- 150 mucous fibrosis Kai-Chiang Yang, Chih-Yuan Fang, Hsing-Yu Chen, Yung-Hsin Cheng PO-28 Analysis of Open-Source Softwares for Ultrasound Tomography Based on 151 Full-Waveform Inversion Lucas da Costa, Nilton Assugeni Neto, João Henrique Uliana, Théo Pavan, Antonio Adilton Carneiro PO-30 Improving diagnostic performance of an automated melanoma dia- 151 gnostic system using fast style transfer data augmentation Takashi Nagaoka, Mitsutaka Nemoto, Yuichi Kimura PO-32 Estimating stride length through a deep learning model utilizing foot in- 152 ertial data in individuals with Parkinson’s disease Hsiao-Lung Chan, Ya-Ju Chang, Ai-Tung Hsieh, Rou-Shayn Chen PO-34 A Convolutional Neural Network (CNN) based Electrocardiogram (ECG) 152 Pattern Analysis under Cycling Movement with the Applications into the Development of an Exhaustion Stratification System Szi-Wen Chen, Yu-Hsuan Huang, Rou-Shayn Chen PO-36 The Prediction of Sleep Quality using Heart Rate Varia-bility Modulations 153 during Wakefulness Andrea Di Credico, David Perpetuini, Pascal Izzicupo, Giulia Gaggi, Nicola Mammarella, Alberto Di Domenico, Rocco Palumbo, Pasquale La Malva, Daniela Cardone, Arcangelo Merla, Barbara Ghinassi, Angela Di Baldassarre PO-38 Comparative Analysis of Depression Detection using EEG Signals 153 Stefana Duta, Georgeta-Mihaela Neagu, Alina-Elena Sultana PO-40 Unveiling Age-Related Patterns in Vocal Expression of Emotions: A Ma- 154 chine Learning Approach with Mel and Gammatone Frequency Cepstral Coefficients Michele Giuseppe Di Cesare, David Perpetuini, Daniela Cardone, Arcangelo Merla PO-42 A Machine Learning Framework for Gait and EMG Analysis for Post-Stroke 154 Motor Dysfunctions Assessment Francesco Romano, David Perpetuini, Daniela Cardone, Arcangelo Merla PO-44 Comet Assay in the Digital Era: A Review of the use of Artificial Intelli- 155 gence for the Analysis of DNA damage based on the results of the Comet Assay Adna Softić, Nejra Merdović, Velid Dlakić, Emina Mrđanović, Lejla Mahmutović, Daria Ler, Lejla Gurbeta Pokvić PO-46 Advancing Pediatric Cardiology Training with a VR-Based TEE and 155 Amplatzer Deployment Simulator David Buyck, Enrique Vergara-Escudero, Susana Arango, Varun Aggarwal, Benjamin Gorbaty, Paul A. Iaizzo PO-48 Automatic Prediction of Pediatric Postoperative Behavioral Disturbance 156 from Neuronal Activity During Anesthesia Javier Gomez-Pilar, Laura Gutiérrez-de Pablo, Carlota Gordaliza, Estefanía Gómez-Pesquera, Jesús Poza, Eduardo Tamayo, Roberto Hornero 29 PO-50 Investigation of the Use of Gait Analysis for Prosthetic Patients: Evaluat- 156 ing the Relationship Between Gait Outcome Measures and Function (K- level) Amy Edwards, Thomas Stone PO-52 Influence of the Subcutaneous Fat Layer Thickness on the Activation of 157 the Phrenic Nerve via Non-Invasive Electrical Stimulation Laureen Wegert, Alexander Hunold, Marek Ziolkowski, Irene Lange, Tim Kalla, Jens Haueisen PO-54 Use of electrical household appliances and risk of all types of tumours: A 158 case-control study Muhammad Waseem Khan, Shabana Noori PO-56 How Many is Enough? The Influence of patient count on structural norm- 158 ative template quality Dorian Vogel, Jérôme Coste, Jean-Jacques Lemaire, Simone Hemm PO-58 Assessing stress level by utilizing heart rate variability and galvanic skin 159 response features Sandro Radan, Jana Bratanić, Ana Sušac PO-60 Sensitivity Analysis of Reconstruction Fiber Tracking Parameters from 159 Diffusion Tensor Magnetic Resonance Imaging of Reanimated Ex Vivo Swine Hearts Neal Duong, Paul A. Iaizzo PO-62 Insights into Magnetomotive Ultrasound: Evaluating Key Parameters for 160 Enhanced Imaging Ariane Sanches, Nicholas Zufelato, David Collazos-Burbano, João Henrique Uliana, Théo Pavan, Antonio Adilton Carneiro PO-64 Unveiling the phenolics of chestnut honey-based propolis and in silico 160 phytoestrogen activity Idris Arslan PO-66 Enhancing Cross-Domain Adaptability of Existing Computer-Aided En- 160 doscopic Lesion Detection Using Plug-and-Play Tracker Yijie Ku, Hui Ding, Guangzhi Wang PO-68 Design and Analysis of a Biomedical Orthosis for Clubfoot Corrective 161 Device Shahrol Mohamaddan PO-70 A complex spinal surgery lifting system for prone positioning 161 Lutong Li, Stuart Watson, Glyn Smurthwaite, John Large, Andrew Weightman, Glen Cooper P L E N A R Y L E C T U R E S ’ A B S T R A C T S Plenary talks session clinicians. AI tools aid radiologists in inter- preting complex fractures, and tumors from Monday morning Plenary CT, MRI, PET-CT, and 3-D ultrasound scans. ML is applied for analyzing imaging data Jun 10, 9:15 - 10:00 with biomarker results for early detection of malignancies reducing human errors and swift PL-1 start of treating cancer patients. New ablation method based on irreversible Applications of AI and ML have been instru-electroporation – a game changer in car- mental in effective and personalized therapy diac ablation to patients based on a comprehensive ana- Atul Verma lysis of historical and current data, genetic McGill University, Canada information, and responses to therapy in sim- Abstract was not received. ilar clusters. AI applications in robotic surgeries minimize invasiveness, improve precision, and reduce complications and recovery times. Plenary talks session AI and ML-powered methods perform intel- ligent searches of chemical and biological Monday afternoon Plenary databases, determine potential drugs, and Jun 10, 13:30 - 14:15 predict their safety and efficacy precisely, thus reducing R&D and product launch costs. PL-2 AI and ML-based technologies support re- Cardiac regeneration: where we stand mote consultations, diagnosis, monitoring, Mauro Giacca and management of patients with con- King’s College London, United Kingdom strained access to timely healthcare services. Applications of AI have resulted in the im- Abstract was not received. provement of overall hospital operations including reduced readmissions, efficient Plenary talks session staffing and scheduling, resource optimization with predictive analytics, cost reduction in Tuesday morning Plenary equipment, maintenance and upgrades, and enhanced safety, and compliance. Jun 11, 8:30 - 10:00 Selected examples of AI applications leading to the improvement in diagnosis and therapy PL-3 associated with cardiovascular, cancer, dia- Applications of Artificial Intelligence in Clin- betes, and neurological diseases and hospital ical Diagnosis, Therapy and Operations operations will be presented. Shankar Krishnan Impressive benefits can be realized by stra- Past-President, IFMBE, United States tegic applications of AI. However, consequent There have always been continuous ef- challenges in AI implementation related to forts to enhance the quality of healthcare. transparency, data sharing, bias, confiden-Recently, applications of Artificial Intelligence tiality, privacy, and security of patient data (AI) have demonstrated promising roles in and ethical and trust aspects will have to be transforming accurate, efficient, and ac- alleviated. Regulatory agencies such as the cessible clinical diagnosis and treatment in FDA are approving AI-based medical devices addition to e improving hospital operations. but proceeding with caution. The objective of this talk is to highlight major In conclusion, the applications of technologies applications of AI in the healthcare domain and processes based on AI and ML predict in synchrony with the mission of biomedical and promise multifaceted enhancements engineers worldwide. in the quality of clinical diagnosis, therapy, By analyzing vast amounts of patient health and operations resulting in better healthcare data including physiological, pathological, and positive impacts on patient care glob- and imaging results, and genomic informa- ally, blending with the mission of the BME tion, AI techniques have increased the overall community. diagnostic efficiency. Machine Learning (ML) algorithms can identify abnormalities and suspicious patterns that may be missed by 33 PL-4 possible. We are developing an advanced Cybersecurity for connected medical form of the technology, high frequency irre-devices versible electroporation (HFIRE), for the treat- Dubravka Maljević ment of cardiac disease, cancer, and other BG Kliniken, Germany malignancies. This new therapy preferentially Abstract was not received. targets cancer cells over healthy cells, pro- motes a positive immune response, and al- leviates the need for a paralytic. Further- Plenary talks session more, with regards to cancer therapy, these pulses have been shown to increase antigen Tuesday afternoon Plenary activity with the release of damage associ- Jun 11, 13:30 - 14:15 ated molecular pathways (DAMPs) and pro- teins. We have demonstrated that fine-tuning the pulsing parameters applied can influence PL-5 The Joy of Point-Based Visualizations biophysical effects as well as modulate thera- peutic outcomes. By modifying the pulse Blaž Zupan parameters and field strength, the activa- University of Ljubljana, Slovenia tion of specific modes of cell death mechan- Two-dimensional embeddings resulting isms can allow for controlling the extent of an from dimensionality reduction techniques electroporation-mediated immune response. such as t-SNE, UMAP, and multidimensional Other critical considerations include modi- scaling are widely used in various scientific fying burst number to balance electropora-fields to visualize high-dimensional data in tion saturation against escalating thermal ef- two-dimensional, point-based visualizations. fects, optimizing therapeutic outcomes. Our In this talk, I will introduce some common group is also developing Electro Antibacterial projection and embedding techniques and Therapy (EAT), an advanced electroporation- show through various examples why I find based technology which targets intratumoral them useful, or rather exciting, for analyzing bacteria in pancreatic cancer by boosting tabular, textual, or image-based data. I will antibiotic uptake in malignant cells harbor- also discuss how AI and other computational ing cancer-promoting bacteria like Fusobac-techniques can help us explain the resulting terium nucleatum. EAT demonstrates effect-visualizations. iveness at clearing out intracellular bacteria, even at the reversible regime, offering prom- Plenary talks session ising therapeutic advantages. Our preclin- ical work focuses on helping canine patients with spontaneous disease, which are excellent Wednesday morning translational models for human disease. Res- Plenary ults of our ongoing trials have been extremely Jun 12, 8:30 - 10:00 positive, supporting that HFIRE is effective for the treatment of cancer, including tumors re- fractory to surgery, radio- and chemother- PL-6 History of irreversible electroporation and apies. This presentation will also discuss recent future clinical applications advances in pulsed electric field therapies to Rafael V. Davalos target various aspects of the tumor microen- Georgia Tech, United States vironment and potential future clinical applic- ations. Irreversible Electroporation (IRE) is a min- imally invasive surgical therapy we invented PL-7 to treat unresectable tumors using low-energy Deep Brain Stimulation - Data Analysis for microsecond pulses. IRE is unique among tis- Clinical Support sue ablation techniques in its mechanism of Karin Wårdell 1, Teresa Nordin1, Dorian Vogel2, only targeting the cell membrane while tis- Simone Hemm2 sue molecules, everything encompassing col- 1Linköping University, Sweden lagen structures to proteins, remain intact; 2University of Applied Sciences and Arts Northwest-thereby making treatment near critical struc- ern Switzerland, Switzerland tures such as major blood vessels and nerves Deep brain stimulation (DBS) is an import- 34 ant therapy for movement disorders such as Front. Neurosci. 16:834026. doi: 10.3389/fn-Parkinson’s disease (PD), essential tremor, and ins.2022.834026 dystonia. The clinical indications for DBS are continuously increasing and are expanding to- Plenary talks session wards psychiatric illness. Following MRI and surgical planning of trajectory and target, a thin DBS lead is implanted in a deep brain Thursday morning Plenary structure using stereotactic neurosurgery. The Jun 13, 8:30 - 10:00 choice of brain target is dependent on the pa- tients’ symptoms, for example the subthalamic PL-8 nucleus is commonly used in PD. Over the last Transferring scientific findings into medical decade, several new DBS-lead designs and devices: the power of simplicity systems have reached the market. This in- Zlatko Matjačić cludes voltage and current stimulation modes, University rehabilitation institute, Slovenia and new configurations of the electrode con- A rehabilitation robot is a device that en- tacts including split contacts for steering of ables physically disabled people to train, re-the stimulation field. Other advances are learn and perform movement. Neurorobot-battery life-length extension, support in post- ics has a special place in rehabilitation robot-operative programming, and DBS systems de- ics, the goal of which is to increase the intens- signed for adaptive stimulation i.e. combined ity of therapy. Wearable exoskeletons, end- recording of local field potentials (LFP) and effector robots that are in contact only with stimulation in a closed-loop. All improvements the walking person’s feet, and cooperative ro-generate more data and increase the demand bots are used for training of walking. Rehab- of technical aids to summarize them and to go ilitation robots generally help physically dis- from the traditional clinician’s “mental imagin- abled people to perform the desired move- ation” to “intuitive visualization” in the defin- ment, but they may only provide assistance ition of target, implantation procedure, and to the minimum necessary extent to maxim- follow-up to gain the best clinical outcome. ize volitional engagement of the users. Re- This presentation will give an overview of the habilitation robots must exhibit soft haptic in- DBS system and supportive techniques for teraction in contact with the patient, which surgical planning, implantation, and patient means that the robotic mechanism must have follow-up. The focus will be set on visualization a high mechanical compliance. Through feed-and our method for electric field (EF) simula- back and virtual reality environments, mod- tions for investigation of the stimulated tissue. ern rehabilitation robots motivate patients to Examples of patient-specific modeling and exercise more intensively, which increases the simulations using the finite element method therapeutic effects of exercise. Rehabilitation (FEM, Comsol Multiphysics AB, Sweden) will be robots must instill confidence in both patients presented, and how these can be further ex- and therapists in the reliability of their opera-plored for probabilistic group analysis to op- tion, and at the same time they must relieve timize the treatment with minimized side ef- the therapist of heavy physical burden and en-fects. Further development of the workflow able the measurement and objective evalu-allows for visualization of traced thin white ation of the effects of training. matter fibers and outlined brain structures to- In the lecture, we will first provide an over-gether with the patient’s MRI and EF. Examples view of selected rehabilitation robots that are of simulation and visualization methods will be used in clinical practice for training and assist- presented. To make the tools available for ance of walking. Even though these rehab-the research community, we have created the ilitation robots have played instrumental role open-access apps: ELMA and DBSim for per- in acceptance of this kind of technology in forming patient-specific simulations. As a next clinical practice they have certain shortcom-step DBviS, an app for visualisation and com- ings, mainly reflected in that they physically parison of EF simulation studies will be intro- over constrain users, thus disabling practicing duced. of propulsion and dynamic balance, which are Reference: Wårdell K, Nordin T, Vogel D, Zsig- crucial for efficient and safe bipedal walk- mond P, Westin C-F, Hariz M and Hemm S ing. A significant factor that impedes the (2022) Deep Brain Stimulation: Emerging Tools wide-spread use of these devices is their cost for Simulation, Data Analysis, and Visualization. that may easily reach up to half a million of 35 euros. Primarily, the reason for such expensive To address these challenges, it is necessary to devices is that they were engineering-driven, rethink the organization of health services for thus having many actuated degrees of free- the well-being of people and the communities dom. However, patients-needs-driven solu-in which they live. To be effective, BME need to tions may be significantly less complex en- work side-by-side with experts of sustainability abling easier fulfillment of the regulatory re- and strength their collaborations with experts quirements. We will present the evolution of of global health, which are the once in charge changes in the rehabilitation robot for assess- for the reorganization of healthcare services ment and training dynamic balance during in most countries. In fact, the conscious use walking from a complex and expensive haptic of enabling and transformative technologies device suitable only for research purposes to can provide safe, effective, and efficient solu-a simple modular system suitable for industrial tions, while simultaneously generating new realization and use in a clinical environment. knowledge and new opportunities for sustain-The lecture will be concluded with presenta- able economic development, in line with the tion of a Balance Trainer, a passive mechanical SDGs. However, the adoption of innovative device intended for balance training during technologies is effective and sustainable, as standing in neurologically impaired individu- well as safe and efficient, if and only if it is als that has been successfully on the market accompanied by an actual reorganization of of medical devices for more than twenty years services. Our community is requested to put and has been sold in excess of 6000 units. global effort in finding innovative and sustain- able technological solutions, which may help PL-9 transforming healthcare systems, following Biomedical Engineering for Sustainable the United Nations Sustainable Development Global Health Goals 2030 (SDG2030) and in line with the 5As Leandro Pecchia principles: affordability, availability, accessib- Campus Bio Medico University of Rome, Italy ility, accommodation, and acceptability. Health systems are facing increasingly In this talk, Professor Pecchia will give an over-complex local (e.g. demographic change, view of the work done in the last few years, population ageing, sandwich generation) as WHO Innovation Manger during COVID and global challenges (e.g. international and in large EU projects (such as GATEKEEPER, health emergencies, antimicrobial resistance, ODIN and EPoCA) in order to bridge research pandemics, climate change, wars, global and global health ecosystems across different warming), which led the WHO to introduce phases of medical devices lifecycle in which among the priorities of the Strategic Plan BME are involved: design, manufacturing, clin- 2025-2028 the development of climate- ical and pre-clinical validation, regulations, resilient health systems and lowering carbon assessment and management. footprint of health systems and societies. There are several factors inspiring this shift in WHO priorities: in May, mosquitoes po- tential carriers of vector of Zika and Malaria were found respectively in Scotland and Italy; European healthcare systems are responsible for the 5% of Co2 emissions, which is remark- able (e.g., transports account “only” for 15%). As a proxy for demographic changes, a report from a world leading medical device manu- facturer estimated that by 2040 one fourth of the European population should work for the healthcare system, to balance the growing request of care due to aging population. This urges the provision of innovative technologies that can protect healthcare workers (HCW) and free their time from repetitive routinary activities, leveraging on key enabling techno- logies such as robots, AI, and IoT. 36 O R A L P R E S E N T A T I O N S ’ A B S T R A C T S Technologies for assistive and focuses on the introduction of a small social preventive healthcare robot into care environment. It can be used both in care facilities and in people’s homes. Monday morning Track A The main goal is to develop an affordable care system that combines smart IoT in the Jun 10, 10:30 - 12:15 home and a robotic toy (PetBot) as a person- alised, friendly and intelligent butler. In the OR-001 context of the project, the PetBot is an assist- Digitalization of elderly care delivery ive companion with autonomous intelligence, Lenka Lhotska, Petr Novak, Jaroslav Cibulka, which should understand its ”master’s” com- Vitezslav Mergl mands and understand how he/she feels. The Czech Technical University in Prague, Czech Repub- robot should react adequately to the current lic situation and, if necessary, offer an adequate Society is aging fast and taking care of the and tailored intervention. Thus, the PetBot elderly is becoming a challenge. The contri- must know the latest activities in the home and bution describes ongoing research work star- collect and interpret information from various ted March 2023. Partners from 7 European sensors. countries (Austria, Czechia, Germany, Italy, The work is supported by the project Poland, Slovakia, and Slovenia) identified that No.CE0100038 DigiCare4CE of INTERREG the digital solutions that may offer support Central Europe programme. to nursing staff are still underused. The pro- ject focuses on innovative solutions that im- OR-002 prove care quality and develops a transna- Design and development of an optical pro-tional strategy for digital transformation of totype system: Preliminary results care facilities. Two pilot actions are defined Teodor Minev, Kristina Bliznakova, Nikolay Dukov, Zhivko Bliznakov for testing deployment of new technologies: Medical University of Varna, Bulgaria digital transformation of care management and delivery; datafication of elderly care de- The aim of this work is to develop and test livery, based on environmental, wearable and a complete platform for an optical im-aging IoT solutions. Our team focuses on the lat- system constructed in transmission geometry ter pilot action in collaboration with Geron- for both in-vivo and in-vitro imaging. The tology Centre Prague where the pilot solution imaging system comprises of a visible light is tested. The solution consists of three sub- source, a CMOS cam-era, and an integrated systems that can be used either separately, or hardware platform, designed for both rota- as an interconnected system. tional and linear movements, accommodat- The first part represents tasks to be used to ing the placement of the imaged object. A train and test cognitive abilities and manual comprehen-sive software framework is de- dexterity. The aim is to design tasks that are signed and implemented to govern the lin-easy to implement and applicable in different ear and rotational motions of the platform. environments (long-term care facilities, day The developed platform was tested with three care centres, home environment). Data can samples to obtain optical images, which are be stored on a smartphone or computer. Data shown together with the images from an op-can be used by both the client and the care tical microscope. Results showed that the pro-staff to provide information about the client’s totype system is capable of obtaining planar condition (e.g. stable condition, tendency to images within a timeframe of less than 40 deteriorate). seconds. Future work will focus on optimizing The second part represents devices and image acquisition speed, conducting extens- control options that can be used for a ”smart ive validation, and expanding the system into home” solution, but with the requirement for three dimensions. This system is intended to very simple and intuitive operation and the use complement the innovative prototype X-ray of simple tools, and reasonable (lower) finan- system currently in development by the ELPIDA cial requirement. The acquired data can be group at the Medical University of Varna. used to analyse user behaviour patterns over the long term and thus reveal any degradation of capabilities. The final part of the proposed pilot study 39 OR-003 Maintaining mobility is crucial, for prevent- Screening prediabetes using 1 hour glucose. ing frailty and sustaining independence in A simulation model to estimate the lifetime older adults. Mobility requires both agility and health and economic outcomes balance during everyday activities. Maintain- Martina Andellini, Melania Manco, Matteo Ritro- ing safe mobility without increasing the risk vato of falls is of concern in almost 50% of older Bambino Gesù Children’s Hospital, Italy adults. As such, there is increasing interest in Background: The current diagnosis method measuring and monitoring mobility and bal-for diabetes patients is based on the 2-h ance in a more dynamic way. Innovation in di-plasma glucose (2-hPG) value during a 75- gital technologies now provides a means of to g OGTT. However, recent robust evidence monitor dynamic balance in community set-demonstrates that 1-hour post-load plasma tings, generating the need for new definitions glucose (1-hPG) ≥ 155 mg/dl in those with nor- of dynamic and static balance during daily mal glucose tolerance during the 2-hPG is activities. To achieve a working definition we highly predictive for T2D incidence. The aim carried out 1) a literature review 2) provisional of this work is to carry out a health eco- data collection 3) a consensus working group nomic analysis to estimate long-term cost- in a sit-to-stand activity, The literature search effectiveness of using 1-hPG compared to 2- analyzed 160 papers to define static balance hPG for screening and for assessing the risk of as the relative positioning of the center of diabetes over 35 years. mass (CoM) to the base of support (BoS) from Methods: A Monte Carlo–based Markov sim- multiple definitions of static balance. The lit- ulation model was developed to simulate the erature highlighted that both static and dy- long-term (35 years) effects of the two screen- namic balance is fundamentally linked to CoM ing strategies, in terms of clinical and cost- positioning relative to BoS, with dynamic bal-effectiveness outcomes. Transition probabil- ance incorporating a velocity component for ities on disease progression, mortality, effects CoM during movement. A 2D analysis of static on preventive treatments and complications and dynamic balance during the sit-to-stand were retrieved from the major diabetes stud- activity was conducted on a healthy individual ies. Direct medical costs were sourced from which split the movement into 4 phases (i) published literature and inflated to 2019 Euros. torso leaning (ii) bottom lifting (iii) hip exten- Results: In the lifetime analysis, 1-hPG was pro- sion and (iv) standing. A semi-quantified dy-jected to increase the number of years free namic balance analysis using the CoM position from disease; to delay the onset of T2D; to relative to the base of support, D, and CoM reduce the incidence of T2D complications velocity, V, was conducted. The consensus and to increase the QALY gained. Even if 1- working group observed that a function which hPG diagnosis method resulted in higher ini- minimizes the vector sum total involving these tial costs for preventive treatments, long-term parameters may be crucial to maintaining dy-diabetes related costs and diabetes-related namic balance during this activity of daily liv-complications costs were reduced. The incre- ing. mental cost-effectiveness ratio was -8,214.7€ per each QALY gained. OR-005 Conclusions: Notwithstanding the higher ini- Design of a Surgical Device for Performing tial costs of screening prediabetes by using 1h- Percutaneous Dilatational Tracheostomy PG due to the incremental preventive treat- Yuan Tang 1, Glen Cooper1, Bruno Adorno1, Brendan McGrath2, Andrew Weightman1 ments, it is feasible and cost-effective resulting 1School of Engineering, University of Manchester, in quality of life gained (QALYs) and reduced United Kingdom overall costs. 2Manchester University Foundation Trust Wythen- OR-004 shawe Hospital Intensive Care Unit, United Kingdom Quantifying static and dynamic balance for Percutaneous Dilatational Tracheostomy the sit to stand movement assessment (PDT) is a routine bedside procedure per- Athia Haron1, Glen Cooper 1, Helen Dawes2, formed on patients managed in Intensive Care Maedeh Mansoubi2, Andrew Weightman1 Units (ICUs). Severe perioperative complica- 1University of Manchester, United Kingdom tions exist, and some, such as major bleeding 2University of Exeter, United Kingdom and tracheal wall perforation, may be caused by operator errors when using PDT instruments, 40 especially inserting and removing the needle tables. Both the therapist and the patient can and dilators. We designed a new device called continuously monitor the results of the training. TrachyPen for performing PDT puncture and The effect of the prehabilitation intervention dilation to reduce the complications associ- was assessed using objective functional tests: ated with current PDT instruments and improve spirometry, Six Minute Walk Test, Borg Rating the procedure. The design followed the re- Of Perceived Exertion, Five Times Sit to Stand quirements developed from a questionnaire test, Hand Grip, ICF - International Classifica-to twenty-one healthcare professionals, which tion of Functioning, Disability and Health. Sub- mapped the mechanical design to the clin- jective tools (Numerical Rating Scale, WHO- ical needs. Either used manually or powered DAS 2.0-WHO Disability Assessment Schedule by a motor, the TrachyPen is designed to com-questionnaire) were also used. bine the PDT puncture, guidewire insertion, and In a pilot study, we tested the applicability dilation into a single step, eliminating the use of the system in patients before urogynecolo- of multiple instruments. Five healthcare pro- gical surgery. Interim evaluations showed sig-fessionals evaluated the conceptual design nificant improvements in most functional tests. through a demo video and an online ques- The use of the telemedicine system contrib-tionnaire. A structural analysis of the key parts uted to improvements in compliance, physical was completed to evaluate the strength of the fitness, as well as selected respiratory para- design. The results reinforced the feasibility meters. Patients found the training motivat-and usability of TrachyPen in terms of match- ing, easy to understand and user-friendly. This ing the design requirements and clinical per- system appears to be a suitable and afford-spectives. able tool for routine use in prehabilitation in urogynecology. Potential use can certainly be OR-006 in other fields of medicine as well. Telemedicine system for prehabilitation in urogynecology OR-007 Petra Sládková 1, Marie Tichá1, Radim Kliment2, Current Challenges and Future Outlook for Jan Mužík2, Marek Doksanský2, Karel Hána2, Pavel Extended Reality as Cutting-edge Assistive Smrčka2 Technology shaping Caring Personnel 1Bulovka University Hospital, Czech Republic Fabiano Bini, Michela Franzò, Alessia Finti, 2Czech Technical University in Prague, Czech Re- Franco Marinozzi public Sapienza University of Rome, Italy Among the most common problems that In 2019, the World Health Organization es- bother women after childbirth and after men- timated that more than 1 billion of people in opause, the descent of the pelvic organs as- the world needed assistive technologies (AT). sociated with incontinence is one of the most Among them 970 million needed low vision common. One option that can be used to aids, 35 million prostheses, 75 million wheel-address this problem is laparoscopic sacro- chairs, 94 million hearing aids, 150 million mo-colpopexy. This surgical procedure is time- bility aids and 150 million cognitive aids. Based consuming and poses a great burden to pa- on a EUROSTAT data from 2020, in Europe tients. around 3.7 million practicing caring personnel Therefore, intensive preparation of patients were registered. Several research sug-gests in the form of prehabilitation is appropri- that new AT are needed and that Extended-ate. It improves cardiorespiratory paramet- Reality may be an adequate option in a con-ers and overall physical fitness before surgery, texts where smart homes are within everyone’s reduces perioperative complications and en-reach. hances recovery. In this study a proof of concept on the possib- A pilot study tested a telemedicine system for ility of use Extended-Reality as an AT for daily-prehabilitation lasting 3 months before sur- life activities is presented. Several and person-gery. The system allows the collection, inter- alized MR strategies are proposed aiming at pretation and sharing of patient data, using a improving the lives of elderly people or young mobile app and a respiratory trainer. The mo- people with learning and cognitive difficulties. bile app provides visual feedback during train- The aim is to reduce discrimination and ining. All measured data is transferred automat- equalities supporting persons with disabilities ically via the web portal. Here, the data is to fully enjoy their rights and freedoms on an stored and displayed in the form of graphs and equal basis with others. In this pro-posal, the 41 figures of the caring personnel and caregiver OR-009 are taken into con-siderations, providing in- Advancing Clinical Decision Support with tegration to the traditional social-care meth- Large Language Models: A Framework for ods. The need for prepared and trained per- Guideline-Compatible Hepatitis C Manage-sonnel to adequately and autonomously man- ment age the new AT is evident as well as the intro- Simone Kresevic 1, Mauro Giuffrè2, Milos Ajcevic1, duction of the social-care professionals with Lory Croce1, Dennis Shung2, Agostino Accardo1 the role of intermediary between the caregiver 1University of Trieste, Italy and caring personnel, who takes care of the 2Yale School of Medicine, Yale University, Italy patient on a daily basis, and the doctor, with Background: Large Language Mod- the aim of providing support in the use of AT els (LLMs) have the potential to optimize actively according to medical indications. healthcare by making clinical guidelines more accessible to providers, thus enhancing guideline adherence and Clinical Decision S21 - The role of Support Systems (CDSSs). However, guideline Large-Language Models and adherence, especially in managing Hepatitis Generative AI in Health C Virus (HCV) infection is suboptimal. The Technology Management development of scalable and dependable interventions is crucial for the implementation Monday morning Track B of medical guideline-conforming care and for Jun 10, 10:30 - 12:00 addressing discrepancies in adherence. Nonetheless, the varied format of clinical guidelines presents challenges in accurately OR-008 interpreting and retrieving relevant informa- Proposal of an XML standard protocol for tion. Evidence-Based Medical Equipment Main- An LLM-based solution in healthcare must tenance adhere to the principles of Honesty, Help- Alessio Luschi1, Fabio Crapanzano2, Francesca fulness, and Harmlessness (HHH). To achieve Satta3, Lorenzo Sani2, Ernesto Iadanza 1 this, strategies like Retrieval Augmented Gen- 1Department of Medical Biotechnologies, University eration (RAG) and Fine Tuning (FT) could be of Siena, Italy considered. RAG is preferred for its flexibility 2School of Engineering, University of Pisa, Italy and efficiency and dynamically integrates 3Departement of Health Technologies, ESTAR, Italy up-to-date information without frequent re- Hospital environments are becoming more training, unlike FT, which is resource-intensive and more complex due to technological ad-and requires retraining for new guideline vancements. The allocation of resources, es- updates. However, heterogeneity in the pecially during events like the SARS-CoV-2 format of clinical guidelines—including text pandemic, is becoming crucial, leading Clin-presentation and the use of visual aids like ical Engineering and Health Technology Man- tables and flowcharts—poses challenges for agement professionals to adopt evidence- the accurate interpretation and retrieval of based strategies for maintaining the reliabil- pertinent information. We aimed to evaluate ity of medical equipment. Real-World Data the feasibility of integrating LLMs into CDSS to (RWD) is leveraged to generate Real-World improve guideline adherence. Evidence for assessing the effectiveness and Materials and Methods: The ablation study safety of health technologies. Evidence- was performed by investigating the influence based maintenance, a systematic process of various steps in formatting the European crucial for superior healthcare services, uses medical guideline on HCV management. Five empirical RWD to identify effective mainten- frameworks based on the LLM developed ance strategies. by OpenAI - Generative Pre-Trained model The study aims to propose a standardized (GPT) 4 - with RAG strategy were developed, XML schema for data exchange, linking main-each differing in their approach to leveraging tenance activities with hospital systems while in-context learning, structured formatting incorporating maintenance data classifica- of clinical guidelines, custom prompt engin- tion. This approach aims to overcome data- eering, and the conversion of tables to lists. sharing challenges while providing insights into A 20-question test set based on the entire maintenance effectiveness and assessment. 42 guideline was defined to assess which is the for consistent comparisons across different optimal framework configuration. The overall languages in future developments. Following accuracy of the various outputs was qualit- tokenization, we selected texts with 10 to 128 atively assessed by two clinical experts. Each tokens to ensure meaningful yet manageable experiment was compared with the baseline translation content, dividing them into train- GPT4. ing, testing, and validation sets in an 80:10:10 Results: The performance across all inquiries ratio. We then employed a cross-entropy loss improved from 43% of accurate responses function for training the model to translate with the OpenAI GPT-4 to 99% of accurate from English to Italian. answers generated by our best-proposed Results: To quantify improvements, we meas-framework (p<0.001). The ablation study high- ured the model’s performance before and lighted enhancements through structured after tuning using a standard metric in the MT formatting of the guidelines, transformation field, the BLEU score. of tables into lists, and refinement in prompt Such a score, which assesses the closeness engineering. of machine-generated text to human trans- Conclusion: Our study showed the capability lations based on n-grams (contiguous se- of a framework based on an LLM to generate quences of n tokens) matching (we employed clinical responses HHH-principle compliant n up to 4), improved from 0.39 to 0.53 post-and consistent with established guidelines. tuning. Additionally, we conducted a qualitat- The results underscore the potential of metic- ive inspection of the translation results, which ulous guideline structuring and sophisticated allowed us to verify the trained system’s cap-prompt engineering in the incorporation of ability to accurately translate words pertain-LLMs into CDSSs. This approach opens av- ing to the domain-specific lexicon. We in- enues for the deployment of LLMs in clinical deed observed successful translations of terms settings, offering a path toward dependable like ”strabismus” to ”strabismo” and ”flukes” to and streamlined applications. ”trematodi”, which prior to tuning the model left untranslated, or translations of terms like OR-010 ”dry eyes” to ”occhi secchi” and ”tear” to ”la- Unveiling a New Resource for Crossing Lin- cerazione”, instead of the less accurate ”oc-guistic Frontiers in Biomedicine: A Prelimin- chi asciutti” and ”lacrima” terms, respectively, ary Study on English-Italian Translation provided by the model before tuning. Luca Bacco, Mario Merone, Leandro Pecchia Conclusions: While the unrefined model has Campus Bio Medico University of Rome, Italy shown a discrete level of understanding and Introduction: In the field of Machine Trans- ability to translate between English and Italian lation (MT), the biomedical domain represents within the biomedical domain, there was a sig-a significant frontier, left partially unexplored, nificant improvement showing the validity of due to its uncharted complexities and the crit- the followed approach and the developed re-ical need for accuracy and reliability. Recog- source. Looking forward, we aim to broaden nizing the growing need for highly reliable and our dataset to include more languages and to domain-specific translation tools in the bio- evaluate additional models, including multilin- medical field, we collected a dataset of over gual options. 28k parallel texts, ranging from brief sentences to longer paragraphs, between English and OR-011 Italian, to provide MT researchers with a valu- Prompt Engineering for Consistent Health-able resource for addressing the unique chal- care Responses in LLMs lenges presented by the technical and spe- Ali Salman cialized language used in biomedicine. University of Siena, Italy Methods: To assess the resource’s poten- Unlock the transformative potential of tial, we conducted a preliminary investigation. Prompt Engineering in Large-Language We utilized the Helsinki-NLP model based on Models (LLMs) for consistent healthcare the MarianMT framework for our experiments, responses. Emphasizing its critical role, we chosen for its efficiency in training and infer- dig into how prompt engineering optimizes ence times compared to larger models and LLMs, rectifying hallucinations, enhancing its extensive suite of language pairs access- symptom analysis, and ensuring accurate ible from the Hugging Face hub, thus allowing diagnostics for personalized treatments. Ex-43 plore prompt design complications and their relevant to the goal of this research. guidance of LLMs to generate reliable health- About 50% of the remaining articles are fo- care responses, improving performance and cused on the use of large language models for consistency. Addressing LLM hallucinations, the extraction of tumor prognostic paramet- engineered prompts mitigate erroneous out- ers for both basal cell carcinoma and melan-puts through case studies, refining prompts to oma. Three articles use natural language pro-correct misleading information and enhance cessing to connect dermoscopic images with response reliability. Discuss achieving consist- language descriptions. The remaining art- ent, evidence-based healthcare responses icles analyze, with large language models, the through engineered prompts, ensuring LLM patients’ comments posted on Reddit (1 on adherence to medical knowledge, and Instagram). In particular, these comments grounding diagnostic decisions and treat- come from people suffering from inflammat-ment recommendations in sound principles. ory skin diseases such as psoriasis and atopic Learn how Prompt Engineering empowers dermatitis, which create great discomfort in LLMs to analyze symptoms effectively, provid- the patient due to itching or pain. These mod-ing clinicians with relevant information for els provide insight into the quality of life and decision-making. Explore personalized treat- mental state of these patients. ments facilitated by tailored prompts, aligning In conclusion, these techniques can be valu- LLM-generated recommendations with in- able from a clinical point of view. Above all, dividual needs, preferences, and medical large language models can make a good con- histories. Join us to unlock Prompt Engin- tribution to a more accurate definition of tu- eering’s capabilities in LLMs and explore its mors, and they can allow us to extract in a profound implications for reliable diagnostics structured way many key features and inform-and personalized treatments in medicine and ation that can only be extracted from histo- healthcare. pathological reports. This allows us to exploit this information to develop further predictive OR-012 or patient management models. The evalu- Natural language processing in dermato- ation of the quality of life of patients is also not logy: an overview to be neglected. It should be considered that Alessandra Cartocci, Alessio Luschi, Linda this parameter is one of the primary outcomes Tognetti, Pietro Rubegni, Gabriele Cevenini, Ernesto used to evaluate a treatment, as inflammatory Iadanza skin diseases are often very disabling. University of Siena, Italy Dermatology is one of the clinical fields most focused on image acquisition; however, S02 - Advances in BME in the most of them are correlated to written text. Asia Pacific Region In addition, for many dermatological condi- tions, a biopsy is required to confirm the dia- Monday morning Track C gnosis, which is then analyzed by dermato- Jun 10, 10:30 - 12:00 pathologists. The dermatopathologist writes a descriptive text of what they have observed OR-013 as well as the diagnosis. This research aims Smart Wearable Haptics for AI Healthcare to conduct a study of the literature about the and the Health Metaverse use of natural language processing in derma- Chwee Teck Lim tology. National University of Singapore, Singapore Using the search terms “((natural language processing OR large language models) AND We explore the transformative synergy (dermatology OR dermoscopy OR dermato- between wearable haptic technologies and scopy)) NOT (chatgpt OR chat-gpt)” a review artificial intelligence (AI) in shaping the future of the literature was conducted on Pubmed of healthcare. As advancements in AI continue and Scopus. 99 original articles were obtained. to revolutionize healthcare, integrating wear-Abstracts, reviews, and non-English articles able sensors and haptic devices into AI and the have already been removed. After reading the health metaverse holds immense potential to abstract or the article if the abstract was not enhance patient experiences, diagnosis, and sufficiently clear, 43 out of 99 articles were fur- treatment modalities. We will delve into the ther excluded since they were they were not innovative applications of our unique wear-44 able haptics, ranging from real-time mon-OR-015 itoring of patients undergoing rehabilitation Advances in BME in Japan: JSMBE leads bio-to interactive simulations for medical training. resilience research to create new scientific Additionally, the Health Metaverse concept principles is introduced, highlighting the integration of Yasuyuki Shiraishi 1, Yumie Ono2, Takeo AI-driven virtual environments into healthcare Matsumoto3, Ichiro Sakuma4, Ryo Haraguchi5, practices. With such applications, we hope to Tomohiro Kuroda6 1 gain insights into the promising intersection of Tohoku University IDAC, Japan 2 wearable haptics and AI, paving the way for Meiji University, Japan 3 more personalized, immersive, effective and Nagoya University, Japan 4 accessible healthcare solutions. School of ENgineering, The University of Tokyo, Ja- pan OR-014 5University of Hyogo, Japan Silk-based Enthesis Regeneration in An- 6Kyoto University, Japan terior Cruciate Ligament Reconstruction Biomedical Engineering serves our physical James Goh and mental health amelioration by physiolo- National University of Singapore, Biomedical Engin- gical and psychological functional extension in eering, Singapore the declining birthrate and super-ageing pop- The fibrocartilaginous type enthesis inter- ulation society. The BME technologies may not facial tissue provides the strength to secure compensate only for the deteriorated or de-ligament to bone. In surgical reconstruc- graded function as final alternatives; the in-tion of ACL with tendon graft, optimal heal- tegrated harmony of biological and biomed- ing of the tendon graft within the bone tun- ical responses from native organs may bring nel is not adequately achieved. Therefore, the with it the amelioration of the quality of life. proposed solution involves the use of a silk Bio-resilience, the activation of natural re- fibroin (SF)-based sheath, loaded with nan- action for remission inherent in the patho- oparticles of low crystallinity hydroxyapatite physiological process by biomedical engineer- (nHA), and sleeved onto tendon autografts ing, is a creation of innovative BME for sus-to complement their use and promote en- tainable societies and ecosystems to realise thesis formation. The sheath was sutured onto a sophisticated quality of daily life. To de- both ends of the graft and pulled through velop well-being BME technologies based on both femoral and tibial bone tunnels in a por- bio-resilience activation, innovative biomed-cine ACL reconstruction model. All animal ex- ical approaches, such as modelling and sim-periments were approved by the respective ulation, which are capable of describing the IACUC. The enhancement in osteointegration biological control methodology and environ-of tendon autograft was evident within the mental responses theories and measuring for femoral and tibial ends of the graft from as electively obtained from humans and the en- early as 1-month post reconstruction. Con- vironment through preclinical and clinical situ- tinuous host integration and bone remodeling ations, that may provide appropriate biolo-were observed through the 9 months period, gical and physiological dynamic responses, will with significant bone tunnel narrowing in the establish great promise in the treatment of test groups observed by the end of the study. various diseases. Mechanical study showed that the sheaths This presentation will highlight case studies on enhanced graft tensile strengths. The SF- bio-resilience research and provide an over- based sheath serves as delivery platform for view of project-based networking. cellular and bioactive components. Progen- itor cells attracted from the host into the por- OR-016 ous sheath could have reconstitute the nat- Automatic ROI Delineation for ive cellular environment of the enthesis by Semi�quantitative Analysis of Brain SPECT differentiating into chondrocytes and osteo- Imaging at Striatum blasts. Consequently, there was enhanced Kang-Ping Lin graft-to-host integration progressively over ChungYuan Christian University, Taiwan the 9 months implantation period, which res-Single-photon emission computed tomo- ulted in overall mechanical properties closer to graphy (SPECT) imaging, a relatively inexpens-that of the native bone-ACL-bone construct. ive nuclear imaging technique, has been used 45 as a diagnostic tool for Parkinson’s disease about the variety of interesting ideas that (PD). The tracer 99mTc-TRODAT-1 is commonly were developed in my group, from fabrica- used in Nuclear Medicine, where the imaging tion techniques, material development, actu- cost is affordable for low and common health ator/sensor mechanisms, to modular, recon- insurance system. However, poor TRODAT- figurable and bio-inspired systems. I will also SPECT imaging resolution is a problem. There- discuss about the technology readiness levels fore, we developed a semi-quantitative pro- and commercialization potential of some of gram that can perform automatic region of these ideas, and how we evolve them from interest (ROI) delineation for semi-quantifying research lab prototypes to now fully func-striatal dopamine transporter activities. A tional industry-grade systems with compelling total of ˜200 subjects were enrolled in this healthcare applications. study, including PD patients and healthy con- trols.. The subjects received a 99mTc-TRODAT- OR-018 1 brain SPECT scan. Some of the subjects also A Novel Direct Left-Ventricular Puncture received a high-resolution magnetic reson- Cannulation Extracorporeal Continuous ance imaging (MRI) scan for the evaluation of Flow Blood Pump for Emergency Cardiac the accuracy of the semi-quantitative assess- Support ment. In addition, the developed automatic Francis Chikweto 1, Yasuyuki Shiraishi1, Hanako ROI delineation method was used to calcu-Suzuki1, Aoi Fukaya2, Kazushi Ishiyama1, Tashihiko late the striatal specific uptake ratio (SUR) in Kijima3, Tomoyuki Yambe1 1 dopamine transporter imaging. The reprodu- Tohoku University, Japan 2 cibility of the SUR obtained using the method Tohoku Bunka Gakuen University , Japan 3 was compared to those in previous studies. C&T medical Labo, Japan Based on linear regression analysis, a good Heart failure (HF) is one of the lead- correlation was found between the results ob- ing causes of death worldwide, presenting a tained using the automatic ROI delineation global health challenge. Mechanical circulat- method and those of the manual method in ory support (MCS) devices play an important striatum with the corresponding MR images. role in providing hemodynamic support to pa- The age-related decline of SUR availability tients with either end-stage heart failure or measured in this study was 6.4% per decade. acute heart failure conditions. Recently out- There was almost no difference between our of-hospital emergency cardiac arrest cases results and those in related clinical studies. The are on the rise as the super-aging society pop- proposed automatic ROI delineation program ulation proportion increases. The develop- for SUR analysis was tested and used to es- ment of short-term compact percutaneous timate the ROIs in TRODAT-SPECT images to MCS devices ideal for emergency acute heart aid PD diagnosis. The program could be help-failure applications is one of the promising ad- ful for physicians who are inexperienced with vances in MCS technologies. In emergency TRODAT-SPECT images. and acute HF conditions, short-term percu- taneous and extracorporeal MCS devices are OR-017 a cornerstone in providing hemodynamic sup- Soft Robotics for Healthcare port where the survival of the patient can Raye Chen-Hua Yeow be secured quickly. Generally, during ex- National University of Singapore, Singapore tracorporeal hemodynamic support prepara- Soft robotics is a rapidly growing field that is tion, the cannula and the pump are configured revolutionizing the way we think about robots. separately and during cardiac arrest, quick ini-From the early days of Ecoflex actuators, soft tiation of circulatory support is very crucial robotics has now exploded into an extensive in order to minimize permanent brain dam-spectrum of topics including advanced ma- age. This study proposes a novel extracorpor- terials, wearable robots, xenobots, sustainab- eal cannulation-type mechanical circulatory ility and learning-based control. Compared to system, a compact centrifugal blood pump traditional robots, soft robots are made of soft capable of quick installation and starting of compliant materials with infinite degrees of hemodynamic support in the emergency car-freedom, which allow them to interact with liv- diac arrest for recovery after cardiac arrest by ing things and delicate objects in a more nat- direct left ventricular apex puncture based on ural and safer manner. In this talk, I will share Seldinger’s technique. Through numerical simulations, elucidation of the pump’s hydraulic 46 performance, flow patterns as well as shear applications in vivo where physical wiring is not stress was achieved. Simulation results showed possible. Several key aspects are observed: that at 7000 RPM rotation speed, the pump’s The global impedance changes in the cellu-flow output was 5.0 L/min against a 120mmHg lar media when conducting materials are im-pressure head with 17.2 % pump efficiency. The mersed even if no percolation does exist; the average wall shear stress on the pump’s vo- material itself, when mixed conducting sys-lute and impeller walls was 63.0Pa and 107.2Pa, tems are chosen, offers a significant redox and respectively. Furthermore, the simulation res- ionic gradient across, that expands stimula-ults showed the capability of the proposed tion in time even after the external field is off; compact blood pump to generate sufficient changes in resistivity occur at the material that pressure in good agreement with the hydraulic create complex dipoles and oscillating beha-test results. In vitro hemolysis tests using fresh viours. Therefore, neurons are exposed to a goat blood for 3 hours, the normalized hemo-variety of voltage gradient profiles, depending lysis index (NIH (mg/100L) of the pump was on the material and the electric field protocol. 0.0052 for 1.0 L/Min and 100 mmHg conditions. As a consequence, different neural behaviours The proposed novel cardiac arrest intervention are observed. In some cases, the speed of approach may be considered in future as an growth is enhanced, while in others, neurite alternative and quick auxiliary mechanical cir- growth turns towards one of the poles. This culatory support modality. work will show a summary of the local resolu- tion studies that evidence the redox and ionic gradients. Finally, additional properties with S07 - Electrochemical Effects in influence on the biological system are con- Bioelectrical Therapies sidered, like the in situ generation of volatile or permanent ferromagnetism on the material Monday afternoon Track A implanted, with possible future applications. Jun 10, 14:15 - 15:30 This strategy emphasizes how nerve growth can be encouraged at injury sites wirelessly to induce repair, and how we may benefit from OR-019 the induced fields in polarized conducting ma- Unwired bipolar electrodes in neural sys- terials to achieve localized therapies. tems. Main aspects of wireless electro- chemistry in biological systems OR-020 Nieves Casan-Pastor Electrochemistry of neural stimulation Institut de Ciencia de Materials de Barcelona, CSIC, Jiří Ehlich Spain CEITEC, Czech Republic Electrical activity underpins all life, but is Electrical neural stimulation, an essential most familiar in the nervous system, where technique in neural engineering, typically long range electrical signalling is essential for employs charge-balanced biphasic current function. The use of external fields can com-pulses to elicit action potentials in excitable pensate for at least some functional deficits, tissues. While the cathodic phase depolarizes if they occur. However, its potential to also cells to initiate action potentials, the anodic promote repair at the cellular level has only phase serves to revert any byproducts of been demonstrated in vitro. Although there electrochemical processes to their original is consistent evidence that external electric state, underpinning the perceived safety fields promote cell growth, not much atten- of these methods for chronic applications. tion is given to the electrode materials. Fur- However, the assumption of complete revers-thermore, electrodes are usually connected to ibility of the electrochemical reactions at the the power source. Recently a new possibility electrode-tissue interface is not entirely ac- has emerged. An external system may polar- curate. Certain byproducts of these reactions, ize a conducting material immersed in the bio-which may be irreversible, pose potential risks electrolytes, creating a bipolar unwired elec- to tissue or may function as signaling mo-trode with induced anode and cathode in op- lecules. The true scope and identity of these posite poles of the material. reactions remains largely unknown. The use of a wireless method to create elec- In this work, we investigate the primary electrical interactions with a biological system rep- trochemical reactions that occur during resents a paradigm shift and may allow new 47 electrical stimulation of neural tissues. We tion, enabling cell death in the surrounding provide a detailed analysis of oxygen reduc- reversibly-electroporated zone. Here, we re-tion reactions, which result in the depletion of view advances in the field of electrolytic elec-dissolved oxygen and the build-up of reactive troporation and look at potential future ap-oxygen species, such as hydrogen peroxide. plications. Additionally, we explore the electrolysis of water, leading to the generation of hydrogen OR-022 and oxygen gases and alterations in local Leveraging the Electrochemical Reac- pH levels. The oxidation of chloride ions, res- tions from Electroporation to Promote ulting in the formation of hypochlorite, and Anti-Tumor Phenotypes in Immune Cells the dissolution of electrodes in chloride-rich Zaid Salameh 1, Manali Patwardhan2, Alana environments are also key areas of our study. Hay2, Joanne Tuohy2, Rafael Davalos1 1 Our methodology utilizes amperomet- Georgia Institute of Technology, United States 2 ric/potentiometric microsensors and spec- Virginia-Maryland College of Veterinary Medicine, trophotometric assays to monitor changes in United States the substrates/products of these reactions. Low pH is a defining characteristic of an This approach enables us to determine the immunosuppressive tumor microenvironment necessary electrode overpotentials required and is a hallmark of cancer. This oncogenic for these reactions to occur on specific elec- trait, which arises due to the metabolic repro-trode materials. We conduct these studies gramming of cancer cells, promotes immune in standard electrolytes and more complex evasion in part through phenotypic changes in mediums designed to mimic real physiological resident macrophages. Although tumor acid-environments. Subsequently, we assess the ity is correlated with a poor prognosis, there extent of these reactions in the context of are minimal therapies which leverage alkalin-standard biphasic stimulation protocols. ity as a possible therapeutic. Electrochemical This research provides vital insights into the reactions occur at the electrode/medium in- often-overlooked electrochemical aspects of terface when a voltage is applied across tis-neural stimulation, highlighting potential risks sue. This elevates and decreases the pH at and laying the groundwork for safer, more the cathode and anode, respectively. Here, effective stimulation strategies. we demonstrate cathodic electrochemical re- actions can be isolated and applied to el- OR-021 evate the pH in liver tissue and promote an Electrolytic Electroporation for Tissue Abla- anti-tumor phenotype in macrophages. In tion our custom in vitro platform, Hep G2 (hepato- Mary Ho cellular carcinoma) cells were exposed to ir- Quinnipiac University, United States reversible electroporation treatment and in- Electrolytic electroporation is a tissue abla- creasing dosages of alkalinity before coculture tion modality that combines the electrochem- with THP-1 macrophages. The macrophages ical effect of electrolysis with electroporation. were then characterized for antitumor sur-Electrolysis alone uses a low-magnitude dir- face marker expression. Our data shows that ect electric current to create chemical spe-a targeted increase in pH through the cath- cies at the electrode-tissue surface which then ode electrode activates naïve macrophages diffuse through the tissue, resulting in extreme in vitro. With appropriate electrical dosing, pH changes and cell death. Electroporation, this technique has the potential to further en-on the other hand, uses pulsed electric fields hance the anti-tumor immune response fol- to create permeabilizations in the cell mem- lowing electroporation-based treatments. brane. Irreversible electroporation results in cell death by loss of cell homeostasis. It is OR-023 hypothesized that electrolytic electroporation Beyond the rainbow: A validated mechan-produces low levels of electrolytic product that istic model of electrochemically produced are able to penetrate within the electropor- pH changes in tissue ated cells, resulting in a new method of tis-Rok Šmerc, Damijan Miklavčič, Samo Mahnič- sue ablation. This method has shown promise Kalamiza for increasing the ablation volume when com- University of Ljubljana, Slovenia pared to traditional irreversible electropora- In various biomedical applications, the oc- 48 currence of pH changes in tissue due to wa- mental observations. ter electrolysis, commonly referred to as hy- drolysis, is a notable phenomenon. One such example is electroporation, wherein the per- S26 - IFMBE Education in BME: meability of a cell membrane is transiently in- Transformation of the BME creased after exposure to an electric field of Education in Asia Pacific Region sufficient amplitude, allowing increased up- take of ions and molecules. In numerous Monday afternoon Track B therapies using electroporation, such as gene Jun 10, 14:15 - 15:30 transfer or drug administration, minimising tis- sue damage is critical, with pH changes be- ing a potential contributing factor alongside OR-024 A new framework of BME education with others. Numerical modelling serves as a valu- medical focus at CYCU, NYCU and NCKU in able tool for the assessment of electrochem- Asia-pacific region ically produced pH changes in tissue. Kang-Ping Lin 1, Tse-Ying Liu2, Ying-Hui Lai2, The aim of this study was to build a compre- Wen-Tai Chiu3, Ming Chen Wang1 hensive mechanistic numerical model of the 1ChungYuan Christian University, Taiwan pH changes in tissue during and after the ap- 2National Yang Ming Chiao Tung University, Taiwan plication of electrical pulses with paramet- 3National Cheng Kung University , Taiwan ers typical for electroporation and to subsequently validate this model experimentally. Department of Biomedical Engineering The model is based on the Nernst-Planck sys- (BME) is uniquely an Inter-disciplinary depart-tem of equations for the concentration of ions ment, which is combined with engineering in a four- (unbuffered) or seven-component and scientific technology applied to biology, (for the bicarbonate buffer) electrolyte. The medicine and medical treatment in university. system consists of four or seven partial differ- In the Asia-pacific, more and more universities ential equations, depending on whether a buf- with engineering subjects have continuously fer is considered, which represent the spatial established BME major department or a BME concentration of ions for each species under department ranging from an undergraduate consideration. In addition, the last equation to a Ph.D. program. In the Asia-pacific, the describes the spatial distribution of the electric areas of expertise for BME education are potential in the electrolyte. The model was de- usually classified as biomedical electronics, veloped using COMSOL Multiphysics software biomedical materials, tissue engineering, bio-and considers two mechanisms of ion trans- mechanics, biomedical imaging, biomedical port: diffusion, driven by the concentration information etc. Regarding course design, gradient, and migration, driven by the gradi- BME education courses are classified totally ent of the electric potential. Convection is as- ˜30% as “foundation core courses” includ-sumed to be insignificant and therefore neg- ing mathematics, anatomy and physiology lected. etc., and ˜40% as “professional required The results of the model were verified through courses” including medical instrumentations, experiments conducted in agarose to which biomaterials, biomechanics, off-campus in- various pH indicators were added: methyl red, ternships and subject practice etc., which bromothymol blue, and phenolphthalein. To are required as graduate of 70˜100 credits accurately monitor the pH changes during the in BME. Course credits from the “subject experiments, we developed and calibrated a practice”, which are required as graduate video recording system. This allowed a direct credits ranging between 4 and 12, and 6 is the comparison between the experimental results most common, in which different graduate and the calculated results of the numerical credits are required in different countries. In model. addition, course credits from the “off-campus Our study presents a successful development internships”, which has been emphasized and of a mechanistic model of pH changes in tis- considered as necessary in the region, are sue during electroporation, along with the es- also required as graduate credits. Before tablishment of an experimental protocol us- graduating, students are required accomplish ing agarose as a mock tissue medium. Fur- off-campus internships in relevant institutions, thermore, our results show a good agreement which is equivalent to required credits ran-between the modelling results and the experi- ging between 2 and 5. In some universities, 49 the “subject practice” and the “off-campus ing can play a huge role in scientific innova-internships” are combined. A subject for the tion and translating invention to practice, so “subject practice” can be chosen from the as to enhance the healthcare interventions. “off-campus internships” to research on cases Therefore, while it is important for BME un- in the internship field, which will be conducted dergraduate degree program to produce en-under the supervision of a supervisor in the gineers with a strong foundation in the relev- field. The subjective is for students to un- ant engineering, sciences and technology, it derstand the role of clinical workflows in the is perhaps as important to emphasize innova- medical institution and their practical oper- tion, enterprise and leadership in the BME curation, to experience the practical operation riculum. Future BME program will need to have in hospital or healthcare related institution, a high degree of flexibility that can provide to obtain clinical professionally knowledge a wide diversity of educational experiences. and applications, to experience the working By providing graduates with a combination environment in the hospital and the way of broad-based fundamentals and special-to work with colleagues. In addition to the ized knowledge, the BME program strives to training content of the above courses, there graduate versatile biomedical engineers that are some new training structures, which es- would be best positioned to lead and be an pecially strengthen the training of medical integral part of the BME industries. This can knowledge for engineering students and the be facilitated by utilizing effective an efficient strengthening of engineering background pedagogical approach in our teaching. We for medical students. These are very special have looked at the used of a blended learn- changes in educational thinking. A summary ing environment, with concepts belonging to of three training methods from CYCU, NYCU flipped classroom, active and authentic learn-and NCKU in this area is presented to sharing ing paradigms applied to a large undergradu- the experiences. ate year 1 biomedical engineering class. The data suggested that firstly, the active learning OR-025 approach led to enhanced learning, deeper Creating an Authentic Learning Environ- impressions of concepts taught, better en-ment in the NUS-BME program gagement of students and additional interest Raye Yeow, James Goh in the subject matter. Secondly, the authen- National University of Singapore, Singapore tic learning component helped correlate the- Biomedical Engineering (BME) is one of the oretical principles with real life, demonstrate fastest growing disciplines in the past few dec- the utility of the concepts taught, and further ades. It has contributed tremendously to the improved students’ engagement and interest. medical field. However, healthcare industries Thirdly, the extensively active and authen-landscape is changing rapidly due to multiple tic learning design generated moderate en- factors, ie healthcare economics leading to hanced enthusiasm of students towards their reformation in the healthcare system, major major and facilitated a slight shift from per- trends in public health, continuing advances formance goal orientation towards learning in our understanding of human biology that goal-orientation. Furthermore, BME program has the potential impact on medical practice needs to address and respond to the rapid and the development of new innovative tech- changes in technologies and ensuring our ... nologies for effective and precise diagnosis, (abstract was cropped - too long) treatment and monitoring. With advances in technology, in particular the development OR-026 wearables, data analytics, IoT, artificial intel- Current status and features of BME educa-ligence and coupled with industry 4.0, the fu- tion in China ture of medicine would be very different. The Hairong Zheng proliferation of health centric devices and di- Shenzhen Institutes of Advanced Technology, gital health will certainly give rise to connec- Chinese Academy of Sciences, China ted health with increased fitness awareness. The inception of Biomedical Engineering Aside from the digital revolution, multi-scale (BME) education in China, which began 40 bioengineering approaches are also making years ago with its pioneering program offered impact in healthcare and medicine. As such by a medical university and leading univer- the field of medical and biological engineer- sities, has since seen a remarkable expan-50 sion. Currently, over 100 universities across scientific meetings. China offer BME programs at various levels, The research fields of this department are thanks to the solid foundation laid by the es- broadly divided into two areas based on tablishment of a national committee dedic- the fundamental academic system: physics, ated to BME’s strategic advancement. China’s electricity and mechanics, and chemistry, BME education, now ranked within the top 1% materials and life sciences. The research fields globally, benefits from an outstanding fac- are further divided into the following six areas: ulty responsible for teaching and leading sig-mechanobioengineering, biodevices, bioelec- nificant research projects. This excellence is tronics, bioimaging, chemical bioengineering, further enhanced by a culture of collabora- and biomaterials. These fields promote aca-tion across academia, industry, and health- demic fusion based on a bird’s-eye view and care sectors. Strategic investments in med- develop innovative medical technologies ical devices, healthcare enterprises, and re- by constructing fundamental technologies search breakthroughs, coupled with the integ- to control the interaction of materials and ration of artificial intelligence in healthcare, systems with living organisms. Examples of highlight the importance of BME in China’s research topics include medical robots, re-strategic planning. The emphasis on train- generative medicine technology that takes ing the next generation of professionals, along into account the mechanical response of with enhancements in the regulatory frame- cells, biochemical measurements using mi-work introducing ethical standards and bio- crofluidic devices, drug delivery systems, safety measures, has attracted both domestic implantable medical devices using hydrogels, and international students and researchers, and functional imaging for the analysis of life thereby increasing the field’s global exposure. phenomena. Collaborative research among With ongoing investments and policy support, laboratories of mechanobiology, chemical the outlook for China’s BME education is highly biology, biomaterials, and bioimaging is positive. Nevertheless, the field’s expanding in- also conducted in such an area of physical terdisciplinary nature calls for continuous cur- stimulus-responsive drug delivery systems. riculum updates and the development of spe- Most of the researches are conducted in cialized areas to keep pace with global ad- collaboration with the School of Medicine, vancements. including clinical trials using the developed technologies. OR-027 Students are encouraged to make full use of Bioengineering education through the pro- the knowledge acquired in the coursework in motion of research collaboration of school their own research activities. The activities of medicine and school of engineering at lead to a deeper understanding of knowledge. The University of Tokyo In addition, through research activities that Ichiro Sakuma involve repeated trial and error and exposure School of Engineering, The University of Tokyo, Ja- to actual biological phenomena, students do pan not simply believe that existing knowledge In the Department of Bioengineering, is complete, but rather discover problems Graduate School of Engineering, The Univer- and plan and conduct research to gain new sity of Tokyo, students are required to take knowledge. Through exposure to basic/clinical courses in the Athe curriculum, which consists medicine, engineering students can acquire of lectures and exercises on basic applications different ways of thinking than medical and of engineering technology and specialized clinical experts. Practical knowledge of risk content related to bioengineering, and to management required for clinical trials will belong to one of the laboratories of the fac- be helpful in the later stages of regulatory ulty members in the department. Students approval of medical devices. are required to conduct research in one of This presentation will present examples of the laboratories of the faculty members of research activities in the Department of the Department for two years in the M.S. Bioengineering at the University of Tokyo and program and three years in the Ph.D. program, discuss how they can contribute to human and to summarize the results in a M.S. thesis resource development. and a Ph.D. thesis, respectively. They are also required to give oral presentations at relevant 51 S03 - Applications of Entropy in land Health Care 2Biomedical Engineering Institute, Kaunas Univer- sity of Technology, Lithuania Monday afternoon Track C The objective of this study was to explore Jun 10, 14:15 - 15:30 whether cross-entropy metrics provide fur- ther insights in comparison to conventional entropy, aiming to enhance the accuracy OR-028 Entropy-Based Analysis of DNA Sequences of mental workload assessment. We initially and IGHV Mutational Status in Chronic filtered and segmented EEG signals from two Lymphocytic Leukemia: Predicting Patient prefrontal channels and decomposed the sig-Survival nals into subbands. Afterward, we calcu- lated a range of cross-entropy and traditional Alexander Martynenko 1, Xavier Pastor2 entropy metrics for each sub-band. Finally, 1V.N.Karazin Kharkiv National University, Ukraine these extracted features were fed into an Ad- 2University of Barcelona, Spain aBoost classifier to evaluate mental work- This article discusses the use of a com- load levels. The comparison of classification bination of mutation status of the immuno- results demonstrated that integrating cross- globulin heavy chain (IGHV) gene and entrop- sample entropy with sample entropy notably ies of DNA sequence for the analysis of leuk- enhanced accuracy by 10%, reaching 84%. emia patients’ survival. Using a chronic lymph- Further, utilizing the complete set of cross- ocytic leukemia (CLL) patient database, the entropy metrics yielded an accuracy of 84.5%. study calculates different entropy measures for each patient’s DNA sequence. An en- OR-030 tropy maximization algorithm is developed to Bayesian Shannon Entropy for assessing estimate the statistical DNA information of patient’s data interrelation in medical ap-each patient, allowing for the classification plications of patients into two groups without relying Alexander Martynenko 1, Xavier Pastor2 on population properties. Survival analysis of 1V.N.Karazin Kharkiv National University, Ukraine leukemia patients is conducted by combining 2University of Barcelona, Spain IGHV subtype analysis and entropy measure- This research explores the application of a ments. Statistical significance is found when proposed interrelation index for analyzing pa- comparing groups with high and low entropy, tient data in medical contexts. The patient as well as different IGHV subtypes. The analysis data is represented as either quantitative time indicates that the combination of a mutated series or qualitative sequences. Conditional IGHV subtype and high entropy of DNA se-Bayesian inference rules were developed for quence can significantly impact the life ex- both types of data. Shannon entropy was em-pectancy of leukemia patients. The results ployed for the initial dataset, while Bayesian are validated using Kaplan-Meier survival ana-rules were applied to the randomized data- lysis, Cox regressions, and a Generalized Lin- set. The interrelation index, calculated as ear Mixed Model (GLMM). Overall, the study the ratio between Bayesian Shannon entropy demonstrates the value of combining IGHV of the randomized dataset and Shannon en- gene mutation status and entropy analysis of tropy of the initial dataset, serves as an indic-DNA sequences for leukemia patient survival ator of data coherence in time series or se-prediction. The findings high-light the poten- quences. The medical applications of this intial for leveraging multiple prognostic factors terrelation index were demonstrated through to improve the identification of patients with the analysis of Heart Rate Variability (HRV) varying lifespan prognoses. in patients with various cardiac diseases and the DNA sequences of a chronic lymphocytic OR-029 Evaluating Mental Workload Through Cross- leukemia (CLL) patient with different mutation Entropy Analysis of Two Prefrontal EEG statuses of the immunoglobulin heavy chain Channels (IGHV) gene. Comparative analyses with nor- mal cases were provided for both medical ap- Matin Beiramvand 1, Mohammad Shahbakhti2, plications. In essence, the study highlights Tarmo Lipping1 1 the utility of Bayesian inference in enhancing The Faculty of Information Technology and Com- the sensitivity and accuracy of entropy-based munication Sciences (ITC), Tampere University, Fin- 52 analysis for quantitative time series and se-pathological conditions, patients displaying quences of qualitative (nominal) variables. heart failure symptoms emphasize the crit- ical need for early detection, facilitating timely OR-031 and appropriate care, enhancing patient out- Relationship of the Correlation between comes, and optimizing healthcare resources. EEG and Heart Rate Variability with Cardi- Heart rate variability (HRV), Left ventricular ovascular Indicators ejection fraction (LVEF) and Global longitud- Merilin Vihmaru1, Laura Päeske1, Hiie Hinrikus1, inal strain (GLS) are prominent parameters Jaanus Lass1, Toomas Põld2, Maie Bachmann 1 that could allow the identification of heart 1Tallinn University of Technology, Estonia failure event. Therefore, the aim of our study 2Meliva AS, Estonia was to develop an interpretable model that The correlation between electroenceph- identify the relation between the occurrence alographic (EEG) and electrocardiographic of heart failure and HRV features, as well as (ECG) signals can provide important informa- LVEF, GLS, sex and age in patients with IHD and tion about cardiovascular regulation. The cur- DCM. The study encompassed two groups: 126 rent study aims to investigate the dependence patients with heart failure (HF group) and 126 of the correlation between EEG and heart rate patients without it (noHF group). GLS, LVEF, variability (HRV) on cardiovascular indicators. and linear and non-linear HRV features were The signals of a group of 30 subjects were di-extracted for each subject. Then, the inter- vided into two groups of 15 subjects accord- pretable model was produced by a logistic re-ing to the lower and higher indicators of blood gression algorithm considering a set of fea- pressure and cholesterol. Relative EEG fre- tures chosen with the univariate logistic requency band powers were calculated in theta, gression method. The univariate logistic re-alpha, and beta frequency bands. From power gression results indicate a significative correl-spectral analysis of HRV, low frequency (LF) ation between the occurrence of heart fail- power, high frequen-cy (HF) power and LF/HF ure events and the following parameters: LVEF, were calculated. In the current study, the cor- age, expBeta, HFn, and LF/HF. The obtained relation between EEG and HRV is detected by classification accuracy of produced model two HRV features, LF and HF, in the the-ta was 73% and the area under the ROC curve band, and also by two features, HF and LF/HF, was 0.77. These preliminary findings showed in the alpha band. The correlation by one HRV that the identified parameters may be useful feature, LF/HF, is revealed in the beta band. for stratification of IHD and DCM subjects with The effect of cardiovascular features is reflec- a risk of a heart failure event. ted by LF and HF features in the theta band and by HF in the alpha band. The novel finding S08 - Electroporation meets that even a small increase in cardiovascular features (blood pressure and cholesterol) can cardiac treatment affect cardio-neuronal regulation is important and needs further investigation. Monday late afternoon Track A OR-032 Jun 10, 16:00 - 18:15 Detecting Heart Failure Relations: A Prelim- inary Study Integrating HRV, LVEF, and GLS in Patients with Ischemic Heart Disease and OR-033 What do we (think we) know about electro- Dilated Cardiomyopathy poration Katerina Iscra 1, Laura Munaretto1, Aleksandar Damijan Miklavčič Miladinovic2, Jacopo Giulio Rizzi1, Marco Merlo1, University of Ljubljana, Slovenia Agostino Accardo1, Milos Ajcevic1 1University of Trieste, Italy Simplistic view on electroporation has 2Institute for Maternal and Child Health “Burlo hindered development and translation of Garofolo” , Italy electroporation-based technologies and treatments from in vitro to in vivo. Scaling Cardiovascular diseases, such as Ischemic treatments to patients has been challenging Heart Disease (IHD) and Dilated Cardiomy- and reaching visceral targets has been limited. opathy (DCM), collectively represent the lead- Recent advances in developing Pulsed Field ing cause of mortality worldwide. In both Ablation for treatment of atrial fibrillation 53 has provided the technology that allows us OR-034 to consider delivering drugs and genes to Initial single centre experience with pulsed cardiac tissue by electroporation. Delivering field ablation for treatment of cardiac ar-high voltage pulses to excitable tissue organ rhythmias with limited to no capacity to regenerate and Matevž Jan 1, Tine Prolič Kalinšek1, Jernej Iršič1, of such vital importance has revealed many Damijan Miklavčič2, Jernej Štublar1 long-known properties and consequences of 1University Medical Centre Ljubljana, Slovenia high-voltage pulse delivery and electropora- 2University of Ljubljana, Slovenia tion (e.g. stunning, interfering with conduction Background: Pulsed field ablation (PFA) is system, causing vascular spasms) but with the an emerging ablation modality for treatment potential to deal with epidemic proportions of cardiac arrhythmias, based on irreversible of atrial fibrillation provided commercial in- electroporation. terest of largest medtech companies that are Methods: We investigated procedural para-providing us with tools to start considering meters and clinical outcomes from our PFA re- delivering drugs and genes with accuracy and gistry. CentauriTM (Galvanize Therapeutics, precision unavailable before. Inc., USA) PFA system was used. It is connec- With this opportunity it is time to review if ted instead of a radiofrequency generator and and what we know of electroporation on thus easily adopted in routine clinical proced- the membrane, cell, and tissue level; and ures, which were performed in deep sedation what is the expected “system” response to between February 21st, 2023 and March 5th, high-voltage pulse delivery. The aim is to 2024. Follow up included scheduled outpa-perform electroporation in a controllable tient clinic visits and ECG recordings guided by and predictable way to either ablate ar- symptomatic arrhythmia episodes. rhythmogenic substrate or deliver genes into Results: We included 43 consecutive pa-cells in the heart while avoiding or at least tients where at least one PFA application minimizing unwanted effects, e.g. coronary was delivered. Mainly persistent (53%) and vasospasms, inflammation, thermal damage paroxysmal (33%) atrial fibrillation (AF) were and off-target effects. treated with pulmonary vein isolation lesion Electroporation, rather than a special phe- set (81%) as recommended by the manufac- nomenon, should be considered as a (non- turer. For persistent AF isolation of left atrial specific) injury inflicted to the plasma mem- (LA) posterior wall (47%) was added. Patients brane and consequently to the cell. As such were male (65%), mean age 60,6 years, over- it triggers cell repair mechanisms and stress weight (29,3 BMI). Totally, 92 ± 60 (mean ± SD) response. When electroporation is performed PFA applications were delivered per patient, in tissue, this injury will cause immune response with skin-skin procedural time 171 ± 64 (mean due to DAMP release. Delivery of high voltage ± SD) min. Acute procedural efficacy was 81% electric pulses will cause transient ischemia for the whole cohort and follow up procedural due to effect on microvasculature, stun- efficacy was 60% and 42% for paroxysmal and ning of cardiomyocytes and lack of action persistent AF respectively. We observed 3 PFA propagation in the tissue due to cardiomyo- related complications. cyte sarcolemma increased conductivity Discussion: Due to proven non-thermal nature and putative effect on gap junctions. These of PFA, LA posterior wall isolation was chosen effects are transient, and cells may recover as per our clinical experience and expert con-and regain their normal function. We do not sensus, nevertheless one patient had post-know however, what is the extent of damage procedural esophagitis, that resolved 2 weeks that the cell/tissue can sustain, and what is later. Follow up analysis showed, that for this the point-of no return. In Pulsed Field Ablation patient higher 25A dose (intended for anterior we would like to know better when and why aspect of LA) was used in LA posterior wall. the cells die, so that we can predict the lesion One patient had acute pericarditis. Coronary size. In gene delivery, however, we need to spasm presenting with acute ST elevation was assure nucleic acid uptake and expression of noticed in one procedure, while ablating at the transgenes, their therapeutic levels and dur- ostium of the left inferior pulmonary vein. Adation of expression, or intended cell (genetic) ditionally, we used PFA in 4 patients after failed modifications. radiofrequency ablation of ventricular tachy- cardia. Despite highest dose (25A) and mul- 54 tiple applications delivered in the same loca-and finds practical application in the devel- tion, only transient effect was achieved in all opment and refinement of medical devices. cases. This could be because CentauriTM sys- The integration of numerical modelling into tem was developed for ablation of thin atrial the design process of catheter-based pul- wall, thus reaching deeper midmyocardial ta- monary vein isolation technologies is a con-chycardia isthmuses only with reversible elec- crete example of how computational insights troporation field. can drive innovation in medical technology. Conclusion: Based on small and heterogen- This presentation will therefore not only adeous patient cohort clinical efficacy compar- dress the theoretical foundations of numerical ison to the established clinical practice is dif- modelling in electroporation therapy, but also ficult. Point-by-point PFA seems moderately highlight its critical role in driving transformat-effective for treatment of AF and ineffective ive advances in clinical practice and medical for intramural ventricular substrates. We be-device development and certification. lieve that the biggest advantage of the PFA system presented here is its possibility of tog- OR-036 gling between different energy sources espe- Machine Learning Models for Predicting cially in the vicinity of neighbouring tissues (oe- Electroporation-Dependent Tissue Proper-sophagus) during extensive ablation of the LA ties and Temperature Distributions from posterior wall. Pulsed Electric Fields Edward Jacobs, Pedro Santos, Rafael Davalos OR-035 Georgia Tech, United States Numerical modelling in pulsed field ablation Pulsed electric field (PEF) therapies de- Bor Kos liver high-voltage, short electric pulses directly University of Ljubljana, Slovenia into tissue to permeabilizes cells through the The use of numerical models in electropor- generation of nano-scale pores (electroporation research has long contributed to the ation). The transitory formation of pores is understanding of the medical application of called reversible electroporation and is used electroporation. Using of these models, the in-to deliver chemotherapeutics, genes, and im- tensity of the electric field, which is the main permeable substances into cells. Larger and prerequisite for electroporation, in the target longer PEFs may induce cell death through loss tissue can be precisely controlled. Such per- of homeostasis, termed irreversible electro-sonalized treatment planning not only ensures poration (IRE). IRE was considered the upper the efficacy of the therapeutic intervention, limit of reversible electroporation but has been but also prioritizes patient safety by taking into developed as a standalone method for tis-account additional safety constraints, such as sue ablation. PEF treatment efficacy depends maintaining safe distances from critical ana- on the application of a critical electric field tomical structures and mitigating the risk of over the targeted tissue volume, but the elec-thermal injury. tric field distribution depends on the tissue- However, the effectiveness of numerical mod- specific electrical properties, which both dif- elling depends on the refinement and adapt- fer between patients in healthy and malig-ation of these computational tools to accur- nant tissues and change in an electric field-ately reflect the intricacies of biological sys- dependent manner from the electroporation tems and clinical scenarios. An important as- process itself. pect explored in this presentation is the nu- We use an in situ method , termed voltage anced impact of myocardial anisotropy and ramp, that applies a series of increasing lethal electric field thresholds during electro- voltages across treatment electrodes and poration on therapeutic outcomes, which are measures the resulting current . Due to the in-particularly important in the context of car- herent non-linearity in the system, we develop diac ablation procedures. By elucidating these a robust deep neural network, trained on finite factors and incorporating them into numerical element model simulations, to unravel the re- models, we aim to improve the precision and lationship between V/I characteristics and tis- reliability of treatment planning processes and sue properties. We found minimal test error ultimately patient outcomes. (p < 0.0001), and our model was validated to Furthermore, the translational potential of this correctly predict the complete dynamic con- research goes beyond the theoretical realm ductivity curve in a previously characterized ex 55 vivo liver model (p < 0.0001). Lastly, we char- ally, we optimized cardioporation paramet-acterize and validate the first reported elec- ers to increase expression efficiency by vary-trical tissue properties of lung tumors from five ing electrode configuration, applied voltage, canine patients. We believe this platform can pulse length, and plasmid vector size. By us- be incorporated prior to treatment to rap- ing a surface monopolar electrode, optimized idly ascertain patient-specific tissue proper- pulsing conditions and reducing vector size, ties paramount in electroporation treatment we were able to prevent ventricular fibrillation, planning models and real-time treatment pre- increase survival, reduce tissue damage, and diction algorithms. Further, this method can significantly increase gene expression levels. be used over current ex vivo methods for in situ Further studies are focused on minimizing ac- tissue characterization . tion potential stimulation in excitable cells with The nonthermal mechanisms for IRE are para- biphasic pulsing parameters yielding minimal mount for treating tissue near anatomic- twitching while enhancing gene delivery. ally sensitive structures. Numerous thermal mitigation protocols have been proposed to OR-038 minimize temperature rise, but intraoperative Gene electrotransfer – the role of electrical temperature monitoring is still needed. We and biological parameters in in vitro exper- demonstrate here that an accurate and ro- iments bust temperature prediction machine learning Alenka Maček Lebar model can be developed using estimated tis- University of Ljubljana, Slovenia sue properties (bulk and dynamic conductiv-Gene electrotransfer (GET), a method in ity), known geometric properties (probe spa- which foreign nucleic acids (NAs) are intro- cing), and easily measurable treatment para- duced into cells using high-voltage electric meters (applied voltage, current, and pulse pulses, is used to treat many individual cells number). We show that the model can predict without the use of additional chemicals and temperature rise within ex vivo perfused por- within a short time in in vitro conditions, and it cine livers, with error <0.5 °C , and is shown to is relatively inexpensive, flexible, and has great predict temperature rise in over 1000 unique potential for clinical use. A better understand- computational test conditions with <1 °C error ing of the GET mechanisms, which allow NAs of and no observable outliers . different types and sizes to enter cells, and the role of experimental electrical and biological OR-037 influencing parameters could improve the ef- In Vivo Cardioporation Enhanced Myocar- ficiency of GET and expand clinical applica-dial Gene Delivery tions. Richard Heller1, Mark Jaroszeski1, Carly Boye2, To explore the current state of knowledge on Michael Francis1, Jacob Hensley1, Alex Otten1, Sezgi the influence of electrical and biological para- Arpag-McIntosh3, Anna Bulysheva 1 1 meters on GET efficiency, we conducted a University of South Florida, United States 2 systematic literature review of GET publica- Wayne State University, United States 3 tions in in vitro conditions and performed a Old Dominion University, United States meta-analysis of reported GET efficiency. A Damage from myocardial infarction (MI) systematic literature search up to mid-2022 and subsequent heart failure are serious public revealed approximately 4600 papers report- health concerns. Current clinical treatments ing the use of GET in various research stud-and therapies to treat MI damage largely ies and biotechnological applications. The do not address regeneration of cardiomyo- experimental electrical parameters included cytes. Challenges to gene based cardiac re- the type and geometry of the electrodes, the generative therapies include efficient deliv- material of the electrodes, the device used ery to the myocardium and minimizing risk. to apply the electrical pulses, the amplitude Our current studies focus on maximizing gene and duration of the electrical pulses, and the delivery efficiency and minimizing risk asso- method used to measure the output pulses. ciated with pulsed electric field application Biological experimental parameters included to excitable cells. We established that it is the type, dosage, size, the promoter of the NA, possible to promote regeneration of cardiac the encoded transgene, the number of cells muscle with plasmid DNA encoding vascu- used for EP, whether the cells were electro-lar endothelial growth factor B delivery dir- transfected in suspension or attached, the cell ectly to the ischemic myocardium. Addition-56 line, the origin and type of cell line, and the proBNP of 1231±1708 pg/mL. At 1 year after cell electroporation medium. Only 23 papers con- therapy, we found a significant improvement tained all the desired information and were in LVEF (+10.5±7.8%, P<0.001), a decrease in therefore reliable enough to be included in the NT-proBNP (-530±1430 pg/mL, P=0.001), and meta-analysis. The meta-analysis resulted in a an improvement in exercise capacity, meas- pooled overall GET efficiency of 25% (95% CI, ured as 6-minute walk test distance ((+31±58 23.8–26.9). The parameters that significantly m, P=0.01). An improvement of LVEF >5% was affected the overall GET efficiency were cell present in 24/30 (80%) of patients, and a con- line type, NA promoter, and pulse paramet- comitant improvement in LVEF, NT-proBNP, ers. Overall, GET was significantly higher in tu-and exercise capacity was present in 18/30 mor cell lines compared to normal and primary (60%). cell lines, and in normal cell lines compared Conclusions. The use of strategies based on to primary cell lines. GET was also signific- informing target individuals with the highest antly higher when the SV40 promoter was used likelihood of regenerative response may sig- compared to CMV or other promoters. Con- nificantly improve the clinical efficacy of cell sidering pulse parameters, exponential and therapy in chronic heart failure patients. combination of short and long pulses resulted in significantly higher overall GET compared to OR-040 all other pulses. The LVAD-CD34+ Stem Cell Combination Reference: Potočnik T et al. Pharmaceutics Therapy in Non-ischemic Dilated Cardiomy- 2022, 14, 2700. opathy Patients Gregor Poglajen, Sabina Frljak, Gregor Zemljič, OR-039 Andraž Cerar, Bojan Vrtovec Precision medicine approach to cell therapy University Medical Centre Ljubljana, Slovenia in heart failure LVAD-stem cell combination therapy has Bojan Vrtovec only been evaluated as a simultaneous treat- University Medical Centre Ljubljana, Slovenia ment approach (transepicardial cell injections Aims. The goal of the study was to develop at the time of LVAD surgery) and it failed a personalized cell therapy approach to be to show additional benefits of cell therapy, used as a clinical management in patients with likely due to detrimental effects of surgery- chronic heart failure. associated systemic inflammation on injected Materials and Methods. In the derivation part cells and failing myocardium. We aimed to of the study we analysed the dataset from 5 evaluate the safety and feasibility of delayed cell therapy clinical trials conducted at UMC intracoronary stem cell therapy in this patient Ljubljana, enrolling a total of 240 patients with cohort. chronic heart failure. We performed machine In a prospective non-randomized single- learning analysis to define individual patient center pilot study we enrolled 3 patients profiles with most clinical benefits after cell with non-ischemic dilated cardiomyopathy therapy. Patient profiles were then used as (NICM), scheduled for LVAD implantation. inclusion criteria in the validation part of the In all patients, cell therapy was performed study. In the validation study, CD34+ stem 2 months after LVAD implantation. CD34+ cells were mobilized by 5-day stimulation with cells were collected from peripheral blood filgrastim, collected with apheresis and im- using GCSF stimulation (5mcg bid for 5 days), munoselection, and injected transendocardi- followed by cytapheresis and immunoselec-ally. Patients were followed 1 year after cell tion. Before injection, cells were labeled with therapy. a technetium-99m radioisotope to allow for Results. In the derivation part we identified myocardial cell retention quantification 18 nonischemic heart failure etiology, lower NT- hours after cell injection. Suspension of 80 proBNP levels, transendocardial cell injection, million CD34+ cells was infused in the tar- and lower end-diastolic volume (LVEDV) as in- get coronary artery, using a microcatheter dependent predictors of favorable response (without balloon vessel occlusion). This is the to cell therapy. Using these criteria in the val- preliminary report of 3-month follow-up. idation part of the study, we enrolled 30 pa- All patients were male with an average age tients (male: 93%), aged 51±13 years, with LVEF of 39.7±13.1 years. A transradial approach of 28.4±5.0%, LVEDV of 224±46 mL, and NT- was used in all patients and patients were maintained on the VKA with a target INR 1.0 57 – 2.0 throughout the treatment protocol. All theory, signal processing and image under-patients received maximal tolerated heart standing. In this talk, this speech will provide failure GDMT: ARNI/ACEI in 100%, Beta blocker, a complete overview of multimodality image and MRA in 67%, SGLT2i in 33%, and vericiguat techniques more commonly used in health in 67% of patients. One patient received care including magnetic resonance imaging CD34+ cell infusion in RCA and two patients (MRI), computed tomography (CT), positron in LAD. At baseline, the average NT-proBNP emission tomography (PET), ultrasound and was 1851±1398 pg/mL, 6’WT 438±39 m, LVEDD biomedical optical imaging. Furthermore, the 5.4±0.2 and LVESD 4.2±0.2. At the 3-month current applications of new biomedical ima-follow-up we observed a decrease in serum ging research and clinical areas in the field of levels of NT-proBNP (-1048 pg/mL), an in- brain, oncology, and cardiology impact the crease in 6’WT (+71 m), and an improvement way medicine. In the era of artificial intelli-of echocardiographic parameters: decrease gence (AI), AI and ML techniques have enorm-in LVEDD (-0,67 cm), LVESD (-0,2 cm). In terms ous potential to convert data into a new gen-of safety, no ventricular rhythm disturbances eration of diagnostic and prognostic mod-requiring cardioversion, access site bleeding, els and to drive clinical and biological discov- VAD thrombosis, or episodes of systemic in- ery. This efficient use of imaging information fection were observed. from modalities is shaping the future of med- In summary, in advanced chronic heart failure ical imaging. patients undergoing LVAD support delayed LVAD-stem cell treatment strategy appears to OR-042 be feasible and safe, and may, in association United States Perspective on Intelligent with mechanical unloading, promote the Health Monitoring in Precision Medicine reverse remodeling in these patients. Shankar Krishnan Past-President, IFMBE, United States S12 - Intelligent Health Generally, physicians provide effective care to a patient based on proven responses ex- Monitoring for Precision pected of an average patient from a selec- Medicine: A Global Perspective ted cluster. In the precision medicine ap- proach, care is customized with the medical Monday late afternoon practices, treatments, processes, and over- Track B all decisions designed specifically for the in- Jun 10, 16:00 - 18:00 dividual patient including factors such as ge- netics, lifestyles, and environment. Progress- ive technologies such as artificial intelligence OR-041 (AI), machine learning (ML), wearable devices, Multimodal Biomedical Imaging for the Ad- and the Internet of Things (IoT) are applied vance Precision Medicine: Past, Present and in intelligent health monitoring for precision Future medicine to acquire, analyze, and interpret Hairong Zheng health data in real-time, enabling accurate Shenzhen Institutes of Advanced Technology, diagnosis, outcomes prediction, and effective Chinese Academy of Sciences, China treatment prescription. Over the past two decades, medical ima- Sensors and wearable devices are instru- ging has been has played an increasingly crit- mental in the continuous collection of essen-ical role in biomedical discovery, diagnosis and tial physiological and biochemical paramet- treatment of various diseases. The principle of ers, physical activity, and sleep. IoT techno-medical imaging is a multidisciplinary science logy connects wearable devices and sensors encompassing physics, computer, information to the internet, enabling the seamless trans-and engineering technology, which is a revolu- fer of health data to the caregivers, patients, tionary technique to capture the complexity of and health information systems ensuring that human health and disease across from struc- patient health monitoring is continuous, com-tural to functional information. Especially, the prehensive, and accessible from anywhere. development of multimodal artificial intelli- AI and ML algorithms analyze huge amounts gence models accelerates medical imaging of data generated by wearable devices to applications in advance precision medicine. identify patterns, predict health outcomes, However, there still exists challenges in imaging 58 and provide personalized health recommend- from affordable portable wearable devices ations. and deep learning offers unprecedented op- Continuous monitoring facilitates early de- portunities for extracting novel information tection of anomalies, potentially preventing form the same biomedical signals. severe health issues. Data-driven insights In this talk, I will present the results of 5 enable healthcare providers to tailor treat- years of studies developing and validating AI ments to the individual’s health profile, improv- models for real-time non-invasive epidermal ing treatment effectiveness and patient out- and blood glucose estimation, run recruit-comes. Robust encryption, secure data stor- ing about 100 participants (paediatric, adult, age solutions, and strict access controls are older) affected diabetic conditions, and ob-essential to protect patient information. served in free-living conditions and in a meta- Precision medicine is practiced at many med- bolic chamber. ical centers in the U.S.A. Mayo Clinic provides Our results demonstrated that it is possible to personalized cancer therapy using genomic detect glycaemic events with sufficient ac- analysis of tumors. At Cleveland Clinic, AI- curacy, specificity and precision, using deep- powered remote monitoring platforms and learning and ECG, minimising the use of finger- genomic information are used to guide the pick. treatment and prevention of heart diseases. Stanford Medical Center uses precision onco- OR-044 logy for personalized cancer treatment. Pre- Precision medicine in Latin America: Univercision medicine is applied at Johns Hopkins for salizing the promise of innovation multiple sclerosis, epilepsy, HIV, and TB. Virgina Laura Ballarin In addition to benefits, intelligent health mon- National University of Mar del Plata, Argentina itoring for precision medicine also faces chal- Advances toward precision medicine have lenges concerning data privacy, scalable in- placed the world on the edge of a true health frastructure, interoperability, and regulatory revolution, where we can provide each indi- frameworks to ensure patient safety, efficacy, vidual with a tailored solution. Medicine has and security. always involved caring for people, however, Future developments will focus on integrating today a series of developments are dramat-advanced AI algorithms for predictive analyt- ically increasing doctors’ ability to understand ics, wearable device accuracy, and patient- the specific factors affecting each patient, centric data platforms to create a proactive enabling much more personalized prevention system to meet emerging health challenges. and treatment plans. In conclusion, intelligent health monitoring These trends include an increasing ability to in precision medicine enables personalized, collect and store information in an integ-data-driven healthcare by leveraging wear- rated manner, especially in patients’ electronic able devices, AI, ML, and IoT, to achieve early health records, which contain all the data detection, personalized treatment, and im- generated by an individual’s interactions with proved patient outcomes. This approach is a healthcare system, as well as by the indi- employed in the U.S.A. with encouraging res- vidual’s own tracking devices; gathering in-ults. Intelligent health monitoring will make formation about the patient also at the mo-healthcare more precise, personalized, effi- lecular level, starting with individuals’ genetic cient, and effective in the future globally. profiles, pathogens, and tumors. This, together with increasingly powerful analytical tools of OR-043 Artificial Intelligence, will enable the diagnosis Precision medicine and AI for pesonalised and treatment as well as the development of estimation of glucose levels via ECG personalized medicines at all relevant phases Leandro Pecchia of the clinical development and implementa- Campus Bio Medico University of Rome, Italy, Italy tion of new personalized health products. Despite the advent of modern digital tech- This transition has only just begun, but there nologies, glucose measures are still mostly are already significant advances foreshadow-based on fingerpick, which is a cumbersome, ing the magnitude of the change that is pos-painful and pollutant disease. sible. The greatest progress so far is seen in Yet, physiological signals (such as electrocar- oncology and the field of rare diseases. Per-diogram) can now be acquired continuously sonalized health interventions are increasing in 59 number and availability in much of the world, OR-046 including Latin America. However, the chal- Bone Metastasis Measurement based on lenge in many developing countries will not be CNN Model for Auto-ROIs supplying precision medicine to those mem- Kang-Ping Lin, Jimmy Hasugian bers of the population with significant finan- ChungYuan Christian University, Taiwan cial resources, as this is already happening. The Bone scan imaging (BSI) is the most widely challenge will be implementing such innova- used technique for detecting cancer that has tion through healthcare systems that serve the metastasized from its original site, such as general population. breast or prostate cancer. However, some As a success story in this regard, there is challenging issues have arisen recently, such the agreement between the Ministry of Pub- as how to use AI technology, as well as the lic Health, represented by the ”Dr. Arturo latest CNN model, to overcome the neces- Oñativia” Hospital in the province of Salta, Ar- sity for nuclear medicine physicians to perform gentina, and the pharmaceutical company BSI inspections to examine the regions of in-AstraZeneca, aimed at developing accessibil- terest (ROI) of each skeletal site every time. ity to precision medicine in oncology and other In addition, changes in brightness informa-rare diseases. tion of image dose, abnormal bone shapes or structures, noisy images, and asymmetric ges- OR-045 ture positions are all things that AI technology Comparative study of machine learning needs to face in practice. There is a high de-methods for the early prediction of adher- sire to correctly determine image hotspot inence to treatment formation in these difficult cases and to make Miguel Rujas1, Beatriz Merino 1, Peña Arroyo1, the confirmed ROI information more clinical Jim Ingebretsen2, Jaime Barrio Cortes3, Ana Isa- meaningful. The purpose of this study is to bel Villimar Rodriguez3, Andrés Castillo4, Ana Roca- Umbert2, Francisco Lupiañez2, Maria Fernanda use 1,500 clinical BSI image data to conduct Cabrera1, Maria Teresa Arredondo1, Giuseppe Fico1 AI technology learning and training, and then 1Universidad Politécnica de Madrid, Spain verify it with 500 image data. The result is to 2PredictBy, Spain provide a simple and accurate ROI marker as 3Fundación para la Investigación e Innovación a prerequisite for clinical application, thereby Biosanitaria en Atención Primaria, Spain helping For clinical observation, the total num- 4Foundation for Research of Hospital Infantil Niño ber of metastases measured on conventional Jesús, Spain bone scan images was analyzed. Experiments on clinical datasets show that the proposed Adherence to treatment is a critical as- framework is a promising approach with high pect of healthcare with a significant impact sensitivity (0.95), specificity (0.94), image con- on patient outcomes. This has led to the tent index (0.91), and acceptable accuracy elaboration of several studies over the years values (0.88). to understand adherence better, evolving to the point of applying Machine Learning tech- OR-047 niques, whether to study the relationship of Trusted quality infrastructure federated different factors with adherence or to make data space for networks in biomedical predictions of levels of adherence. However, engineering and medical diagnostics due to the diversity of techniques, evaluation Tomasz Sołtysiński metrics and adherence measures utilized, no Physikalisch-Technische Bundesanstalt, Germany conclusions have been drawn as to which al- gorithms are best suited to address prediction Connected ecosystems require to ex- problems in this domain. This paper aims to change and confirm trust and identity shared apply the three most widely used algorithms automatically among participating fed- in the literature to a database obtained from erations to establish responsibilities. Such a primary care centre. The study evaluates federations are able to point towards each the performance of these algorithms with and other using secured generally available public without applying a feature selection method communication channels to exchange digital and with three different adherence measures, assets, manage the processes, establish in-using four different evaluation metrics. teroperable exchange of data and protocols and connecting different stakeholders. It is also the case of networks connecting devices, 60 institutions and the society, for instance for European Metrology Dataspace and the re-medical diagnostics. cently published Quality-X white paper. Such a challenge nowadays, starting from evaluation of devices and technologies up to evaluation of data quality purely deals with Biomedical instrumentation, measurement of any kind. In this sense, med- signal processing and imaging ical diagnostic is nothing but a metrological process securing to provide verifiable verific- Monday late afternoon ation of performance, to quantify an inform- Track C ation content in the data and to guarantee keeping the best possible standards and ac- Jun 10, 16:00 - 18:15 curacies. Despite a plethora of different data sources, acquisition methods and quality is- OR-048 sues, a precise extraction of real information A System for Multimodal Ultrasound Vascu-content from the data, covering the most ac- lar Assessment curate and adequate representation of the Marcin Lewandowski, Piotr Karwat real object or process remains difficult. A pro- us4us Ltd., Poland ficiency test to validate a method or perform-Ultrasound may be a fast, non-invasive, ance of a participant may be designed to and easy-to-use tool for vascular assessment tackle this challenge. The generality of such and cardiovascular diagnostics. There is an a tool offers enormous opportunities to com- ongoing pursuit of new ultrasound modalities pare the skills of a human investigator to a and biomarkers that can achieve this goal. machine implemented, semi- or fully auto- Our objective is to develop a demonstrator mated approach to gain highest measure- and algorithms for the non-invasive mul-ments standards. timodal ultrasound assessment of arteries us- To solve the problem, a selective disclos- ing a combination of parameters, in particular: ure within a quality infrastructure realized by pulse wave velocity (PWV), vascular elasticity, applications of W3C decentralized identifi- and blood flow. ers (DID) and verifiable credentials (VC) in We exploit the software-defined ultrasound a federated international data spaces (IDS) paradigm applied on a portable program-approach is proposed. To realize a select- mable research system to build an integrated ive disclosure on demand between federa- system solution. tions, a trust and identity work package of We propose a fusion of the following methods: the Gaia-X environment has been adapted ultrafast acquisition, plane wave imaging and along with the newly developed Trust mAn-Doppler techniques. We hypothesize that the agement INfrastructure (TRAIN) concept to application of these methods at a single site facilitate individual trust decisions in a self- and at the same time will provide new insight sovereign identity (SSI) ecosystem. Such an into the arterial system. We are currently fo- open-source-based approach, utilizing the cusing on two easily accessible sites: the com- tools of Cross Federates Services developed mon carotid artery and the radial artery. by Gaia-X and the Eclipse Foundation demon- We have developed a set of signal pro- strates a secure, restricted, and interoperable, cessing algorithms tested on datasets collec-automated framework to exchange only re- ted using three angle plane-wave and ultra-quired information of interest. fast acquisition (2000 frames/sec). The ex- The proposed approach is readily scalable, perimental setup consists of a portable ultra- extendable, and accessible to any authorized sound research system us4R-lite™ (us4us, Po-body and industrial partner. Any legal parti- land) and a linear-array probe Ultrasonix L9-4 cipant can join Gaia-X ecosystem and use the (128-element) at the frequency of 6.5 MHz. We offered services exposed in a federated cata- acquired in-vivo data on ourselves. logue. Through application of SSI based on Preliminary verification has shown that the DIDs, VCs and secured communication, we de- implemented algorithms can track arterial veloped a scenario adaptable to a biomed- motion on B-mode images and provide in- ical engineering endeavor, starting with a use put for the estimation of pulse wave velocity case in a metrology network. Such an ap- (PWV). For the 5-second recording, we cap- proach is easily adaptable to any quality in- tured seven pulse waves for which the aver- frastructure, especially considering a potential age PWV was equal to 4.1 ± 0.3 m/s (mean ± 61 standard dev.). Therefore, the obtained res- index (CI) and indices related to left ventricu-ult is slightly below typical values for healthy lar contractility, including systolic time inter-adults. The standard deviation, as for such a vals: velocity index (VI), acceleration index simple implementation of the local PWV es- (ACI), preejection period (PEP), left ventricu-timation, is reasonably low. lar ejection time (LVET), ejection time index We have demonstrated that the method for (ETI) and flow time corrected (FTc). The measurement of the local PWV can be effect- spontaneous variability of selected measures ively implemented on the portable research was measured in time and frequency do-ultrasound scanner. The ultrafast acquisition, main and baroreflex function was measured access to raw data, and a high-level of pro- employing causal frequency domain method. grammability will enable us to exploit mul- For statistical analysis of between groups timodal approach for vascular assessment. differences, Mann-Whitney U-test was used Our long-term goal is the development of with Benjamini–Hochberg procedure for mula complete prototype point-of-care ultra- tiple testing correction. While α 2A-AR poly-sound scanner for vascular examination. It morphism (rs1800544) was associated with will provide multimodal vascular assessment baroreflex control changes during MA only for clinical research that could lead to the (p=0.010-0.013), HUT demasked the effect successful monitoring of cardio-vascular bio- of the α 1A-AR polymorphism (rs1048101) on markers. vasomotion (vascular resistance variability) (p=0.008-0.027) and cardiac contractile func- OR-049 tion (SV, SI, LVET, FTc) (p=0.006-0.018). Poly- Cardiovascular control alterations associ- morphism rs5443 in G protein subunit beta ated with genetic polymorphisms of pro- 3 was associated with cardiac contractility teins involved in sympathetic pathway sig- measures (VI, ACI) (p=0.004-0.035) and renal transduction activity of several systolic time intervals to Michal Javorka, Lenka Matušková, Barbora orthostasis. (p=0.001-0.026) Polymorphism Czippelova, Jana Čerňanová Krohová, Zuzana rs5751876 was associated with heart rate and Turianiková, Zuzana Kolková, Zora Lasabová its variability at rest (p=0.001-0.016). Further- Comenius University in Bratislava, Slovakia more, this polymorphism influenced vasomo- From the family of G protein coupled re- tion at rest (p=0.008). Hemodynamic para-ceptors (GPCRs) involved in the signal trans- meters (SV, SI) and contractility measures (VI, duction, the adrenergic receptors (ARs) me-LVET) were also associated with this poly- diate the effects of catecholamines. While morphism. (p=0.002-0.025) We conclude that α-ARs play an important role in the regu- complex analysis of cardiovascular control lation of blood pressure, adenosine A2A re- can reveal subtle differences in cardiovascular ceptor exerts potent vasodilatory effects. It sympathetic nervous control associated with is assumed that single nucleotide polymorph- genetic polymorphisms. Perspectively, geno- isms of proteins involved in signal transduc- typing of polymorphisms could contribute to tion might influence cardiovascular function. risk stratification of patients with cardiovascu- However, the information on their influence lar disease. on cardiovascular control at rest and under Supported by grant VEGA 1/0283/21. stress in human is still very scarce. We aimed to explore the associations between four se- OR-050 lected gene polymorphisms and cardiovas- Control Unit for Phototherapy Sources cular control related variables in 119 young Jan Havlík, Lenka Maierová healthy adolescents at rest and during applic- Czech Technical University in Prague, Czech Repub-ation of two stressors (orthostasis – head up lic tilt (HUT) and cognitive load – mental arith- Bright-light therapy is one of the current metics (MA)). Subjects were genotyped for techniques for preventing or supporting the α 1A-AR (rs1048101), α 2A-AR (rs1800544), G pro-treatment of several psychological disorders, tein subunit beta 3 (rs5443) and adenosine such as seasonal affective disorder, bipolar A2A receptor (rs5751876) gene polymorphisms. disorder, postnatal depression etc. The pa- We measured selected cardiovascular para- per deals with the design and realization of a meters, including heart rate, blood pressure, control unit for phototherapy sources. Gen-hemodynamic measures – stroke volume (SV), eral requirements on the control unit are spe-stroke index (SI), cardiac output (CO), cardiac 62 cified and the principal approach to imple-especially useful when scarce data in the des- ment these requirements is described. Three tination domain was available. different control unit designs are prepared for three specific devices - a light therapy cham- OR-052 ber, a portable phototherapy device and a Simple electrochemical sensor for measur- biodynamic lamp. The goal of the control unit ing oxygen tension in blood or respiratory is handling DALI LED controllers, wireless user gases communication and setting the therapy para- Tadeusz Pałko, Kazimierz Pęczalski meters up. Faculty of Mechatronics, Warsaw University of The unit is realized as a simple embedded Technology, Poland system equipped with the ESP-8266 or the Measurements of the oxygen content in flu- WSP32-WROOM-32 microcontrollers. The sys- ids and gases (oximetry) have found many ap-tem implements transmitting DALI commands plications in various fields, and are particu- and receiving the answers from the bus, user larly desirable in medicine and environmental communication using the web based Graph- protection. The standard method for meas-ical User Interface and therapy control. The uring of oxygen tension, also known as oxy- specific implementation of the control unit gen pressure (pO2) in fluids and gases, includ- for the phototherapeutic light chamber is de- ing blood and respiratory gases, is the Clark’s scribed in more detail in the article. The design electrochemical method. The authors have and implementation of the device were evalu- developed and tested a simple electrochem-ated both in laboratory conditions and in real ical sensor based on the Clark method for operation. measuring oxygen pressure in blood and res- piratory gases. A sensor, developed by us for OR-051 the assessment of pO2 by the amperometric Transfer Learning from the Domain of Dia- method, meets the basic metrological require-betic Retinopathy to Aid in the Detection of ments for use in direct oxygen tension meas-Age-Related Macular Degeneration urements of blood and respiratory gases. Roberto Romero-Oraá 1, María Herrero-Tudela1, Roberto Hornero2, María I. López1, María García1 OR-053 1University of Valladolid, Spain Comparison of forehead and finger pho- 2Universidad de Valladolid and Centro de Investiga- toplethysmographic waveform index for ción Biomédica en Red en Bioingeniería, Biomater-stress assessment iales y Nanomedicina (CIBER-BBN), Spain Kristjan Pilt, Ivo Fridolin Age-related macular degeneration (AMD) Tallinn University of Technology, Estonia is the leading cause of vision loss in the eld-The assessment of acute mental stress is erly population. Transfer learning has proven important in daily life for implementing relaxa-useful in fundus image analysis for early dia- tion techniques and preventing mental health gnosis. Previous models were pre-trained on conditions. Augmentation index from the the ImageNet database. However, a source finger photoplethysmographic (PPG) signal domain related to retinal diagnosis would fa- waveform (PPGAI) has been utilized for stress cilitate model learning. Our objective was to assessment. However, a reduction in environ-apply transfer learning from the domain of mental temperature can induce vasoconstric-diabetic retinopathy (DR) to aid in the detec- tion and decrease skin perfusion, leading to a tion of AMD (binary classification). The pro- reduction in finger PPG signal-to-noise ratio. posed model was based on the ResNet-RS The forehead is relatively insulated from fluc-architecture. Pre-training aimed at DR dia- tuations in external temperature due to the gnosis was conducted using the Kaggle data- presence of the skull and scalp. The aim of base. Then, fine-tuning was performed using this study was to determine whether the PPG the Automatic Detection challenge on Age- waveform registered from the forehead could related Macular degeneration (ADAM) data- be used instead of the finger for stress assess-set. We carried out 3 experiments with differ- ment using the PPGAI. ent number of images used for fine-tuning. As The PPG signal was registered from the the main result, our method showed a much forehead and finger using reflectance and faster convergence than the corresponding transmittance mode sensors, respectively, models pre-trained on ImageNet. Addition- with infrared LEDs. The PPG signals were ally, the proposed source domain was proven 63 registered synchronously and digitized with a Measuring the IAP with periodic intervals 24-bit ADC with a sampling rate of 1kHz. is essential for timely interven-tion, i.e., for The PPG signals were recorded while the keeping the IAP at normal levels. The cur- subject was at rest in a sitting position with rent clinical method is in-vasive and offers eyes open for 4 minutes, followed by a 2- only intermittent assessments, limiting its minute long stress test. Stress was induced effectiveness in con-tinuously monitoring using either the serial seven, the trail mak- the IAP. Here, we present a continuous mul-ing or the stroop test. The recorded signals timodal IAP monitoring system (IAP-CMM), were analyzed in MATLAB. Each period of the which is non-invasive and provides continuous PPG signals was resampled to a 1-second long readings. The IAP-CMM device is equipped segment, and the harmonic components were with bioimpedance measurement (BioZ) limited to 6. The forehead and finger PPG and a mechanical muscle contraction force waveforms were averaged for rest and stress (MC) sensor. Electrical meas-urements are states, and the PPGAI was calculated. performed via dry electrodes. Following per- Signals were recorded from 16 subjects (7 formance and safety verifications, the device male, 9 female) with an average age of 30.2 was tested on 4 patients who underwent years. The PPGAI for the finger and fore- laparoscopic surgery. The preliminary results head was 0.99±0.148 and 1.34±0.193 for the suggest MC has a high linear relationship rest and 1.05±0.157 and 1.43±0.217 for the stress with IAP (r=0.99; r2=0.987) and the BioZ ex-state, respectively. The difference in PPGAI hibits a second order relationship with IAP between the rest and stress states was statist- (r2=0.95). An extensive clinical study recruiting ically significant for finger (p=0.019) and fore- more patients is needed to draw statistically head (p=0.025) using a paired t-test. The Pear- meaningful conclusions. son’s correlation coefficient between finger and forehead PPGAI was 0.62 (p=0.01) and 0.26 OR-055 (p=0.32) for the rest and stress states, respect- Estimation of middle ear characteristics by ively. an innovative Pressure-Less Acoustic Im- Forehead PPGAI differs between the rest mittance (PLAI™) device and stress states similarly to the finger. The Francesco Bassi, Agostino Accardo forehead PPGAI is higher than the finger PP- University of Trieste, Italy GAI due to the different signal registration Tympanometry is a gold standard method sites. The forehead and finger PPGAI are re- for the evaluation of hearing function, allow-lated during the rest state; however, during the ing the identification of pathological altera- stress state, no linear relationship was found. tions of the outer and middle ear by a non-Therefore, the forehead PPGAI could be used invasive approach. However, in order to make for stress assessment, however, it is not inter- the measurement it is necessary to alter the changeable with the finger PPG signal calcu-pressure of the outer ear, which limits the use lated PPGAI. of the technique for newborns and people with tympanic perforation. To overcome this OR-054 problem, a complete pressure less technique Non-Invasive Continuous Measurement of (PLAI™ ) was proposed. This paper aims to the Intra-Abdominal Pressure present the statistically significant correlations Josias Wacker 1, Srdjan Djordjevic2, Blaž between the resonance frequency measured Trotovšek3, Simon Krašna4, Jan Žumer2, Etienne with PLAI™ and the estimated volume and Haenni1, Grégoire Banderet1, Patrick Richard1, Gürkan Yilmaz1 compliance obtained through tympanometry 1Centre Suisse d’Electronique et de Microtechnique in both healthy subjects and patients af-CSEM SA, Switzerland fected by Otitis Media with Effusion (OME). 2TMG-BMC d.o.o., Slovenia The tests, conducted on 57 adult subjects, in- 3University Medical Centre Ljubljana, Slovenia dicated a significant linear relation between 4University of Ljubljana - Faculty of Mechanical En- the volume measured with tympanometry and gineering, Slovenia the resonance frequency obtained with PLAI™ that could be used to calculate a value of Abdominal compartment syndrome volume comparable with the gold standard. (ACS) is characterized by progressive intra- Moreover, an inversely proportional relation abdominal organ dysfunction resulting from between the compliance from tympanometry elevated intra-abdominal pressure (IAP). 64 and the resonance frequency from PLAI™ was without a MSR. Such a system should allow found, even if affected by several outliers, to use MCG as a complementary method to which hinder a practical use of this specific the generally used ECG. Wide application of relation. These two preliminary findings show such a system may improve the safety margin that it is possible to use this new technique as for cardiological patients and may reduce the a preliminary test in subjects deemed to be at number of sudden cardiac death cases. risk with tympanometry, eliminating the draw- [1] Antzelevitch C. Heterogeneity and cardiac backs related to the pressure changes but of- arrhythmias: an overview. Heart Rhythm. fering a comparable meas-urement of volume 4(7):964-972 (2007). and possible future chances of research. [2] Peczalski K., Sobiech J., Buchner T. et al. Synchronous recording of magnetocardio- OR-056 graphic and electrocardiographic signals. Sci Synchronization of magnetocardiographic Rep 14, 4098 (2024). and electrocardiographic recordings ob- [3] Tang Q., Liang J. & Zhu F. A comparative tained with two separate systems review on multi-modal sensors fusion based Kazimierz Pęczalski, Teodor Buchner, Judyta on deep learning. Signal Process., 213, 109165 Sobiech, Gerard Cybulski, Tadeusz Pałko (2023). Warsaw University of Technology, Poland Spatial heterogeneity of cardiac tissue, es- pecially during repolarization phase has long S15 - Optimizing Healthcare: been considered a potential risk factor, re- Patient-Centered Technology lated to the probability of life-threatening ar- Innovations rhythmias [1]. However, due to the recent ad- vancement in design of new magnetomet- Tuesday morning Track A ers, such as spin exchange relaxation free Jun 11, 10:30 - 12:15 (SERF), and design of innovative electrodes that are neutral to the magnetic field [2], the sensor fusion approach [3] may embrace a OR-057 Patient-centered Innovation in Hospital Pa- new area of application in cardiological dia- tient Monitoring gnosis. A system was developed to synchronize Jens Muehlsteff the parallel electrocardiographic (ECG) and Philips, Netherlands magnetocardiographic (MCG) recordings ob- tained from two separate measurement sys- Healthcare systems worldwide aim at mak- tems. The synchronization was based on a ing care more accessible, convenient, and manual generation of a marker by the oper- sustainable. Socio-economic trends, such as ator. The marker was generated in the MCG the rise in chronic diseases and an increas-system, with software on the LabView plat- ing number of multi-morbid patients, coupled form. Then the marker in digital form, via a with staff shortages, require new and endur- USB connection, was subjected to the digital- ing technology solutions. These solutions shall analog conversion. The value of the analog enhance the cure and care process for indi- marker was fed to the low-sensitivity channel vidual patients, guided by the Quadruple Aim. of the ECG system. The new system was in- Patient-Centered Innovations in Healthcare tended to replace the previous one. In the represent a paradigm, where rather than a former version numerical synchronization of one-size-fits-all approach, these innovations recordings was used. Validation of the system focus on solutions driven by the individual pa-was carried out by performing recordings of tient needs. This approach considers aspects both signals in a group of volunteers located in from a 360° perspective, e.g. including the the magnetically shielded room (MSR) with an patient’s lifestyle, preferences to tailor ther- active magnetic compensation. An interesting apies, improving measurement accuracy via finding, related to parallel recordings, was the the inclusion of a patient’s specific anatomies independent occurrence of artifacts in MCG or creating supportive care environments for and ECG waveforms, which is caused by the the recovery process considering mental and different mechanisms of their formation. Fur- emotional aspects. In our patient-centered ther work should be concentrated on devel- innovation process for new patient monitoring oping the tools for obtaining the diagnostic solutions, we begin by thoroughly understand-MCG recordings under clinical conditions, i.e. ing unmet needs. This understanding serves as 65 the foundation for outlining solution options, OR-059 de-risking concepts, and ultimately conduct- Consensus on best practices for the man-ing clinical studies to verify effectiveness. This agement of Large-Scale Pilot projects: res- innovation process is outlined in detail and dis- ults from a Delphi study cussed by examples. Davide Piaggio 1, Jordi De Batlle2, Alba Gallego3, Gloria Cea3, Pedro Checa Rifa1, Alessia Maccaro1, OR-058 Giuseppe Fico3, Leandro Pecchia1 Shaping the healthcare of tomorrow 1University of Warwick, United Kingdom through patient-centered digital tech- 2Institut de Recerca Biomèdica de Lleida, Spain nologies 3Universidad Politecnica de Madrid, Spain Josias Wacker The Gatekeeper EU project, funded under Centre Suisse d’Electronique et de Microtechnique the Horizon2020, is a pioneering project aimed CSEM SA, Switzerland at addressing the healthcare challenges asso- For the past few decades, health systems ciated with Europe’s aging population. With have been challenged by a series of funda-an emphasis on digital health solutions, Gate- mental societal trends, e.g., ageing and an in- keeper aims to improve the quality of life for crease of illnesses related to lifestyle, resulting older adults by promoting independence and in a continuous increase of health costs. Ef- active aging. Gatekeeper is a large-scale pi-ficient solutions to these challenges shall be lot project, encompassing eight pilot imple- centered on patients’ needs and therefore re- mentations across various European regions, quire precise knowledge about the well-being serving as crucial testbeds for innovative tech-of individuals and groups, i.e., reliable and rel- nologies and services tailored to the needs of evant health data. older individuals. With the rise of digital health technologies Currently, there are no guidelines in literature - based among others on wearable health relative to the management of large-scale pi- sensors, ubiquitous mobile internet, and ma- lot projects. For this reason, the large-scale chine data processing - we now have tools pilot team of Gatekeeper decided to evaluat hand which allow us collecting these data ate the lessons learned, and to find consensus in various situations, including individuals’ daily beyond the Gatekeeper consortium, involving lives, and extracting and synthesizing action- the other Horizon2020 large-scale pilot pro- able parameters. The resulting devices and jects. systems surpass the state of the art in terms To this purpose, an online Delphi study was or-of information content, autonomy, ergonom- ganised based on a questionnaire of 18 ques-ics, and signal quality. tions, with some collecting professional in- In this talk we will present how the Swiss Center formation, and others asking the panelists to for Electronics and Microtechnology (CSEM) rank the best practices divided into differ-puts patients at the center of its develop- ent themes, namely engagement, interven-ments and thereby shapes tomorrow’s digital tion, large-scale pilot monitoring and con-health technologies with groundbreaking in- trol, planning, recruitment, and other. Borda novations such as a multi-signal ambulatory counting was applied to calculate the overall, shirt for non-obtrusive functional lung ima- while consensus was measured using Kendall’s ging, wearables for cuffless blood pressure W. monitoring, and algorithms for non-invasive Twenty-eight panelists took part to the Delphi and continuous measurement of the pulmon-study, and two rounds of the study were ary arterial pressure. The transformative po- enough to reach consensus. This talk will tential of these developments is anticipated present a list of prioritised best practices for to materialize in reduced healthcare costs, im-large-scale pilot management. proved quality of life of patients and care- givers, and highly individualized illness preven- OR-060 tion. Let’s Play a Cognitive Game Anton Nydal, Ankica Babic University of Bergen, Norway This research involved the development of artifacts for testing cognitive decline using a medically validated test, namely “Mini-Cog”. 66 The first artifact was based on literature and quality of the mammographic image which design studies that included user feedback. It depends on the correct positioning of the is a straight-forward test with minimal visual breast during the radiological examination. clutter, intended to be used by clinicians. The From this perspective, automatic tools sup- second artifact was an extension of the first porting technologists and radiologists and artifact that included gamification elements, guiding them in evaluating breast positioning such as a narrative, rewards, and Emojis. from mammographies can be very useful es- The development process involved assessing pecially if they can be based on quantitat- and comparing relevant development frame- ive parameters. However, there is no clear works such as Unity, JavaScript, and SvelteKit. consensus concerning the evaluation of breast The project necessitated attention to prin-positioning from mammographies: several ciples of web development, Human Computer scores have been proposed and commercial Interaction, and international guidelines for software exists with their implementations. In technology based cognitive test development general breast positioning scores take into ac-and gamification. count a small number of parameters on which The second artifact mirrors old school text- a consensus has emerged. based adventure games, in which players are In this study, we compared an automatic prompted with situations, and must make de- and semi-automatic method for calculating cisions. For instance, the simple task of adjust- the posterior nipple line (PNL), which rep-ing a clock in the first artifact becomes part resents an important quantitative parameter of the game narrative when the mayor of a useful for guiding positioning. From the com- fictional town requests the test taker’s char- parison of the automatic method with the acter to climb the city hall building to adjust semi-automatic method, an excellent agree-the town clock. Upon accepting and complet- ment emerged in the case of the CC view (ICC ing this task, the mayor presents their personal = 0.92) and to be improved in the case of the pen as a reward. Should the test taker deny MLO view (ICC = 0.73). This study therefore of-the mayor’s request, the game provides an al- fers encouraging results for the use of the PNL ternative narrative which also involves adjust- as a quantitative parameter to guide correct ing a clock, and consequently an alternat- breast positioning. ive reward. This allows the test taker to exert agency and collect rewards, while performing the cognitive test. Biomedical engineering in Development consisted of incremental im- developing countries provements based on user feed-back from hands on testing. The second artifact was Tuesday morning Track B assessed through a Usability Scale (SUS), and Jun 11, 10:30 - 12:00 a dedicated set of questions for assessing gamifica-tion: perceived engagement, enjoy- ment, and interest, on numeric scales from 1 to OR-062 Smartphone Technology: Manual Validation 100. of Two Eye-tracking Algorithms Preliminary results indicate high system usabil- Wanzi Su 1, Damon Hoad2, Davide Piaggio1, ity. Leandro Pecchia3 1University of Warwick, United Kingdom OR-061 2 Comparison of automatic and semiauto- South Warwickshire NHS Foundation Trust , United matic approach for the posterior nipple line Kingdom 3 calculation Campus Bio Medico University of Rome, Italy, Italy Francesca Angelone 1, Alfonso Maria Eye movement disorders serve as sig- Ponsiglione1, Roberto Grassi2, Francesco Amato1, nificant indicators of neurodegenerative Mario Sansone1 diseases, with recent literature exploring their 1University of Naples Federico II, Italy association with long-COVID patients. Novel 2University of Campania ’Luigi Vanvitelli’, Italy automated eye feature evaluation methods Early diagnosis of breast cancer is an im- offer a non-invasive and objective approach portant prerequisite for a better prognosis and to assess neurodegeneration. Leveraging improved life expectancy. smartphone technology, we aim to develop An accurate diagnosis also depends on the a personalized toolbox for ophthalmological 67 and neurological research and eHealth ap-offered at the University of Siena, focused on plications. Medical Devices Design and Certification. The Our study builds upon previous findings course integrates key components such as the demonstrating the feasibility of using smart- European Union Medical Devices Regulation phones for pupillometry. A two-phased trial is (MDR) and practical design strategies tailored underway, collecting eye signals from healthy for the healthcare sector. Students gain a subjects (Phase 1) and neurological disorder deep understanding of risk management and patients (Phase 2). Ad-hoc image processing the importance of aligning design with the algorithms and artificial intelligence tech- context of use, ensuring devices are both safe niques will be used to analyze eye features and effective. Human factors engineering is and evaluate the presence, severity, and emphasized to optimize user-device interac- progression of neurodegenerative diseases. tion, while heuristic evaluations and usability In parallel, we developed an eye-tracking tests are leveraged to refine user interfaces. algorithm based on smartphone video data, The course also explores the burgeoning realm comparing Circular Hough Transform (CHT) of Software as a Medical Device (SaMD), em- with Template Matching (TM). Manual meas- phasizing its increasing significance in health-urements served as ground truth, validating care delivery. Sustainable design principles the algorithm’s accuracy. Results demon- are taught, aiming to equip students with the strate that TM significantly improves the knowledge to create medical devices suitable original eye-tracking algorithm, reducing for Low and Middle-Income Countries (LMICs), execution time by 79% while maintaining thereby addressing global health disparities. high accuracy. Specifically, TM achieved an Through project work, students apply their average MPE of 0.34% and 0.95% in the x knowledge to design simple medical devices, and y directions, respectively, across the nine fostering a blend of innovative thinking and manually validated videos, outperforming practical skill development. This course pre-CHT’s MPE of 0.81% and 0.85% (although both pares students to be industry-ready profes-low error). This advancement holds prom- sionals capable of navigating and contribut-ise for enhancing eye movement tracking, ing to the future landscape of medical device facilitating early screening and diagnosis of development. neurodegenerative diseases. Moreover, our study contributes to the growing OR-064 body of literature highlighting the benefits of Demand for Biomedical Engineers in Tan-smartphone-based eye-tracking technology. zania – A View from the Outside By utilizing the ubiquity and affordability of Mirjam Bodenstorfer smartphones, our research overcomes barri- UAS Technikum Wien, Austria ers associated with traditional methods and We tried to assess the actual demand makes sophisticated eye-tracking technology for biomedical engineers (BMEs) in Tanzania accessible and efficient. Our research paves comparing it with the figure of 7,000 BMEs re-the way for transformative advancements in ported by the Tanzanian Minister of Educa- healthcare, promising a future where early tion in 2017. To estimate the present demand screening and diagnosis are seamlessly integ- and whether existing degree programs would rated into routine clinical practice. Ultimately, cover it an extensive literature review was con-this research has the potential to revolution- ducted. In addition, both a quantitative and ize neurodegenerative disease diagnosis and a qualitative survey answered by 30 Tanzanian monitoring, ushering in a new era of accessible BMEs was analyzed. Local staffing guidelines, and efficient healthcare technologies. the number of registered hospitals and results of the survey were used to further calculate the OR-063 current staffing requirements considering de- A Novel University Course on Medical gree courses offered in this field and the asso-Devices Design and Certification at the ciated number of graduates. Our results show University of Siena an actual need of at least 2,500 BME gradu- Ernesto Iadanza ates which can be covered within 11 years Department of Medical Biotechnologies, University by the existing degree programs. However, of Siena, Italy it remains open whether the present training This paper details the innovative course courses will meet the demand also in the face 68 of the actual rapid growth of technologization tial biomarkers for disease detection and in the healthcare sector. We conclude that our monitoring. The systematic literature re- solidly based results highlight the imperative view identified 50 relevant studies out of 1118 for further measures in order to enhance the items, with glaucoma being the most studied future development of the biomedical engin- condition (25 papers), followed by diabetic eering sector in Tanzania. retinopathy (10 papers), refractive error (6 pa- pers), macular degeneration (7 papers), and OR-065 cataracts (2 papers). Quality appraisal cat- The systematic literature review of auto- egorized studies as good (42%), fair (50%), or mated pupillometry system on five common poor (8%), which indicates the need for cau-ophthalmological diseases tious interpretation. Glaucoma received the Wanzi Su 1, Katia Polsi2, Jacopo Vitale2, Em- most attention, with consistent findings of re- manuel Nathaniel3, Alvin Vyapooree1, Daniel Fox1, duced constriction amplitude and prolonged Damon Hoad3, Leandro Pecchia2, Davide Piaggio1 duration in affected patients. Diabetic retino- 1University of Warwick, United Kingdom pathy studies highlighted impaired pupillary 2Campus Bio-Medico University, Italy function. Macular degeneration research 3South Warwickshire NHS Foundation Trust, United showed decreased constriction amplitude Kingdom and velocity. Refractive error papers presen- The pupillary light reflex (PLR) is controlled ted different responses influenced by stimulus by the autonomic nervous system. It plays type and severity. Cataract studies revealed a crucial role in regulating light entering the reduced pupillary constriction associated with retina and serves as a potential indicator severity and age-related changes. of neurological and ophthalmological ab- Stimulus technologies demonstrate diversity, normalities. PLR abnormalities have been including LCD screens and Maxwellian view reported in various medical conditions, includ- optical systems, each offering unique ad- ing neurodegenerative diseases and acute vantages. Pupillometer technologies involve medical emergencies. The traditional clinical manual and automated devices such as assessment using a pen torch is subjective NeurOptics’ NPi, providing objective meas-and may overlook subtle abnormalities, while urements. Infrared cameras were widely used automated pupillometry offers a more ob- for precise image capture. Quality appraisal jective and precise method for evaluating PLR. emphasized the importance of considering This review aims to explore the applications of confounders and sample size limitations, automated pupillometry in five common oph-encouraging data sharing for future research. thalmological conditions: glaucoma, diabetic retinopathy, macular degeneration, refractive OR-066 error, and cataracts. A frugal smart tool for screening for diabetic Following the Population, Intervention, Com- neuropathies parison, and Outcome (PICO) criteria and Davide Piaggio 1, Rossana Castaldo1, Gianluca the Preferred Reporting Items for Systematic Garibizzo2, Ernesto Iadanza3, Leandro Pecchia1 1 Reviews and Meta-Analyses (PRISMA) state- University of Warwick, United Kingdom 2 ment, a comprehensive search strategy was Independent researcher, Italy 3 implemented in SCOPUS. The search string University of Siena, Italy includes keywords related to pupillometry and Diabetic neuropathy, a nerve damage the five selected eye diseases. Eligible articles linked to diabetes mellitus, poses significant written in English and published between 2002 risks of disability, morbidity, and mortality if not and 2023 were included based on established promptly diagnosed. It presents a consider-criteria. Data extraction and quality appraisal able economic burden and is increasingly pre-were conducted using ad-hoc Excel sheets valent in sub-Saharan Africa, affecting up to and the Mixed Methods Appraisal Tool (MMAT), 61% of diabetic individuals. Recognizing the respectively. The narrative synthesis method urgency, the United Nations has prioritized rewas used to summarize findings from different ducing diabetes-related mortality in its Sus-study designs. tainable Development Agenda. While existing Studies in the literature have investigated pu- solutions for diabetic management primarily pillary responses to light stimuli and evaluated focus on lifestyle adjustments and glycemia variables such as baseline pupil diameter, monitoring, the screening for diabetic neuro-constriction velocity, and latency as poten- pathies, particularly in the realm of digital 69 health, remains limited. Cutting-edge screen- rat models with a defect size of 1.5 mm dia-ing methods are resource-intensive and time- meter; and 4.0 mm depth healed at a faster consuming, whereas traditional approaches rate suggesting inadequate defect size. In are effective but require specialized know- contrast, the rat model 2.4 mm diameter; and ledge often lacking in low-resource settings. 7.0 mm depth defect emerged as the suitable These settings, especially in low-income coun-model with evidence of newly formed bone tries, struggle with limited expertise, finan- signifying the process of mineralization at the cial resources, and access to medical supplies, defect site. The Hematoxylin and Eosin (H&E) compounded by challenging environmental staining of bone tissue demonstrates substan-conditions that impede the safe deployment tial formation of bone tissues (osteoid) and of medical devices. vascularized areas, consequently supporting To address these challenges, this paper the efficacy of this model. Therefore, this study proposes a smart screening tool for diabetic finds that the 8-week timepoint with a 2.4 mm neuropathies, integrating three established diameter and 7.0 mm circular defect is ideal for methods with 3D-printed accessories and a assessing bone regeneration of an engineered smartphone app. This innovative approach scaffold in rat bone model. aligns with the UN’s Sustainable Development Goal 3 and advances the fourth industrial re- S17 - Regulatory Learning volution in healthcare. Additionally, an ongo- ing on-field evaluation of this smart tool in- volves a pilot study with 11 normosubjects. Pre- Tuesday morning Track C liminary results suggest its potential as a vi- Jun 11, 10:30 - 12:00 able solution to enhance the standard of dia- betic care, particularly in diabetic neuropathy OR-068 screening, on both a global scale and within Living Labs as Regulatory Learning Tools for resource-constrained settings. Biomedical Engineering and Digital Trans- formation in Healthcare OR-067 Giuseppe Fico Establishment of Femoral Bone Defect Universidad Politécnica de Madrid, Spain Model in Sprague-Dawley Rat for Engin-In this talk, we delve into the paradigm eered Scaffold Implantation: A Pilot Study of open innovation, focusing particularly on Amira Raudhah Abdullah, Intan Maslina Musa the concept of Living Labs as experimental Department of Basic Medical science I, Faculty of environments fostering collaborative learning Medicine and Health Sciences, Universiti Sains Islam across various stakeholders. Drawing inspira- Malaysia (USIM), Malaysia tion from the framework of the quintple helix, Animal models undeniably offer advant- we explore how Living Labs serve as hubs for ages for studying bone regeneration in bone tackling complex, industry-specific challenges tissue engineering. Currently, a lack of docu- while engaging with stakeholders from gov- mented and standardized critical size defects ernment, academia, industry, civil society and (CSD) protocol exists for femoral bone. This the environment. study established a femoral bone rat model In the context of this decade’s regulatory land- to evaluate engineered scaffold and its effect scape, marked by the adoption of landmark on bone regeneration. Eight Sprague-Dawley policies such as the General Data Protection rats were divided into four groups, each in-Regulation (GDPR), the impending European duced with specific sizes of circular femoral Health Data Space regulation, the AI Act, defects: 1.5 mm diameter; 4.0 mm depth the Health Technology Assessment (HTA) reg- (Groups A and B), and 2.4 mm diameter; 7.0 ulation, and the Medical Device Regulation mm depth (Groups C and D). Rats were eu- (MDR), new opportunities emerge alongside thanized at 4- and 8-weeks post-induction. challenges. These regulatory developments Observations revealed that the 4-week period underscore the imperative for the biomedical was insufficient for initiating the bone heal- engineering community to actively contribute ing process. Notable signs of bone healing to the creation of secure and protected en- and remodelling become apparent only after vironments conducive to sustainable growth 8 weeks with normal morbidity scoring at week within the European healthcare sector. 5 onwards. Gross examination indicated that As we navigate this evolving regulatory land- scape, it becomes increasingly clear that col- 70 laborative efforts and innovative approaches tems and infrastructures enabling real-world are essential for fostering responsible innova- testing and develop-ment of new technolo-tion and addressing emerging challenges ef- gies and approaches. fectively. Biomedical engineering education Within the context of the H2020 VITALISE plays a pivotal role in equipping future profes- project funded by the European Commission sionals with the knowledge, skills, and ethical Living Lab infrastructures have been used to frameworks necessary to navigate these com-support research infrastructures and remote plexities and drive positive change within the digital access to datasets (virtual access) of healthcare ecosystem. rehabilitation, transitional care and activities Through initiatives such as regulatory exper- of daily living through har-monized processes imentation and the integration of real-world and common tools. case studies into curricula, we can empower The LifeSpace at Universidad Politécnica BME to become proactive agents of change, de Madrid has been used as a controlled capable of addressing regulatory challenges laboratory-like environment for the genera- with creativity and ingenuity. Moreover, lever- tion of new knowledge and the creation of aging the infrastructure of Living Labs, we can new innovative products and services espe- create immersive learning experiences that cially in the field of transitional care. bridge the gap between theory and prac- tice, enabling BME to gain firsthand insights OR-070 into the complexities of regulatory compliance Bridging Digital Transformation through and ethical decision-making. EU-LAC Cooperation In conclusion, the convergence of open in- Maria Fernanda Cabrera novation principles, regulatory dynamics, and Universidad Politécnica de Madrid, Spain educational initiatives presents a unique op- The SPIDER project aims at harnessing the portunity for the biomedical engineering com- unexploited potential of the newly established munity to shape the future of healthcare in- BELLA network, thereby facilitating the real- novation in Europe. By embracing collabor- ization of Europe - Latin America and the ation, innovation, and a commitment to eth- Caribbean (EU-LAC) dialogues’ outcomes in ical practice, we can build a more resilient and the digitalisation and Research & Innovation sustainable healthcare ecosystem that meets (R&I) sectors. This initiative is in line with the the needs of society while fostering economic European Union’s commitment to prioritize di- growth and technological advancement. gital cooperation within its relationship with LAC, with the goal of mutual economic ad- OR-069 vancement through new technologies, innov- Integrating Living Labs for Harmonized ation, and digitalization. Data Collection in Transitional Care The longstanding tradition of EU-LAC cooper- Beatriz Merino, Gloria Cea, Ivana Lombroni, ation on digital transformation led to the Irene Mallo, Diego Carvajal, Alejandro Medrano, design and deployment of the interconnectiv- Maria Fernanda Cabrera, Maria Teresa Arredondo, Giuseppe Fico ity infrastructure between both regions in the Universidad Politecnica de Madrid, Spain framework of BELLA Programme. Under the leadership of RedCLARA and GEANT, BELLA The concept of transitional care has be- supported the construction of a submarine come increasingly critical as the demographic fibre-optic cable linking Lisbon (Portugal) with landscape shifts towards an aging population, Fortaleza (Brazil) as well as an onward ter- along with the rising prevalence of chronic dis- restrial connection with several countries in the eases across all age groups. Various studies region. More recently, the ongoing BELLA II in the sector have shown that data on pa-project seeks to expand the digital ecosystem tient mobility and functional status can pre- to new countries, especially in Central Amer-dict successful transitions and reduce the risk ica and the Caribbean, enabling relations of adverse events. Collection of data from and exchanges between companies, research clinical and patient-reported outcome meas- centres, educational institutions and national ures may support better decisions during care research and education networks. Neverthe-transitions. However, there is no harmonized less, despite the high-capacity data connec-way to combine this information or to exploit tion and long-term interconnectivity between it in these contexts. European and Latin American R&I communit- Living Labs emerge as innovative ecosys- 71 ies, the full potential of the BELLA network ive for fostering transparency, ac-countability, has not been exploited so far. Consequently, and enhancing patient outcomes within the the potential for EU-LAC cooperation on di-swiftly evolving landscape of AI in healthcare. gital transformation is hindered, which is hav- Through a comprehensive examination of key ing a negative impact on the economic and regulations and a structured validation ap- social development of many countries in the proach, this research underscores the critical LAC region. Recent years have seen the initi- need for meticulous scrutiny and validation of ation of various digital dialogues between the AI models to en-sure their reliability and ef- EU and key LAC partners like Brazil, Mexico, ficacy in improving healthcare delivery. This and Argentina, and agreements with the Pa- study aims to lay the foundation for further cific Alliance to enhance R&I cooperation. The exploration and advancement in this pivotal EU-LAC Digital Alliance, inaugurated in March area, offering a roadmap for stakeholders, re-2023, marks the first regional digital partner- searchers, and poli-cymakers to navigate the ship advocating for a human-centered di-complexities of AI integration in healthcare gital transformation. Nevertheless, a signific- while prioritizing patient safety and quality of ant divide between the political commitments care. The work has been done in the frame-made during these dialogues and their effect- work of the GATEKEEPER project, funded by ive execution persists, undermining the expan- the European Commis-sion under the Horizon sion of digital cooperation in R&I. 2020 program. SPIDER proposes to bridge this gap by provid- ing a structured framework to translate polit- OR-072 ical agreements into actionable strategies at The BEAMER Lab: conceptualizing a living bilateral, regional, and multilateral levels, while lab frame-work to develop predictive mod- also engaging stakeholders comprehensively. els, tools, and support programs to improve By aligning digital dialogue commitments with adherence to treatment national policies and strategies, SPIDER in- Beatriz Merino 1, Miguel Rujas1, Peña Arroyo1, tends to lay the groundwork for implement- Rodrigo Martín Gómez del Moral Herran1, Jim ing dialogue outcomes throughout the BELLA Ingebretsen2, Francisco Lupiañez2, Giuseppe Fico1 1 countries in LAC, thus enhancing EU-LAC co- Universidad Politécnica de Madrid, Spain 2 operation in digital transformation and R&I for PredictBy, Spain sustained economic and social progress. This study introduces the BEAMER living lab for the aggregation of data, scenarios, and OR-071 ecosystems to improve adherence to medical Regulatory Frameworks and Validation treat-ments. Recognizing the complexity of Strategies for Advancing Artificial Intelli- adherence behaviours across different health gence in Healthcare conditions, the living lab environment enables Laura Lopez-Perez1, Beatriz Merino1, Miguel 1) the management of real-world data, 2) the Rujas1, Alessia Maccaro2, Leandro Pecchia2, Sergio definition and simulation of real-world scen-Guillen 3, Maria Fernanda Cabrera1, Maria Teresa arios, and 3) the implementation of real-world Arredondo1, Giuseppe Fico1 solutions where an intervention to ad-dress 1Universidad Politécnica de Madrid, Spain adherence is introduced. The framework may 2University of Warwick, United Kingdom have potential in bridging the gap between 3ACTIVAGE Association, Spain, Spain theoretical research and real practice, offer- As AI technologies progress rapidly, there ing a scalable so-lution to improve health outis an increasing need for tailored regulations comes by optimizing adherence to treatment. that effectively address data provision, shar- The work has been carried out as part of ing, utiliza-tion, and knowledge generation. the activities of the BEAMER pro-ject funded This paper delves into the essential regula- through the Innovative Medicines Initiative 2 tions and emphasizes the crucial role of AI Joint Undertaking under grant agreement No model validation in guaranteeing the depend- 101034369. ability and effectiveness of AI-driven solu- tions. An innovative approach is introduced, detailing an organized four-phase method- ology for external validation. The integra- tion of these frameworks and the implemen- tation of a DataLab are deemed imperat- 72 S08 - Electroporation meets signaling cascade for each protein isoform. cardiac treatment Major alterations in the contractile apparatus were highlighted by immuno-fluorescence Tuesday afternoon Track A staining of selected sarcomeric proteins. Major differences in proliferation between Jun 11, 14:15 - 15:30 healthy and TTNtv hiPSC-CMs were detected by Edu incorporation, consistent with the OR-073 differential activation of YAP and NOTCH Modelling Dilated Cardiomyopathy through pathways, revealed by AAV6- based fluor-Induced Pluripotent Stem Cell-Derived Car- escent reporters. Moreover, the contractile diomyocytes phenotype of patient-derived hiPSC-CMs was Rebecca Artioli 1, Rossella Cannilla1, Cecilia characterized using fibrin-based engineered Thairi2, Matteo Dal Ferro3, Elisa Di Pasquale2, Gian- heart tissues (EHTs), an in vitro model of an franco Sinagra3, Mauro Giacca4, Chiara Collesi3 auxotonic contractile syncytium. 1International Centre for Genetic Engineering and Collectively, our in vitro evidences, arising Biotechnology (ICGEB), Italy from a combined biological, biochemical 2Humanitas Clinical and Research Center-IRCCS, and electrophysiological approach, revealed, Italy in hiPSC-CMs bearing specific mutations in 3University of Trieste, Italy the Titin protein, structural alterations of the 4King’s College London, United Kingdom contractile apparatus and the intracellular Non-ischemic dilated cardiomyopathy biochemical machinery, possibly enabling (DCM) is the leading cause of progressive the dissection of the molecular mechanisms cardiac failure, arrhythmia and sudden death. leading to the clinical phenotype, in view of a Dominant truncating mutations of the sar- patient-tailored translational approach. comeric protein titin [TTN-truncating variants (TTNtvs)] are the most common genetic cause OR-074 of DCM, occurring in approx. 20% of familial Exploring the Effects of Electroporation on or sporadic cases. Titin, the largest human Primary Rat Ventricular Cardiomyocytes: protein, is an essential component of the Insights into Action Potential, Calcium Re- sarcomere, providing most of the passive lease, and Contraction Dynamics strenght and modulating the active contract-Vid Jan, Marko Stručić, Tina Turk, Jernej Jurič, ile force in striated muscles. TTNtvs have been Monika Kos, Matej Reberšek, Martina Perše, Lea Rems, Damijan Miklavčič identified in each protein segment, but TTNtvs University of Ljubljana, Slovenia in DCM patients are markedly enriched in the A band. Atrial fibrillation (AF) is the most common Human induced pluripotent stem cell-derived cardiac arrhythmia that poses a significant cardiomyocytes (hiPSC-CMs), together with challenge in clinical practice. It is estimated Engineered Heart Tissues (EHTs), offer new that 1 in 4 adults above the age of 40 will ex-promising approaches for cardiac disease perience a form of AF. Catheter ablation that modeling and mechanistic investigation. isolates pulmonary veins from heart tissue, In this project, we successfully generated stands as a primary treatment for paroxysmal functional cardiomyocytes from hiPSCs, de-AF. Pulsed Field Ablation (PFA) emerges as rived both from healthy donors and two TTNtv a promising new ablative approach that patients (TTNTS_3 and TTNTS_5), selected utilizes irreversible electroporation to cut off from of the Heart Muscle Disease Registry electric triggers within pulmonary veins from of Trieste, bearing the c57847+4delGTAA and entering heart tissue and thus enables heart c.69783G>A mutations respectively. to beat normally again. Even though PFA was In order to investigate the molecular signature proven to be comparably effective and safer of the TTNtv variants, we biochemically dis- than conventional thermal ablation meth-sected the molecular anatomy of each variant ods, questions persist regarding its cellular by an optimized Western Blot approach, using impact and the recurrence of arrhythmias Titin-specific antibodies matching different post-procedure. regions of the protein. Of note, the protocol In normal physiological conditions, heart will allow to identify specific intracellular muscle contraction is headed by Ca2+ release, transducers, enabling to characterize at caused by an action potential. In this study, a molecular level the specific intracellular we delve into the cellular responses of primary 73 rat ventricular cardiomyocytes to electro- trophysiological response of excitable cells in poration. Through optical electrophysiology, a complex way. During medical treatments, we investigate the effects of different pulse even non-targeted cells can be affected. parameters on action potentials (AP), calcium Therefore, it is important to better understand transients (CaT), and sarcomere shortenings the underlying mechanisms of electroporation (SS). When cells were exposed to regular elec- and electrostimulation. trostimulation that mimicked physiological Electroporation-based treatments use elec-pacing all three signals were appropriately tric pulses of different pulse parameters synchronized and consistent in amplitudes (amplitude, duration, number, repetition rate). and dynamics. Afterwards, cells were subjec- Therefore, we studied the effect of electric ted to either eight 100 µ s pulses (monopolar), pulses of different durations on excitability eight bursts of 25 x 2 µ s pulses (bipolar) with and electroporation of excitable cells in vitro. interphase and interpulse delays of 2 µ s, or As a model, we used genetically engineered S-eight bursts of 40 x 200 ns pulses (monopolar) HEK cells that express sodium and potassium with interpulse delay of 100 µ s. Bursts and 100 channels, a minimal complement needed �s pulses were applied at a frequency of 1 Hz. for excitability of otherwise non-excitable With all three evaluated protocols we noticed HEK cells. We exposed S-HEK cells grown in an uncoupling of AP-CaT-SS with pulses that monolayers to electric pulses of 500 ns, 1 �s, 10 �s were below the lethal threshold. When we and 100 �s and optically monitored transmem- exposed cells to 100 µ s monopolar pulses, brane voltage (TMV) with a potentiometric uncoupling was attained at lower electric dye ElectroFluor630 under a fluorescence fields in cells that were oriented parallel to microscope. We were able to trigger single or electric field, compared to those oriented multiple action potentials in cells with pulses perpendicularly. For cells exposed to either of all selected durations, however, the shape 2 µ s bipolar or 200 ns monopolar pulses the of action potential became affected by pulse opposite was true. duration. At lower electric fields we detected Interestingly, our results demonstrate that a delay between the pulse exposure and the even in the absence of action potentials, occurrence of the action potential, which was contractions can occur due to uncoupling, significantly longer after the shortest (500 suggesting an alternative pathway for cal- ns) pulse delivery compared to 100 �s pulse cium ion entry. We further elucidate these delivery. At higher electric fields, we achieved effects using the Luo-Rudy model, highlight- a prolonged depolarization which we consider ing the increased sarcolemma conductivity a hallmark of electroporation. By monitoring induced by supraphysiological electric pulses. intracellular calcium using the Fura-2 dye, we These insights into the cellular dynamics also observed a complex calcium response following electroporation shed light on the for all pulse durations. These results provide a complexities underlying PFA procedures and useful basis for developing theoretical models underscore the importance of considering describing the complex interplay between contraction dynamics beyond intracardiac excitability and electroporation. electrogram disappearance. OR-076 OR-075 Localized electroporation on a chip for ef- Excitation and electroporation in excitable fective DNA electrotransfer S-HEK cells exposed to electric pulses of dif-Pouyan Boukany ferent durations Department of Chemical Engineering, Delft Univer- Tina Batista Napotnik, Tina Cimperman, Lea sity of Technology, Netherlands Rems The electrotransfer of nucleic acids and University of Ljubljana, Slovenia proteins has become a pivotal aspect of bi- Electroporation is used in many medical otechnology, especially in gene modification treatments. Some of them are used for treat- and genome editing applications. Traditional ing excitable tissues (nerve, muscle, cardiac), bulk electroporation methods face significant such as cardiac pulsed field ablation for treat- limitations in delivering precise amounts of geing heart arrhythmias, brain tumour ablation netic cargo into living cells without comprom- based on irreversible electroporation, and ising cell viability during electric pulse adminis- nucleic acid delivery into muscle tissue for tration. We have developed a localized single-gene therapy. Electroporation alters the elec- 74 cell electroporation chip designed to deliver model depended on the pulse protocols; and exogenous biomolecules with exceptional ef- (v) different electrode configurations (plate or ficiency while preserving high cell viability. Our needle electrodes with various arrays). microfluidic device employs a microtrap array Common to all 3D models was that successful to immobilize cells via flow, followed by the GET could only be ascertained on the surface introduction of target molecules for electro- of the 3D constructs, indicating that DNA dif-transfer to the cells under electric pulses. This fusion is impaired in the 3D cell model. In ad- system allows real-time monitoring of exogen- dition, expression was only detected in a cap ous biomolecule electrotransfer. Our numer- of the spheroids facing the cathode, show-ical simulations confirm that localized elec- ing that the negatively charged DNA moves troporation is the primary mechanism of per- towards the positive electrode in the electric meabilization in the microtrap array electro- field. poration device. We demonstrate the sim- For our study, we built a 3D cell model (CHO plicity and accuracy of this microtrap tech- cells embedded in a collagen matrix) in which nology for electroporation by delivering both we investigated different DNA applications small molecules using propidium iodide and and different pulse parameters. The GET effi- large molecules using plasmid DNA for gene ciency was higher when the DNA was injected expression. This showcases the immense po- into the 3D cell model, showing that the DNA tential of this minimally invasive method for has impaired diffusion in the 3D cell environ-precise intracellular delivery, applicable across ment. Therefore, it is important that the DNA is various domains from bioprocess engineering injected in a dense extracellular matrix (e.g. in to therapeutic applications. tissue) at multiple sites to achieve an appropri- ate local DNA concentration that enables ef- OR-077 ficient GET. We have also shown that GET ef- Gene electrotransfer into cells in hydrogels ficiency in the 3D cell model depends on pulse as extracellular mimicry protocols and that electrophoretic mobility of Saša Haberl Meglič DNA is impaired in 3D. University of Ljubljana, Slovenia Gene electrotransfer (GET) enables the S27 - Perspectives of BME transfer of DNA into cells using electric pulses. Education The efficiency in in vivo, where there is a dense extracellular matrix, is usually low due to the re- latively low diffusion and electrophoretic mo- Tuesday afternoon Track B bility of DNA. Jun 11, 14:15 - 15:30 In order to unravel GET mechanisms and limit the number of animals used for transfection OR-078 studies, the use of different 3D cell models has IFMBE project on Survey of BME Programs been proposed: (i) different cells embedded worldwide by IFMBE regions in an allograft, in a self-secreted extracellular Ratko Magjarević matrix, in a barium alginate membrane, in a University of Zagreb, Croatia hyaluronic acid-based matrix or in a collagen The session is aimed to facilitate exchange hydrogel; and (iii) different cancer entities ag- of academic and professional experiences of gregated in multicellular spheroids. a group of invited panelists in the field of Many parameters were investigated, such as: Biomedical Engineering and Bioengineering. (i) different DNA sizes, where smaller DNA was IFMBE distinguished academics will present transfected with higher efficiency; (ii) different their diverse perspectives of the future of de-DNA concentrations, where transfection effi- velopment of biomedical engineering educa-ciency increases with DNA concentration up tion. Professor from IFMBE affiliated BME so- to a certain point; (iii) different mode of DNA cieties will be invited to participate in this ses- application (application of the DNA to the 3D sion. model or injection into the 3D model to mimic the in vivo experiment where the DNA is injec- OR-079 ted into the tissue); (iv) different pulse con- How the BME study programs survey is ditions (pulse strength, pulse number, HV LV planned to be organized: A proposal pulses...), where the GET efficiency in the 3D cell Nicolas Pallikarakis Institute of Biomedical Technology (INBIT), Greece 75 The approach for the new Global Sur- (UNAM), Mexico 8 vey on Biomedical Engineering (BME) studies Human Assistive Technologies, Mexico 9 that are offered today worldwide, is proposed Nalecz Institute of Biocybernetics and Biomedical to be based on previous experience of sim- Engineering Polish Academy of Sciences, Poland 10 ilar surveys performed in the past in Europe, Universidad de la Salle, Colombia taking also advantage of new internet data The Summer School model was organized searching tools. The initial step should be by the Department of Electronics at Pontificia to achieve involvement of the IFBME mem- Universidad Javeriana in Bogotá, Colombia, ber national societies, as well as the regional in collaboration with IFMBE, CORAL, ABIOIN, and international organizations. An interna- and technical support from IEEE/EMBS, has tional project steering committee should be emerged as a transformative force in Biomed- established with a priority to decide on the cri- ical Engineering education since its inception teria/requirements that BME educational pro- in 2017. Drawing over 300 participants across grams should fulfil in order to be included in 13 countries in Latin America, including Ar- the BME survey. Given the variety of the pro- gentina, Bolivia, Brazil, Chile, Colombia, Costa grams that are appearing under a BME um-Rica, Cuba, Ecuador, El Salvador, Honduras, brella, this is a critical issue. A second task is to Guatemala, Mexico, Peru, and Panama, the reach an agreement on the way the inform- program is committed to leveraging emerging ation should be collected. The collaboration technologies to enhance the quality of life for of regional and national societies is very im-the elderly population. portant in performing this work. Commonly The Summer School cultivates a collaborative accepted approaches are critical in this pro- environment, emphasizing interdisciplinary cedure to avoid heterogeneity of results. A de- teamwork through various activities such as tailed guide should be prepared and explained plenary lectures, seminars, workshops, liv-to the participants that will perform informa- ing labs, technical industry meetings, and a tion collection. Hackathon Challenge. These activities were Verification through internet searches meticulously designed to develop soft skills should be the next step for the final inclusion like collaborative work, creativity, decision- of the BME programs meeting the criteria to making, and communication, which are the survey list. Universities should then be in- essential for effectively addressing healthcare formed and asked to provide their comments challenges. and potentially their agreement to appear Central to Biomedical Engineering is the on the list. A management committee should resolution of healthcare issues through user- be responsible for the whole project and a centered solutions, particularly in addressing coordinating organization should be assigned the needs of vulnerable populations like the for the overall project management. A more elderly. Thus, the program fosters an in- precise plan will be deployed during the first terdisciplinary and sustainable approach, phase of the project. incorporating frameworks like design thinking and scrum to promote critical thinking and OR-080 empathy among participants. The Sum- Emerging Technologies to Support Health mer School significantly enhances technical Care and Independent Living Summer knowledge through its emphasis on digital School: An Innovation Education Model in transformation. Latin America Beyond academic pursuits, the Summer Martha L. Zequera 1, Ratko Magjarević2, Virginia School serves as a platform for cultural ex-Ballarin3, Luis Zamudio4, Shankar Krishnan5, Juan change and networking opportunities, creat- Pastore3, Eduardo Pinos6, Rosa Flores-Luna7, Alvaro ing a collaborative platform for participants Poveda8, Piotr Ładyżyński9, Nubia Arroyo10, Mauricio from diverse backgrounds. This inclusive Cubides1 environment encourages the exchange of 1Pontificia Universidad Javeriana, Colombia experiences and ideas between countries, 2University of Zagreb, Croatia academia, industry, and international or- 3Universidad Nacional de Mar del Plata, Argentina ganizations, with the goal of developing 4Instituto Politecnica Nacional - CITEDI, Mexico innovative technological solutions for the 5Past-President, IFMBE, United States elderly. 6Universidad Politécnica Salesiana, Ecuador The Summer School model exemplifies an 7Universidad Nacional Autónoma de México 76 approach to Biomedical Engineering edunational and international levels poses ad- cation, blending technical training with soft ditional hurdles. However, harmonization will skill development and cultural exchange. By yield realizable benefits to the BME students, fostering collaboration and innovation, the faculty, and researchers. program equips participants with the tools The major advantages of harmonization and mindset needed to address healthcare for students include increased mobility, en-challenges and create a meaningful impact hanced recognition of qualifications, and ex- on society. posure to diverse approaches. For faculty, harmonization contributes to the streamlined OR-081 curriculum, shared standards, and learning What does harmonization bring to students, outcomes providing a framework for new teachers, and researchers in BME course developments, reducing redundancy, Shankar Krishnan and promoting consistency. The major ad- Past-President, IFMBE, United States vantages of harmonization for the researchers Harmonization is defined in different ways: include faster research progress, faster devel- “the act of making different people, plans, opment of new medical technologies, greater situations, etc., suitable for each other”; “the comparability to new findings leading to more act of making systems or laws similar in dif- robust scientific conclusions, enhanced na-ferent companies, countries, etc. so that they tional and international collaborations, and can work together more easily; the act or pro- access to a wider pool of knowledge and re-cess of producing a pleasing visual combina- sources. tion. Harmonization in higher education may In conclusion, harmonization creates a be considered to be, “a process to ensure ho- more cohesive global landscape for BME edu- rizontal and vertical articulation between pro- cation and research. By streamlining pro- grams and institutions among various higher cesses, supporting collaboration, and ensuring education systems.” quality, harmonization empowers students, Biomedical Engineering (BME) is the fusion teachers, and researchers to advance the BME of physical, mathematical, and life sciences field to improve healthcare solutions globally. with multiple engineering disciplines to solve intricate problems in medicine and biology. OR-082 In the complex interdisciplinary BME field, BME educational programs - Between his- harmonization plays a catalytic role in nav- torical foundations and the need for inter-igating the students, faculty, and researchers disciplinarity to reach new horizons of understanding, Damijan Miklavčič innovation, understanding, and coopera- University of Ljubljana, Slovenia tion. Harmonization fosters an environment The development and rise of Biomedical facilitating collaborative efforts applying Engineering and growing MedTech Industry common standards and accepted practices has resulted in need to educate profession-with a seamless flow of knowledge to thrive als to support and provide further the devel-and succeed. While harmonization offers opment of ever more complex and diversified meaningful benefits for students, teachers, “biomedical engineering”. But also, to enable and researchers, multiple challenges are use of complex devices and systems in hospit-encountered in its implementation and realiz- als and at home. Heath care professionals and ation phases, especially in BME. Experienced supporting professions like clinical engineers academics will agree with the difficulties and medical physicists, are involved and re- involved in the harmonization of curriculum, sponsible for operation of these devices, qual- credits for transfer/substitution/advanced ity control and safety. placement, distribution of academic credits, There were several attempts in Europe and USA lab/project/research/internship work to meet to define (essential components) of Biomed-the graduation requirements, and accredita- ical engineering curriculum. Several European tion criteria of the program by the appropriate projects (many lead by Nicolas Pallikarakis) agencies. The cited factors are essential for provided guidelines and made and effort to harmonization at the undergraduate level. define the minimum course including course Harmonizing the coursework and research material for BME 1st and 2nd degree accord-at the master’s and doctoral levels at the ing to Bologna structure of higher education 77 implemented in Europe. Similarly, efforts sup-trodes. We discuss the differences between ported by Whitaker foundation, which has en- both dry electrode concepts and present abled growth and development of Biomedical validation results in different application Engineering Educational programs through contexts and metrics: comparison between out of USA, lead to several exercises in devel- gel-based and Multipin dry recordings as well oping common curriculum. as between Multipin and Flower dry electrodes It soon became obvious that some pro- using different signal quality metrics in time, grams grew from chemistry, some from elec- frequency and spatial domain; comparison of tronics/electrical engineering and some from application time and wearing comfort; repro-mechanical engineering – depending on the ducibility in a multi-center study; impact of history, local environment and strength and operator training; interrelation of electrode- background of key faculty members. There is skin impedance and channel reliability. Our no unified or commonly accepted curriculum results show evidence that novel Flower dry of Biomedical engineering. At the same time, electrodes provide equivalent signal qual- we have to admit that graduates from BME ity compared to gold-standard gel-based programs are well received even outside nar- electrodes and can be applied for extended row professions – mainly because of their in-periods of time with high wearing comfort terdisciplinary knowledge and ability to work both in sitting and supine position. We also with other professions. Perhaps the profession demonstrate that increased electrode-skin and area of Biomedical Engineering still needs impedances of dry electrodes are not con- to mature, or the essence of Biomedical En- siderably influencing channel reliability and gineering is interdisciplinarity. Before we be- signal quality in state-of-the-art biosignal come a discipline, we are not well defined. Is amplifiers. Finally, we provide evidence that this good or not is something that remains to dry electrodes can be applied in different be seen. populations, for varying head geometries and by operators without prior experience S16 - Out-of-the-lab dry EEG in gel-based EEG. Nevertheless, we highlight that training and experience in dry electrode application influence the channel reliability Tuesday afternoon Track C during application. Jun 11, 14:15 - 15:30 Our findings highlight the impact and the high potential for dry electrodes enabling mobile OR-083 and out-of-the-lab EEG, with new fields of Dry electrode technologies for comfortable application like home-side monitoring, neur-and reliable out-of-the-lab EEG ofeedback, BCI and social and hyperbrain Patrique Fiedler 1, Abdumumin Olimzoda1, studies in naturalistic settings. Indhika Warsito1, Silvia Comani2, Jens Haueisen1 1Technical University of Ilmenau, Germany OR-084 2University G. d’Annunzio of Chieti-Pescara, Italy A flexible soft cap for neonatal EEG-NIRS re- Dry electrode technologies have enabled cording produced by additive manufactur- new fields of application for electrophysiolo- ing gical measurements, specifically EEG. In Mohammed-Reda Mejbar 1, Ralf Hauffe2, Jens Haueisen1, Patrique Fiedler1 contrast to conventional gold-standard 1Technical University of Ilmenau, Germany gel-based electrodes, dry electrodes directly 2eemagine Medical Imaging Solutions GmbH, Ger-contact the scalp of the user through the hair many layer, without the need for electrolyte gels or pastes between electrode and skin. The Neuromonitoring in neonates is essential dry application principle reduces preparation for assessing brain maturation and detect-effort, complexity and the need for specially ing early signs of brain damage. The integ- trained medical personal. No preliminary skin ration of long-term electroencephalography gelling or subsequent cleaning is required and (EEG) and functional near-infrared spectro- there is no risk for gel-bridges between ad- scopy (fNIRS) into a single monitoring system jacent electrodes, especially in high-density may support the use of these neuromonitor- setups. ing modalities in neonatal intensive care units We present an overview of our recently (NICU). Such integration needs to address the proposed novel dry Multipin and Flower elec- challenges posed by the neonates’ small head 78 size, sensitive skin, and hygiene requirements. OR-085 Our objective is to develop a flexible, custom- Comparison of methods for bad channel izable cap for neonates that facilitates sim-detection in dry EEG ultaneous EEG and fNIRS acquisition, ensures Milana Komosar, Patrique Fiedler, Jens Haueisen good fit, maintains contact pressure below 3N. Technical University of Ilmenau, Germany We developed a compliant cap utilizing ad-Electroencephalography (EEG) is a non- ditive manufacturing techniques to produce invasive method for studying brain function. a flexible structure. The structure was op- In addition to traditional gel-based record- timized for various head sizes and shapes ings, there is a growing interest in dry electrode by comparison and implementation of spe- EEG applications. EEG recording techniques cific design patterns, incorporating expand- are susceptible to artifacts and noise, lead-able lattice structures. We selected a final ing to the inevitable presence of bad chan- design derived from auxetic rota chiral pat- nels in the recordings. Detecting and man-terns which ensures a stable fit and precise aging these bad channels is a standard pre-control over sensors positioning and contact processing step. Due to its inherent mechan-pressure. We employed parametric design ical properties, dry EEG is particularly suscept- principles to facilitate rapid and efficient in- ible to artifacts such as movement. Despite tegration of varying electrode and sensor lay- this, dry EEG is promising due to its ease of outs. The design parameters enable adjust- use and potential for self-application in a vari-ments in structure thickness, the curvature of ety of settings, from laboratories to sports re-the pattern, and scaling in order to adapt the search and home environments. While vari- basic design to different head regions. We im- ous algorithms have been developed for bad plemented a semi-automated design process, channel detection, no single optimal choice using an MRI-based 3D model of a neonate’s has emerged for dry EEG applications yet. head. This allows using individual MRIs to gen- We conducted an analysis of bad chan- erate individualized head caps produced with nel detection in EEG recordings using seven thermoplastic polyurethane powder with Multi methods: pop_rejchan (kurtosis), pop_rejchan Jet Fusion. (probability), FASTER, pop_clean_rawdata, A validation protocol has been established PREP-Pipeline, HAPPE, and ISD. These methods to assess the cap’s mechanical perform- were used to identify bad channels in 17 volun-ance, sensor position accuracy, and adduc- teer datasets recorded with a 64-channel dry tion. This involved mechanical testing on head EEG setup during resting state. The analyzed phantoms of circumferences ranging from 27 performance metric was the accuracy of the to 39 cm, equipped with a network of flexible method. The manual ranking of the channels capacitive sensors to measure contact pres- by an experienced EEG operator has been sure accurately for each electrode. used as a reference. Through the manufacturing and validation of The results showed that the ISD method four neonatal cap prototypes, we have suc- was the most accurate, with an average ac-cessfully fine-tuned the cap’s flexibility by ad- curacy of 91.7%. Statistical analysis indicated justing the structural parameters of the pat- significantly lower accuracy for all other meth-tern (pattern shape, strut thickness, curvature ods analyzed. angle, and material). The sensor position- The performance of different bad channel ing accuracy was 1.5± 0.5 mm, while the force detection algorithms may vary depending on measurements resulted in a contact pressure electrode type, channel layout, and acquis-of 2.15 ± 0.9 N. ition protocol. In our evaluation, seven bad Our developed iterative optimization process channel detection methods were compared. has resulted in a novel cap design compat- Considering the results obtained, we suggest ible with additive manufacturing and enabling using the ISD method for bad channel detec- simultaneous EEG and fNIRS on neonates. The tion in dry EEG recordings with a 64-channel use of a flexible sensor matrix has demon- layout. strated the cap’s ability to conform to head curvatures accurately, providing good con- tact pressure and maintaining comfort. The design process may also be extended to adult cap versions for mechanical validation and signal quality testing in real-world scenarios. 79 OR-086 cluding minimum table tennis rally duration, Technology and study protocol for the mul- sampling rate of kinematic data, and total timodal and ecological investigation of joint duration of the neurophysiological recordings, action in table tennis were defined according to the requirements Silvia Comani 1, Gabriella Tamburro1, Antonio of a multilevel analytical approach, including De Fano1, Khadijeh Raeisi Nafchi1, Mohammad neural, cognitive-behavioral, and social levels Khazaei1, Filippo Zappasodi1, Ricardo Bruña2, Ed- of investigation. son Filho3, Hannes Oppermann4, Patrique Fiedler4 Our described technical setup provides a novel 1G. d’Annunzio University of Chieti-Pescara, Italy solution for integrating multiple devices from 2Universidad Complutense de Madrid, Spain different vendors in a multimodal setup al- 3Boston University, United States lowing the simultaneous monitoring and re- 4Technische Universität Ilmenau, Germany cording of neurophysiological and kinematic Activities of daily living are permeated by signals during the performance of ecological spontaneous or deliberate, conscious or un-and naturalistic dyadic joint actions. The pro- conscious inter-individual interactions, which posed solutions go beyond the state-of-the- can be independent (e.g., accidental encoun- art both on the technological and scientific ters) or interdependent (e.g., a couple dan- level. Our novel paradigm and our specific- cing). The latter type of interpersonal inter- ally developed hardware solutions may con-action, called joint action, aims at achieving a tribute to advance the understanding of both shared common goal. Elucidating the mech- overt and covert processes occurring during anisms underpinning joint action may have an joint actions, and could support the defini- impact in a variety of social contexts involving tion of systems able to predict cooperative mates, colleagues, competitors, genders, up or competitive behaviors before being overtly to the understanding and improving of the expressed. mechanisms of decision-making. So far, joint action studies have been mainly performed OR-087 using hyperbrain scanning methods to invest- Analysis of hyperbrain data underpinning igate inter-brain synchrony and the brain-to- naturalistic joint action brain functional connectivity of two or more Gabriella Tamburro 1, Ricardo Bruña2, Patrique interacting individuals by means of multiple Fiedler3, Antonio De Fano1, Khadijeh Raeisi Nafchi1, synchronized neuroimaging devices. Mohammad Khazaei1, Filippo Zappasodi1, Silvia Comani1 We designed a novel multimodal experimental 1G. d’Annunzio University of Chieti-Pescara, Italy setup and dyadic study protocol to investig- 2Universidad Complutense de Madrid, Spain ate the neurophysiological underpinnings of 3Technical University of Ilmenau, Germany joint action through the synchronous acquisition of brain activity (EEG), muscle activa- In many daily activities, we need to co- tion (EMG), heart activity (ECG), respiration, ordinate our actions with those of other and body movements from two individuals en-people in time and space. Joint action is a gaged in ecologic and naturalistic joint ac- form of interpersonal interaction character-tions, and incorporating psychological factors. ized by a shared and public goal that cannot The EEG systems include a new class of elec- be achieved by acting individually. Joint ac-trodes specifically designed to monitor brain tion is also characterized by individual goals of activity during free full body movements, and the interacting people, which can be comple- the multiple devices of different vendors were mentary (i.e., cooperation) or mutually exclus-synchronized through a single-board micro- ive (i.e., competition). Investigating the neural computer and custom Python scripts. The mechanisms underpinning both cooperative novel dyadic study protocol permits to best and competitive joint actions may have a rel-exploit the multimodal data acquisitions. We evant impact in many social contexts of hu-selected table tennis for the dyadic motor task man daily life. However, no consensus on because it allows naturalistic and face-to- the ideal method(s) to characterize global face interpersonal interactions, free in-time brain dynamics and brain-to-brain coupling and in-space full body movement coordina- has been reached so far, especially for joint tion, cooperative and competitive joint ac- action electroencephalographic (EEG) data tions, and the implementation of two task acquired in a naturalistic context. Moreover, difficulty levels to mimic changing external such data may be affected by motion-related conditions. The recording conditions, in- technical and biological artefacts. 80 We propose an effective and generaliz- (HD) were monitored during 88 treatment able pipeline for the analysis of hyperbrain EEG sessions, including HD and haemodiafiltra-data acquired in a naturalistic context. EEG tion treatments. Pre- and post-dialysis blood signals were recorded simultaneously from two samples and spent dialysate samples, from the dyads of semi-professional table tennis play- drain outlet of HD machine, were collected ers involved in cooperative and competitive during each session. HPLC analysis was used exchanges. Hyperbrain networks of synchron- as a reference to determine UA concentra-ized activity in the individual brains and across tions in the samples. An optical sensor was brains were reconstructed using the correc- connected to the drain outlet of the HD mated imaginary part of the Phase Locking Value chine and ultraviolet light absorption of spent (ciPLV) between EEG signals. This analysis dialysate was measured online for each HD was performed separately for three frequency session at four different wave-lengths. A lin-bands: theta (4-8 Hz), alpha (8-12 Hz) and ear interactions regression model was used to beta (12-30 Hz). A k-means clustering ap- estimate UA concentrations in spent dialys- proach and graph-theoretical measures were ate with the optical sensor based on the light used to disentangle specific features of the absorption. Sensor estimated spent dialysate hyperbrain, within-brain and between-brains UA concentrations were highly accurate with functional patterns and differentiate between standard error (SE) < 3.50 �mol/L and strongly cooperation and competition. correlated (R2>=0.983) to the ac-tual UA con- Results show that the graph metrics out- centrations of spent dialysate in the train-lining the roles of nodes in the functional net- ing and the test dataset. A linear regression works and the different cluster topographies model, using treatment settings and UA con-identified for the within-brain and between- centration deter-mined with optical sensor, brains patterns can effectively differentiate was employed to estimate concentration of between cooperation and competition in the UA in blood samples. Sensor estimated blood three frequency bands. These results suggest UA concentrations were similarly highly accur-that the proposed approach for the analysis of ate (SE < 32.1 �mol/L) and strongly correlated hyperbrain data recorded within a naturalistic (R2>0.95) to the actual UA concentrations of paradigm could lead to valuable insights on blood in the calibration and the validation the brain mechanisms underpinning cooper- dataset. In conclusion, this study showed that ative and competitive joint action. UA levels in blood can be estimated noninvas- ively with the optical sensor in real time with IFMBE Young Investigators higher accuracy than shown previously. Competition OR-089 Comparison of Different Configurations Tuesday late afternoon for the Implantable Capacitive Intrabody Track A Communication on a Two-Layer Phantom Jun 11, 16:00 - 18:00 Matija Roglić 1, Luka Klaić1, Ziliang Wei2, Yueming Gao2, Željka Lučev Vasić1 1 OR-088 University of Zagreb, Croatia Online Uric Acid Concentration Estimation 2Fuzhou University, China in Blood from Spent Dialysate Measure-Implantable capacitive intrabody commu- ments Using an Optical Sensor nication is a wireless communica-tion method Joosep Paats 1, Jürgen Arund1, Kristjan Pilt1, An-that utilizes capacitive coupling as a way for nika Adoberg2, Liisi Leis2, Merike Luman1, Jana the implants to communicate between each Holmar1, Risto Tanner1, Ivo Fridolin1 other as well as with the devices that are 1Department of Health Technologies, Tallinn Uni- placed on the body. Since there are sev-versity of Technology, Estonia eral possible configurations, such as in-body 2Centre of Nephrology, North Estonia Medical to on-body (IB2OB) communication and on-Centre, Estonia body to in-body (OB2IB) com-munication, it is This study aimed to estimate concentra- important to investigate how the communic-tion of uric acid (UA) in blood from online ation channel behaves for each configuration. spent dialysate measurements during hae- Therefore, in this paper, a two-layer 3D model modialysis (HD) with an optical sensor non- of the upper part of the leg between the knee invasively. Twenty-two patients on chronic and ankle has been cre-ated for the measure-81 ment purposes. The phantom consists of an The flow vortex and shear stress fluctuations outer layer that mimics the fat tissue with a were hardly avoidable in pulmonary arteries. very low conductivity and an inner layer that In conclusion, the appropriate and equival- has a conductivity like the muscle. During the ent pulmonary flow could be well realized by measurements, one pair of electrodes with an a larger-sized shunt when changing the shunt insulated ground electrode was inserted into location of MBTS to the third aortic branch, the muscle layer and connected to the trans- compared with the MCS. The pulmonary flow mitter, while the other pair of the electrodes distribution is greatly affected by the shunt was placed on the phantom and connected to location and vascular anatomy. The assess-the receiver (IB2OB). After the IB2OB measure- ment and monitoring of thrombosis risk is rements were completed, the transmitter and quisite during perioperative management and receiver devices were swapped to measure the postoperative follow-up no matter what shunt OB2IB scenario. The results show that changes design is applied. in the configuration lead to negligible differ- ence in the measurement results, i.e. the sys- OR-091 tem is reciprocal. Moreover, a high-pass pro- Accurate and interpretable deep learning file was observed for the frequency range from model for sleep staging in children with 100 kHz to 84 MHz. sleep apnea from pulse oximetry Fernando Vaquerizo-Villar 1, Daniel Álvarez2, OR-090 Gonzalo C. Gutiérrez-Tobal1, Adrián Martín- Computational hemodynamic Evaluation of Montero1, David Gozal3, Eduardo Tamayo4, Different Surgical Designs of Systemic to Roberto Hornero2 Pulmonary Arterial Shunt under Controlled 1University of Valladolid, Spain Pulmonary Flow 2Universidad de Valladolid and Centro de Investiga- Jiwen Xiong, Qi Sun, Jinfen Liu, Jinlong Liu ción Biomédica en Red en Bioingeniería, Biomater- Shanghai Children’s Medical Center, School of iales y Nanomedicina, Spain 3 Medicine, Shanghai Jiao Tong University, China Joan C. Edwards School of Medicine, Marshall Uni- versity, United States Systemic to pulmonary arterial shunt in- 4Biomedical Research Networking Center in Infec- cluding modified central shunt (MCS) and tious Diseases (CIBERINFEC), Carlos III Institute of modified Blalock-Taussig shunt (MBTS) are Health, Spain widely applied surgeries to increase pulmonary perfusion. It’s crucial to ensure the pulmon- Identification of sleep stages is crucial in ary flow was well-controlled within a reason- the diagnosis of sleep-related disorders but able range, since both excessive and insuffi-relies on the labor-intensive and manual scor- cient pulmonary perfusion leads to poor pro- ing of overnight polysomnography (PSG) re-gnosis. The pulmonary flow could be well- cordings. To simplify the sleep staging pro-controlled by selecting various shunt designs, cess, deep learning (DL) algorithms have been e.g. shunt location and size. However, proposed to automatically analyze pulse rate the hemodynamic performance of different (PR) and blood oxygen saturation (SpO2) sig-designs is less explored under equivalent and nals from pulse oximetry in children with ob- reasonable pulmonary flow. Here, the in- structive sleep apnea (OSA). However, exist- dividualized vascular model was reconstruc- ing approaches are perceived as black boxes, ted. To realize appropriated and equival- limiting their implementation in clinical set-ent pulmonary flow, three virtual surgeries in- tings. Accordingly, we develop a DL architec- cluding 4 mm left MBTS, 4 mm MCS, and ture based on a U-Net to automatically per- 5 mm right MBTS were implemented using form 4-class sleep stage classification (wake, computer-aided design (CAD). The postoper- light sleep, deep sleep, and rapid-eye move-ative hemodynamic parameters were calcu- ment sleep) using entire-night PR and SpO2 lated by computational fluid dynamics (CFD). recordings. Furthermore, Semantic Segment-The results showed the postoperative mod- ation via Gradient-Weighted Class Activation els had equivalent pulmonary flow. The left Mapping (Seg-Grad-CAM), an eXplainable Ar-MBTS had a better performance in balan- tificial Intelligence methodology, is proposed cing the pulmonary flow distribution. The to provide an interpretation of the sleep scor- right MBTS had a lower wall shear stress (WSS) ing process. PR and SpO2 from 1,633 PSG and time-averaged WSS region in the shunt. recordings obtained from the Childhood Ad-enotonsillectomy Trial database were used for 82 these purposes. The U-Net model showed a ods are very different. MEG measures mag-high performance for the 4-stage classific- netic field changes at frequencies above 1 Hz ation procedure in an independent set, with and fNIRS below this frequency since vascu- 78.2?curacy and 0.696 Cohen’s kappa. The lar processes are generally slow. In this work, Seg-Grad-CAM heatmaps revealed that the we show that with OPM-MEG we can meas- PR signal has a higher contribution than SpO2 ure sustained activity of the brain (below 1 Hz) towards sleep staging, while also showing the if the measurements are made in a unique key roles of mean and variance in PR amp- magnetically shielded room with a high shield- litude, along with changes in the content of PR ing factor below 1 Hz and if we use advanced spectral bands, in the sleep staging process. preprocessing methods, such as independent These findings suggest that an explainable DL component analysis (ICA). model to analyze pulse oximetry signals could be integrated in the clinical environment for OR-093 automatic sleep staging in abbreviated pedi- Frequency Domain Cluster Analysis of Hu- atric OSA tests. man Activity Using Triaxial Accelerometer Data OR-092 Krunoslav Jurcic, Goran Šeketa, Ratko MagjareCombined measurement of brain activation vić during a motor task using fNIRS and OPM- University of Zagreb, Croatia, Croatia MEG This paper presents cluster analysis of tri- Urban Marhl 1, Stanislaw Wojtkiewicz2, Piotr axial accelerometer data aquired from various Sawosz2, Vojko Jazbinšek1, Zvonko Jagličić3, Adam human physical activities as well as simulated Liebert2, Tilmann H. Sander4 falls using frequency domain features. Clus- 1Institute of Mathematics, Physics and Mechanics, tering was performed using K Means, Gaus-Department of Physics, Slovenia sian Mixed Model and Fuzzy C-Means meth- 2Nalecz Institute of Biocybernetics and Biomedical ods. In our analysis we focused on two prob-Engineering, Poland lems: the first clustering problem being activ- 3University of Ljubljana, Faculty of Civil and Geo- ity recognition and differentiation from simu-detic Engineering, Slovenia lated human falls, while the other problem fo- 4Physikalisch-Technische Budesanstalt Institut Ber- cused on distinction between single jerk events lin, Optical Metrology, Germany (e.g. jumping, falling) and continuous activity Neurovascular coupling (NVC) serves as the signals (e.g. running, walking). fundamental mechanism supplying energy to active regions of the brain. NVC is most OR-094 easily investigated by the simultaneous ac- Electrochemical detection of desorbed cal-quisition of vascular and neuronal responses cium ions from bovine serum albumin bind-with two distinct brain imaging methods. In ing our work, we utilize two non-invasive mod- Tamara Boscarino, Antonio D’Ambrosio, Andrea alities: functional near-infrared spectroscopy Palermo, Vincenzo Piemonte, Leandro Pecchia (fNIRS) and magnetoencephalography (MEG). Università Campus Biomedico di Roma, Italy fNIRS measures cortical hemodynamic activA preliminary detection technique of cal- ity by tracking changes in oxy- and deoxy- cium ion is proposed in this work. It com-hemoglobin concentration. Meanwhile, MEG bines the calcium ions desorption from the measures cortical neuronal activity by meas- bovine serum albumin (BSA) binding in buf-uring magnetic fields near the head. Un- fer solution and the electrochemical detec- til recently the only suitable sensor for MEG tion by cyclo-voltammetry (CV). For this pur-was the superconducting quantum interfer- pose, various concentrations are tested, with ence de-vice (SQUID), but in the last dec- particular focus on physiological and patho-ade, optically pumped magnetometers (OPM) logical values of calcium levels (0 mM, 1 mM, have advanced to the point where they have 1.5 mM). Solutions at different concentrations similar sensitivities as SQUID mag-netometers. are analysed on the electrode surface by ap-In this work, we present simultaneous meas- plying a triangular input voltage. The out-urements of brain activa-tion during a motor put voltage is represented as a function of task with an integrated system consisting of the applied voltage by creating a cyclic wave-both modalities: fNIRS and OPM-based MEG. form, the cyclic voltammogram. Focusing on Traditionally, the time scales of the two meth-83 the closed-loop curve, the measured voltage these techniques to optimize bone tissue scaf-value progressively increases as the concen- fold designs. tration of calcium ions in solution rises, particu- larly in the range of interest between 0.3V and 0.7V with a Limit of Detection (LOD) of 0.08V. S24 - Wireless Wearable The promising results indicate effective per- Networks for Physical Function formance. The forthcoming challenge involves Rehabilitation Monitoring utilizing a screen-printed electrode with a car- bon/calcium ionophore working electrode. Tuesday late afternoon OR-095 Track C Computational analysis of large bone de- Jun 11, 16:00 - 18:00 fect healing using bone tissue scaffolds, de- gradation, and growth factor delivery: a OR-096 mechanobiological model of bone tissue Revealing Statistical Patterns in Shoulder formation Rehabilitation Exercises Characteristics Adel Alshammari, Fahad Alabdah, Lipeng Song, Martina Sassi, Margherita Anna Grazia Matar- Glen Cooper rese, Umile Giuseppe Longo, Leandro Pecchia University of Manchester, United Kingdom Campus Bio Medico University of Rome, Italy, Italy Large bone defects are a significant med- Tele-rehabilitation has the potential to ical challenge. 3D printed synthetic biocom- transform the way patients are monitored patible and biodegradable bone tissue scaf- from home, overcoming geographical barri-folds offer a possible clinical solution. De- ers, enhancing accessibility, and promoting gradability of these bone scaffolds within the patient autonomy. The development of a body precludes the need for additional in- tele-rehabilitation system capable of auto- terventions allowing native bone to replace mating the recognition of performed exercises the implant during healing. Degradation rates may significantly impact rehabilitation out-can be altered by chemical composition but comes. Implementation of machine learning the optimal design of these rates to maxim- algorithms combined with magneto-inertial ize bone growth is unknown. Many research- measurement units (M-IMUs) has enabled ers have conducted experiments on degrad- remote home-based rehabilitation therapy ation, but this is time consuming and difficult through wearable systems. Thus, in this study to measure particularly in vivo. Mechanobi- sixteen healthy participants and sixteen pa-ological models to evaluate degradation and tients with rotator cuff injuries were enrolled bone healing for bone tissue scaffolds would to perform six shoulder rehabilitation exercises be an attractive option to design bone tissue while wearing a wearable system based on scaffold degradation but these are not widely three M-IMUs. This study aimed to conduct available. This study aims to model both bone a thorough analysis of the features extracted healing and scaffold degradation for a bone from time-series data collected by these three tissue scaffold integrating agent-based mod- sensors during these exercises. The statist- elling and finite element analysis (FEA). It intro- ical analysis indicated statistically significant duces an innovative computational method to differences in task features, but not between examine the effects of scaffold degradation participant groups. Three features, identi- rates (slow, medium, and fast) and the incor- fied as the most representative and distinct-poration of embedded bone morphogenetic ive among all tasks, were subsequently, used proteins (BMPs) on bone regeneration. The to train the Support Vector Classifier in classi- results indicated that scaffold degradation at fying the six exercises. The obtained classific- a medium rate of 6% volume reduction per day ation results are promising for the application resulted in the greatest volume of regener- of this wearable device in remote monitoring ated bone by 18.2mm3, in contrast to slow and of patients with shoulder musculoskeletal dis- fast degradation by 16.4mm3 and 18.1mm3, re- orders during home-based rehabilitation ex-spectively. The case using medium scaffold ercises. Further studies will involve the imple-degradation with embedded BMPs enhanced mentation of the Principal Component Ana- bone regeneration, resulting in a further in- lysis (PCA), along with the training of additional crease of 2 mm3 of bone volume. Further work machine learning models. is planned to calibrate the model and to apply 84 OR-097 proach included different signal labelling de- Body Movement Analysis during Sleep pending on its continuity, exploring the pos-Based on Ultra-Wideband Communication sibility of recognizing and separating a single Channel Impulse Response Measurement occurrence (e.g. falling or jumping) event from Ivana Culjak 1, Željka Lučev Vasić2, Domagoj a continuous activity such as running or walk- Vidović3, Mario Cifrek2 ing, questioning whether we can gain new in- 1KITE-Toronto Rehabilitation Institute, University sights in the field of human activity recogni-Health Network, Canada tion and fall detection. Clustering was per- 2University of Zagreb, Croatia formed using three different clustering meth- 3University Psychiatric Hospital Vrapče, Croatia ods: K Means, Gaussian Mixed Model and Fuzzy The study introduces an innovative non- C-Means clustering. For the time domain fea- invasive method using Ultra-Wideband (UWB) ture models’ performance, Gaussian Mixture technology for monitoring body movements Model (GMM) proved to perform best regard-during sleep. Conducted at the Center ing both the first (89?curacy, 92? score for for sleep and wake disorders, the research ADL and 86? score for falls) and the second involved positioning UWB devices with two clustering problem (68?curacy, 64? score for transmitters and receivers on either side of the continuous activities and 70? score for single bed, approximately 0.5 meters from the pa- events). As for the frequency domain mod- tient. Ground-truth measurements were done els, there is no significant differences in per- using a polysomnography (PSG) system. The formances among the clustering methods. K analysis was conducted in 30-second epochs, Means and Fuzzy C Means models performed involving 17 participants over 103.41 hours. This with slightly higher results for the first clustering method leverages UWB’s Channel Impulse Re- problem (77?curacy and 85 ? score compared sponse (CIR) to monitor body positions and to 75 ?curacy and 80? score of GMM), while the movements, serving as an alternative to tra- models perform similar regarding the second ditional radar-based approaches. The re- clustering problem. search highlights UWB’s potential in under- standing sleep patterns and quality, showing OR-099 the interplay between body movements, sleep Dynamic Equivalent Circuit Models for phases, and respiratory events. This study Intracardiac Communication in Leadless can contribute to sleep medicine by offering a Pacemakers new method for sleep disorder management, Dongming Li 1, Jiamei Wang2, Pedro Antonio Mou1, Yadong Yin2, Hung Chun Li3, Yueming Gao2, marking progress in non-invasive sleep monit- Peng Un Mak1, Sio Hang Pun1, Mang I. Vai1 oring technologies. The comparison with con- 1University of Macau, China ventional polysomnography provides valuable 2Fuzhou University, China insights into the efficacy of UWB technology in 3Lingyange Semiconductor Inc., China improving sleep monitoring. The leadless multichamber pacemaker OR-098 offers significant advantages in treating ar- Comparison of Time and Frequency Do- rhythmias and enhancing cardiac function. main Features For Cluster Analysis of Human Intracardiac communication is crucial for en-Activity Using Triaxial Accelerometer Data abling synchronous pacing of multi-chamber Krunoslav Jurcic non-leadless pacemakers. However, the University of Zagreb, Croatia dynamic changes in the intracardiac chan- This abstract presents a summary of two nel present a significant challenge. Hence, studies focusing on comparison of the impact mastering the dynamic characteristics of in- of both time and frequency domain features tracardiac channels and accurately modeling in analysis of tri-axial accelerometer data ac- them is crucial. In this paper, we propose an quired from various human physical activities intracardiac channel equivalent circuit model, as well as simulated falls. The goal was to get which is based on the electrical properties insights on the performance of several cluster- of cardiac tissue and the volume changes ing algorithms when dealing with different hu-of cardiac compartments, and is experi- man activities, and whether they can distin- mentally validated. The results demonstrate guish activities of daily living from falls, treating that the agreement coefficient between the it as a binary clustering problem. Another ap-circuit model and the in vitro pig heart model reached 80.7 85 OR-101 Impact of the Tightness of a Sensorized Top OR-100 in the Quality of the ECG it Records Cardiac Phantoms for Signal Transmission Marta Guzmán-Alarcón, Natalia Guzmán Characteristics Analysis of Leadless Pace- Rodríguez, Pablo Pérez-Tirador, Constantino A. makers García, Abraham Otero Han Wang1, Jiamei Wang1, Yang Shuang 1, Zili- University CEU San Pablo, Spain ang Wei1, Željka Lučev Vasić2, Jiejie Yang1, Zhimeng The telemonitoring of physiological para- Xu1, Mario Cifrek2, Yueming Gao1 meters can provide useful information for 1Fuzhou University, China chronic diseases management and for the 2University of Zagreb, Croatia identification of life-threatening situations. The emergence of leadless pacemakers However, to achieve continuous adherence marks a revolutionary breakthrough in cardiac to the technological solution, it must be rhythm management. However, current lead-comfortable and easy to use. To this end, less cardiac pacemakers (LCPs) can only pace the sensorization of commercial garments at a single location, which fails to meet the through the integration of conductive textiles demand for multi-ventricular sequential pa- is a promising solution. In a previous work cing in most patients. To address this limita- three sensorized garments (pants, shirt and tion and achieve multi-chamber pacing with top) to record ECG were compared [1]. The LCPs, it is essential to analyze the signal trans- top provided a higher quality, but it also was mission characteristics of intracardiac con-the tightest. However, since three different duction communication.Many existing studies garments were used, it is difficult to draw analyze the signal transmission conclusions about the impact of the tightness characteristics of intracardiac channels on the ECG quality. based on static experimental measurements. This work studies the impact of the tightness However, our previous research revealed that of a sensorized top on the quality of the ECG. the signaling characteristics of intracardiac Three sensorized tops of different sizes (M, channels vary with the cardiac beat cycle. S, XS) were built from the same commercial Consequently, it is inaccurate to consider the garment. The choice of sizes was based on channel gain measured in any arbitrary state the test subject (20-year-old woman, 1.59 m, as representative of the channel character- 47 kg). Size S would be her natural size, with M istics during the establishment of intracardiac being slightly loose and XS being too tight and signaling communication. Thus, the meas- somewhat uncomfortable. urement results cannot be reliably used as A silver-plated conductive fabric (Adafruit a reference for LCP design.In this study, we Industries LLC; New York, USA) and a steel developed a semisolid spherical mimic us- conductive thread (SparkFun Electronics ing a mixture composed of 91.94?ionized wa- Inc; Colorado, USA) were used for the senter, 2.63% agar, 0.18% potassium chloride, sorization. Using the conductive thread, a and 5.25% hydroxyethylcellulose (HEC). This connection was created between the elec- mimic was designed to simulate the chan- trodes and the data acquisition unit (Plux nel characteristics during the establishment Wireless Biosignals S.A.; Lisbon, Portugal). The of intracardiac signal conduction communic- data was recorded with a mobile terminal. ation. The experimental measurements of Four 8-hours recordings were made for each the mimic demonstrated a difference of no garment while performing daily living activ- more than 2 dB compared to the dynamic ities. The mSQI ECG quality index [2] was measurements of the isolated porcine heart used to compare the recordings. The average in the frequency range of 1 MHz to 20 MHz. value for the 4 days (approximately 36 hours) This mimic offers a highly accurate, reprodu- of msQIs in the M top was 0.55±0.14, in the cible, and stable method to analyze the sig-S 0.74±0.08 and in the XS 0.71±0.08. These nal transmission characteristics of intracar- values suggest that it is possible to achieve a diac channels in conductive cardiac commu-balance between the quality of the signal and nication(CIC) studies. Moreover, it provides an the comfort of the garment: while the excess- effective reference for CIC transceiver design ively loose top (size M) does produce a signal studies. of a markedly lower quality, the quality of sizes S (the subject’s natural size) and XS (too tight) is equivalent, presenting no increase in quality 86 associated with the tighter garment. or time series causality methods (e.g. Granger causality) to predict the occurrence of critical [1]A. Grech, C.A. García, A. Otero Evaluation health incidences. of three sensorized garments with conductive In our laboratory we used ICG derived textiles for ECG recording. CASEIB 2023, Spain. stroke volume and systolic time intervals [2]A. Grech, C.A. García, P. Perez-Tirador, A. (STI) changes (including their variability) to Otero. A metric index to assess the quality analyze the hemodynamic response to active of wearable ECG recordings. IEEE BSN, 2023 and passive orthostatic maneuver, handgrip, Boston, EEUU. dynamic exercise, Valsalva maneuver. We applied our described earlier ambulatory OR-102 impedance cardiography wearable system Impedance Cardiography signals in applic- (Reomonitor). This device was constructed for ations for cardiovascular information con- non-invasive acquisition of central hemody-tent analysis namic data during everyday activity. Stroke Gerard Cybulski, Tadeusz Pałko, Kazimierz volume was evaluated using the Kubicek Pęczalski formula. The validation of impedance car- Faculty of Mechatronics, Warsaw University of diography has been performed many times. Technology, Poland Reomonitor was verified using echocardio- The analysis of cardiovascular system graphy in both the supine and tilted position. response to various physiological stimuli in- Some studies has shown the usefulness of cluding active orthostatic maneuver, head-up ICG in monitoring the effects of neurode- tilt test, handgrip or dynamic exercise may generative processes, analysis of sleep apnea provide some supplementary diagnostic data disorders, monitoring the effectiveness of on autonomic control activity and might physical training and predicting orthostatic help to predict the occurrence of some crit- intolerance. ical events, e.g. orthostatic syncope. Heart The aim of this study is to deliver the review of rhythm/rate variability (HRV) has been in- some possibilities and prospects of extract- tensively studied in supine and other body ing diagnostic information from analysis of positions in humans. However, only lim- impedance cardiography signals (ICG). ited study results were published describing rhythms, fluctuations and generally dynamics OR-103 of hemodynamic parameters, especially mon- An Everyday Hat for Detection of Eye Blinks itored during transient phase of physiological and Forehead Clenching tests. Moreover, they were in much smaller S. M. Musfequr Rahman, Henna Mattila, Asif groups of subjects. Impedance cardiography, Shaikh, Pasi Raumonen, Johanna Virkki which allows non-invasive, continuous mon-Tampere University, Finland itoring of mechanical activity of the heart This paper presents an early prototype of (including ambulatory conditions) seems to an everyday hat capable of detect-ing eye be the method suitable to perform this type blinks and forehead clenching and turning of analysis. them into inputs for digital actions. In fabrica- The analysis of cardiovascular hemodynamics tion, commercially available standard dry and appears to be a potential source of diagnostic semi-dry Ag/AgCl electrodes are compared to information on mutual coupling between dry and semi-dry electrodes fabricated from the cardiovascular system and autonomic copper and silver-based textiles and silver-nervous system dysfunction or modification based thread. According to measurements induced by several environmental and/or from six participants, the created hat proto-pathological factors So far researchers used type with integrated electrodes is fully com- either descriptive statistics method or power parable to a commercially available electrode spectrum analysis for analysis of absolute holder band, while the copper-based textile values, their changes of stroke volume (SV), electrodes have detection accuracy for eye stroke volume output velocity (SVOV), ejec- blinking and forehead clenching comparable tion time (ET), pre-ejection period (PEP) and to the standard Ag/AgCl electrodes. Among additional systolic time intervals or other all materials, the semi-dry electrodes have parameters derived from ICG signals. They better detection accuracy than the dry elec- also tried to apply measures of complexity trodes. Finally, a proof-of-concept applica- (sample entropy, approximate entropy, etc.) tion of the hat, turning light on/off by eye blinks 87 and forehead clenching, reveals more than based on previous quality control data collec-80 % success rate, when tested by four parti- tion and analysis results, leveraging the pro- cipants. fessional resources of experts at the provin- cial medical equipment and instrument man- agement quality control center. The 2022 edi- S09 - Health Technology tion of the quality control indicators has been Assessment of Medical Devices - optimized from the 54 items in the 2018 edi- Advances and Challenges tion (13 quarterly reporting indicators and 41 annual indicators) to 15 items (9 quality con- Wednesday morning Track trol indicators reported each quarter and 6 A quality control indicators reported each year). Conclusion Since the release of the Jiangsu Jun 12, 10:30 - 12:00 Province Medical Equipment and Instrument Management Quality Control Indicators (2022 OR-104 Edition) in November 2022, after specialized Exploration and practice of quality control training through online and offline quality con-index system of medical device manage- trol meetings, 4 sets of quarterly data rement ports and 2 sets of annual data reports have Yueqi Yang 1, Yingxin Xu1, Xin Li1, Xiangnan been completed, with overall positive feed- Geng1, Xiang Xu1, Hong Gao1, Yuzhi Yang2, Wei Jin3, back. Quality control entities can compare Hui Zhong4, Ying Qian1 their reported quality control data with regu- 1Jiangsu Province Hospital, China larly published quality control reports to pos- 2Nanjing Drum Tower Hospital, China ition their own quality control work and carry 3Wuxi People’s Hospital, China out continuous improvements. 4Subei People’s Hospital of Jiangsu province, China Objective Quality control indicators are an OR-105 important component of the medical quality Cost-effectiveness of Continuous Gluc- management and control system and are cru- ose Monitoring For Paediatric Patients cial tools for quality management. It is essen- With Type 1 Diabetes Compared With tial to explore the establishment of a quality Self-Monitoring Of Blood Glucose control indicator system for the management Martina Andellini 1, Riccardo Schiaffini1, Matteo of medical equipment and instruments to en-Ritrovato1, Leandro Pecchia2 1 sure the safety and effectiveness of clinical Bambino Gesù Children’s Hospital, Italy 2 applications of medical devices. Method The Università Campus Biomedico di Roma, Italy selection and establishment of quality con- BACKGROUND Type 1 diabetes (T1D) is one trol indicators should follow three principles: of the most common chronic disease in pedi- scientificity, standardization, and operability. atric age. Frequent blood glucose measures This can be achieved through extensive re- are required to reduce the risk of severe ad-search, expert consultation, and centralized verse events and the onset of long-term com-discussions. The selected indicators should plications. Continuous glucose monitoring cover multiple dimensions such as structure, (CGM) has been showed to improve glycemic process, and outcomes. The consultation for control in adults compared to finger prick medical equipment and instrument manage-measures (SMBG), however scarce evidence is ment quality control indicators should include available on pediatric patients. This study aims experts in the field, as well as experts from at determining the cost-effectiveness of CGM related professions within medical institutions compared to SMBG in a pediatric population. (such as medical, nursing, hospital manage- METHODS A four states Markov Model (T1D, ment, etc.), and opinions from competent au- ketoacidosis, cerebral injury, death) was de- thorities (such as health commissions at vari- veloped by using data inputs retrospectively ous levels). For some special technical indic- collected at Bambino Gesù Children’s Hospital. ators, opinions and suggestions from repres- Data from patients who have been using CGM entative equipment and instrument manufac- for at least three years were included in the turers should also be sought. Results The Ji-analysis and compared with patients who use angsu Province Medical Equipment and Instru- SMBG. Hba1c level, glycemic variability and ment Management Quality Control Indicators quality-of-life were estimated by administer- (2022 Edition) have been revised and improved ing ad hoc questionnaires to evaluate clinical 88 benefits. Transition probabilities were estim- model to predict possible LSM inaccuracies ated by retrospectively collecting data from in a cohort of 509 HCV-positive treated pa-hospital registries and the costs from avail- tients. The dataset, characterized by 55 variable literature and national sources. The ana- ables, underwent feature reduction and bal- lyses were performed by simulating 100 pa- ancing to mitigate class imbalance to train the tients over 16 years horizon. The main outcome predictive algorithm. The developed model was cost per quality-adjusted life-year (QALY) can identify inaccuracy in LSM and achieves gained. an accuracy of 88.0% on the training set and RESULTS Results showed that all health out- 92.0% on the test set. Furthermore, it exhib-comes improved during follow-up when pa- ited a consistent mean Area Under the Curve tients used CGM instead of SMBG. Namely, the One-vs-One (AUC-ovo) of 0.97 across both number of hospitalizations, DKA and cerebral datasets. The model’s performance in pre- edema reduced for patients who use CGM as dicting abnormal LSM may enable healthcare a consequence of a better T1D management. providers to tailor treatment plans more pre-Managing T1D with CGM was associated with cisely, optimize patient follow-up, and reduce lower disease-related costs and higher utilities unnecessary invasive procedures. These find- as compared to SMBG. The cost per quality- ings highlight the potential of machine learn-adjusted life year gained by using CGM is ap- ing in improving patient care in the context of proximately 12,373.54€/QALY, that represents chronic HCV management. the expected cost per QALY by relying on the CGM instead of the SMBG. OR-107 CONCLUSIONS Each patient who manages Advances in Health Technology Assessment T1D with CGM cost 43,726.91€ more than pa- of Wearable Devices: A Clinical Engineering tients who use SMBG over 16 years, with an in- Perspective crement of 3.53 QALYs. Neverthless the higher Fabiola Martinez-Licona initial cost of CGM is counterbalanced by the Universidad Autonoma Metropolitana, Mexico health benefits which lead to the reduction Wearable devices have revolutionized of the number of severe acute events, with a healthcare delivery by continuously monitor- consequent reduction of the related complic- ing vital signs and other health parameters in ations costs resulting in cost-effectiveness. real time. As these devices become increas- ingly sophisticated, rigorous health tech- OR-106 nology assessment (HTA) to evaluate their Optimizing Liver Stiffness Assessment in effectiveness, safety, and cost-effectiveness HCV Patients: A Machine Learning Ap- becomes paramount. Here, we explore recent proach to Identify Confounding Factors in advances in HTA methodologies specific to Fibrosis Estimation wearable devices from a clinical engineering Simone Kresevic 1, Mauro Giuffrè2, Milos Ajcevic1, perspective. Lory Croce1, Agostino Accardo1 First, an overview of wearable devices and 1University of Trieste, Italy their growing importance in healthcare is 2Yale School of Medicine, Yale University, Italy provided, highlighting their potential to im- Hepatitis C Virus (HCV) infection is a sig- prove patient outcomes, enhance remote pa- nificant global health con- cern with approx- tient monitoring, and facilitate early disease imately 1.5 million new infections yearly. The detection. It then delves into the challenges choice of the most appropriate HCV treat- of assessing wearable devices, including data ment depends on several factors, including accuracy, privacy concerns, and regulatory liver fi- brosis status. Current guidelines re- requirements. commend liver fibrosis evaluation using non- Next, a brief exposition of the current state invasive techniques such as Liver Stiffness of HTA methodologies for wearable devices Measurement (LSM) using liver elas- tography. emphasizes the importance of multidisciplin-Although LSM revolutionized patient care in ary collaboration between clinical engineers, the last decade, allowing biopsy-free treat- healthcare providers, and researchers. The ments, several factors can lead to overestim- role of clinical engineering in designing robust ation or underesti- mation of liver stiffness val- evaluation protocols, conducting clinical trials, ues, affecting management strategies. This and analyzing real-world data to assess the study pre- sents a machine-learning ap- clinical utility and value of wearable devices is proach using an eXtreme Gradient Boosting 89 discussed. this paper. Furthermore, the abstract explores emer- ging trends in HTA, such as the integration of OR-109 artificial intelligence and machine learning Exploring the Potential of Health Data: algorithms to analyze wearable device data, EHDS, Secondary Utilization, and Stake-predictive modeling techniques to forecast holder Perspectives in Czech Healthcare health outcomes, and patient-centered ap-Petra Hospodková, Martin Budil proaches to HTA that incorporate patient Czech Technical University in Prague, Czech Repub-preferences and experiences. lic Finally, future directions and challenges in HTA The European Health Data Space (EHDS) of wearable devices, including the need for introduction has raised a discussion on the use standardized assessment frameworks, valid- of health data (HD) for secondary purposes ation studies to establish the reliability and that could essentially improve health systems accuracy of wearable device data, and long- efficiency. This research examines how key term outcomes research to evaluate their stakeholders in the Czech Republic feel about impact on population health and healthcare this process. In qualitative analysis (interviews), delivery, are discussed. two major themes were identified with regard to the use of health data: Risks and Barri- OR-108 ers, as well as Health Data Access theme. We Alarms Early Detection in Dialytic Therapies investigated views and perspectives of 8 key via Machine Learning Models stakeholders from healthcare sector (insur- Alessia Nicosia 1, Nunzio Cancilla1, Marco Siino1, ance companies representatives, Ministry of Michele Passerini2, Francesca Sau2, Ilenia Tinnirello1, Health, IT technologies, data applicants etc.). Andrea Cipollina1 1 These opinions were analyzed using software Università degli Studi di Palermo, Italy 2 MAXQDA 24 and compared to better under- Mozarc-Medical company, Italy stand the challenges and benefits associated Hemodialysis (HD) is a clinical treatment for with using health data for secondary purposes. patients affected by Chronic Kidney Dis-ease According to our study, the benefits surpass (CKD). The goal of a treatment is to purify the risks and barriers shown by secondary util- the patient’s blood using dialysis machines, ization of health data. The stakeholders highly devices that act as artificial kidneys. However, appreciated making this data accessible to a common problem is the alteration of the pa-enhance quality care and efficiencies in the tient’s health status due to side effects or to healthcare system. Therefore, this study’s out-machine malfunctions that may occur during comes are an aid in understanding and pro- treatment. A dialysis machine is a complex ap- motion of use of health data within EHDS con- paratus consisting of a control system of sev- fines. Furthermore, they will be valuable not eral quantities (e.g., pressure, flow rate, tem- just to researchers but also policymakers, busi-perature, conductivity, etc.) capable of alert- nesses, and the public interested in harnessing ing medical operators when an alarm occurs. health data to enhance the healthcare sys-In the present work, a Machine Learning (ML) tem. predictive model able to act in advance with respect to the dialysis alarm system was de- veloped. Several machine learning mod-els S22 - Tools, technologies and were tested and a comparison study was car- computing for point-of-care ried out. Datasets for training and testing the person-centered health and mod-els came from treatments performed by care delivery dialysis machines manufactured by Mozarc- Medical®. Among the models tested, the Ran- dom Forest (RF) classifier was identified as the Wednesday morning Track more promising one and was then used to per- B form a parametric sensitivity study. By using Jun 12, 10:30 - 12:00 a time window of 10 seconds, the RF model provided a Recall of 79% and an F1-Score of up to 85% on test data, demonstrating the good generalization ability that is always required by predictive models such as the one analysed in 90 OR-110 OR-111 Pointwise reliability of machine learning Towards Wearable Continuous Point-of-models: application to cardiovascular risk Care Monitoring for Deep Vein Thrombosis assessment of the Lower Limb Jorge Henriques1, Teresa Rocha 2, Simão Eleni Kaldoudi 1, Vaidotas Marozas2, Pavlos Paredes2, Paulo Gil3, João Loureiro1, Lorena Moustakidis3, Dmitry Novikov4, Jurkonis Rytis2, Petrella1 Andrius Sakalauskas4, Nicolas Pousset5, Math- 1University of Coimbra, Portugal ieu Legros5, Marco Kircher6, Babajide Ayinde3, 2Polytechnic Institute of Coimbra, Coimbra Institute Lara Alessia Moltani7, Susann Balling8, Antti of Engineering (IPC/ISEC), Portugal Vehkaoja9, Niku Oksala9, Andrius Macas10, 3CTS, Dep. of Electrical and Computer Engr. NOVA Neringa Balciuniene10, Maria Bigaki11, Michail Potoupnis11, Stella-Lida Papadopoulou11, Elvira School of Sci. and Technology, Portugal Grandone12, Maxime Gautier13, Sabrina Bouda13, Machine learning has made significant ad- Cord Schloetelburg14, Thorsten Prinz14, Pietro vances in many areas, particularly in the Dionisio15, Spyridon Anagnostopoulos16, Ioanna healthcare domain. However, despite the ad- Drougka16, Frans Folkvord17, George Drosatos1, vances, the implementation of these models in Stylianos Didaskalou1 1 clinical scenarios is still limited due to several Athena Research Center, Greece 2 challenges, including the lack of trust. Stand- Kaunas University of Technology, Lithuania 3 ard performance measures, such as sensitiv- EchoNous Inc, United States 4 ity, specificity and confidence intervals can be TELEMED UAB, Lithuania 5 used to evaluate the reliability of a model, but VERMON SA, France 6 these are overall performance metrics and do Fraunhofer Institute for Photonic Microsystems, not provide insight into the performance of Germany 7 individual instances. Moreover, these estim- ComfTech SRL, Italy 8 ates are typically calculated during the train- medis Medizinische Messtechnik GmbH, Germany 9 ing phase and are not easily generalized to Tampere University, Finland 10 new, unseen instances, occur-ring in the de- Lietuvos Sveikatos Mosklu Universitetas, Lithuania 11 ployment phase. Papageorgiou General Hospital, Greece 12 As result, besides the prediction outcome, the Fondazione Casa Sillievo Della Sofferenza, Italy 13 existence of a measure of reliability in the pre- Groupement Hospitalier Eaubonne Simone Veil, diction of individual estimations would add a France 14 layer of security, increas-ing trust in human- VDE Verband der Elektrotechnik Elektronik In- AI interaction, as well as it could also be help- formationstechnik eV, Germany 15 ful to support the improvement of the model. Medea SRL, Italy 16 This study proposes a reliability measure, com- Phaze Clinical Research and Pharmaceutical bining density and local fit principles, to estim- Consulting SA, Greece 17 ate the confidence of individual predic-tions PredictBY Research and Consulting SL, Spain in the deployment phase. When applied to a Deep vein thrombosis (DVT) is the forma- machine learning model in the cardiovascular tion of a blood clot within the deep veins, most risk assessment context, the method demon-commonly those of the lower limbs, causing strates the ability to distinguish between reli- obstruction of blood flow. In 50% of people able and unreliable predictions, as well as aid-with DVT, the clot eventually breaks off and ing in the stratification of the population. travels to the lung to cause pulmonary em- bolism. Clinical assessment of DVT is notori- ously unreliable because up to 2/3 of DVT episodes are clinically silent and patients are symptom free even when pulmonary embol- ism has developed. Early diagnosis of DVT is crucial, and despite the progress made in ul- trasound imaging and plethysmography tech- niques, there is a need for new methods to enable continuous monitoring of DVT at the point of care. This paper presents the con- ceptual design and methodology towards a novel wearable diagnostic device for point- 91 of-care, operator-free, continuous monitor- Focusing on key pacing sites, including the ing in patients with high DVT risk. The device right ventricular (RV) lateral, apical, and septal will combine novel wearable hardware for ul-zones, alongside intrinsic waveforms, our trasound imaging and impedance plethysmo- graphical user interface (GUI) facilitates the graphy with autonomous, AI driven DVT detec- precise definitions of time intervals, providing tion, to allow continuous monitoring for blood a valuable tool for investigating temporal clot formation in the lower limb. Activity and variations in ECG amplitudes. Observations other physiological measurements will be used based on the heatmap morphologies enable to provide a continuous assessment of DVT risk distinct identifications of ventricular depolar- and guide the automated scanning via an in- izations and timings. Shown here, traditional telligent decision support unit that will provide RV apex pacing exhibits a dual RV-LV depolar-accurate monitoring and alerts. The work is ization, while RV lateral wall pacing manifests supported by the Horizon project ThrombUS+ wider QRS morphology with a right-dominant co-funded by the European Union. (Grant depolarization. Conversely, RV septal pa- Agreement No. 101137227). cing resembles intrinsic signals but displays a fragmented RV depolarization. These signals OR-112 have the potential to democratize 12 lead Attention Theory Based 12 Lead Visu- ECG interpretations and improve procedure alization of Ventricular Function During efficiencies towards more physiologic pacing. Ventricular Pacing Lead Implants Beyond potential improvements in procedure Neal Duong 1, Jessica Lee2, Paul A. Iaizzo1 times, our developed QRS representation 1University of Minnesota, United States could also provide insights as to relative ana- 2Arizona State University, United States tomic locations, hence reducing reliance on Optimal cardiac pacing lead implantation fluoroscopy during lead placement proced-is a critical yet complex procedure neces- ures. Thus, such could not only contribute to sitating the electrophysiologist’s attention enhanced patient outcomes and reduced to ensure accurate placement, appropriate procedure times, but also address concerns hemodynamic responses, and an elicited regarding ionizing radiation exposures to both electrical morphology confirmation for im-patients and clinicians. Our study underscores provement of patient outcomes. This study the evolving landscape of cardiac inter-introduces an innovative, low-computational ventions, emphasizing a balance between methodology based on attention theory technological advancements and patient- to transform the conventional 12-lead EKG centric care. Overall, this research could format into a more readily visual response. significantly benefit implanter’s precision, Leveraging both object-based theory and as well as the safety of cardiac proced- the proximity compatibility principles, the ures, ultimately enhancing patient care in developed method converts the visually cardiovascular medicine. separated V1-V6 precordial leads’ wave mor- phologies and amplitude information into a OR-113 colormap format; to condense the complex Integration of in-vitro technologies, in-silico data into a more observable signature of biophysical and data driven models towards ventricular function. better stratification and care for hyper- Implemented in Python algorithms, our trophic cardiomyopathy patients method offers a dynamic ECG data visualiza- Jari Hyttinen tion, utilizing a notch filter for signal cleaning Tampere University, Finland and converting absolute lead values into a Hypertrophic cardiomyopathy (HCM) is the responsive color map. The resulting heatmap most common inherited cardiac disorder. It representation allows the implanting elec- is a multifactorial and multiorgan disease trophysiologist to identify subtle changes in with multiple manifestations, giving rise to a ongoing pacing morphologies more efficiently range of symptoms and complications includin a clinical environment. The colormap, with ing chest pain, arrhythmias, and sudden car- high amplitudes depicted in yellow and lower diac death. It has various comorbidities im-amplitudes in red along a gradient, offers pacting e.g., the circulatory system and mula visually intuitive means for the real-time tiple physiological organs, which may lead to assessment ventricular functions. stroke or many neurological and psychological 92 complications. It affects all age groups from OR-114 children to the elderly, it is a leading cause Development of an Explainable Deep of death among young athletes. Implantable Learning-Based Decision Support System cardiac defibrillators are used for HCM pa-for Blood Glucose Levels Forecasting in tients to prevent sudden cardiac death. Mul- Type 1 Diabetes Using Edge Computing tiple causal mutations and variations in cel-Isabel Longo1, Federico D’Antoni 2, Lorenzo lular processes lead to highly diverse phen- Petrosino1, Vincenzo Piemonte1, Mario Merone1, otypes and disease progression that are not Leandro Pecchia1 1 fully understood. HCM is still diagnosed as Università Campus Bio-Medico di Roma, Italy 2 one single disease, leading to suboptimal care. Fondazione Policlinico Universitario Campus Bio- Better tools for deep phenotyping of patients Medico, Italy are needed to produce better assessments of Maintaining blood glucose levels risk and care options. within the euglycemic range to avoid The SMASH-HCM EU HORIZON-HLTH-2023- hypo/hyperglycemic events is a very im-TOOL-05-03 project will develop and valid- portant and complex challenge for people ate a digital twin platform to support patient with Type 1 Diabetes. Current solutions em-self-management and aid clinicians in de- ploy complex deep learning models requiring livering optimized cost-effective stratification substantial computational resources, often and patient management strategies tailored necessitating cloud-based processing, raising to individual patient pathophysiology, geno-privacy concerns and significant CO2 emis- type and phenotype. SMASH-HCM unites 8 sions. Furthermore, the ’black-box’ nature of research partners, 3 hospitals, 3 SMEs, and a these neural networks obscures the reasoning global health-technology corporation in col- behind their predictions, impeding user trust laboration with patients to advance the state and understanding. of the art in human digital-twins. This paper introduces a decision support SMAASH_HCM integrates in-vitro cell and tis- system leveraging a Long Short-Term Memory sue models, in-silico modelling from molecu- (LSTM) neural network for glycemic fore- lar manifestation of the disease to patient casting that utilizes solely Continuous Gluc-cellular, tissue, whole heart and finally car- ose Monitoring (CGM) data, using edge-diovascular systemic level models. The in- computing for overcoming some of the above vitro and in-silico model insight on the dis- limitations. We present a streamlined model ease and drug mechanisms and their data with an architecture optimized for edge and structured clinical and unstructured data devices, ensuring data privacy and reducing analysis will lead to data-based models with CO2 emissions by eliminating the need for emphasis on explainable artificial intelligence. transmitting and storing more than needed All these will be integrated into a three level sensitive data. The system encompasses a decision support solution for both healthcare full communication pipeline from CGM data with deep phenotyping and risk stratification acquisition and transmission to on-device tools to support clinical workflows and de-prediction, culminating in the display of results cision making enabling the development of for patient consultation. effective personalized treatment and for the Our findings show that the model’s perform- patients for their self-management. SMASH- ance, with an average Root Mean Square Error HCM will provide tools for detailed personal- of 14.22 mg/dL and Clarke Error Grid Analysis ized assessment of the need, risks, and be-with 91.95% of predictions in zone A and nefits involved, making the identification of 7.05% in zone B for a prediction horizon of 30 the need for ICDs substantially easier. In ad- minutes, is on par with current high-standard dition, SMASH-HCM presents a well-defined models. Crucially, we integrate a renowned pathophysiology and servers as a paradigm to explainability algorithm to elucidate the pre-understand the broader spectrum of hyper- dictive process, offering valuable insights into trophy, translating to other myopathies and the model’s decision-making framework. This cardiac diseases and beyond. In reaching its transparency aims to bolster user confidence goals, SMASH-HCM has potential to serve as a and facilitate a deeper understanding of basis for future digital-twin platforms for other the predictive outcomes, marking a signific-cardiac diseases integrating models and data ant advancement in personalized diabetes from various scales and sources. management. 93 OR-115 project). Views and opinions expressed are Enhancing Balance Rehabilitation through however those of the author(s) only and do use of Augmented Reality not necessarily reflect those of the European Aleksandra Vulović 1, Đorđe Ilić2, Nenad Union or the granting authority: European Filipovic1 Health and Digital Executive Agency (HADEA). 1University of Kragujevac, Serbia Neither the European Union nor the granting 2Bioengineering Research and Development Cen- authority can be held responsible for them. ter (BioIRC), Serbia Authors also acknowledges the funding by the Falls bear profound physical, psychological, Ministry of Science, Technological Develop-and societal repercussions alongside imposing ment and Innovation of the Republic of Serbia, substantial healthcare costs and diminishing contract number [451-03-65/2024-03/200107 quality of life. Among the primary risk factors (Faculty of Engineering, University of Kraguje-contributing to falls is impaired balance, stem- vac)]. ming from diverse conditions such as aging, stroke, or vestibular disorders. Balance rehab- ilitation is one of a crucial measure to prevent S19 - IFMBE Industry Committee: falls and enhance balance function, and one Medical Technologies - From possibility to improve that rehabilitation pro- Concept to Commercialization cess is through telerehabilitation. Telerehab- ilitation allows the utilization of technology to Wednesday morning Track surpass traditional barriers, with virtual coach- C ing systems utilizing augmented reality as a Jun 12, 10:30 - 12:00 notable approach to guide and motivate pa- tients through exercises. The Virtual Coach- ing Platform (VCP) implemented within the OR-116 Research collaboration between academia, TeleRehaB DSS system is designed to actively healthcare and industry. Experience from engage patients and facilitate their rehabil- the research profile Embedded Sensor Sys-itation journey. Leveraging augmented real- tems for Health and an adjunct industrial ity, VCP integrates holographic elements into graduate school developed exergames and cognitive games to optimize patient experiences. TeleRehaB DSS Maria Lindén, Mats Björkman project employs VCP to deliver home-based Mälardalen University, Sweden training programs for patients with balance One of the cornerstones at Mälardalen disorders, leveraging augmented reality to en- University, Sweden, is research in close collab- hance reality perception by overlaying vir- oration with society and companies, trying tual elements onto the real environment. De-to find solutions to real problems. In 2013, veloped using Unity, VCP encompasses various a research profile called Embedded Sensor exergames and cognitive games, adapting Systems for Health was established, in which to individual patient needs, thereby creating researchers at the university were working immersive and engaging scenarios to chal-in close collaboration with healthcare and lenge patients and facilitate treatment. Aug- industry. The initiative was financially suppor-mented reality-based virtual coaching sys- ted by the Swedish Knowledge foundation, tems exhibit substantial promise by provid- Mälardalen University, and in total 20 coming challenging exercises that encourage reg- panies in form of in-kind financing. The focus ular participation, offering a potential av- of the research profile was sensor systems for enue to address balance disorders and en-health monitoring of humans, independent of hance patient quality of life. Despite the ad- time and place. vantages, it’s crucial to acknowledge associ- As a result from the profile, the research en- ated challenges and continue refining aug- vironment in medical and health engineering mented reality-based rehabilitation interven- at Mälardalen University grew to include 25 tions to optimize their effectiveness within clin- people. 5 patents were received, in addition ical settings. to more traditional academic results, as 50 Acknowledgement: This research is funded by journal publications, over 160 conference the European Union, under the Horizon Europe publications, and PhD dissertations. An ad-programme, GA number 101057747 (TeleRehaB junct industrial PhD school was started with 11 PhD students, which was an excellent way to 94 perform collaboration with companies, giving vanced techniques and ML/AI algorithms. In the opportunity to understand each other’s the first step, a demonstrator can be built us-organization. ing ready-to-use programmable research sys- Since the research profile has ended 2022, tems (us4R / us4R-lite). The provided open-Mälardalen University has initiated a cross- source SDK featuring Python/Matlab/C++ in-disciplinary center together with several terfaces makes it easy to implement cus-healthcare organizations. In this center, tom ultrasound acquisition and processing. researchers from engineering, health and wel- Available GPU libraries may provide computa-fare, organizational research, and innovation tion acceleration and integration with ML/AI and product development come together. frameworks required for real-time processing. The center invites companies for collabor- Next, the demonstrator can be tested and val-ation. Especially, the need from healthcare idated in laboratory and/or (pre)clinical envir-is in focus, and information driven care has onment. been identified as a main topic. This seems a For medical device development, our OEM ul- promising way to continue the work started trasound modules are embedded and integ-within the research profile Embedded Sensor rated into the final product. As the very same System for Health. modules and software are used in the research platforms, the risk of migration is minimized. OR-117 This strategy enables a fast commercializa- Innovative ultrasound devices — from re- tion path from research, through proof-of-search to commercial product concept, to real medical device development. Marcin Lewandowski Today, we work with several start-ups de- us4us Ltd., Poland veloping innovative ultrasound systems using Us4us is a spin-out company created in the presented approach. We are confid- 2015 by academic researchers to commercial- ent that the SDU paradigm implemented on ize ultrasound solutions for biomed and indus- our efficient acquisition/processing hardware trial markets. Our research and OEM ultra- can drive future ultrasound developments and sound platforms exploit software-defined ul-device commercialization. trasound (SDU) technology. The SDU is a major paradigm shift enabled by GPU-accelerated OR-118 processing. Today, new innovative ultra- Development and commercialization of a sound modalities require access to raw echo novel stimulation device for denervated data and advanced real-time processing al- muscles – challenging under MDD, most gorithms. Moreover, ML/AI methods may be likely impossible under current MDR used to extract even more diagnostic inform- Winfried Mayr ation out of raw ultrasound signals. Medical University of Vienna, Austria Legacy ultrasound scanners implement B- The European R&D-Project RISE, collabora- mode imaging and color-flow by applying tion of 20 European partner institutions, had scan-line acquisition and hardware-based addressed the special subtopic “denervated beamforming. Most of the new modalities muscles / flaccid paraplegia” from basic re- (elastography, super-resolution imaging, ad- search to clinical study and transfer to mar- vanced flow, etc.) require ultrafast acquisition ket and clinical application of specific novel and synthetic aperture processing methods. stimulation technology. Research under MDD Only a new system architecture and massively conditions was administratively challenging as parallel processing can stand up to those new even inside the European Community rules of techniques. small and large animal research as well as clin- Traditionally, translating academic research ical studies deviated, sometimes quite sub-and innovations into a commercial product stantially, among different national states and is a struggle. I will show how our SDU solu-led to increased administrative burdens and tions are applied to research and building delays. Another quite difficult problem oc- early demonstrators (proof-of-concept), and curred, as the electrical charge transfer across later, easy migration to real medical device stimulation electrodes, necessary for effect-development. Open software interfaces and ive muscle activation, had to exceed limita- high-performance GPU processing open un- tions given in standards for stimulators (IEC limited possibilities for implementation of ad- 601) significantly. Finally, sufficient solutions 95 were found, and the project outcome was a inherent within global health systems. This ab-validated stimulation prototype and clinical stract underscore the need for interventions application protocol with proven safety and that support the inception, development, and efficacy, which could be transferred to com- adoption of appropriate health technologies. mercial availability in the form of a certified Despite the allure of imported technologies medical product, which is now stably on the and the seemingly benevolent donation of market as efficient problem solver to main- medical devices, the local compatibility and tain muscles, after temporary or chronic loss of sustainability of these technologies are often nerve supply, in healthy physiological state. questioned. Dependence on imports neglects On the other hand, a critical look on actual the specific environmental and infrastructural MDR-conditions makes clear, that a similar at- idiosyncrasies of low- and middle-income tempt to develop such a device got kind of un-countries. To effectively counter this, the thinkable, beginning with administrative bur- health technology landscape must be reden and costs for Article 62 and Article 82 stud- shaped to prioritise solutions that resonate ies and strict limitations to harmonized stand- with the context-specific requirements of ard in the notification process. Even if all individual health systems. obstacles could be mastered, the cost in rela- The journey towards innovation is further chal-tion to the given small market segment made lenged by the prevailing risk-averse climate it unrealistic to make such important devices that stymies local ingenuity. A paradigm shift available for patients, who badly need them. is necessary to value entrepreneurship within Unfortunately, we have already a number sad the healthcare technology arena. Successful examples of very useful rehabilitation hard- global models have shown that developing ware, available, safe and effective for many a thriving innovation ecosystem requires a years, that got now off the market due to harmonious interplay between academic too high recertification costs, like meanwhile mastery, public governance, and private more or less all controllable multichannel elec- sector agility. tro stimulators for movement rehabilitation in Academic institutions emerge as vital cata- paralyzed limbs. lysts in this innovation ecosystem, nurturing a So, we urgently need to reconsider, who knowledge pool conducive to research and much regulation is acceptable, necessary and development while simultaneously anchoring meaningful to provide best possible patient the local industry. The combined forces of safety without hindering availability of innov- academia, public and private sectors are ations for best possible medical care, in par- pivotal in bridging the multiple innovation ticular for small target groups. Such develop- valleys. Together, they can advance tangible ments can lead to unacceptable discrimina-business incentives that ultimately promote tion of minorities in our society, with particu- the broader adoption and implementation of lar health related needs, and it is our strong health technologies. ethical responsibility to counteract and get To circumnavigate the valleys of death in inadequate regulatory restrictions readjusted innovation, a systemic transformation em-within shortest possible time. bracing locally oriented health technology solutions is essential. This change not just OR-119 fuels affordable healthcare, it also fosters Navigating Innovational Valleys: Advancing job creation, fortifies the healthcare infra-Health Technology Innovation for Sustain- structure, and most crucially, ensures the able and Affordable Care relevance and effectiveness of the healthcare Sudesh Sivarasu solutions to the populations they intend to University of Cape Town, South Africa serve. Sustainable and affordable healthcare The path to affordable healthcare is is thus a multifaceted endeavour that hinges fraught with ’valleys of death’ a critical chasm upon the continuous support and evolution of in the innovation cycle that often hinder appropriate health technology innovations. the transition of health technologies from This abstract discusses the potential of cre-concept to clinical application. Sustainable ating such interventions required at various innovation in healthcare demands not only stages of the innovation life cycle of a medical the reduction of manufacturing costs but also technology. a strategic alignment with the complexities 96 OR-120 techniques are gaining prominence across Understanding Regulatory Requirements: A various application domains, several studies Postmortem Analysis of Tremitas’ Bank- demonstrated that the electric pulses can af-ruptcy in the Medtech Sector fect cell excitability, leading to undesirable Tibor Zajki-Zechmeister effects. However, comprehending the inter- CommuModo GmbH, Austria play between electroporation and the elec- This paper examines the regulatory chal- trophysiological response of excitable cells lenges encountered by Tremitas GmbH, an poses a significant challenge. In this study, Austrian startup that developed the Tremi- genetically engineered human embryonic kid-pen, a device for quantifying tremor, within ney (HEK) cells were employed as a simple cell the European Union’s medical device sector. model to investigate the effects of electro-Transitioning from the Medi-cal Device Direct- poration on excitable cells. These cells were ive (MDD) to the Medical Device Regulation exposed to nine 100 µ s pulses of increasing (MDR) intro-duced complexities that dispro- electric field strength. The objective of this portionately affect startups due to stringent research is to develop an algorithm for the compli-ance requirements. Through a de- automatic classification of cell responses. The tailed postmortem analysis, we identified crit- dataset consists of 2758 binary masks and cor-ical regulatory missteps by Tremitas, includ- related other experimental information. Three ing inadequate understanding of the regu- distinct approaches were implemented for cell latory framework, mismanagement of clinical response classification, namely, a residual net- trials, and ineffective outsourcing of regulat- work (ResNet), traditional machine learning ory activities. These challenges culminated models, and an hybrid combination of the two. in delayed market entry and contributed to The Nu-Support Vector Classifier, using fea-the company’s eventual insolvency. Key res- tures extracted from the image contours, ex-ults indicate that early integration of regulat- hibited superior performance among all mod-ory knowledge, strategic planning for clinical els. validation, and careful selection of develop- ment partners are essential for navigating the OR-122 Optimizing Electroporation Responses in medtech regulatory landscape. The paper un- Genetically Engineered HEK Cells: An En- derscores the importance of regulatory affairs semble Learning Approach in the success of medical device startups and Francesco Bassi 1, Simone Kresevic1, Alessandro suggests that a proactive approach to com-Biscontin2, Aleksandar Miladinovic2, Milos Ajcevic1, pliance can mitigate risks and enhance mar- Agostino Accardo1 ket success. This analysis aims to serve as a 1University of Trieste, Italy cautionary tale and guide for future startups 2Institute for Maternal and Child Health-IRCCS in the medical device in-dustry, emphasizing “Burlo Garofolo”, Italy the critical role of regulatory strategy in entre- preneurial success. Use and understanding of electroporation have grown in recent years, revo-lutionizing various fields. However, optimization for IFMBE Scientific Challenge stimulation techniques is still needed. In this context, the introduction of genetically mod- Wednesday after lunch ified cell cultures allows a dramatic increase Track B in simulation capabilities, but also intro-duces the necessity of more advanced and human Jun 12, 13:30 - 14:15 independent analysis methods. We aimed to identify features, including OR-121 morphological characteristics and other ex- A Machine Learning Approach for Predict- perimental parameters, to develop models ing Electrophysiological Responses in Ge- for predicting the responses of genetically netically Modified HEK Cells engineered HEK (Human Embryonic Kidney) Jacopo Vitale, Martina Sassi, Leandro Pecchia cells to pulsed electric fields. Università Campus Bio-Medico di Roma, Italy This subset of predictive features, including Electroporation involves exposing biolo- the presence of K+ channels, electric field gical cells and tissues to pulsed electric fields strength, experiment number, initial fluores- (PEFs). While these electroporation-based cence, and cell mor-phology characteristics 97 was identified. and the initial curative effect appears posit- A machine learning approach based on ive. Based on this, for the coming of the Space ensemble learning techniques de-ployed Age, a forward-looking space experiment is through the XGBoost algorithm was utilized. designed and performed with dLhCG for fu-This approach involves sequentially building ture space medicine. For this purpose, dLhCG numerous weak decision trees, where each samples have been delivered onto Chinese subsequent tree aims to correct the errors Space Station via Tianzhou-6 cargo space-made by the ones before it. Considering the craft for a six-month space experiment. unbalanced frequencies of the cell response types, we adopted different strategies to OR-124 balance the training set and avoid bias were Multicomponent printing inks for the adopted. designing of 3D extracellular matrix- The produced XGBoost model trained with mimetic biomaterials a combination of real and syn-thetic data Adriana Lungu 1, Raluca Dobrisan1, Valentina exhibited an accuracy of 66.0 %, a mean Ciobanu1, Minodora Marin1, Sorina Dinescu2, AUC of 0.89, and an average F1 score of 0.66 Izabela-Cristina Stancu1 1 when evaluated against the internal test set National University of Science and Technology Po- compris-ing solely real data. Further analysis litehnica Bucharest, Romania 2 on an external test set revealed an F1 score of Department of Biochemistry and Molecular Bio- 0.57. logy, University of Bucharest, Romania In conclusion, we identified predictive features The research interest in 3D printing tech- and produced models that may contribute nologies has increased lately, due to their use to predicting the responses of genetically in multiple fields, including biomedical applic-engineered HEK cells to pulsed electric fields. ations. Nanocomposite inks are considered a great promise in 3D (bio)printing to fabricate Cellular and tissue engineering scaffolds that mimic the extracellular matrix (ECM) of the lost tissues. The combination of Wednesday after lunch different types of nano-fillers into a polymeric matrix is a very modern strategy providing Track C biocompatible scaffolds for tissue regenera- Jun 12, 13:30 - 14:15 tion. In the current study, several biomimetic nano- OR-123 composite inks were developed by combining Decellularized Tissue Engineering Hyaline two different types of nano-fillers: 1.5-3 % Cartilage Graft for Articular Cartilage Re- montmorilonnite nanoclays (Cloisite®Na / pair and Its Forward-Looking Test for Space Cloisite 15A) and cellulose nanofibrils (CNF), Medicine embedded within a soft-polymer matrix Dongan Wang based on protein-polysaccharide combin- City University of Hong Kong, Biomedical Engineer- ation. Extrusion printing tests were used to ing, Hong Kong investigate the potential of the nanocompos- Articular hyaline cartilage, a tissue articu- ite hydrogels as 3D printing inks. Some relevant lating skeleton at joints, is highly prone to dam- structural, physico-chemical, mechanical and ages caused by trauma, diseases and ageing; biological characteristics of such scaffolds once injured, its self-regeneration is difficult were defined. and slow due to the avascular nature. To repair All the 3D printed materials reached their and regenerate damaged articular cartilage, maximum degree of swelling after 24 h of we have innovatively developed a continu- immersion in PBS buffer and then remained ous methodology to directly set up a macro- in equilibrium (600-700% for samples doped scaled 3D decellularized tissue engineering hy- with Cloisite®Na and 500-600% for samples aline cartilage graft (dLhCG). Good osteo- doped with Cloisite 15A). Stability tests were chondral defect healing and complete integ- also conducted for identifying product deration with adjacent native cartilage in in-situ gradation. All the scaffolds exhibited a slower implantation of dLhCG samples in large an-dissolution rate after immersion in PBS and imal models demonstrated the competence incubated at 37°C (about 20% in 7 days), which of dLhCG as a cartilage graft. Investigat- could be a sign of the high crosslinked degree ive clinical trials have been initiated in China of these biopolymer composites 98 The compositional features of the prin-tion. The nanocomposite nature was con- ted scaffolds were determined using FTIR, firmed by scanning electron microscopy (SEM) whereas the morphology was studied by and micro-computer tomography (micro-CT). SEM. The reference materials exhibited an In vitro biomineralization was tested. To uniform and highly-porous interconnected assess scaffolds’ biocompatibility and ability structures (with large pores of about 100±30 to support osteogenic differentiation in vitro, micrometers diameter) , while the incorpor- MC3T3-E1 cell line was used. ation of nanoclays dramatically modifies the Macroporous scaffolds with nanocomposite morphology. The nanoclay-doped samples filaments were obtained. Composition ad- present more compact structures with small justments allowed improved deformation un- clusters (about 300 nm) due to the nanoclay der controlled compression and a modulation agglomeration. of the bulk and surface mechanical behavior, Preliminary in vitro biological tests per- leading to enhanced adhesion of preosteo-formed with pre-osteoblast cells revealed blasts. Formation of a new mineral phase was optimal biocompatibility of the double nano-detected after cell cultures. reinforced scaffolds. During seven days of ACKNOWLEDGMENT: This work was supported culture, the cells tend to proliferate and from the project PN-III-P4-PCE-2021–1240, no. populate all 3D printed scaffolds, especially PCE 88/2022. Cloisite®Na-enriched scaffolds (MTT cell vi- ability 55%). According to our findings, the double nanostructuration of the inks clearly S20 - System design for contributes to the improvement of the cellular biomagnetic applications using component response. optically pumped Acknowledgement: This work was suppor- magnetometers ted by a grant from the National Program for Research of the National Association of Wednesday afternoon Technical Universities - GNAC ARUT 2023. Track A OR-125 Jun 12, 14:15 - 15:30 Artificial Bone Extracellular Matrices based on 3D-printed nanostructured hydrogels OR-126 Adriana Lungu 1, Izabela-Cristina Stancu1, Elena Active compensation for OPM-MEG inside a Olăreț1, Filis Curti1, Carmen-Valentina Nicolae1, two-layer magnetically shielded room Sorina Dinescu2, Alexandra-Elena Dobranici2, Bog- Michal Władziński 1, Anna Jodko-Władzińska1, dan Stefan Vasile1 Tilmann H. Sander2 1National University of Science and Technology Po- 1Warsaw University of Technology, Poland litehnica Bucharest, Romania 2Physikalisch-Technische Budesanstalt Institut Ber- 2University of Bucharest, Romania lin, Germany In the current paper we engineered new Measuring brain responses by magne- nanostructured hydrogel precursors for 3D toencephalography (MEG) is currently in a printing, bioinspired from the nanocomposite hardware development phase since a new bone extracellular matrix. type of easy-to-use sensor be-came avail- Natural origin hydrogels (based on gelatin, al- able. These optically pumped magnetometers ginate, cellulose, and gellan gum) were used (OPM) have many advantages in comparison to provide the hydrated matrix in which nan- to the long-employed SQUID sensors, but ofillers mimicking the mineral phase of hard the re-quirements for magnetic shielding tissues were added. Paste-like 3D printable are more demanding. One is the need to inks with different nanofiller loadings were compensate for field fluctuations below 1 Hz, formulated. Calcium carbonate from cuttle which are a consequence of the decreasing bone fragments, wollastonite and biosilica efficiency of magnetic shielding towards lower from diatomite were used for their reinforcing frequencies. Here a cost effective and simple effect and for the potential to biomineral- active compensation is developed, which uses ize. The scaffolds were characterized through a Helmholtz-like compensation coil together swelling degree, water contact angle, stabil- with a simple digital Proportional-Integral (PI) ity under simulated physiological conditions, controller. The signal from a single OPM is morphology evaluation, and nanoindenta- used as the control loop input variable. An 99 attenuation of 21 dB is achieved in z-direction are pulsed sequentially for 200 ms, while at 100 mHz ena-bling linear operation of the inbetween each pulse the relaxation signals OPMs. Linear operation is then investigated are recorded for 400 ms. After extensive by measuring MEG while a subject is listening noise removal, the relaxation parameters are to tones. The experiment is per-formed twice extracted from the measured data and the with active compensation OFF and ON and MNP distribution is reconstructed. higher auditory M100 amplitudes are observed With our setup we were successfully able to for the ON condition. reconstruct point-like MNP sources inside the region of interest. OR-127 We will discuss the challenges towards unshiel- Quantitative magnetic nanoparticle ima- ded operation of an MRXI setup and present ging with magnetorelaxometry in unshiel- first proof-of-principle reconstruction results. ded environments The imaging parameters of our system are Aaron Jaufenthaler, Daniel Baumgarten currently being investigated. UMIT TIROL - Private University for Health Sciences and Health Technology, Austria OR-128 Magnetic hyperthermia is a promising Improved degaussing procedure for a mag-technique for tumor treatment. In magnetic netically shielded room hyperthermia, magnetic nanoparticles (MNP) Peter Koss 1, Jens Voigt2, Allard Schnabel2, Ronja Rasser1 are injected into the tumor and are then 1Fraunhofer IPM, Germany heated by applying a e.g. 100 kHz, 20 kA/m 2Physikalisch-technische Bundesanstalt, Germany magnetic field, leading to tumor regression. However, for a safe and efficient treatment, A series of measurements was conducted quantitative spatial imaging is required. In to demagnetize a magnetically shielded room magnetorelaxometry imaging (MRXI), the (MSR) located at Fraunhofer IPM. The MSR relaxation of the MNP’s net magnetic mo- is used for applications such as magnetoen- ment after previous magnetization using cephalography (MEG) and materials testing, pulsed DC excitation coils, is measured with which require a low-noise and low residual field very high sensitive magnetometers, e.g. op- environment. To operate optically pumped tically pumped magnetometers. The MNP magnetometers (OPM) inside the MSR, a large distribution is then reconstructed by solving volume with a residual field of less than 5 nT an ill-posed inverse problem. Recently, we is required. Therefore, we aimed to establish demonstrated the upscaling of MRXI to the an adequate fast and reproducible degauss- human torso and head inside a magnetically ing procedure. shielded room. In order to enable a broad The MSR at Fraunhofer IPM is constructed out use of MRXI in clinical settings, the shielding of 2 layers of mu-metal and provides a pass- requirements need to be reduced. ive shielding factor for quasi-static fields of Here, we report our first steps towards small- above 100. Both mu-metal shells can be de- scale unshielded imaging. Our setup is com- magnetized individually by three degaussing posed of two dual channel optically pumped coils. Thus, the MSR has a total of six independ- magnetometers (OMG, Twinleaf), 6 triaxial ent degaussing coils. Each one of the coils en- fluxgate reference sensors, 32 excitation coils, circles the four walls of one MSR shell around a 3D Helmholtz coil system and a 3D-gradient one of the three coordinate axes of the MSR. coil system. The region of interest of our setup The residual field in the central 1 m3 was meas-is 36 mm x 36 mm x 36 mm and our phantom is ured on a 25 cm grid at 5x5x5=125 points. When composed of gypsum-immobilized RCL-MNP all demagnetization coils are connected in with a clinically relevant iron concentration. series an overall demagnetization is possible We first use our fluxgate reference sensors in less than 4 minutes. This configuration still and the Helmholtz and gradient coil sys- leads to the desired residual magnetic field of tems to compensate environmental fields <5 nT within the scanned 1 m3 volume. and field gradients. This is needed to op- In this unexpectedly low field region, other in- timize sensor performance and to avoid the fluences are no longer negligible. Concern- background-field induced altering of the ing the field measurement itself, the DC-offset relaxation behavior of the MNP. The MRXI drift of the used fluxgate and the influence of sequence is then started: The excitation coils field changes of the environment outside of the MSR observed by a second fluxgate should 100 be corrected. The obtained field maps re- for about 10 years, proving the sensors’ reliabil-vealed weakly magnetic objects in the MSR ity and robustness. The sensors are placed on a construction, which must be replaced to ob- flexible helmet with about 90 different sensor tain significantly lower and more homogen- positions allowing for a whole-head configur- eous residual fields. ation or denser arrays. The system computer includes a visualisation and acquisition soft- OR-129 ware allowing the user to display and record Full MEG system based on helium OPMs for data in real time. The system produces files medical imaging in an fiff-like format widely compatible with Rudy Romain, Jaroslaw Rutkowski, Sergey Mitry-common data processing tools. The proto- ukovskiy, Kevin Arth, Matthieu Le Prado, Agustin type based on 5 He-OPMs have been used in Palacios-Laloy, Etienne Labyt various contexts (epilepsy recordings, somato- MAG4Health, France sensory and visual tasks, and magnetocardi- The first type of Optically Pumped Magne- ography). First results using the full 48-sensor tometers (OPMs) available on the market has system will be presented. represented a technological leap to tackle the drawbacks of SQUID-based MEG systems in OR-130 terms of cost and adaptability to the subject. Full OPM-MEG system with limited sensor A second type of OPM is now available: He- counts lium OPMs (He-OPMs) are sensors using meta- Urban Marhl1, Rok Hren1, Tilmann H. Sander2, stable helium-4 atoms as sensitive element Vojko Jazbinšek 1 1 contrary to their existing counterparts that Institute of Mathematics, Physics and Mechanics, use rubidium atoms. As a consequence both Slovenia 2 sensors have different functioning principles Physikalisch-Technische Budesanstalt Institut Ber- and thus have their own advantages and lim- lin, Germany itations. MAG4Health develops MEG system Magnetoencephalography (MEG) is a based on helium OPMs and aims to democrat- non-invasive functional imaging technique for ize the use of MEG by making it more cost- measuring weak magnetic fields generated by effective and expand its range of usability. A electrical currents in the brain. Standard MEG standard system consists in a set of 48 sensors, system is using a superconducting quantum a flexible helmet, and a control unit driving interference device (SQUID). These systems the sensors and running an acquisition soft- need expensive cooling and have a fixed, large ware. He-OPMs operate at room temperat- array of sensors to measure MFMs (magnetic ure, they can then be placed in direct con-field maps) around the whole head. As an tact with the subject scalp and can be used alternative, OPMs (optically pumped magne- for extended period of time. They operate in tometers) operate at room temperature and a 3-axis closed-loop mode making their dy- can be placed individually much closer to the namic range virtually unlimited (current elec- head surface. In the standard SQUID-MEG trics compensation up to 300 nT), which re- recordings, we typically encounter the prob-lease by a great deal the specification on the lem of redundancy and uniqueness of signal magnetic shielding room. In addition, their information contained in a dense array of bandwidth of 2 kHz makes it possible to re- sensors covering the whole head. In this study, cord fast brain activities. The sensor principle we explored the possibility of obtaining most leads to a vectorial measurement with 2 axes of the relevant information using OPM-MEG more sensitive than the third one. The sens- with limited sensor counts. We employed an itivity of this two axes (radial and tangential iterative statistical technique (IST), previously to the scalp) are below 40 fT/rtHz. This fig-developed for body surface potential map- ure is larger than their alkali counterparts is ping, on large data sets of whole-head MEG counter-balanced by the closer proximity to measurements of audio-evoked fields (AEFs) magnetic sources within the brain, leading to to select an optimal subspace of recording an overall similar signal-to-noise ratio to al-sites and to determine a transformation kali OPMS and better than SQUIDs. He-OPMs matrix, which projects measurements from have been developed for space exploration selected to unselected sites to estimate the at CEA (French Technological Institute) and full MFM. Since we have not had enough OPMs have been used in the SWARM mission com- yet to measure whole-head OPM-MFMs, we missioned by ESA (European Space Agency) employed a database of 18 SQUID-MEG AEF 101 measurements and project it to both radial the Erasmus/Tempus IV, CRH BME (Curriculum and tangential components of the mag-Reformation and Harmonization in the Field netic field on 80 OPM measuring sites close of Biomedical Engineering) project, revealed to the head surface using current distribu- that this number was almost triple. A more re-tion on the brain surface obtained by MNE cent survey was performed by INBIT in 2019, (minimum norm estimation) from measured in collaboration with EAMBES. In 40 european SQUID-MFMs. We have tested IST on various countries included in this survey, 346 BME study protocols, like using the whole head MFMs programs were identified in total, offered by or only one hemisphere, measuring only one 184 Universities in 31 countries. 115 BSc, 161 component or a combination of both. We cal- MSc degree and 60 PhD degree. An important culated measures like correlation coefficient factor that should be taken into consideration (CC), root-mean-square and relative errors is the content of programs offered since the to evaluate the quality of full MFM estim- title alone could be misleading. Similar growth ated from optimally selected measuring sites. as in Europe, is expected for most of the other Additionally, we checked source localization world regions and the worldwide survey under errors for the largest audio response (M100) preparation by IFMBE, that is joined by INBIT. by calculating equivalent single current dipole for the one hemisphere or dual current dipoles OR-132 for the whole head M100-MFMs. In all cases, EU students, teachers and researchers ex- we found that most of the information con- change programs tent is in the first 15 to 20 optimally selected Ratko Magjarević measuring sites (CCs>0.95, localization errors University of Zagreb, Croatia < 1 mm). The main finding of our study is that In recent years, BME and BME education for event-related fields, like AEF in this study, have been influenced by two major factors: which have mainly focal origin, a markedly the transition of health care to digital health smaller number of recording sites, in compar- care and the significant impact of the pan-ison to that currently employed in systems for demic on the way of teaching in general, MFM recordings, may be sufficient to extract including in BME. In this session, we would relevant information. like to explore how these factors influenced on core BME curricula, interdisciplinary topics, S28 - BME Education in Europe discipline-specific sub-specialties, and stu- dents’ capacity for problem solving and com- Wednesday afternoon munication. As a reference, we use the Tempus Track B project on BME education in Europe conduc- ted a decade ago. Jun 12, 14:15 - 15:30 OR-133 OR-131 The CRH-BME Tempus project – results and Overview of the BME study programs – pre- experience a decade after vious European wide survey results Tomaž Jarm Nicolas Pallikarakis University of Ljubljana, Slovenia Institute of Biomedical Technology (INBIT), Greece Historically, the origins of Biomedical En- It is widely accepted today, that modern gineering (BME) at our school (University of health care delivery is the outcome of the pro- Ljubljana, Faculty of electrical engineering; gress made in biomedical engineering. An in-UL-FE), date back to the late 1960s. BME dicator of this fact is the continiously grow- was first introduced into a formal undergradu-ing number of universities offering BME stud- ate study program in 1977, as a specialty sub-ies at all levels: BSc, MSc and PhD. Three sur- track Cybernetics in Medicine within one of veys performed by INBIT give a general view of four major study tracks under the umbrella of this evolution. The first one under the EU fun-Electrical Engineering, called Control Engin- ded MELETI Project (Medical Engineering Lis- eering. About 3 decades later the running ted Education & Training Information) in 2000, of European Tempus Project CRH-BME coin-found 67 programs that were offered by 50 cided roughly with the formation of the new Universities in Europe: 33 undergraduate (BSc) Department of BME at UL-FE and the Bo-and 34 postgraduate (MSc and PhD) pro- logna reform of the 1st and 2nd (Masters) level grams. The second performed in 2010 under 102 study programs at our school, resulting in a The Erasmus+ project Ethical Engineer aims new Masters-level study program of Biomed- to enhance AI education by promoting Trust- ical Engineering, which was run for the first worthy AI in Europe, emphasising ethical, so-time in the academic year 2012-2013. cial, and legal aspects in engineering educa- Our study program of BME was rooted in the tion and training. Focusing on higher edu- earlier program, but it was also influenced by cation engineering and robotics students, we the CRH-BME Tempus Project, of which we seek to instil ethical principles in AI develop-were an integral part. In the formation of our ment. Our goal is to empower future AI ex- study program, we followed the recommend- perts to create systems aligned with European ations arrived at during the duration of the values, ethical guidelines, and societal well- Project in terms of the number of credits al- being. In the Ethical Engineer project, the located to different categories of courses and work is divided into 4 activity-based WPs: De- in inclusion of specific courses in the program velop Case Studies & Starter Kit: providing that covered in sufficient quantity and quality AI-focused educational materials; Establish the so called ”core” BME topics, recognized in Problem-Based Learning OERs: creating flex- the CRH-BME recommendations as being es- ible, interactive open resources; Initiate Eth-sential for any BME study program. But at the ical Engineer Bootcamps: offering immersive, same time, our curriculum also included spe- interdisciplinary learning; Engage in Sharing cial topics reflecting the long tradition and ex- & Sustainability: leveraging networks and en-pertise in the field of BME which existed and hancing project presentation. The project will continues to be developed in research groups produce the following outcomes: An Ethical AI at UL-FE. Case Studies & Starter Kit, to teach engineer- Over the last 11 years, the study program of ing students about the ethical implications of BME has established itself as a stable study their work; Ethical Engineer OERs, flexible open track, one of 7 tracks that currently make up education resources based on real-world case the Masters-level study program of Electrical studies that lecturers can use to teach AI & Ro- Engineering at UL-FE. Except for a couple botics Ethics in their programs; Bootcamps, an of shared elective courses, the program of interdisciplinary approach to teaching ethical BME has its own curriculum and courses. The considerations in AI to students that will be sys- total number of students enrolled in the whole tematised in a guide for others to replicate. Masters-level study program is 120 per year The project aims to target four groups: EN- (after the entrance exam), and the BME track GINEERING STUDENTS - Undergraduate and attracts on average about 15 students per graduate students enrolled in engineering and year. Even though the majority are the 1st robotics courses need to be equipped with the level graduates of Electrical Engineering at our ethical knowledge and skills necessary to nav- school, we also attracts 1st level graduates igate the ethical challenges posed by AI and of Physics, Mechanical Engineering, Chemical robotics technologies. ENGINEERING PRO- Engineering and from Biotechnical Faculty. FESSORS AND LECTURERS - Many profess-The curriculum remains largely unchanged in ors and lecturers in engineering faculties have terms of courses, but the topics included in the varied experiences and knowledge of ethics courses reflect the changes and rapid devel- in practice. While they may be well-versed opment of new technologies that inevitably in the technical aspects of AI and robotics, affect the field of BME. they require evidence-based approaches and proof of effectiveness before investing their OR-134 BME in EC research projects time in new practices that integrate ethical considerations into their teaching. LEARN- Leandro Pecchia ING/EDUCATION SCIENTISTS - Academics and Campus Bio Medico University of Rome, Italy, Italy practitioners whose interest lies in the study Abstract was not received. of education and the improvement of instruc- tional methodologies for more effective learn- OR-135 ing and teaching. Their shared interest is Ethical Engineer: Enhancing AI Education in the use of innovative methodologies to im-Engineering prove competence development within tradi- Lenka Lhotská, David Macku tional academic courses. AI ETHICS EDUCA- Czech Technical University in Prague, Czech Repub- TION STAKEHOLDERS - The project is of dir- lic 103 ect relevance to national education authorit- prostate cancer patient care. However, ad-ies, networks of HEIs and membership associ- dressing existing challenges, including ensur- ations interested in improving AI Ethics educa- ing user-friendliness, effective integration into tion. healthcare systems, and protecting personal Our project seeks to fill a gap in engineering health data, is imperative. These findings education by integrating ethical considera- provide an overview of the current state and tions into the development and deployment of outline potential trajectories for future devel-AI and robotics technologies. This project has opment and implementation of m-health in the potential to make a significant contribu- oncology. tion to the development of engineers who are not only technically proficient but also ethic- OR-137 ally responsible and able to navigate the com- Research on dental materials for their plex challenges posed by AI and robotics tech- suitability in building anthropomorphic nologies. phantoms Nikolay Dukov, Minko Milev, Todor Todorov, Zhivko Bliznakov, Kristina Bliznakova BME Miscellaneous topics Medical University of Varna, Bulgaria Wednesday afternoon The main goal of the study is to investig- ate materials from dental practice and as- Track C sess their suitability for manufacturing an- Jun 12, 14:15 - 15:30 thropomorphic phantoms designed for X-ray imaging modalities. For this purpose, we in- OR-136 vestigated nine commercial materials com- M-Health in Prostate Cancer: Professional monly employed in dental procedures for tasks and Patient Perspectives such as restoration, duplication, and imita- Petra Hospodková, Irina Klubarská, Matyáš tion. They underwent scanning at a clinical Mašek, Martin Budil Computed Tomography (CT) facility using six Czech Technical University in Prague, Czech Repub- anode voltages. Subsequently, we measured lic their CT numbers, expressed in Hounsfield Units In the current study, an analysis of mo- (HU). The findings of our investigation indic- bile health (m-health) technology usage in ate that three specific materials, namely Temp managing prostate cancer, focusing on pa- silic, Gingifast Elastic, Ellite Double A, exhibit tients undergoing radiotherapy, was conduc- promising characteristics for the production of ted. Amid increasing incidence and mortal- anatomically accurate bone structures in an-ity rates of prostate cancer in the Czech Re- thropomorphic phantoms. public, exploring m-health potential in onco- logical care is of critical research interest. The OR-138 aim was to identify the benefits of m-health Analysis of redox processes from in silico in radiotherapy, assessing current market dy- perspective in colorectal cancer cell line namics and stakeholder roles to inform the Milica Nikolic 1, Nevena Milivojevic Dimitrijevic2, conceptualization of future m-health applic- Ana Miric2, Marko Zivanovic2, Nenad Filipovic2 1 ation devel-opment in oncology. A qualitat- Steinbeis Advanced Risk Technologies Institute ive approach utilizing semi-structured inter- Kragujevac, Serbia 2 views (using software MAXQDA 22) with pa- University of Kragujevac, Serbia tients and medical staff was employed to gain Oxidation and reduction processes (redox) deeper insights into personal experiences and are essential for the proper functioning of preferences regarding m-health applications healthy cells, where the environment creates a and devices. The analysis showed limited pa- prooxidant-antioxidant balance. Disturbance tient engagement with m-health technolo- of this balance, such as the increase of oxid-gies; some were interested in health monitor- ative stress due to the reduction of antioxid-ing, while others were worried about complex- ant systems, leads to pathological changes in ity and usability. Medical professionals recog- cells and disease occurrence, including can-nized m-health’s potential in treatment op- cer. Glutathione (GSH) is an antioxidant cap- timization but noted integration and data se- able of preventing damage to cellular com- curity challenges. In conclusion, m-health rep- ponents caused by reactive oxygen species resents significant potential for innovation in 104 (ROS), such as superoxide anion radical, hy- OR-139 droxyl radical and hydrogen peroxide. GSH Modeling the conductive characteristics of protects the cells by neutralising ROS in a way the human lower back using medical ima-that it changes form from reduced (GSH) to an ging and FEM modeling to optimize tSCS oxidised state (GSSG). The final maintenance therapy of the ratio between GSH-GSSG is a key factor Jón Andri Árnason, Ragnhildur Guðmundsdóttir- in the defence of the cell from exposure to ox- Korchai, Þórður Helgason idative stress and oxidative damage. Colon Reykjavík University, Iceland cancer cells (HCT-116) with disrupted balance Transcutaneous spinal cord stimulation of prooxidative-antioxidative systems can be (tSCS) is a promising non-invasive therapy for treated with noble metals to recover the en- managing chronic pain and enhancing motor vironment of healthy tissues. In this research, function in individuals suffering from spinal two types of gold complexes were applied to cord injuries (SCI). The precise mechanisms study in vitro the disruption of the redox bal-of tSCS are currently not fully understood ance in colorectal cancer. In silico models with the medical modality suffering from were developed in the first phase to replic- being highly patient-specific with differences ate experimental observations and provide a in patient body composition needing to be good matching of the results (validation). In accounted for with electrode placement and the second phase in silico models were used current intensity. to predict temporal responses, by creating a The purpose of this research project is to fitting curve, and to perform sensitivity ana- provide some insight into the distribution of lysis to highlight the influence of each com- electric current within the human body by ponent in the determined molecular mech- developing a novel process of creating a Finite anism of oxygen reactions, including the an- element model (FEM) of the human body tioxidant enzymes’ role. Estimation of GSH- using CT images. High-resolution CT images GSSG ratios in time was based upon the im- were taken of the lower ventral body cavity. plementation of the bi-bi mechanism for en- The images were then imported into Mater- zymes. The mathematical model was cre- ialize Mimics where a 3D model was created ated and solved in Matlab commercial soft- for each major structure e.g. skin, fat, muscle, ware. The responses were collected for cases skeleton, spinal cord, etc. The 3D models when gold complexes were not applied (con- were pre-processed in Autodesk Meshmixer trol group) and when different concentrations to remove any unwanted geometry. The 3D of gold complexes were applied during the first models were imported into Ansys SpaceClaim 72h. Used concentrations of gold complexes where the models were converted into solid were 0.1, 1, 10 and 50 [�M]. Calculated errors re- CAD objects. There each organ model was vealed that the range of predicted and exper- recombined to create a single model of the imental values have the same order of mag-human lower ventral body cavity. The com- nitude. Cytotoxicity results revealed the sig- pleted model was then imported into Ansys nificant influence of one gold complex, espe- Maxwell 3D to be solved. cially at a concentration of 10 �M and higher, Five simulations were performed. Four of those causing a higher production of free radicals simulations had the cathode electrode at T10, and cytotoxic effect on the cancer cell line, T12, L2, and L4 and the anode electrode fixed which is confirmed by in silico results as well. at S2. The fifth simulation had electrodes at all Acknowledgement: This research was fun- previous placements, but the stimulation was ded by the Ministry of Science, Technological introduced through the T10 electrode. The Development and Innovation of the Republic simulations took on average 47.4 hours. of Serbia, contract number 451-03-66/2024- The simulation results conformed to the cur-03/200378 (Institute for Information Tech- rently understood pattern of current density nologies Kragujevac, University of Kraguje- distribution during TSCS with a measurable vac) and TOXBOX H2020 project, ID:101138387, difference in electrical current density of the 2024-2028. spinal cord depending on electrode position. The T10 to S2 configuration had the highest density in the spinal cord with L4 to S2, being the lowest. This simulation method, utilizing Ansys Max- 105 well, has the benefit of easily changing the OR-141 electrode configurations, as is useful to target Difference in Mentalizing Function between specific areas for high electrical current dens- Face-to-Face and Online Communications ities, but suffers from long simulation times Joohyeong Kim, Hayato Watanabe, Koichi and undesirable mesh densities. Future work Yokosawa can be used to reduce simulation time and Hokkaido University, Japan validate the process. Coronavirus disease 2019 triggered the worldwide spread of online communication OR-140 tools such as Zoom, which are now used for Assessing Ergonomic Compliance in In- meetings, classes, and interviews. It has been dustrial Environments with Markerless 3D reported that long-term online communica-Camera-Based Systems tion can induce a psychological disorder (so- Jindřich Adolf 1, Ilona Kačerová2, Kateřina called Zoom fatigue), suggesting differences Jurčová3, Vladimíra Lipšová3, Jaromír Doležal1, between online communication and face-to- Lenka Lhotská1 face communication. During communication, 1Czech Technical University in Prague, Czech Re- mentalizing functions are engaged, referring public our ability to understand ourselves and oth- 2State Institute of Health, Czech Republic ers in terms of intentional mental states, such 3Occupational Safety Research Institute, Czech Re- as feelings, desires, and attitudes. Therefore, public in the current study, we investigated whether In this research, we introduce a marker- mentalizing function differs between online less, 3D camera-based system for ergonomic and face-to-face communications. Magne-assessment in industrial settings, utilizing two toencephalograms were recorded while par-standard RGB cameras and the OpenPose ticipants engaged in a question-and-answer framework for real-time 2D skeleton mapping. dialogue either in face-to-face or online con- This novel approach overcomes the limitations ditions. Amplitudes of spontaneous brain of traditional, subjective ergonomic assess- rhythms, alpha-, beta-, and theta-bands were ment methods and sensor-based systems by compared in 16 brain regions which have been offering objective posture evaluations without reported to be involved in mentalizing func- physical markers. The dual-camera setup en- tion. Significant differences were observed ables accurate 3D reconstruction of worker only in the theta-band rhythm. Normalized movements. Our feasibility study validates the amplitudes in the right temporo-parietal junc- system using the Intraclass Correlation Coef- tion and the right anterior temporal lobe were ficient (ICC), achieving a ground truth agree- significantly lower, and those in the right pos-ment among ergonomic experts with an ICC terior rostral frontal cortex were significantly of 0.97, and the Pearson Correlation Coeffi- higher in the online condition compared with cient (PCC), demonstrating a strong correl- those in the face-to-face condition. These ation (PCC of 0.81) between expert evalu- results suggest that low visual reality (or visual ations and our camera-based measurements information context) and reduced attention for neck and upper limb flexion. This highlights (eye-contact) in online communication mod-the system’s reliability and potential as a cost- ulate mentalizing function. effective solution in industrial ergonomics. OR-142 Track-EP: A Semi-automated Tool for En- S14 - Non-invasive methods for hancing Interictal EEG Interpretation monitoring electrophysiological Margherita Anna Grazia Matarrese 1, Simonetta and hemodynamic brain activity Filippi1, Leandro Pecchia1, Christos Papadelis2 1Campus Bio Medico University of Rome, Italy, Italy 2 Wednesday late afternoon Jane and John Justin Institute for Mind Health, Neurosciences Center, Cook Children’s Health Care Track A System, United States Jun 12, 16:00 - 18:00 Inspecting epileptic activity in EEG record- ings can be time-consuming for clinicians, es- pecially when recordings are performed with a high number of channels. Identifying the 106 epileptogenic zone (EZ) –the area responsible ditional EZ approximations that can be used for causing seizures– can be difficult and relies to shorten and improve surgical planning. on the integration of several diagnostic tests. Typically, clinicians identify the channels that OR-143 record multiple stereotyped seizures as the The Neuromodulation Evaluation with EZ. However, seizures are unpredictable and Functional Near-Infrared Spectroscopy may not occur during monitoring. Therefore, for Transcutaneous Auricular Vagus Nerve interictal biomarkers, such as spikes (1-70 Hz) Stimulation on Upper Extremity Rehabilita- or ripples (80-250 Hz), recordable in the period tion between seizures, can be used to approximate Chien-An Chen 1, Jia-Jin Chen1, Wei-Cheng the EZ. Intracranial EEG (iEEG) studies show Sun2, Chao Chen Lo2 1 that spikes and ripples propagate over time National Cheng Kung University, Taiwan 2 and that the onsets of these propagations Shu-Zen Junior College of Medicine and Manage- (spike-onset and ripple-onset) are more spe- ment, Taiwan cific to the EZ than the areas of spread. Yet, The preliminary evidence suggests that VNS their definition is challenging. Here, we aim to and rehabilitation training must be precisely present a semi-automated pipeline, Track-EP, coordinated in time to produce rehabilita-that can assist clinicians in identifying these tion effectiveness. Several benefits taVNS areas in clinical EEG data, comprising iEEG with rehabilitation have been illustrated , such and scalp-EEG recordings. Track-EP stream- as:(1)Recovery of motor function, it has been lines the visualization and analysis process observed that intervention with taVNS leads and can track the onset and propagation of to a significant improvement in various clinical interictal activity for spikes and ripples. assessments such as Fugl-Meyer and WMFT Utilizing custom MATLAB® detectors, spikes scores, and these improvements are main-and ripples are automatically identified on tained or further enhanced in subsequent each channel using filtered data in the spe- follow-ups. (2)Reduction of muscle tension: cified frequency band. Subsequently, an there is a more significant improvement in automated algorithm categorizes them to muscle tension in the wrist and hand muscles define their temporal occurrence and extracts compared to the control group. (3) Restor-the time position of the first channel, showing ation of sensory function: There are also sig-that biomarker. We reconstruct cortical and nificant differences in sensory function for in- scalp surfaces from the patient’s MRI and dividuals receiving taVNS. This may contrib- compute a Boundary Element Model (BEM) ute to a positive feedback loop for both mo- as forward problem estimation. We localize tor and sensory recovery, ultimately promot-spike/ripple propagation patterns and on- ing neuroplasticity. Whether the mechanism set regions in the source domain with dSPM of the taVNS for rehabilitation modulation re- using a time-sliding window approach. We mains unsupported by research. further validated the spike-onset and the In our previous study, the functional near in- ripple-onset areas by comparing them with frared spectroscopy(fNIRS) monitoring dur- the clinically defined EZ. ing active or passive pedaling exercise found Track-EP generates color-coded volume- that during passive pedaling exercise, activa-based visualizations in the patient’s MRI space tion in the sensorimotor cortex, supplement- by reconstructing spike/ripple propagation. ary motor area, and premotor cortex was sim-The visualization ranges from red to blue, ilar to that during active pedaling exercise where red volumes represent the brain regions without physiological feedback, but the activ-that initiate spike/ripple propagation, while ation level in the contralateral supplementary blue volumes indicate those that occur later in motor area was smaller. Furthermore, in act- the propagation. Track-EP then dichotomizes ive pedaling training with physiological feed-brain area involved in propagation as either back, there was a reduction in speed vari- onset or spread areas. Finally, we found that ation, an increase in EMG symmetry index, spike-onset and ripple-onset areas overlap and an increase in activation of the premotor with the EZ more than areas or spread. cortex. These findings support that physiolo- This integrated methodology may aid presur- gical feedback helps people to self-regulate gical evaluation, offering an efficient, reliable, muscle strength and achieve neuromodula-and objective interpretation of iEEG and tion effects. scalp-EEG recordings. It also provides two ad- In this study, to evaluate the neuromodulation 107 effect of taVNS on motor rehabilitation. The our participants improved with real tACS, we upper limb pedaling exercise is selected as the expected to see such a decrease in oxygen motor task. Pealing serves both as a method consumption after the real intervention. Thus, of physical conditioning and as a method of the results highlight fNIRS potential in evalu-rehabilitation exercise. The participants was ating the effectiveness of tACS treatment on asked to maintain the certain speed by the people with AD. visual feedback. The taVNS was conducted during the motor task for 10 minutes. The elec- OR-145 trodes were placed at the left ear on cymba Multiparametric measurement of cerebral concha, cavum concha, and tragus. The con-blood flow in neurointensive care tinuous theta burst stimulation (cTBS) wave- Karin Wårdell 1, Sofie Tapper1, Stina Mauritzon1, form is selected which is validated in previous Fredrik Ginstman2, Peter Zsigmond2 1 study. The initial standard intensity of taVNS Linköping University, Sweden 2 was set to 1000uA, and adjust to maximum in- Linköping University Hosptial, Sweden dividual subject tolerance. The EMG on the bi- Measurement of cerebral blood flow in the ceps and triceps on both arms and the fNIRS neurointensive care unit (NICU) is challenging on primary motor cortex is monitored during and most available techniques give indirect the whole session. The symmetry of the EMG measures or provide snap-shot information. and fNIRS was compared before and after Both the micro- and macro blood flow com-the intervention. The activation mapping of ponents should ideally be captured over time hemodynamic response difference was repor- and space and be visualized in real-time at ted in the SMC,SS and PMC. The result shows bedside. Our research group moves towards taVNS as the potential tool for motor rehabil- this goal by setting up a concept combining itation. laser Doppler flowmetry (LDF) with magnetic resonance imaging (MRI). The aim was to OR-144 present the first examples of this multipara- Can fNIRS be used as a treatment monitor- metric approach on patients being monitored ing tool for Alzheimer’s disease? in the NICU after suffering subarachnoid Samaneh Azarbarzin, Zahra Moussavi hemorrhage (SAH). University of Manitoba, Canada An LDF system (Perimed AB, Sweden) and Functional near-infrared spectroscopy related in-house developed software and (fNIRS) is a non-invasive technique used to fiber-optical probes were adapted for clinical measure brain activity by monitoring changes use as previously described [1]. The flexible in blood oxygenation levels. In this study, we probe was implanted in relation to routine investigated whether fNIRS can be used as a surgery in seven patients (EPN 2018/322-31). monitoring tool to explore the effect of trans- In the NICU, LDF data was recorded mean-cranial alternating current stimulation (tACS) while patient movements and care were treatment on cognitive function of people monitored together with routine paramet- with Alzheimer’s disease (AD) in a cross-over ers. At Linköping University Hospital, a 3T MR designed study. Data of 27 individuals with system (Siemens Healthineers, Germany) is AD who finished the study, were analyzed. available at the NICU for use together with Study participants received real/sham tACS the LDF system. Simultaneous LDF and MRI paired with cognitive exercises for two 20-min measurements were performed in one SAH sessions/day, 5 days/week for 4 weeks; there patient. The MRI protocol included T1- and T2 was 2-5 months washout period between the weighted imaging, 2D flow measurements in two treatment blocks of the study. An eight- larger cerebral arteries, and perfusion ima-channel fNIRS device was utilized to measure ging using pulsed arterial spin labeling (ASL). blood oxygenation levels in the prefrontal MRI measurements were performed on day 6 cortex before and after each treatment block, and 8 after the initial hemorrhage. Analyses during a mental task. Compared to baseline, included calculation of the total inflow to the results show a greater reduction of oxyhemo- brain, quantification of ASL maps using an globin levels at post-intervention for real tACS openly available flow territory atlas [2], and compared to sham. Patients with AD often inspection of LDF recordings. exhibit increased oxygen consumption during A clean LDF signal, with artifacts noted < 5 % of cognitive tasks as compensation. As most of the total time, was acquired for up to 10 days of monitoring in the NICU. Also, during MRI 108 the LDF signal quality was stable and without OR-147 distortion nor interference from movement Data processing and visualization for mul-artifacts. With the LDF kept in the control timodal imaging of neuro-vascular coupling room, a 5m long cable passed through a hole Stanislaw Wojtkiewicz 1, Urban Marhl2, Piotr in the wall to the patient, which made the Sawosz1, Vojko Jazbinšek2, Tilmann H. Sander3, recording safe from an MR perspective. The Adam Liebert1 1 tip of the probe was visible in the T2 images Nalecz Institute of Biocybernetics and Biomedical and did distort the image quality in the vicinity. Engineering, Poland 2 MRI showed a reduction of the cerebral flow Institute of Mathematics, Physics and Mechanics, between the two measurement days, both Slovenia 3 seen globally and regionally. As a next step, a Physikalisch-Technische Budesanstalt Institut Ber- multiparametric comparison between macro lin, Germany and microcirculatory parameters will be done, The neurovascular coupling (NVC), a mech- and more patient data collected. anism linking neuronal activity with vascular responses, is crucial for understanding brain References: function. Methods like functional resonance 1. Mauritzon, S., et al., Front Neurosci, 2022. 16 imaging (fMRI) offer an insight on the vas- 2. Liu, C.-F., et al., Scientific Data, 2023. 10(1) cular response. Combination of electroen- cephalography (EEG) and functional near- OR-146 infrared spectroscopy (fNIRS) can provide a Measuring Anxiety Quantitatively in a De- more comprehensive view to the NVC. This pression population study introduces a multimodal approach to Zahra Moussavi, Brian Lithgow image the NCV utilizing magnetoencephalo- University of Manitoba, Canada graphy (MEG) combined with the fNIRS, im- Electrovestibulography (EVestG) has been plementing recent advancements in fNIRS im- applied to Bipolar Disorder and Major De- agers and optically pumped magnetometers pressive Disorder patients each divided into (OPM). Unlike standard MEG systems (SQUID those with Anxiety Disorder and those without. magnetometers) requiring cryogenic cooling, EVestG is a recording of vestibulo-acoustic OPMs offer portability and enhanced signal-activity from the external ear canal. It incor- to-noise ratio, enabling imaging during nat-porates a complex wavelet analysis of phase ural tasks. across 6 bands. Fifty Bipolar (26 with Anxi- The aim of this study is to provide a spatial ety Disorder) and 42 Major Depressive Disorder distribution of the neurovascular coupling effi- (24 with Anxiety Disorder) were analysed. Dif- ciency and develop tool for neurophysiological ferences in a frequency component within the research and clinical medicine. Hippocampal Theta frequency range as well Data is registered using 15 OPMs (QuSpin, as an asymmetry between left and right-side USA) and in-house developed, high-dynamic recordings in this same frequency band were range fNIRS capable of high-density optical characteristic of Anxiety Disorder in both pop-diffuse tomography. Optical signals are re- ulations. Whilst the presentation of additional gistered using 15 source fibres and 9 de- spectral energy components within the Hip- tection fibres fixed above the motor cortex pocampal Theta frequency range were rep- of a healthy subject resulting in 135 source-resentative of Anxiety Disorder they are not detector pairs working simultaneously at two identical for Bipolar Disorder and Major De-wavelengths: 750 nm and 850 nm. Both pressive Disorder. The results indicate comor- OPMs and fibres are mounted within the same bidities such as Anxiety within a Bipolar or Ma- space using a 3D-printed and subject-specific jor Depression population can impact classi-helmet. The experiment is carried out in a fication accuracy as well as treatment effic- magnetically shielded room (the single most acy. Further, they highlight the need for meas- magnetically-shielded room on Earth located uring treatment efficacy for both the patho-at the PTB in Berlin). The protocol involved right logy and comorbidity. hand ball squeezing followed by rest periods, synchronized across both modalities: MEG and fNIRS. Data processing included MEG pre- processing using FIR filtering and independent component analysis and at the fNIRS side in- 109 tensities conversion to haemoglobin changes dynamics measured by fNIRS during cognitive with the modified Beer-Lambert law. Further, task from generalized anxiety disorder (GAD) a block averaging is carried out to denoise patients and healthy subjects was processed. the NVC responses. Results on healthy vo- A total of 25 subjects and 34 healthy adults lunteers show a visible delay (˜8 s) between were recruited in this study. All the participants fast MEG and slower fNIRS responses. Here were instructed to finish 4 cognitive tasks in-we propose a method of topographical cluding verbal fluency, N-back, Go or No-Go visualization of both modalities. Further, we and Serial Subtraction tasks. The optodes propose a method of spatial visualization of placements of fNIRS were covered the ventro-the NVC parametrized with a slope between medial PFC and DLPFC on both hemispheres. the vascular (fNIRS) and the neuro (MEG) The hemodynamic response in both time and responses. frequency domains including lateralization in- This study introduces the approach for un- dex calculation, inter-hemispheric correlation, derstanding neurovascular coupling dynam- and wavelet phase coherence were adopted ics. By leveraging advanced imaging tech- in this study. niques, it offers insights into the complex rela- Our study showed that the blood oxygen con-tions between neuronal activity and vascular centration required by the prefrontal cortex responses, with implications for studying brain under the cognitive task of GAD patients was function and pathologies. higher than that of healthy adults indicating that the cognitive task has a greater cognit- OR-148 ive burden on GAD patients. This implies that Applying near infrared spectroscopy for as- severe GAD patients have poorer prefrontal sessment of anxiety lobe function. In addition, the time-frequency Chien-An Chen 1, BO-TSEN Lin1, Yi-Jing Huang2, domain analysis based on wavelet transform Yu-Jui Huang3, Pao-Huan Chen3, Jia-Jin Chen1 indicating a significant energy change in ex- 1National Cheng Kung University, Taiwan tremely low frequencies band observed from 2National Taiwan University, Taiwan high-anxiety patients during tasks compared 3Taipei Medical University Hospital, Taiwan to the resting state. This phenomenon was not Studies have found that people with higher observed in healthy subjects. In summary, our levels of anxiety have asymmetric hemody- study of limited population showed that fNIRS namic responses in the dorsolateral prefrontal has the potential to be used as an assessment cortex (DLPFC) on both hemispheres. This tool for anxiety symptoms. The future adapt- lateralization phenomenon is one of import- ation of customed made NIRS system can be ant indicators of anxiety disorder which can used for evaluation of brain neuromodulation provide quantitative assessment of anxiety as as well as for neuro-biofeedback training for well as for homecare neurofeedback scheme. subjects with GAD. Functional near infrared spectroscopy (fNIRS) provide noninvasive brain image approach to monitor the brain activity based on the neur- S25 - IFMBE Education in BME: ovascular coupling. Currently, clinical stud- BME Education in Latin America ies have shown the efficacy using fNIRS for and BME in Africa the assessment of anxiety disorders. Stud- ies have found that the index of social anxi- Wednesday late afternoon ety and depressive rumination in adolescents Track B is significantly negatively correlated with the Jun 12, 16:00 - 18:00 hemodynamic response of the cognitive con- trol network in the frontal lobe. In general anxiety-related research, the lateralization in- OR-149 dex is used as one of the indicators to evalu- ARCUSUR SYSTEM: supporting the way to ate the degree of anxiety during experiment of accreditation programs subjects facing threats. The aims of this study Virgina Laura Ballarin are to explore the utilization of fNIRS record- IFMBE AC Member - National University of Mar del ings from relevant brain areas to understand Plata, Argentina neural mechanisms and neuromodulation of In 1991 the MERCOSUR Education Sector anxiety disorder. created the Meeting of Ministers of Education To assess the anxiety level, the cerebral hemo- of the member countries of MERCOSUR (MER) 110 as the entity responsible for the coordination programs in Biomedical Engineering (BME) of education policies in the region. In June through the lens of a university with a rich his-of 1998, in Buenos Aires, the MER approved tory of providing such programs for over five the Memorandum of Understanding on the decades. The Metropolitan Autonomous Uni- Implementation of an Experimental Mechan- versity (UAM) is one of the first institutions in ism of Accreditation of Careers for the Re- Latin America to include an academic pro-cognition of University Degrees in the MER- gram in biomedical engineering. The evol- COSUR Countries. Advisory Commissions on ution of the curriculum at the undergradu- Agriculture, Engineering and Medicine, pro- ate level, the creation of the first postgradu-duced the documents ”Dimensions, Compon- ate program in this discipline, the incorpora- ents, Criteria and Indicators for the Accredit- tion of professors trained in foreign universit-ation of MERCOSUR” in their respective areas ies in the most important BME research top-but recently in November 2006, the Meeting ics in the different decades, the influence in of Ministers of Education approved the Op-the region reflected in enrollment of students erational Plan for the creation of the System from countries such as Argentina, Colombia, of Accreditation of University Careers of MER- Paraguay or Bolivia and the consolidation of COSUR (ARCU-SUR). the profession in Mexico through its gradu- Currently, Bioengineering, Biomedical Engin- ate students, has given the university a unique eering and Clinical Engineering Education Pro- personality and made it a Latin American grams have been growing up significantly in benchmark. Delving into various aspects of Latin American Region. After 10 years of the BME education, the presentation sheds light accreditation process implementation in Latin on the journey of the educational institution, America, there is a need for actualization. In-the evolution of curriculum development, and teresting alternatives could arrive looking at the dynamic realm of BME research. It nav- the restructuring process of higher education igates through the milestones achieved and adopted in Europe, starting with the so-called hurdles in shaping BME education in Mexico. Bologna Declaration. Its objectives of increas- Key discussion areas include the inception and ing the comparability, compatibility, transpar- expansion of BME programs, innovative cur-ency and flexibility of higher education sys- riculum design, impactful research endeavors, tems coincide with the concerns and interests and fruitful collaborations with industry part-already raised in various Latin American coun- ners and healthcare organizations. By reflect- tries and with the proposals on quality as- ing on past endeavors and analyzing current sessment and accreditation exposed. On the trends, the presentation provides valuable in- other hand, the under graduate and graduate sights into the accomplishments and chal-programs must be based on the global health lenges encountered in BME education in Mex- needs of the region. ico. Furthermore, it seeks to identify promising It is necessary to deepen the debate on our avenues for future growth and advancement conceptions of quality in higher education in the field, thereby fostering a vibrant eco- beyond Latin America, and scientific societ- system for BME education and research in the ies can smooth the way to new and modern country. accreditation process. Improving the defini- tion of the dimensions, criteria and indicators, OR-151 working together, young countries and coun- Integration of health technology manage- tries with experience, better methodological ment in healthcare service delivery systems approaches to evaluation and accreditation in Kenya could be proposed. Salome Mwaura Association of medical engineering of Kenya, Kenya OR-150 Medical are devices developed to solve a Academic Programs in Mexico: The Current health problem and improve quality of life. State of Biomedical Engineering 50 Years Health care providers require Medical Devices Later for effective and efficient preventive, dia- Fabiola Martinez-Licona gnosis, treatment and rehabilitation services. Universidad Autonoma Metropolitana, Mexico Due to the importance of Medical Devices the This presentation offers a comprehensive WHO expanded its scope and replaced the examination of the landscape of academic term with Healthcare technologies. 111 WHO has a mandate, as outlined in the as Kenya’s vision 2030. The idea involved World Health Assembly (WHA) Resolution leasing of medical equipment to address key 60.29 “to encourage member states to draw areas in specialized healthcare targeting Non-up national or regional guidelines for good Communicable Diseases (NCDs), such as can- manufacturing and regulatory practices, cer, renal, diagnostic, radiological and crit-to establish surveillance systems and other ical care services. chronic diseases like cancer measures to ensure the quality, safety and and diabetes. Kenya, along with other United efficacy of medical devices and, where ap- Nations member states, is working towards propriate, to participate in international achieving Universal Health Coverage (UHC) by harmonization” the year 2030. The World Health Organization Target of Health Technology defines UHC as the ‘ability for people to re- • Improve quality of Health delivery service ceive health services without suffering finan- • Improve efficiency cial hardship. The Kenya Vision 2030 aims to • Improve effectiveness transform Kenya into a middle-income coun- • Reduce costs try that provides better quality life to all its • Improve accessibly of health services citizens. Good health promotes economic Health Technology Planning growth by boosting productivity and contrib- • If you fail to plan you are planning to fail uting to poverty reduction. Managed Equip- • For proper Management of HT there must be ment Service arrangement ensures that pub- a work plan lic hospitals have access to modern health in- • Work plan must consist of at least the follow- frastructure, equipment and/or services over ing an agreed period , with the government mak- Random sampling of few public hospitals ing regular, pre-arranged payments based on raised the following issues of concern; agreed performance parameters. • Biomeds are still not fully integrated into our The Areas of focus included: Theatre Equip- health systems ment , Sterilization Equipment and Theatre In- • Budgetary allocation for maintenance and struments, Renal Dialysis Equipment ,Intens- investment in medical equipment/health ive Care Unit (ICU) Equipment and X-ray and technology is very low and not based on any other Imaging Equipment. data (irrational) The Implementation of this project has how- • Technical competence of Biomeds to repair/ ever done great to the Kenya’s population and provide maintenance service is questionable many can attest to this fact. The Project has (level of training) seen hospitals benefit from specialized, mod- Conclusion ern and state-of-the-art medical equipment • Health technology will bring a revolutionary and by extension, upgrading of hospitals to change in the health sector. Therefore Em- accommodate these equipment, training of bracing Health technology for Health workers hospital staff both Technical and users some is inevitable. in factory training and other on site train- • Patients are also increasingly becoming ing continuously. Citizens having access to aware of their health rights and this is a threat uninterrupted, quality, specialized Healthcare to doctors. (Information available in the services regardless of location within Kenya, internet, information centers, etc) Regular service, maintenance, repairs and re- • There is need for Biomeds to apt in the game placement of spare parts for all equipment of quality healthcare technology manage- throughout the programme at no additional ment. cost to the counties, Improved services in the hospitals due to reduced equipment faults and OR-152 malfunctions, Reduced referrals and import- Managed equipment service in Kenya antly Reduced Mortality rates greatly. Symon Mbakah Although the Projects had its own challenges Association of Medical Engineering of Kenya, Kenya we cannot forget to underscore that the gains The Managed Equipment Service in Kenya achieved from this iniative are ernoumous and is an Idea that was initiated in the year 2015 thus important for developing nations, low with the aim of improving healthcare ser- resource countries can try this approach of vice delivery in Kenya and to enable in real- medical equipment manaagement. ization of Universal health Coverage as well 112 OR-153 ies, which are much larger and have a bigger Needs assessment in Kenyan healthcare fa- user pool, the most critical units are better cilities for context-sensitive design of med- maintained than others. Critical issues were ical devices found in the procurement of both spare parts Florinda Coro 1, Bazil Mlamba2, Augustine (52%) and consumables (43%). Furthermore, Waswa2, Carmelo De Maria1, June Madete2, maintenance was carried out by internal staff Valentina Mangano1, Margaret Keraka2, Arti who are not adequately trained in 50% of Ahluwalia1 cases. The human resources employed were 1University of Pisa, Italy rated insufficient by 35% of the participants. 2Kenyatta University, Kenya Overall, our results show widely different prob- Medical devices (MDs) are essential to lems, perceptions, and constraints (economic, deliver healthcare services effectively, but human resources) across structures even in most of the world’s population cannot access the same region, indicating that a more sys- safe and appropriate MDs. According to the tematic methodology for context-sensitive World Health Organization, in low-resource development and deployment of MDs is settings (LRS), 80% of the medical equipment needed. Our current studies are geared in this is donated and, less than one-third of it be-direction. comes operational. Providing equitable access to MD techno- OR-154 logy is a challenge; it requires comprehensive Standardisation of Medical Devices: A Con-knowledge of local constraints and needs. Our tended Practice with Potential Benefits for objective was to conduct a needs assessment South African Healthcare on-field study to maximise the effectiveness Sudesh Sivarasu of medical equipment and design context University of Cape Town, South Africa compatible MDs in LRS. Stakeholder dialogues within the Western The study, approved by the Kenyatta Univer- Cape’s medical equipment maintenance sity Ethical Review Committee, was carried sector have sparked a debate regarding out from June to August 2023 in Nairobi in standardising medical devices in South Africa. collaboration between the University of Pisa Advocates from both public and private and Kenyatta University. Six healthcare fa- health sectors suggest that unifying medical cilities representing different levels (primary, equipment practices can enhance patient secondary, and tertiary) and sectors (private, safety and lead to cost savings. However, this public) were included in the study, to capture stance faces opposition, notably in the public generalizable constraints of LRS and to obtain finance domain and in some approaches from a checklist of characteristics that biomedical the global north, which perceive standardisa- engineers must consider when designing MDs tion as a potential hindrance to competition to be used in these environments. The first and innovation. qualitative phase was carried out through Our investigation critically assesses the ar-a focus group (N=4) and semi-structured guments for and against standardisation, interviews (N=3) with biomedical engineers. considering its potential as a recommended Then, a questionnaire on different aspects of practice in South Africa and comparable MD management (to be rated from 1 to 5) was economies. The literature on the subject delivered to healthcare workers employed in remains limited, with some critiques citing the facilities selected for the study (N=54). concerns that cost-driven standardisation The results of focus groups and semi- could compromise patient safety by neglect- structured interviews were used to refine ing specific device requirements. Moreover, the questions to be included in the question- international advocacy for standardisation is naires for quantitative research. Answers were lacking as a strategic approach to achieving more heterogenous in the public than in the Universal Health Coverage or the Sustainable private sector. For example, the availability of Development Goals. electrical security systems in private facilities In South Africa, policies such as the Public Fin-was rated consistently (75% of participants ance Management Act clearly oppose stand-5/5, 25% 4/5), whereas in one public facility, ardisation, framing it as anti-competitive. the opinions of the respondents were at odds Despite these challenges, certain hospitals with each other: 85% rated the systems as 5/5, and clinical departments, especially within while 15% rated them as 2/5. In public facilit- 113 the private sector, proactively implement issues faced by healthcare workers in using standardisation for its reported safety, train- the malaria diagnosis test kits. ing efficiency, and maintenance benefits. Methods: To ensure adherence to malaria This research illuminates anecdotal evidence case management protocol of ‘Test, Treat suggesting that standardisation can reduce and Track (T3s), a mobile phone installed with costs and enhance patient safety through the Fionet App was deployed. Under this pilot uniformity of equipment, ease of mainten- study, 28 pharmacy shops were selected from ance training, and availability of spare parts Greater Accra Region and 13 Over the Counter in settings where healthcare resources are Medicine Shops (OTCMS) from Central Region constrained and the care burden is heavy, ad-were selected. The pharmacists and Medical vocating for standardisation deserves further Counter Assistants from these facilities were scrutiny. Such exploration should be balanced trained to use the Fionet mobile phone App for with thorough health technology assess- malaria testing. The shops were also supplied ments, allowing for a nuanced approach that with RDTs through the National Malaria Con- prioritizes patient welfare over competitive trol Programme. Inventory of RDTs supplied market principles. to these shops were tracked to ascertain References: Hyman, W.A. ”Device stand- whether some tests were conducted without ardisation can be dangerous.” Infec- using the Fionet App and to understand the tion Control Today, 2018. Available at: motivations behind the bypass through inter- https:/ /www.infectioncontroltoday.com/view/device- views. standardization-can-be-dangerous Results: A total of 5312 RDT tests were recorded [accessed on May 24, 2023]. Pub- through the Fionet platform, out of which 3972, lic Finance Management Act, Act representing 74.7% were malaria antigen neg- No. 1 of 1999. Available at: ht- ative and 1286 (24.2%) were malaria antigen tps:/ /www.treasury.gov.za/legislation/pfma/act.pdf positive. The uploaded records also reported [accessed on May 26, 2023]. 1.05% tests as error. 91.91% of malaria positive cases were treated with ACTs, where as 14.85% OR-155 of malaria negative cases were also treated Digitizing Malaria Case Management Pro- with ACTs. Of the malaria negative cases tocol in Ghana: A pilot Study that were treated, 9.1% of patients insisted George Boadu they want ACT treatment and 91.8% were Komfo Anokye Teaching Hospital, National Centre given ACT treatment because the patients for Radiotherapy and Nuclear Medicine, Ghana were showing malaria symptoms according Introduction and Background: It has been to the caregivers. For the RDTs that were a common practice in Ghana for patients not processed through the Fionet platform, to seek first aid medical attention from the caregivers indicated that the incubation community pharmacies or chemical shops. period is too long when they use the Fionet Management of malaria cases is no exception platform, as it compels them to stick to the to this, as many patients who experience incubation period. symptoms of malaria visit pharmacy shops to Conclusion/Recommendation: The gold seek medical attention. The Ghana Health standard for malaria treatment should be Service guidelines for malaria case man-followed to ensure continuous drug. agement ensure that all persons presented with malaria symptoms at any health facility OR-156 including pharmacy and chemical shops From Emergency to Enhancement: Stream- are tested through microscopy or malaria lining Ventilator Procurement in Africa for Rapid Diagnostic Test (mRDT). Malaria case Improved Healthcare Outcome management interventions such as point of Sudesh Sivarasu care testing using mRDT kits ensures accurate University of Cape Town, South Africa diagnosis and effective treatment and can In the wake of the COVID-19 pandemic, the be well monitored through a real time data management of health technology in Africa capturing solution. This study evaluated the has come under scrutiny, highlighting a re- use of a mobile health technology platform, liance on the ”if it works, let it work” philo-known as the Fionet Platform to remotely sophy. This perspective, though it emphasises monitor malaria RDT testing and understand operational stability, neglects the progress-114 ive nature of healthcare necessities, particu-effectively expands dataset size while main- larly evident in the procurement of ventilat- taining HRV pattern realism. This aids in devel-ors. Traditional bureaucratic barriers, exempli- oping and refining models used in AF research, fied by South Africa’s stringent B-BBEE rating overcoming challenges associated with lim-system, exacerbate the lag in accessing these ited sample sizes. Emphasizing privacy consid- life-saving devices, directly impacting health- erations, this approach showcases the poten- care quality and responsiveness. tial of classic statistical methods in synthetic The abstract delves into the ventilator case data generation for advancing AF research study, exploring the advancements and chal- within the constraints of small datasets. lenges associated with their procurement in African health systems. It proposes the har- OR-158 monisation of medical device regulations as a An explainable deep-learning model to aid pivotal solution for expediting the introduction in the diagnosis of age related macular de-of locally manufactured health technologies generation into the market. With a specific focus on vent-María Herrero-Tudela1, Roberto Romero-Oraá 1, ilators, our research suggests that streamlin- Roberto Hornero2, Gonzalo C. Gutiérrez-Tobal1, ing approval processes could be revolutionary, María I. López1, María García1 1 fostering the medical device industry’s growth University of Valladolid, Spain 2 and sustainability within the continent. Universidad de Valladolid and Centro de Investiga- Through this abstract, we aim to spark con- ción Biomédica en Red en Bioingeniería, Biomater- structive discourse on eliminating procure- iales y Nanomedicina (CIBER-BBN), Spain ment roadblocks. By adopting a unified reg-Age-related macular degeneration (AMD) ulatory framework, innovation can flourish, is the most frequent cause of blindness in and health technologies, like ventilators, can people of an advanced age. As AMD is asymp- become more accessible. Such advance- tomatic in its early stages, this condition is nor-ments are crucial for creating a resilient and mally identified in advanced stages of the dis- equitable healthcare infrastructure, ensuring ease, when treatments are less effective. To all African nations are better equipped to im- address this challenge, automated AMD im- prove health outcomes and strive toward uni- age assessment systems offer the potential versal health care coverage. to significantly reduce the time, costs, and effort involved in screening. While previous works have demonstrated success for AMD Artificial intelligence in detection using convolutional neural networks, healthcare their lack of explainability mechanisms lim- its their use in clinical settings. To address Wednesday late afternoon this limitation, we propose an explainable Track C deep-learning approach using Local Inter- Jun 12, 16:00 - 18:15 pretable Model-agnostic Explanations (LIME). Our model, based on RegNetY-320, achieved 86.5?curacy, 85.21% sensitivity, and 91.01% spe- OR-157 cificity on the Automatic Detection challenge Insights in Data Generation: A Synthetic on Age-related Macular degeneration data-Data Approach for Enabling Small Datasets set. Through the LIME technique, we identified in Atrial Fibrillation Research the specific areas in retinal images that influ- Ali Salman 1, Francesco Goretti2, Alessandra ence the prediction of the model, providing a Cartocci1, Ernesto Iadanza1 tool for clinical interpretation and enhancing 1University of Siena, Italy diagnostic confidence. 2University of Florence, Italy This study explores the Gaussian Copula OR-159 Synthesizer’s (GCS) utility in addressing the lim- Wearable sensors and artificial intelligence itations of a small dataset (58 real patient re- for blood pressure estimation during daily cords) in Atrial Fibrillation (AF) research, focus- life and sleep, a pilot study ing on Heart Rate Variability (HRV). Leveraging Katy Stokes 1, Salman Haleem2, Rossana this method, we generated a realistic synthetic Castaldo3, Francesco Cappuccio1, Leandro dataset of 1, 000 records, replicating the fea- Pecchia1 tures observed in the original records. The GCS 1University of Warwick, United Kingdom 115 2Queen Mary University of London, United Kingdom diagnostics. The research aims to invest- 3IRCCS SDN-SYNLAB, Italy igate the effect and potential advantages High blood pressure is the leading risk factor transformer models have on binary semantic for development of cardiovascular disease segmentation, the method utilized for the (CVD). It is now understood that 24-hour and project. Hence, I employed the SegFormer nocturnal measures are better predictors of model, for its lightweight architecture, as the CVD than one-off measurements. Despite this, primary deep learning model, alongside the devices are costly, uncomfortable, inaccurate Unet. A custom 2D computerized tomography and disrupt sleep. Machine learning has been (CT) scan dataset was assembled, CT-Org2D applied to the electrocardiograph (ECG) for through meticulous operations. Extensive ex-blood pressure estimation with a view to de- periments showed that, in contrast to the se- velopment of non-disruptive continuous blood lected models, the task’s simplicity required pressure monitors. Prior studies have focused a redesigned Unet architecture with reduced on data collected under highly controlled set- complexity. This model yielded impressive restings in hospitalised populations, this study as- ults: Precision, Recall, and IOU scores of 0.91, sesses machine learning’s viability with wear- 0.92, and 0.85 respectively. The research serves able ECG sensors in real-life scenarios. as a starting point, motivating further ex- In this study, a protocol was developed, and ploration, through different methodologies, to data were prospectively collected from 30 vo- achieve even greater efficiency in organ seg-lunteers for a 24-hour free-living period of mentation. continuous ECG, ambulatory blood pressure OR-161 monitoring and continuous glucose measure- Rare Eye Diseases Automatic Classification: ments. Machine learning models, emphas- A Deep Learning Approach ising understandable ECG features and in-Jacopo Vitale 1, Maria Elisabetta Pagnano1, terpretable models, were trained and tested Margherita Anna Grazia Matarrese1, Rosa Boccia2, through a thorough holdout process. Differ- Paolo Melillo2, Francesco Testa2, Francesca ent combinations of ECG morphology fea- Simonelli2, Leandro Pecchia1 tures and activity metrics were used as inputs. 1Università Campus Bio-Medico di Roma, Italy The final model achieved a mean absolute er- 2University of Campania “Luigi Vanvitelli”, Italy ror of 7mmHg and Pearson’s’ correlation coefficient of 0.48. Statistical and feature import- Two uncommon genetic eye disorders, Ret- ance analysis identified key ECG regions for initis Pigmentosa (RP) and Stargardt Disease blood pressure prediction. (STGD), both classified as Inherited Retinal This research improves the understanding of Diseases (IRDs), are subjects of investigation the feasibility of continuous blood pressure for potential treatments employing advanced monitoring during daily activities and sleep us- technologies, including artificial intelligence ing machine learning and wearable ECG sig- (AI) with fundoscopy. These IRDs underscore nals. Inherent data quality challenges as- the complexities of genetic eye disorders, em-sociated with wearable sensors must be ad- phasizing the imperative for sustained re- dressed in consideration of the aspects of in- search efforts and innovative interventions, in- terest of the ECG signal. cluding AI applications, to mitigate the impact of vision loss. The aim of this work is to create OR-160 an algorithm for automatically classifying the Unlocking robotic potential through modern fundoscopies of 74 paediatric eyes. Three arti-organ segmentation ficial intelligence algorithms are developed, a Ansh Chaudhary, Robail Yasrab Custom Residual Net, a Custom Vision Trans- CCIR, United Kingdom former and a YOLO Net. Among the three mentioned algorithms, only the YOLO Net suc- Deep learning has revolutionized the ap- cessfully classifies the test set samples without proach to complex data-driven problems, committing any misclassification errors. The specifically in medical imaging, where its tech- ultimate goal of this study is to introduce the niques have significantly raised efficiency in most effective classification algorithm to aid organ segmentation. The urgent need to in the diagnostic process of rare diseases. enhance the depth and precision of organ- based classification is an essential step to- wards automation of medical operation and 116 OR-162 and behaves. This illness can be treated with Empowering Colorectal Cancer Research the aid of Transcranial Direct Current Stimu-through Advanced Data Integration and lation (tDCS), which can help to reduce the Analysis: A case study of the DIOPTRA pro- symptoms of depression. The level of illness ject is typically evaluated using the Hamilton De- Marilena Tarousi, Stavros-Theofanis Miloulis, pression Rating Scale (HDRS). The focus of this Maria Haritou, Konstantinos Bromis, Ioannis Kouris, paper is the prediction of the HDRS score after George Botis, Ioannis Kakkos, George Matsopoulos a tDCS course. By predicting the result of tDCS, Biomedical Engineering Laboratory, School of Elec- psychiatrists can provide better counseling to trical and Computer Engineering, National Tech- the patients about their future conditions after nical University of Athens, Greece the treatment and decide wisely about the The escalating incidence of colorectal can- treatment method. We used different kinds of cer (CRC), paired with the progressively de- demographic information, treatment inform- creasing age threshold and the low screening ation, and the HDRS score before the treat-adherence by citizens, underscores the urgent ment as predictors and supervised Machine need for innovative approaches to screen- Learning (ML) algorithms for the prediction. ing and early detection. In this light, the DI- The analysis is conducted on 169 patients with OPTRA EU project presents a pioneering initi- depression. Our preliminary results show that ative aimed at harnessing advanced data col- the accuracy can be up to 63% when predict- lection, integration, and analytics methodo- ing the value of HDRS after tDCS treatment logies to unlock critical insights into CRC. This sessions as a binary variable using Gradient paper provides a comprehensive overview of Boosting. the DIOPTRA platform’s conceptual architec- This is encouraging on such a small data set. ture and its potential implications for address- Moreover, our results provide insight into the ing CRC research challenges. Leveraging data predictors pivotal to this outcome. They show integration, artificial intelligence, and state- that the HDRS score at baseline, the age, and of-the-art biomarker analysis, DIOPTRA offers the gender of the subject are the three main a promising solution for enhancing early de-predictors. The results suggest this methodo- tection and thus improving patient outcomes. logy may yield very interesting results. The integration of the DIOPTRA Back-end, fa- cilitating data ingestion, curation, and stor- OR-164 age, with the Front-end’s user-friendly inter- Evaluation of Hydrogel Flow into Osteo-faces and a dedicated anonymization tool, porotic Trabecular Bone: A Computational offers the required link between clinical prac- Fluid Dynamics Study tice and innovative research in CRC screening Fahad Alabdah 1, Adel Alshammari1, Araida programs. By fostering interdisciplinary part- Hidalgo-Bastida2, Glen Cooper1 1 nerships and embracing continuous innova- School of Engineering, University of Manchester, tion, DIOPTRA has the potential to revolution- United Kingdom 2 ize CRC screening practices, reduce mortality Department of Life Sciences, Manchester Metro- rates, and shape the future of healthcare de- politan University, United Kingdom livery. In this regard, this paper stresses the im- There is significant potential in the treat- portance of ongoing efforts to advance can- ment of osteoporosis using hydrogels tis- cer screening technologies, highlighting the sue scaffolds impregnated with cells, growth role of integrated platforms like DIOPTRA in im- factors and antibiotics. However, the deliv- proving public health outcomes through mul- ery of hydrogel into the bone is yet to be de- tidisciplinary research within clinical settings. veloped. Non-invasive delivery is desirable for both patients and healthcare systems. Aim: OR-163 evaluate hydrogel flow mechanisms and how Predicting Depression Status After Trans- that can affect the bone structure as well cranial Direct Current Stimulation Treat- as the impregnated cells to begin the un-ment Using Machine Learning derstanding of injection delivery. Methods: Sayna Rotbei, Giordano D’Urso, Alessio Botta hydrogel flow dynamics are investigated us- University of Naples Federico II, Italy ing a linear laminar flow computational fluid Depression is a serious medical illness that dynamics model. Three cases with different adversely affects how a person feels, thinks, Reynold’s (Re) numbers were simulated and trabecular bone wall shear stress, and hydro-117 gel flow velocity were extracted to evaluate of classification performance. bone/cell damage and velocity profile. Res- ults: higher Re number which is proportional to higher inlet velocity is associated with higher S11 – IFMBE Education in BME: bone wall shear stress, and hydrogel pressure. Education in Biomedical Although the shear stress and pressure do not Engineering for Students pose any risk regarding the bone structure, the cell viability is compromised in the case Thursday morning Track B of Re=6. Conclusions: the out-come of this study provides insights that will contribute to Jun 13, 10:30 - 12:00 the design of a hydrogel injection procedure. Further work is needed to improve the model OR-166 to further inform hydrogel delivery mechan- An introduction to Ventra; a programmable isms. abdominal phantom for training, educa- tional, research, and development purposes OR-165 Salar Tayebi 1, Ashkan Zarghami1, Manu Enabling COVID-19 detection from multiple Malbrain2, Johan Stiens1 audio recordings: a preliminary comparison 1Vrije Universiteit Brussel, Belgium between cough, breath, and speech signals 2Medical University of Lublin, Poland Alfonso Maria Ponsiglione1, Francesca An- In recent years, several concepts of med- gelone 1, Rossella Sparaco1, Salvatore Piccolo1, Amy ical equipment have been recommended to Parrish2, Andrea Calcagno3, Guillaume Fournier2, measure intra-abdominal pressure (IAP). Ac- Ayana de Brito Martins4, Fulvio Cordella4, Arianna cording to the research guidelines of the Arienzo4, Lorenzo Castella3, Vincenzo Norman Abdominal Compartment Society (WSACS), Vitale1, Francesco Amato1, Maria Romano1 1 every novel method to measure IAP should University of Naples Federico II, Italy 2 be strictly validated against the actual gold Laboratoire National de Santé, Luxembourg 3 standard of the IAP measurement technique, University of Turin at ASL “Città di Torino”, Italy 4 which is an indirect measurement method VoiceMed S.a.r.l., Luxembourg through the bladder. Therefore, further valida- Patients with COVID-19 experience severe tion studies should be conducted to systemat-respiratory and vocal difficulties as well as ically investigate different criteria and the ro- symptoms that give rise to unique audio char- bustness of a novel alternative for IAP mon-acteristics in their voices. The present study itoring. Clinical studies provide the highest takes advantage of vocal biomarkers extrac- level of evidence and insight, but they are ted from cough, speech and breathing audio often expensive and not always feasible for recordings obtained through personal smart- investigating novel concepts. Preclinical ex-phones from both SARS-CoV-2 virus-infected periments involving cadaveric or animal spe-persons and non-infected participants ac- cimens encounter various challenges. Hu- cessing two different healthcare facilities. The man cadaveric tissue availability is limited, results provide findings on the use of acoustic and ethical concerns surround animal experi-feature sets taken from low-level feature rep- ments. In previous years, abdominal phantoms resentations for COVID-19 recognition from have been used for different purposes includ-cough, breath, and speech patterns. Machine ing medical imaging research and develop-learning models were trained on datasets from ment, radiotherapy dosimetry studies, her-individual vocal exercises and on a dataset nia repair, laparoscopic and robotic surgery from combined exercises (cough, breath, and training, and abdomen biomechanical stud- speech). The classification models provided ies. Nevertheless, no abdominal phantom up to 68.6?curacy, 86.7% sensitivity, and 66% has been developed to simulate the IAP in specificity, whose values and most significant an artificial abdominal compartment. This features vary according to the type of vo- study aims to introduce the “Ventra” as an cal pattern examined and the type of model abdominal phantom that is able to simulate adopted, indicating that audio characterist- intra-abdominal pressure, abdominal compli-ics may be used to detect COVID-19 symp- ance, respiration-related pressure variations, toms and that the combined use of multiple and urine input and output, and is compat- audio patterns from different vocal tasks can ible with intra-vesical and intra-gastric meas- achieve the most encouraging results in terms urement devices for validation, educational, 118 and training purposes. The abdominal com-AI’s integration into medical diagnostics of- partment is designed as an inflatable balloon fers a pertinent example. Machine learning al- with a total volume of 4000 mL. An artifi- gorithms, fueled by vast datasets, have shown cial bladder was made using 500 mL styrene promise in areas like disease detection and ethylene butylene balloon (and placed inside treatment planning. However, concerns about the abdominal compartment. Subsequently, bias, data privacy, and algorithmic transpar-Foley catheters and nasogastric tubes were ency underscore the critical need for eth-connected to make the phantom compatible ical considerations within AI-driven biomed- with intra-gastric and intra-bladder pressure ical applications. For instance, the reliance on measurement devices. Vacuum DC pumps biased datasets in AI diagnostics could per- were used to inflate and deflate the abdom- petuate disparities in healthcare outcomes if inal compartment and to increase and de- not addressed ethically. We see that eth- crease the IAP. Two peristaltic pumps were also ics should be seamlessly integrated into en- used to instill and remove fluid to and from the gineering courses to emphasize its relevance bladder. As the next step, a graphical user in- in real-world applications. By providing stu- terface was made to enable the user to con- dents with ethical frameworks and tools to trol the phantom. Several validation analyses navigate complex ethical dilemmas inherent were done to ensure the functionality and per- in AI-driven biomedicine, engineering educa-formance of the phantom. Reviewing the val- tion can cultivate a generation of respons-idation results, we can see that Ventra shows ible technologists. Recently we started an excellent reliability and accuracy in simulating Erasmus+ project whose goal is to integrate different IAP, compliance, and respiration vari- teaching of ethical considerations related to ations, and marks a significant step forward in AI into engineering faculties at higher educa- the realm of medical training, education, and tion institutions, utilizing innovative pedago- research. gical approaches. The project will produce three concrete results. The first is Ethical AI OR-167 Case Studies & Starter Kit, which includes case Ethics as part of biomedical engineering studies and starter kits to teach engineering and informatics education students about the ethical implications of their Lenka Lhotska work. The second is Ethical Engineer OERs, Czech Technical University in Prague, Czech Repub- flexible open education resources based on lic real-world case studies that lecturers can use Ethics education in engineering has under- to teach AI & Robotics Ethics in their programs. gone significant evolution since its inception Finally, the project will create Ethical Engineer in the 1970s, yet its integration into curricula Bootcamps, an interdisciplinary approach to remains inconsistent across disciplines. As a teaching ethical considerations in AI and ro- research discipline, engineering ethics educa- botics to students that will be systematised in tion appeared in late 1990s and early 2000s. a guide for other institutions to replicate. While biomedical engineering (BME) programs The work is supported by the project No. 2023-at institutions such as the Czech Technical Uni- 1-CZ01-KA220-HED-000160663 Ethical Engin-versity in Prague began incorporating ethics eer: Integrating teaching ethics in artificial in-education in the 1990s, other fields like elec- telligence and robotics into Engineering Edu-trical engineering and information technology cation of Erasmus+ programme. have been slower to adopt similar initiatives. Although ethics is present to a certain de- OR-168 gree in BME and biomedical informatics edu- Implementation of a Pattern Classifier on cation, it has not yet fully reflected the is- Thermograms from Plantar Region sues connected with artificial intelligence (AI) Santiago Humberto Ramirez Martinez, Martha L. Zequera, Francisco Carlos Calderón Bocanegra applications in biomedicine. As AI becomes Pontificia Universidad Javeriana, Electronics De- more prevalent, it introduces complex ethical partment - Footlab, Colombia concerns related to privacy, security, fairness, and social discrimination. Addressing these This study aims to implement a pattern challenges requires a workforce that can un- classification algorithm for plantar thermo- derstand and navigate the ethical implica- grams, focusing on identifying altered tem- tions of AI development and implementation. perature zones in the feet of diabetic patients. Utilizing a database of 334 thermo- 119 grams, various classi-fication algorithms in-urnal HRV parameter trends were analyzed cluding Support Vector Machine (SVM), Lo- to identify potential indicators for monitoring gistic Re-gression, Artificial Neural Network the prognosis of heart failure patients. Using (ANN), Random Forest, and K Nearest Neigh- commercially available Holter monitors, one bors (KNN) were evaluated. Features ex- set of 24-hour ECG data was collected for tracted from the literature, such as number each group: normal, hypertension, and heart of pixels, maximum entropy, variance, mean failure. Every half hour, a 5-minute segment of value, correla-tion, contrast, energy, and ho- ECG was selected for short-term HRV analysis. mogeneity, were utilized for training and eval- Analyzed parameters included SDNN, LFP, uation. The classification task involved assign-and HFP. Results showed significant diurnal ing thermograms to 5 classes based on the differences in SDNN, LFP, and HFP indicators thermal change index (TCI). among normal subjects, whereas diurnal dif- Performance evaluation was conducted using ferences in HF subjects were not significant. an information theory metric approach based This might be due to the disappearance of di- on mutual information, measuring the align- urnal rhythm in sympathetic and vagal nerves ment between predicted and true classes. The of HF patients or a decrease in regulatory neural network achieved the highest mutual capacity. Hypertension subjects exhibited information score of 2.69 out of 5, indicat- diurnal differences in indicators, but the dif-ing that approximately 53.8% of the informa- ferences were opposite to those of the normal tion obtained from model predictions aligned control group, indicating that the heart still with the true classes. Additionally, a database has regulatory capacity but is under load was established in the Footlab BASPI labor- during rest. Through analysis of diurnal HRV, atory, comprising 20 thermograms from the this study demonstrated distinct patterns of plantar region of 10 control subjects. A novel HRV in HF subjects compared to the other protocol, incorporating additional elements to two groups, serving as one of the reference the STANDUP base protocol, was proposed. indicators for the prognosis of HF patients. Finally, classification using the ANN on data acquired from the Footlab - BASPI database OR-170 yielded satisfactory results, successfully distin- Proposal for the Application of Blockchain in guishing between 1.7 classes, representing an Predictive Management in Medical Devices 85% success rate in classifying thermograms. Mariana Brandão, Renato Garcia Institute of Biomedical Engineering (IEB-UFSC), OR-169 Brazil Using 24-hour Heart Rate Variability indices The veracity and reliability of information for prognosis monitoring in heart failure pa- in the health sector is essential for security tients in the Predictive Health Technology Manage- Shi-Yi Wu, Shao-Hung Lu, Mei-Fen Chen, Wen- ment (HTM) and quality in patient care in Chen Lin, Wen-Chi Lin, Cheng-Lun Tsai, Kang-Ping health care establishments. One technology Lin that has been contributing to the traceabil- Chung Yuan Christian University, Taiwan ity and security of data is Blockchain, which Heart failure (HF) patients experience im- consists of a decentralized network for stor- paired contraction or relaxation of the heart, ing transactions in an immutable way. The ob- leading to inadequate blood circulation, jective of this work was to develop a study of which can easily result in complications such the intersection between blockchain techno- as pulmonary edema or difficulty breathing. logy and the application of HTM by Clinical En-Due to the influence of sleeping posture at gineering. A rapid literature review of block- night, pulmonary edema is prone to worsen. chain applications in medical devices was de-Normally, heart rate variability (HRV) should veloped and presented two models of integ- exhibit a diurnal rhythm, but after the onset of ration of this technology in hospitals. The heart failure, the body’s sympathetic nervous first proposal was about the incorporation of system regulation capability decreases, po-Blockchain in the traceability of information tentially causing disruption in the diurnal incorporated in the historical record in the life- rhythm of the heart. Based on this theory, this cycle of the technology and the second in the study collected 24-hour electrocardiogram application in calibration certificates and test (ECG) data from three groups of subjects: reports on medical equipment. normal, hypertension, and heart failure. Di- 120 OR-171 Slovenia EMG Rehab: An interactive platform con- 2Faculty of Mathematics and Physics, University of trolled by EMG signals to improve adher- Ljubljana, Slovenia ence to rehabilitation therapy in patients 3Department of Experimental Oncology, Institute who have suffered lower body muscle injur- of Oncology Ljubljana, Department of Experi-ies mental Oncology, Slovenia Leslie Yessenia Cieza Huané, Ana Cristina Hyperspectral imaging (HSI) holds poten- Aldana Palomino, Sergio Enrique Moreno Elescano, Andre Jesus Cruces Chancahuaña, Angel Eduardo tial in clinical applications, provid-ing valu- Dianderas Jorge, Pablo Cardenas Caceres able information about perfusion using vis- Facultad de Ciencias e Ingenieria, Universidad ible light. The notable advantage of HSI is Peruana Cayetano Heredia, Peru that it is non-contact, non-ionizing, and non- invasive, while also eliminating the need for Physical rehabilitation is crucial to maximize contrast agents. In medical practices, HSI has function and minimize disability in those who shown promising results in detecting differ- have suffered patellar tendon , however it de- ent types of carcinoma and assessing perfu- pends not only on the method, but also on how sion in healing processes like transplants and it is performed as depending on this the treat- wound management. Considering the crit-ment may be more effective than another in ical role of perfusion in oncology, we have re- the patient. This paper introduces EMG Re- viewed the utility of HSI in quantifying perfu-hab, an interactive platform for the physical sion changes during clinical interventions. Our rehabilitation of young college students who review primarily focused on neoplasms loc- have suffered lower body injuries. Employing ated in anatomical sites such as the kidneys, electromyography (EMG) in conjunction with breasts, eyes, brain, and gastrointestinal (GI) the Bitalino device and Python programming tract. Through our examination, we have es-language, quantifies muscle intensity during tablished that HSI demonstrates remarkable therapy. Through interactive gaming, EMG potential as an imaging technique within a Rehab fosters more efficient and engaging clinical oncology setting. Notably, we found treatment, characterized by a patient-centric the evaluation of mastectomy skin flaps per-approach that balances challenge and enter- fusion after the breast reconstructive surgery tainment. The proposed platform utilizes EMG and anastomotic perfusion during GI conduit to assess muscle electrical activity, comple- reconstruction to be particularly promising. mented by a dynamic rehabilitation environ- ment featuring video games designed to mo- OR-173 tivate patients and facilitate targeted exer- OHIO: Integrating IoT Technologies for En-cises. EMG Rehab offers personalized feed- hanced Clinical Engineering and Dynamic back by recording and analyzing muscle activ- Tracking of Medical Equipment ity, empowering patients to track progress and Alessio Luschi 1, Gianpaolo Ghisalberti2, Gio-improve performance. This interactive game- vanni Luca Daino2, Vincenzo Mezzatesta2, Ernesto based platform is designed to engage young Iadanza1 college students actively in therapy, thereby 1Department of Medical Biotechnologies, University enhancing rehabilitation outcomes. of Siena, Italy 2Azienda Ospedaliero Universitaria Senese, Italy Clinical engineering The paper outlines the OHIO project, fun- ded by the European Union’s Horizon 2020 re- Thursday morning Track C search and innovation action programme, via the ODIN – Open Call issued and executed un- Jun 13, 10:30 - 12:00 der the ODIN project (GA 101017331), which fo- cuses on the enhancement of hospital safety, OR-172 productivity, and quality. The main object- Assessing Perfusion Changes in Clinical ive of OHIO is to provide a solution to em-Oncology Applications using Hyperspectral power the management of hospital facilities Imaging in terms of clinical engineering and logistics Rok Hren 1, Jošt Stergar2, Urban Simončič2, for the use-case hospital “Le Scotte” in Siena, Gregor Serša3, Matija Milanič2 Italy. The OHIO framework integrates various 1Institute of Mathematics, Physics and Mechanics, technical solutions to address possible man-121 agement challenges which may arise during OR-175 the maintenance of medical equipment, such Towards the Creation of a National Medical as the unavailability and the untraceability of Equipment Inventory a device, or the technicians’ lack of knowledge Aris Dermitzakis, Spilios Zisimopoulos, Nicolas on how to reach it. The paper also describes Pallikarakis a set of Key Performance Indicators (KPIs) de- Institute of Biomedical Technology (INBIT), Greece signed for measuring OHIO’s impact on main- During the last decades, medical tech- tenance timings and efficiency. Preliminary nology has become a key factor in modern outcomes show promising results in prevent- healthcare delivery systems. Despite that, as ing failures, improving scheduling, and provid- showcased in the WHO 2022 Global atlas of ing efficient indoor navigation. The OHIO pro- medical devices, most countries worldwide do ject demonstrates potential enhancements in not have any form of national medical equip- hospital operations and maintenance through ment (ME) inventory, and in most cases where digital IoT solutions, successfully responding to data is available, it refers to high capital value a set of previously unmet needs. equipment. As a result, no immediate inform- ation about the operation, status and costs OR-174 of ownership of the equipment is available to An Innovative Solution for Efficient Work- national healthcare authorities, health policy flow Management in Healthcare and decision makers and other stakehold- Alessio Luschi, Ernesto Iadanza ers. Tending to the above issues, the Institute Department of Medical Biotechnologies, University of Biomedical Technology (INBIT) has under- of Siena, Italy taken a nationwide project regarding the cre- Hospitals are challenged to provide a wider ation of an electronic ME inventory of all hos- range of services due to a growing patient pitals belonging to the public healthcare sec-population, leading to a faster deterioration tor in Greece. rate compared to other buildings. Recent The procedure for creating the inventory en- years have seen a focus on improving main- lists a room-by-room approach, where every tenance management in hospitals through piece of equipment is assigned a unique na-strategies, performance measurements, tional device number, in the form of an in-and Information Technology. Challenges delible label bearing the number and a QR include resource allocation, communication code. Essential data that will be included in gaps, and workflow monitoring. Healthcare the electronic record (ER) of each equipment workflow management involves various stake- contain photos, GMDN and UDI codes, man-holders and aims to ensure safe, efficient, and ufacturer, model/type, year of manufacture, effective patient care while minimizing waste area of installation, operation status and ac-and reducing costs. Workflow Management quisition method. Information like the tech-Systems (WFMSs) are promising solutions, nical support provider, acquisition costs and automating administrative procedures to en- maintenance data can also be incorporated. hance efficiency and effectiveness in health- A comprehensive and updated ME inventory care services. The paper presents a RESTful provides a clear picture of the technology WFMS developed in Microsoft .NET 6, utilizing available, is essential for strategic procure- the open-source .NET library Elsa Workflows. ment planning and constitutes the corner-It introduces pre-typed sub-workflows, each stone for the use of a Medical Equipment defined in JSON format, allowing users to Management System (MEMS) and a Vigilance customize workflows using Graphic User Inter- system. Another collateral benefit of a ME faces (GUIs). The system stores information in inventory can be the establishment of vari- a Microsoft SQL Server database and inter- ous ME related centralized databases. A acts with the hospital Computer Aided Facility ME group codification database, using the Management (CAFM) system via APIs. The total GMDN groups will enable the seamless platform enables managers and department transition to a new system, like the EMDN. A heads to create customizable workflows in a model/manufacturer database will be funda- no-code environment. The system is used in mental, as it will allow checking at any time an actual healthcare environment at the ”Le whether the devices installed in the country Scotte” University Hospital in Siena, Italy. meet the necessary standards and specifica- tions, as well as their involvement in adverse 122 events. Finally, a department database can a part of the project. help in the redesign and central organization This project aims to improve various aspects of the health system, providing insight into of the healthcare system. The computerized the geographical distribution and operational MEMS facilitates the tasks clinical engineers status. are responsible for performing and provides In conclusion, the availability of such reliable reliable data enabling evidence-based de-data regarding the installed ME is necessary cision making. Strategic planning of invest- for evidence-based decisions regarding in- ments will be feasible, knowing the installed vestment planning, the rational management ME and its operational status, as well as the of the equipment and its safe operation. This overall capacity of healthcare units in differ- will also form the basis for a centralized MEMS ent regions. A redistribution of ME can also be at a regional or national level. made, taking into consideration the real needs of health care units that may appear in spe- OR-176 cial circumstances, like in the period of covid- Implementing a Centralized Medical Equip- 19. In addition, equipment downtime is expec-ment Management System on a National ted to be reduced. The various automated level and streamlined processes encompassed by Aris Dermitzakis, Spilios Zisimopoulos, Nicolas the MEMS not only limit the need for paper-Pallikarakis based documentation, but also facilitate pre- Institute of Biomedical Technology (INBIT), Greece ventive maintenance schemes, thus reducing The use of medical equipment (ME) is device malfunctions. prominent and essential in every modern In conclusion, the implementation of a cent- healthcare environment. ME management ralized MEMS is expected to improve the vi-covers a wide range of important issues, in- ability and robustness of the national health- cluding acquisition, evaluation, maintenance care system, through informed decision mak-and adverse event reporting. Due to the com- ing, while at the same time enhancing access- plexity of modern technology, the use of med- ibility to healthcare, patient safety and service ical equipment management systems (MEMS) quality. is necessary, resulting in many and important benefits. An indicator of the above is the wide- OR-177 spread usage of MEMS in European healthcare Calibration of medical equipment - why is it units during the last decades. However, ac- critical? cording to data from the WHO 2022 Global Symon Mbakah atlas of medical devices, most countries lack Association of Medical Engineering of Kenya, Kenya a national inventory and MEMS. The Institute CALIBRATION OF MEDICAL EQUIPMENT - of Biomedical Technology (INBIT) has under- WHY IT IS CRITICAL? taken a pioneering project for the implement- Calibration of medical equipment is an activation of a centralized MEMS (web-Praxis) on a ity carried to determine the conventional truth nationwide level in Greece. This system will al- of the value of the designation of measur- low for a complete and updated picture of the ing tools and materials. This is done by com- available equipment on the whole territory, its paring traceable measuring standards to na- technical characteristics, geographic distribu- tional and international standards. It is actu-tion, the operational status, the cost of owner- ally a procedure for detecting and fixing the ship and other economic data among others. uncertainties in measurements and bringing Web-Praxis consist of various modules. The them to an acceptable level. The accuracy equipment electronic record module plays a of the device has a great deal of importance, pivotal role, as it encompasses information as it can seriously affect the diagnostic pro-from other modules like Corrective and Pre- cedure and endanger patients’ life including ventive maintenance, Contract management, the users as well. The deviation in output is Geographical distribution, Vigilance, KPIs and not necessarily the fault of the device itself, Statistics. Relevant information includes iden- nor can it represent the manufacturer’s incom- tification data, photographic record, main- petence. Most commonly, it results from con- tenance history, contracts and economic tinuous device usage, which over time affects data. For the MEMS to be feasible, the exist- its internal components, resulting in a devi-ence of a ME inventory is a prerequisite and ation from the device standard measurement will also be established at a national level, as 123 output. The reason why it is important calibration of medical equipment as a norm is simply to ensure that the results generated from the various medical equipment are ac- curate and therefore do not in any way harm the patient or the health worker. As a medical professional, you have a responsibility to your patients to make certain that your medical equipment is safe, accurate, and reliable at the point of use. Calibration is done to provide that the item is functioning properly and to ensure that the results or readings it provides are accurate. Individuals completing calibra- tion need to be trained and are required to document the process to ensure any mistakes can be traced back to the source. When Clini- cians are subjected to various results from this equipment as diagnosis, they proceed to pre- scribe drugs as well as administer procedures as dictated by the equipment results. How- ever, some clinicians who have a wealth of ex- perience occasionally may doubt the results and therefore will request a second opinion at the cost of the patients from other and in most cases the doubt is usually confirmed. To have medical equipment calibrated at some interval of time and the results circulated to the clinicians. There is therefore the need for trained Biomedical engineers, use of appropri- ate tools and adherence to calibration pro- cedures at all times. The paper will therefore give a highlight of the importance of calibration of Medical equip- ment, give incidences of misdiagnosis caused by lack of calibrations, guide on importance of allocating resources for calibration, training of continuous Biomedical engineers, monitor- ing and documentation of the entire process of calibration. 124 P O S T E R P R E S E N T A T I O N S ’ A B S T R A C T S Poster session Machine Interface (BMI) and the decision lo- gic of the AI. First, RP was measured in an Tuesday morning poster electrostatic shielded room for each left and right foot under following three conditions. In session the first and second conditions, participants Jun 11, 10:00 - 10:30 imagined stepping on a button (Image) and stepped on a button (Move), and in the third PO-01 condition, they stepped on a button at an Neural network based fetal ECG extraction arbitrary timing (Voluntary). The waveforms from abdominal signals were then observed by an additive average of Dragoș-Daniel Țarălungă1, Radu Botezatu2, 50 trials. As the results using ANOVA, signific- Alina-Elena Sultana1, Titus Mihai Vasile2, Georgeta-ant differences between the movement terms Mihaela Neagu 1 for the electrodes located in the frontal lobe 1National University of Science and Technology Po- were revealed, but not for the motion sides litehnica Bucharest, Romania or the interaction effects. Multiple comparis- 2Carol Davila University of Medicine and Pharmacy, ons indicates that in the movement terms, the Romania peak means were larger for Image, Move, and Non-invasive foetal ECG extraction from Voluntary, in that order, and the 95% confid-abdominal signals might provide significant ence intervals were independent only for Vol-information for long-term foetal monitoring, untary. In the motion sides, the peak mean being very attractive for physicist. Neverthe-was larger for the left foot, but the confid- less, accurate extraction of the foetal ECG is a ence intervals overlapped each other. Fur- challenging task, due to the disturbing signals, thermore, LSTM using the raw data was able which overlap the signal of interest in the fre- to classify a total of six conditions, taking into quency domain. Among the current denoising account the movement terms and the motion methods, neural net-works are very attractive sides, with a final value of 98.4% high accuracy due to their performance. The current paper and 0% loss. These results suggest that LSTM is pro-poses a linear feed-forward neural net- promising for application to the RP based BMI work that estimates very accurately the ab-by LSTM.As a consideration, the difference in dominal mECG, the strongest disturbing sig- RP peaks by movement terms may be due to nal, based on two thoracic mECG, removing the fact that the prefrontal cortex was activit thereafter. The obtained results are very ated more strongly in the image-directed than promising, al-lowing the further investigation in the non-directed Voluntary. Furthermore, of the fHR, for the foetal well-being evaluation. as for the motion sides, since the dominant The comparison with the event-synchronous foot of all participants was the right foot, the interference canceller shows the advantage right foot was easier to move due to functional of the neural network in preserving the fECG sharing on the motion sides. This may have morphology, with the cost of higher compu-resulted in a slightly smaller peak value for the tation time. Both methods require the prepro- right foot than for the left foot. In conclu-cessing of abdominal signal in order to remove sion, ANOVA and multiple comparisons iden-the power line interference and the baseline tified significant differences in prefrontal ac- wander. tivation based on the presence of instructions for movement and functional sharing among PO-03 motion sides. In addition, the highly accurate Fundamental study of Readiness Potential classification in LSTM shows promise for ap-elicited by foot movement plication of LSTM to BMI using EEG. Reon Takahashi, Puwadej Leelasiri, Tomoya Oi, Tatsuhiro Kimura, Hiroshi Ohshima, Kiyoyuki PO-05 Yamazaki, Fumitaka Aki Automatic detection of sympathovagal re- Tokai University , Japan sponse using HRV analysis. Case study: res- In this study, we measured the Readiness ident surgeons during training and their first Potential (RP) in the left and right foot under laparoscopic surgery following three conditions, analyzed by AN- Maria Elisabetta Pagnano, Jacopo Vitale, Mar- gherita Anna Grazia Matarrese, Gianluca Mascianà, OVA and multiple comparisons, and classified Marco Caricato, Leandro Pecchia the conditions by LSTM, with the aim of elu- Università Campus Bio-Medico di Roma, Italy cidating the appropriate control of the Brain 127 Monitoring the physiological response dur- knowledge are required in order to accurately ing training can provide valuable information diagnose CVDs from cardiac sounds acquired about how surgeons cope with the stress as- by either acoustic or electronic stethoscopes. sociated with this activity. Heart Rate Variab- In many developing countries, like Zambia, ility (HRV) analysis is a widely used method of with low doctor to patient ratios, there is a assessing the autonomic nervous system (ANS) big shortage of experts which leads to limited response, providing important knowledge on access to quality healthcare. the stress state of subjects and how the ANS This study proposes a preliminary develop- intervenes in the response to such stress. Dur- ment and evaluation of a low-cost digital ing a stressful task, the ANS releases hor- stethoscope which leverages on machine mones such as cortisol and adrenaline into the learning to distinguish between normal and bloodstream, triggering an increase in heart cardiovascular disease (CVD) related heart rate, increased perspiration, sharper concen- sounds. A prototype of a digital stethoscope tration and other physiological processes. In is designed and developed with Bluetooth this study, 6 resident surgeons were monitored and universal serial bus (USB) technology. at rest, during training with a surgical sim- preliminary results show that the system re- ulator, and during their first surgery. Cor- cords heart sounds with good quality and relations between HRV features and surgeon low noise. Machine Learning models such as performance were investigated. Six Machine K-Nearest neighbor (KNN), support vector Learning (ML) models were trained to detect machines (SVM) and artificial neural networks mental stress using a public dataset compris- (ANNs) are trained and tested on features ing 84 5-minute ECG recordings labelled as derived from the heart sounds or phonocar- stress/rest. Subsequently, these models were diogram signals (PCG) by Wavelet scattering tested on data collected from the resident sur- and Fast-Fourier Transform (FFT). The overall geons. In conclusion, pre-trained ML mod-validation accuracy range was 89.9%-98.5% els demonstrated significant efficacy, recog- and the models can effectively classify normal nising mental stress in surgeons with a spe-and abnormal heart sounds. These promising cificity and accuracy exceeding 80%. results will be of great importance for early de- tection of cardiovascular diseases, enabling PO-07 timely intervention and disease management Preliminary development and evaluation of in low resource countries like Zambia. a low-cost digital stethoscope Muyanga Kampekete 1, Francis Chikweto2, PO-09 Foster Munsanje1, Boyd Mubanga1, Conix Sunday1, Trade-Off Between Real-Time and Classi- Jimmy Sikwese1 fication Performance in Motor Imagery BCI 1Evelyn Hone College, Zambia Aleksandar Miladinovic1, Milos Ajcevic2, Kater- 2Tohoku University, Institute of Development, Japan ina Iscra 2, Francesco Bassi2, Alessandra Raffini2, Joanna Jarmolowska2, Uros Marusic3, Agostino Cardiovascular diseases (CVDs) are among Accardo2 the leading causes of deaths globally. The 1Institute for Maternal and Child Health “Burlo World Health Organisation (WHO) estimates Garofolo” , Italy that over 17 million lives from the patient pop-2University of Trieste, Italy ulation are lost due to cardiovascular diseases 3ZRS Koper, Slovenia and this is a major concern. Cardiac aus- Brain-computer interfaces (BCIs) offer dir- cultation using either acoustic stethoscopes ect communication between the brain and or electronic versions still remains the most external devices, holding immense potential widely method for the assessment of cardiac across various applications. This study focuses functionality conditions and for the detec- on Motor Imagery-based BCIs (MI-BCI), de-tion of holosystolic (mitral regurgitation and coding neural patterns associated with men- ventricular septal defect), systolic ejection tally rehearsed motor actions. Despite their (aortic stenosis or hypertrophic cardiomy- promise, BCIs face challenges in real-world opathy), and diastolic (aortic insufficiency applications, primarily in reliability and com-and mitral stenosis) murmurs that can precip- plexity. While classification accuracy is a itate the presence of abnormal heart sounds standard metric for BCI performance, the lit- indicating underlying pathological conditions. erature often overlooks real-time responsive- However, experience and domain expertise ness. Many studies report classification out- 128 comes offline, disregarding the prompt trans- ference in mean temperature between restlation of EEG signals into actions. The accept- ing and sleeping, and no decrease in nasal able delay from EEG signal to action should skin temperature was observed during sleep. not exceed 1 second; however, numerous stud- We hypothesized that this is because a depth ies employ time windows exceeding 4 seconds, of 3 cm allows measurement to the cereb- affecting user control perception. This article ral cortex. We considered that the parasym- aims to compare the trade-off between time- pathetic nervous system became active be- window length and classification accuracy in fore sleep onset, which had a relaxing effect MI-BCI, using three linear classifiers (LDA, MLP, and increased blood flow. The right hemi- SVM). Participants include stroke patients and sphere is responsible for functions such as spa- subjects from the BCI IVa dataset. Results tial attention and landscape perception, and demonstrate time-frequency plots indicating it is assumed that these functions are related MI-related EEG changes, revealing a trade-off to driving. Males entered sleep in a short time between accuracy and responsiveness. Our and showed rapid fluctuations, while females findings underscore the importance of ad- took longer to fall asleep and showed gradual dressing real-time responsiveness in BCI evalu- fluctuations. We consider it possible that men ations, proposing a balance for practical sys- are accustomed to their environment and wo-tem utility. In conclusion, this study enhances men are not accustomed to their environment. our understanding of the delicate balance The nasal temperature did not decrease be- needed for optimal real-world application of cause of the shallow depth of sleep. We in-MI-BCIs, emphasizing the trade-off between vestigated changes in cerebral blood flow us-accuracy and responsiveness. ing the NIRS system and verified its relation- ship to sleep. It was found that measurement PO-11 at deep depth was effective and that cerebral Evaluation of Precursors to Momentary blood flow increased before sleep. To examine Sleepiness in Automobile Driving Using the possibility of detecting the antecedents of Cerebral Blood Flow Variability and Ther- sleep, data before and after sleep should be mography taken in detail. Future work is to examine the Tomoya Oi, Reon Takahashi, Puwadej Leelasiri, relevance of the NIRS device by recording fluc-Kiyoyuki Yamazaki, Hiroshi Ohshima, Fumitaka Aki tuations in arousal levels by EEG and analyzing Tokai University, Japan the correlations in detail. In this study, we used a portable NIRS device and a thermal imaging camera to eval- PO-13 uate the stage of drowsiness in a simplified Use of ECG gating approach for high fre- manner for the purpose of elucidating the quency ultrasound vector flow imaging antecedents of drowsy driving. Seven par- Chi-Hung Yang, Chih-Chung Huang ticipants drove on a driving simulator while National Cheng Kung University, Taiwan wearing an NIRS device on their foreheads High frequency ultrasound (HFUS, > 30 MHz) and monitoring their skin temperatures with Doppler imaging has been used widely in many a thermal imaging camera, and a YouTube small animal and human applications due to video was played to simulate driving on a high-its high-resolution ability. Vector Doppler ima- way. Dozing was defined as prolonged eye ging (VDI) also exhibits the advantage for closure, and skin temperature of the nose was visualization of complex flow patterns without measured and filmed when judging dozing. considering the Doppler angle. However, there Subjects were cautioned to concentrate on is no commercial HFUS VDI machine avail- driving and not to try to force themselves to able to date, therefore, several studies have sleep. At the point when the subjects judged used the ultrasound research platform (Vera- that they fell asleep, Hb concentration de- sonics Vantage 256™) connect with a HFUS ar- creased from the starting point and returned ray transducer for HFUS VDI. Unfortunately, the to the original value when they awoke. Com-maximum frame rate of Verasonics is only 10 parison by depth showed that the amount of kHz at 40 MHz operational frequency due to change was greater at 3 cm. The amount the limitation of hardware data transmission, of change in the right prefrontal cortex was which reduces the maximum detectable velo-also greater. Men’s Hb concentrations fluctu- city of Doppler measurement due to the insuf- ated rapidly in relation to sleepiness, while wo- ficiency of frame rate. To address this draw-men’s increased slightly. There was little dif- 129 back, in the present study, an ECG-gating polygon meshes, the UV mapping proced-HFUS VDI was proposed to keep the maximum ure was applied. First, contours (slices) of frame rate of 10 kHz in Verasonics for data the scan body were obtained by rotating a acquisition to avoid the Doppler flow aliasing vertical cross-section plane around the scan by aligning all the tilted angle plane waves body’s longitudinal axis by specific degrees, to the ECG R-wave, which solves the trade- then identifying the closest point on the mesh off between frame rate and multi-tilted angles for each rotated position. Second, for each in conventional VDI. The performance of the vertex, three scalar values were calculated: proposed data acquisition method for HFUS the shape index, latitude angles, and the area VDI was verified firstly through a steady flow of the Voronoi cell. Contours, which contain phantom, which all the estimation errors are vertices, are then organized in ascending or-less than 10% under different flow settings. In der based on their angle sequence. This ar- animal studies, the peak velocities were meas- rangement of vertices, distinguished by their ured through the ECG-gating HFUS VDI from scalar values, encodes specific features of mice carotid artery, left ventricle, and aor- the scan body into a 2D format. To achieve tic arch, which were approximately 55 mm/s, a fully standardized format-matrix, nearest-655 mm/s, and 765 mm/s. Compared to con- neighbor interpolation was applied to the ver- ventional method, no Doppler aliasing occurs tices. After further normalization images were in the proposed method because the frame analyzed using Shannon entropy and image rate is sufficient. All the experimental res- histograms. For the classification problem ults demonstrated the proposed ECG-gating concerning each image type, Gradient Boost-HFUS VDI has the potential to become a useful ing Classifiers were utilized. Results: The ac-tool for vector flow visualization in small anim- curacy estimate for latitude angles demon- als even under a high flow velocity. strated the most contrasting outcomes based on entropy, with E4 and PRIMESCAN exhibit- PO-15 ing the lowest entropy values (<3 bits). In con- Intraoral scanner classification and accur- trast, the remaining scanners showed values acy evaluation using UV mapping of scan above 3.5 bits. The entropy of the shape in-bodies dex presented similar results, with values ex- Mykolas Akulauskas 1, Darius Jegelevičius1, ceeding 5 bits for all scanners, except for E4, Vygandas Rutkūnas2, Liudas Auškalnis3, Justinas which was at 4.8 bits. In terms of classification Pletkus2 performance, the entropy of the Voronoi cell 1Kaunas University of Technology, Lithuania area provided the most distinguishable results. 2Vilnius University, Lithuania Histogram-based classification achieved ac- 3DIGITORUM Research Center, Lithuania curacies of 0.86 for the longitudinal angle, 0.89 Background: The crucial stage of dental for the shape index, and 0.92 for the Voro-implant placement with a digital workflow in- noi cell area. Conclusions: Employing image- volves scanning the patient’s dental arches based representations of scan bodies for the using an intraoral scanner (IOS), where scan classification of scanners allows for an effect- bodies (implant-positioning-transfer devices) ive comparison of relative accuracy among are present. Despite IOS’s advantages over them or with known dimensional shapes. traditional impressions, their accuracy evalu- ation is complicated by varied protocols and PO-17 reliance on black-box solutions. The object- Multi-input CNN based Classification of ive of this study is to develop algorithms that EEG and NIRS signal during voluntary hand can distinguish between different IOS systems movement and to evaluate their accuracy using digit-Puwadej Leelasiri 1, Reon Takahashi1, Fumitaka ized scan bodies in a standardized format. Aki2, Hiroshi Ohshima2, Kiyoyuki Yamazaki2 1 Materials and Methods: The study’s data- Graduate School of Engineering, Department of base includes 2222 scan bodies extracted from Biomedical Engineering, Tokai University, Japan 2 698 dental arches, each digitized using four School of Engineering, Department of Biomedical different IOS (PRIMESCAN, Carestream 3600, Engineering, Tokai University, Japan TRIOS3 and TRIOS4) and one laboratory scan-The brain-computer interface (BCI) is a ner E4. For reference, the CAD model of popular communication method between the the CARES RC Mono scan body was used. brain and computer or electronic devices. To standardize the scan bodies presented in An electroencephalography (EEG) signal is a 130 popular way to create a traditional BCI sys- PO-19 tem. However, when the brain signal from Combining quantitative MRI, fluorescence, EEG is sensitive and inter-feres with external and neuropathology in frameless brain tu-and internal noise, the system design is com- mor needle biopsies plicated, and Malfunctions may occur in the Elisabeth Klint, Anders Tisell, Ida Blystad, Martin BCI system. Near-infrared spectroscopy (NIRS) Hallbeck, Teresa Nordin, Johan Richter, Karin Wår-that is strong in external noise supports EEG dell work in BCI systems. However, the sampling Linköping University, Sweden frequency from two of these artifacts is sig- Introduction: Quantitative MRI (qMRI) nificantly different. It needs some complic- holds the promise of probing tissue micro-ated process to combine them or design a sys- structure non-invasively but has been limited tem that considers the characteristics of both by long acquisition times and incomplete val-data. This study examines the system design idation with standard techniques. Recently, with a flexible convolutional neural network multidimensional multi-echo MRI sequences (CNN) setting to reduce the effect of differ- have enabled acquisition and calculation of ences in sampling frequency for the significant relaxometry maps within clinically accept-accuracy classification of the movement side able times. Our group has previously set up of left- and right-hand movements by using a workflow for acquisition of tissue fluores- raw EEG and NIRS data on multi-input CNN. cence during frameless brain tumor needle In this study, nine healthy right-handed parti-biopsies. The fluorescence measurements cipants between the ages of 20 and 30 per- were included in a pipeline combining analysis formed studies involving human participants in of pre-, intra- and postoperative imaging accordance with the ethical standards of the data with neuropathological assessment. This institutional and/or national research com- study aimed to add qMRI to the workflow and mittee and with the 1964 Helsinki Declaration compare the findings with fluorescence and and its later amendments or comparable eth- neuropathology in the biopsied volume. ical standards. This study was approved by Methods: Ten patients undergoing tu- the Review Board of Tokai University. None of mor needle biopsy were included (EPM 2020-the experimental participants have ever used 01404). In addition to the clinical MRI nav- the BCI. Participants were seated comfortably igation protocol (3T, Siemens, Healthineers, and squeezed left and right hands Separately Germany), qMRI multidimensional multi-echo to measure EEG and NIRS during hand squeez- imaging (voxel size = 0.7x0.7x4 mm3; slice ing. Each participant’s EEG and NIRS wave- gap 1 mm) was acquired before and after forms were averaged to observe the appear- gadolinium (Gd) administration. From the ance of RP and cerebral blood flow before in- qMRI sequence, synthetic images and R1 re- put into the proposed model. In this study, in- laxation rate maps were calculated (SyMRI put size on EEG and NIRS was set at 7x250x1 v0.45.38, SyntheticMR AB, Sweden). During and 4x20x1. Kernels size and stride on each in- surgery, neuronavigation (StealthStation S8, put was set differently before combine to clas- Medtronic Inc, USA) with real-time measure-sify. This model was trained using 80 epochs ments of 5-aminolevulinic acid-induced fluor- with 0.005 learning rate and 16 mini batch size. escence was performed using an in-house de-According to the result, the proposed model veloped system and probe. After surgery, could make classifying a task with 90 percent postoperative CT or MRI were acquired. accuracy on the participant-dependent 9- Processing was done in Python utilizing a fold cross-validation method possible. There- wrapper (Nipype) including FMRIB’s Software fore, the proposed model can classify a la-Library and Advanced Normalization Tools. bel with high accuracy. The multi-input CNN The pipeline includes corrections for magnetic model makes it possible to support utilization field inhomogeneities, skull stripping, and re-in BCI systems about artifact problems such as gistration (either rigid or affine) to the clin- temporal resolution. EEG is superior to NIRS in ical T1w-Gd image space, i.e., the image used estimating the localization of brain function. for neuronavigation during surgery. The fi- On the other hand, NIRS is better at estimating nal biopsy volume was defined on postoperat-relative changes in blood flow. Creating a sys- ive imaging. The defined volumes were trans-tem that exploits both advantages is essential formed to native R1 and R1Gd space. Me-for BCI applications. dian and interquartile range of the voxel values were compared, and the corresponding fluor- 131 escence signals extracted. crocontroller using the open-platform Ardu- Results: Eight biopsy volumes showed 5- ino IDE, and the reading and analysis will be ALA-induced fluorescence and contained tu- sent to the smartphone via the Bluetooth app. morous tissue. Final neuropathological dia- Finally, the ventilation rate is compared with gnoses were seven glioblastomas and one WHO guideline reference values, which are a lymphoma. R1 values (median (IQR)) in the tu- minimum of 60 L/s per patient. morous volumes were 0.68 s-1 (0.48-0.87) be- The small-scale box test was conducted in a fore and 1.1 s-1 (0.83-1.4) after Gd. In the two box with a volume of 10.692L that was used as non-tumorous volumes, R1 values were 0.66 s-1 a single-zone study area, which has different (0.56-0.76) before and 0.79 s-1 (0.73-0.88) after openings mimicking the real environment and Gd and no 5-ALA-induced fluorescence was has different small windows and doors. A fan found. is used to mix the tracer gas during the sat- Conclusion: A workflow for combined ana- uration phase. And when the saturation point lysis of qMRI, fluorescence, and neuropatholo- is reached, it will be stopped, the window and gical assessment was implemented for frame- door of interest opened, and the discharge less brain tumor needle biopies. Through this phase measured by the device we have de- combination, further insights on tissue micro- signed. We also measured the speed of air flow structure could be gained beyond what is by using an anemometer to compare the res- available with clinical MRI. ult, which is strongly in agreement with WHO guidelines. PO-21 Preliminary findings from a study on health- Indoor Natural Ventilation Assessment in care facility ventilation systems suggest us-Healthcare Facilities in Low-Resource Set- ing tracer gas dispersion analysis for real-time ting monitoring and optimizing ventilation. This Nahimiya Husen Ibrahim, Vincenzo Piemonte, method offers a simple way to estimate vent- Leandro Pecchia ilation performance without complex systems, Università Campus Biomedico di Roma, Italy emphasizing data-driven decision-making to In low-resource settings (LRSs), keeping ad- improve ventilation strategies and enhance equate ventilation in healthcare facilities is es- infection control in low-resource healthcare sential for ensuring air quality and prevent- settings. ing healthcare-associated infections among patients, healthcare workers, and visitors. PO-23 While the World Health Organization (WHO) Towards Equity in Healthcare: Designing has established ventilation guidelines, imple- an Affordable Spirometer for Low-Resource menting effective ventilation assessment sys- Settings tems in LRSs remains challenging. Utilizing Pedro Checa Rifa, James Wallace, Davide natural ventilation strategies, such as open- Piaggio ing windows and using architectural design University of Warwick, United Kingdom to promote airflow, can significantly and Pulmonary diseases contribute significantly cost-effectively improve indoor air quality in to global mortality, accounting for almost 4 resource-limited healthcare facilities. This million deaths in 2016. Within this stagger- study addresses the need for affordable and ing number, 90% of them occur in low- and practical methods to assess ventilation in such middle-income countries, highlighting the settings. clinical need for the design and develop- Ventilation assessment was tested within the ment of an affordable and frugal spirometer, Plexiglas box-controlled environment by us- capable of accurate measurements whilst ing the mass balance equation of the con-providing enough resilience to withstand the centration decay method of tracer gas (eth- conditions encountered in such settings. anol), which poses no risk to patients’ health. Currently, over 75% of the medical device use The tracer gas sprayed the substance into a can be accounted for by less than 15% of room model designed from a box with various the world population, suggesting inequitable openings and measured its dispersion using a healthcare access. Nonetheless, the right to sensor characterized by high sensitivity to the health is a fundamental right of every human gas molecule. Real-time ventilation rate and being, as recognized in the World Health air change, as well as the number of people al- Organization Constitution of 1948. This right lowed in the room, will be processed by the mi- is also at the basis of Goal 3 of the United 132 Nations Sustainable Development Goals, General Self-Efficacy Scale (GSES) were meas- published with the aim of “leaving no one ured 1 week before and after training. The behind”. This unequal situation can be further Mann–Whitney U test was used for continuous undermined by the expected exponential variables, Fisher’s exact test was used for vari-growth in terms of populations of low- and ables with two categories, and Kruskal–Wallis middle-income countries. H test was used for variables with three cat- The primary focus of this project is to en- egories, to compare the difference of demo- hance the set of medical devices available graphics between the experimental and confor respiratory tract diseases and conditions trol groups. A Wilcoxon signed-rank test was in low-resource settings, with a specific focus used to compare the difference of outcome on those in low-income countries and lower- measures before and after training for each middle-income countries. group in two Hospitals. The results revealed The device was designed following the user that immersive video games, including action and expert-driven contextualised approach, and puzzle videogames, and real-world activ-with a commitment to sustainable and ities incorporating whole-body physical activ- environmentally-conscious ethos. The device ities had a positive and significant impact (p was developed based on Arduino, a differ- < 0.05) on the cognitive function, self-efficacy, ential pressure sensor MPX5500DP, Matlab, motor function, and balance of the schizo-and 3D-printed parts, relying on Bernoulli’s phrenia patients. Among the different inter-principle. Usability studies and risk assessment ventions, IVRAVGs were the most beneficial and management were performed based on for cognitive function, followed by conven- IEC62366 and ISO14971 and guided the design tional real-world interventions incorporating phase. The device was tested for accuracy physical activities. TVRPVGs and conventional performing tests both with a pump and sub-sedentary-only interventions had the weak- jects. est effects on the measured outcomes (p > The poster will present the results of this 0.05). The effects of interventions that primar- project. ily involved sedentary activities may have had weak effects and, in some cases, even negat- PO-25 ive effects on outcomes such as balance abil- Efficacy of a virtual reality–based video ity. game intervention for individuals with schizophrenia spectrum disorder PO-27 Jen-Suh Chern 1, Yu Yang2, Yi-Ching Tsai1 Range of Motion and Morphological Char- 1National Taiwan Normal University, Taiwan acteristics of the Sacroiliac Joint 2Tri-Service General Hospital Beitou Branch, Taiwan Seonjin Shin, Dai-Soon Kwak The purpose of this study was to examine The Catholic University of Korea, Korea, Republic of the effects of immersive virtual reality action The aim of this study is to investigate how video games (IVRAVGs) and tabletop virtual the morphological characteristics of the au- reality puzzle VGs (TVRPVGs) on cognition, mo- ricular surface may influence the SIJ range of tor function and self-efficacy of schizophrenia motion (ROM) to improve the diagnosis and patients. This study was conducted in two treatment of SIJD. hospitals. Patients from hospital A underwent We measured SIJ ROM using motion-analysis an IVRAVG training (Audioshield, table tennis cameras in 24 fresh cadavers of Korean adults and tennis), while those from hospital B under- (13 males and 11 females). Using three- went a TVRPVG training (Brain Age and Puyo dimensional renderings of the measured auri-Puyo). Each hospital has its own matched par- cular surface, we investigated the correlations ticipants in experimental and control groups. between the morphological characteristics of The training regimen of both hospitals was the auricular surface and the ROM of the SIJ. three 40-minute sessions each week for 12 The SIJ ROM was between 0.2° and 6.7° and weeks. Time to complete Color Trails Test 1 was significantly greater in females (3.58° ± and 2, number of blocks moved with domin- 1.49) compared with males (1.38° ± 1.00). Divid- ant and non-dominant arm in Box and Block ing the participants into high-motion (3.87° ± test (BBTD and BBTND), distance reached in 1.19) and low-motion (1.13° ± 0.62) groups based Functional Reach Test (FRT), time to complete on the mean ROM (2.39°) showed no signi-Timed Up-and-Go Test (TUG), and scores of ficant differences in any measurements. Ad-133 ditionally, bone defects around the SIJ were showed a similar tendency. It was strong at the identified using computed tomography of the distal area from the femoral head. Thus, the high-motion group. In the low-motion group, present results indicate that a risk factor for calcification between auricular surfaces and periprosthetic fracture is high in the cemented bone bridges was observed. polished tapered femoral stems, of which the The morphological characteristics of the au- friction coefficient is small. In the future, the ricular surface do not affect SIJ ROM, which response will be investigated for cyclic loads. was greater in females than in males. In terms of auricular surface measurements, in- PO-31 cluding width, height, inferior length, superior Effect of Tissue on Stress-Strain Character- area, inferior area, total area, and surface me- istics of Pulmonary Pleura dial groove, males had significantly larger val- Hitomi Sakai 1, Hirosane Hayashi1, Sumiko ues than females (p < 0.05). However, when Maeda2, Noriyuki Takano1 1 comparing the high-motion and low-motion Kanazawa Institute of Technology, Japan 2 groups, there was no significant difference in Dokkyo Medical Univercity, Japan shape, size, and overall morphology of the au- Primary spontaneous pneumothorax is ricular surface. This study provides valuable in- caused by the breakdown of the brevis, which sights into the differences in SIJ ROM between occurs when the visceral pleura is detached females and males, as well as the anatomical from the lung surface. Although the mech-structures surrounding the SIJ that influence its anism by which the brevity occurs has not ROM. Our findings have the potential to con- yet been elucidated, it has been reported tribute to the enhancement of SIJD diagnosis that it often occurs at the apex of the lung, and treatment. suggesting that not only the shape and characteristics of the lung itself, but also the PO-29 mechanical properties of the pleura of the Effect of friction coefficient on subsidence lung may have an effect. Although there have of cemented polished stem in THA not been many studies on the mechanical Noriyuki Takano 1, Yuuta Nakajyou1, Ayumi properties of the lung, it is known that the Kaneuji2 stress-strain characteristics of the lung pleura 1Kanazawa Institute of Technology, Japan show nonlinear changes due to the influence 2Kanazawa Medical University, Japan of elastin and collagen, and that it is generally It is well known that cemented polished isotropic. However, in the tensile test of por- tapered femoral stems are stable by their sub- cine visceral pleura conducted in this study, sidence, but a risk factor for periprosthetic the results varied greatly depending on the fracture increases by compressive stress indirection of loading, confirming anisotropy. duced by the subsidence. The subsidence is Tissue residue was observed on the surface of affected by the friction between stem and ce- the specimens used in the tensile test. There ment, which depends on their materials and is connective tissue between the lung paren- surface condition. The present paper used chyma and the visceral pleura. We expected the nonlinear finite element method to estim- that the interlobular septa connected to ate the effect only of the friction coefficient this tissue would remain on the pleural side between stem and cement on the subsidence when the visceral pleura was peeled off from and stress of cement mantles. A polished the lung to prepare the specimens, and thus tapered stem, which was assumed to be made affecting the measurement results. The resid- of titanium alloys, was inserted into a femur ual microstructure and the anisotropy of the model truncated femoral head from CT data stress-strain curves in each of the tensile test of a female patient with hip osteoarthritis. specimens were therefore checked. Therefore, A cement mantle layer was made between the residual microstructure and the aniso-the stem and the femur model. The cement tropy of the stress-strain curves in each of the layer and the femur were fixed. The friction tensile test specimens were checked. The res- coefficient between the stem and the cement ults showed that the anisotropy of the pleura mantle varied from 0.1 to 0.9. The static load is influenced by the residual microstructure on 1817 N was applied to the tip of the head at 15° the surface of the specimen, while the pleura downward to the bone axis. The subsidence itself is isotropic. of the stem increased as the friction coeffi- cient. Compressive stress to cement mantle 134 PO-33 PO-35 Evaluation of the biocompatibility of Dual growth-factors delivery (VEGF and interface-free polymer-ceramic fusion BMP-2) with alendronate improve cell ad- technology using laser-driven penetration hesion in osteoinductive bone repair synthesis Amira Raudhah Abdullah 1, Mohd Riduan Minseong Chae 1, Kang-Sik Lee1, Yu-Chan Kim2, Mohamad2, Ka Liong Tan1, Nadia Mohd Effendy1, Hojeong Jeon2, Seung-Hoon Um2, Hwachul Jung3, Nur Nabilah Abu Bakar1, Intan Maslina Musa1 Dongkyu Koo3, Dong-Ho Lee1 1Universiti Sains Islam Malaysia (USIM), Malaysia 1University Of Ulsan, Korea, Republic of 2Universiti Teknologi Malaysia, Malaysia 2Korea Institute of Science and Technology (KIST), The dynamic mechanisms inherent in bone Biomaterial Research Center, Korea, Republic of homeostasis yield invaluable insight for ad- 3Innosys Co.,Ltd., R and D, Uijeongbu, Gyeonggi-do, vancing scaffold biomaterials in bone regen-Korea, Republic of eration. The increasing recognition of drug Bioactive hydroxyapatite (HA) coated im- delivery systems and the release of bioactive plants on inert Polyether ether Ketone (PEEK) substances is significantly elevating their im- substrates are widely utilized in orthopedic in- portance in bone tissue engineering. This ap- tervertebral disc prostheses and joint replace- proach not only supports bone tissue form-ments. However, the adhesion of HA coatings ation but also enhances the scaffold’s abil- to PEEK surfaces is hindered by a known weak- ity to facilitate bone ingrowth. Bisphosphon-ness in the bond between the two materials, ates (BPs) play a crucial role in bone remodel-resulting in frequent delamination. ing subsequently affecting bone regeneration. In an effort to address this issue, a spe- Despite this, there is a scarcity of studies ad- cialized nanosecond laser-based process was dressing the systematic delivery of BPs within utilized to synthesize interface-free HA/PEEK bone defect models. In this study, integra- surfaces. This method involves the controlled tion of bisphosphonates pamindronate (Pam) melting of the PEEK interface, enabling the co-and alendronate (Aln) into a hydroxyapatite existence of HA and molten PEEK. (HA) scaffold with MC3T3-E1 cells and growth As a result, a robust bond forms between factors (VEGF and BMP-2), is expected to yield the implant and bone interface, preventing a synergistic effect, intensifying osteoinduc-the detachment of the coating after it is im- tion and efficient bone regeneration. Cell vi- planted. The primary objective of this study ability was measured using 2,5-diphenyl-2H-was to assess the bonding strength of in- tetrazolium bromide (MTT) assay and mor- terface free PEEK specimens within a rabbit phological assessment was documented us-model. Thin rectangular samples (5 x 8 x 1 mm) ing the inverted microscope. Characteriza-of HA/PEEK and pure PEEK, manufactured us- tion of engineered HA scaffold was performed ing a laser fusion process, were bilaterally im-using Field emission scanning electron micro- planted into the tibia tuberosity of New Zeal- scopy (FESEM), and its elemental analysis was and White Rabbits (2.5 kg). done using energy-dispersive X-ray (EDX) ana- Following an 8-week observation period, lysis. The mineralization rate was assessed by the rabbits were euthanized, and Micro CT analysing alkaline phosphatase (ALP) expres-imaging was used to examine the implant and sion. Data demonstrated that Aln offers better bone interface. Subsequently, tissue morpho- potency on osteoblast cells as compared to metric analysis was conducted around each Pam. FESEM micrograph revealed that the en- implant site using Goldner’s trichrome stain. gineered HA-VEGF+BMP-2/Aln scaffold facil- The results demonstrated that the bone- itated osteoblast attachment and spreading, to-implant contact (BIC) for HA/PEEK was forming a concrete connection with HA scaf-32.778%, while pure PEEK achieved a BIC of fold. Engineered HA-VEGF+BMP-2/Aln also 20.637%. Consequently, the HA/PEEK grafts significantly increased ALP expression, indicat-exhibited a 12.141% higher BIC value than ing that the extracellular matrix is advancing pure PEEK, highlighting the superior bonding into the mineralization phase. To conclude, strength of the HA/PEEK interface. our investigation unveils the synergistic ef- fects of combining dual growth factors (VEGF and BMP-2) with BPs, specifically Aln, resulting in enhanced cell adhesion on hydroxyapat- ite scaffolds. This emphasizes the substantial promise of employing such a strategy in pro- 135 moting the regeneration of bone tissue. opment of the Republic of Serbia, contract number [451-03-65/2024-03/200107 (Fac- PO-37 ulty of Engineering, University of Kragujevac) Cardiac Hypertrophy Simulations Using and 451-03-66/2024-03/200378 (Institute Echocardiography-based LV Model of Information Technologies, University of Bogdan Milicevic 1, Miljan Milosevic2, Milos Kojic2, Kragujevac)]. The authors acknowledge Nenad Filipovic1 support from the City of Kragujevac, Serbia. 1Faculty of Engineering, University of Kragujevac, Serbia PO-39 2BioIRC, Serbia Evaluation of brain activity using NIRS to ex- In our research, we introduced the utiliza- amine the antecedents of microsleep tion of shell elements in creating left ventricle Fumitaka Aki, Tatuhiro Kimura, Hiroshi Ohshima, models based on echocardiographic data. Kiyoyuki Yamazaki We conducted simulations to analyze the Tokai University, Japan effects of cardiac hypertrophy and monitored Every year, Traffic accidents caused by the progression of the disease. Hypertrophy drivers falling asleep at the wheel are frequent affected changes in both wall thickness and on Japan’s highways. In this experiment, we displacement field, which we adjusted to focused on microsleeps, which are particularly reflect hypertrophic effects. Using our left dangerous. Microsleep is sleep that goes un-ventricle models, we simulated both eccentric noticed for a few seconds during a meeting or and concentric hypertrophy effects and ob-while driving. This type of brief, unconscious served alterations in ventricle shape and wall sleep is called microsleep. The purpose of this thickness. Concentric hypertrophy led to wall experiment is to examine whether the charac-thickening, whereas eccentric hypertrophy teristics of microsleep and its precursors can resulted in wall thinning. To represent passive be detected from Brain activity in frontal lobe stresses within the left ventricle model, we em-regions. Brain activity in frontal lobe regions ployed a recently developed material model is assessed using near-infrared spectroscopy based on Holzapfel experiments. Additionally, (NIRS). The NIRS device has a simple structure our specialized shell composite finite element that can be worn like a headband and is very models allowed for more efficient simulation lightweight, so it does not impose a burden of heart mechanics compared to conven- on the user. First, volunteers were instructed tional 3D models. Furthermore, our approach to sit on a chair with the NIRS device on. Vo-to modeling echocardiography-based left lunteers were than presented with a less in-ventricle mechanics can be applied practic- formative video. The experiment was terminally, as it relies on patient-specific geometry ated when microsleep was visually observed. and experimental constitutive relationships. Cerebral blood flow variations obtained from Our model provides insights into hypertrophy 10 volunteers were additionally averaged and development in realistic heart geometries graphed. As a result, a characteristic spike-and offers potential for testing medical hypo- like waveform was observed 100 seconds be-theses regarding hypertrophy evolution from fore the onset of microsleep. Furthermore, a healthy state to disease progression, under the waveform at a distance of 3 cm showed various conditions and parameters. greater cerebral blood flow variability than Acknowledgement: The research was sup- that at a distance of 1 cm. This spike-like wave- ported by the STRATIFYHF project that has form corresponds to behavior resisting drowsi- received funding from the European Union’s ness and is considered a precursor to dozing Horizon Europe research and innovation off. Since unconscious anti-drowsing beha-programme under Grant Agreement No vior is an activity of the hypothalamus and 101080905. Views and opinions expressed are other deep brain regions, blood flow at a dis-however those of the authors only and do tance of 3 cm may be increased. By detect- not necessarily reflect those of the European ing such spike-like waveforms with NIRS, the Union or granting authority - European Health antecedents of microsleep can be determ-and Digital Executive Agency (HADEA). Neither ined. However, the large individual differences the European Union nor the granting authority in these waveforms make classification based can be held responsible for them. This re- on nonadditive data difficult. The future chal-search was also supported by the Ministry of lenge is to improve the accuracy of NIRS de-Education, Science and Technological Devel- tection and microsleep identification using AI 136 and apply it to the field of traffic safety. ages successfully. Two translucent packages were not correctly recognized if it was sep- PO-41 arated by more than above 5 mm from the An IoT-based Medication Calendar for photo reflector. We confirmed that the med-Home-Visiting Nursing ication information was recorded in a Google Yasushi Yamauchi, Junto Hosono Spreadsheet when the medication was taken Graduate School of Science and Engineering, Toyo out, and that the information was automat- University, Japan ically updated depending on whether the Introduction: Home-visiting nurses play medication was taken correctly or not. We a coordinating role in determining the ap- also confirmed that the system sent a noti- propriate method of medication administra- fication corresponding to the correctness of tion based on symptoms and reporting it to the medication. Our system is expected to the physician, and also in collaboration with strengthen cooperation among nurses, doc- people in other professions and patients’ fam- tors, and other medical professionals. ilies. However, nurses cannot be present at all times of the day to administer medications PO-43 to their patients, they do not have a com- Investigation of the use of 3D modeling and plete grasp of their medication status. Previ- predictive analysis for optimizing the man-ous works to solve this problem include a med- agement of the Miljacka River ication support robot (CareBot) and a smart Madžida Hundur, Merima Smajlhodžić-Deljo, Faruk Bećirović, Naida Babić Jordamović, Lejla medication system (WW. Chang et al.), that Gurbeta Pokvić can automatically record medication inform- Verlab Research Institute for biomedical engineer-ation. Both systems require the nurse to place ing, medical devices and artificial intelligence, Bosthe medication in dedicated container. nia and Herzegovina Purpose: Medication calendars are commonly used for medication management in in-home Introduction & Aim: Recent advancements nursing care. We propose an IoT-based med- in hydrological modeling technology offer a ication calendar system that records medic- promising avenue for optimizing management ation information of home care patients as strategies aimed at enhancing the sustainab- online data and sends notifications to nurses. ility of aquatic ecosystems, such as the Mil-Specifically, a photo reflector is attached to jacka River. Employing advanced modeling each compartment of the medication calen- techniques for precise visualization and simu- dar, to detect whether and when medication lation of river hydrology emphasizes the sig- has been taken. nificance of understanding pollution sources Methods: A 3D printer and an acrylic board and their impact on aquatic ecosystems, spe-were used to fabricate the medication calen- cifically focusing on the Miljacka River in this dar, with each compartment designed with a study. The aim of the study is to util- slope so that the medication slides down to ize three-dimensional (3D) modeling techno-the center, without any dedicated container. logy to identify pollution hotspots, assess their A photo reflector is mounted on the center causes, and propose remediation strategies, of each compartment, to detect whether the ultimately promoting the sustainability of the medication is taken or not. In this study, the river’s aquatic ecosystem. system manages a week’s worth of medica- Methods: Initially, a two-dimensional (2D) tions (morning, noon, and night) and identifies model based on satellite imagery was cre-21 doses of medication. Accordingly, 21 photo ated to provide a foundational understand-reflectors are used. ing of the river’s hydrology. This 2D model When the medication is taken out, two micro- was then upgraded to a more sophisticated controllers (Arduino MKR WiFi 1010) record and three-dimensional (3D) model using tools like send the medication information to a Google SketchUp, Revit, and Twinmotion. The pro- Drive spreadsheet via IFTTT. In addition, the cess involved integrating topographic char- contents are notified to the nurse’s optional e- acteristics and physical data, including length, mail address. width, and depth of the river. Additionally, Results: Of the 15 types of counterfeit medi- the methodology included mapping signific- cines and medicine packages, our system re- ant pollution areas along the riverbed, partic-cognized 13 counterfeit medicines and pack- ularly from City Hall to Otoka, to identify pollution hotspots. The identification of these hot- 137 spots involved a thorough analysis of potential Jupiter, has brought microscopic-level high-pollution sources. resolution imaging of the whole body, achiev- Results: The analysis reveals that these hot- ing the clinical ability to ”see the unseen” and spots are primarily caused by streams and col-leading the forefront exploration in precision lectors, highlighting the importance of ac- medicine. By integrating AI into the scan- curately pinpointing pollution sources for ef- ner, the MRI achieves ultra-fast imaging, ultra-fective remediation. Furthermore, the study high resolution, and a fully intelligent work-demonstrates the capability of 3D modeling to flow, helping to enhance diagnostic capabil-provide detailed insights into the river’s hydro- ities and productivity. logy, facilitating the identification of optimal In the field of molecular imaging, the world’s sites for algae colonization for bioremediation first total-body PET/CT scanner, uEXPLORER, purposes. which is known as the ”Hubble Telescope” for Conclusion: This strategy provides an effect- the human body, has led the international ive approach to water purification and pre- wave of total-body PET technology and ex-servation of aquatic life. This research not panded the application boundaries of nuc-only provides a foundation for an extensive lear medical imaging. By integrating AI into database that enhances the understanding of PET/CT, the capabilities of low noise and high river pollution but also increases awareness of contrast can be empowered to other clinical the vulnerability of aquatic life, crucial for the systems, further promoting the technological development of a sustainable water resource evolution and exploring clinical panoramas of management strategy. Healthy aquatic eco-the nuclear medicine industry. systems are vital for supporting biodiversity With Al-empowered technologies, CT scan-and ecosystem services, which in turn can ner offers precise imaging and ease of use have indirect impacts on human health. The throughout the entire clinical spectrum, in- implications of this research extend to the field cluding functions such as smart patient po- of water resource management, encouraging sitioning, AI empowered iterative reconstruc-the exchange of knowledge and information tion, and AI empowered motion correction. among experts dedicated to improving the These functions greatly expand the bound- sustainability of aquatic ecosystems. aries of clinical practice and makes the im- possible possible such as One-beat Cardiac PO-45 Solution and ultra-low-dose solutions. The Development of Medical Imaging in In addition to enhancing the capabilities of China and Its Impact on the Clinical Practice all full-modality medical imaging and treat-Kui Cai1, Guobin Li2, Junan Zheng2, Lei Chen2, ment equipment, AI also has great applica- Qing Mao2, Ruixin Jiang2, Shuheng Zhang2, Taomin tion value and exploration prospects in clinical Zhang2, Wenjing Ma2, Xiaoyue Gu2, Ying Qian 3 quality control and scientific research such as 1Beijing Hospital, China clinical data man agement, versatile annota- 2Shanghai United Imaging Healthcare Co., Ltd., tion tools, and advanced research tools. China In summary, the progress in clinical practice 3Jiangsu Province Hospital, China and the advancement of medical imaging im- Thanks to growing innovations and dis- pact and improve each other, and we believe coveries, medical imaging devices have been that they will help patients to healthier lives widely used not only as a diagnostic tool and contribute for human health enterprise for diseases, but also used to treat, man-through numerous technological innovations age, and predict illnesses in clinical prac- and new findings from clinical studies. tice. During this process, radiologists are fa- cing many challenges, for instance, ”how to PO-47 achieve high-resolution imaging”, ”how to get An explainable XGBoost model to predict scanning done faster ”, and ”how to make pediatric sleep apnea resolution after the system more user-friendly”. To address treatment from new phenotypic informa- these issues, imaging practitioners from China tion have got certain positive results through a Gonzalo C. Gutiérrez-Tobal 1, Javier Gomez- series of new products design, development, Pilar1, Daniela Ferreira-Santos2, Pedro Pereira-manufacturing with advanced technology.In Rodrigues2, Daniel Álvarez1, Félix del Campo1, David Gozal3, Roberto Hornero1 the field of magnetic resonance, the world’s 1UniversidaddeValladolidandCentrodeInvestiga-first whole-body ultra-high-field 5T MRI, uMR 138 ción Biomédica en Red en Bioingeniería, Biomater-new phenotypes can be used to help predict iales y Nanomedicina (CIBER-BBN)„ Biomedical En- whether a child will respond to surgical remov- gineering Group, Spain ing of adenoids and tonsils in the context of 2CINTESIS@RISE, MEDCIDS, Faculty of Medicine of pediatric OSA. Furthermore, both anatomical the University of Porto, Porto, Portugal and clinical characteristics are relevant to 3Joan C. Edwards School of Medicine, Marshall Uni- conduct these predictions. versity, United States Grants. This research has been developed Pediatric obstructive sleep apnea (OSA) under the grants PID2020-115468RB-I00 and causes recurrent breathing pauses during the PDC2021-120775-I00 funded by ’Ministerio night of the affected children, termed apneas de Ciencia e Innovación/Agencia Estatal de and hypopneas. These apneic events lead to Investigación/10.13039/501100011033/’ and an inadequate gas exchange and, eventually, ERDF, A way of making Europe; and by ‘CIBER fragmented and restless sleep. Up to 5.7% en Bioingeniería, Biomateriales y Nanomedi-of children are suggested to be affected by cina (CIBER-BBN)’ through ‘Instituto de Salud OSA, hampering their normal development, Carlos III’ co-funded with ERDF funds. and leading to neurocognitive affectation, PO-49 decreased quality of life, and increased risk Analysis of age-related variations in pho- for cardiovascular diseases. However, children toplethysmography: a Machine Learning show a heterogeneous range of symptoms approach and responses to treatment, being OSA Sara Lombardi, Federica Tavernise, Piergiorgio phenotypes, not yet completely understood, Francia, Leonardo Bocchi one plausible reason for these differences. University of Florence, Italy A recent study automatically defined 3 new pediatric OSA phenotypes with different Vascular aging is an important indicator in odds of recovering after treatment based cardiovascular risk assessment. In this study, on 26 anthropometric and clinical features. we used a machine learning approach to es-In this study, the objective is to use the new timate the cardiovascular age of subjects us-phenotypic information to predict before ing the photoplethysmographic signal (PPG). treatment (surgical removing of tonsils and From PPG, acquired in 115 healthy subjects adenoids) whether a child will recover from aged 18 to 66 years, we extracted a set of OSA after treatment. Accordingly, we used morphological features and Heart Rate Vari- data from 199 children (5-10 years old) to ability parameters. These parameters were train and validate a Gentle Adaptive Boosting used in a cross-validation approach to pre-ensemble model (XGBoost implementation). dict the cardiovascular age of the subjects us-We also used the relative importance of the ing the GradientBoostingRegressor algorithm. variables to conduct an explainable artificial Quantitative performance evaluation showed intelligence (XAI) analysis. A leave-one-out promising results, yielding a mean absolute er-cross-validation method was used for hyper- ror of (6.81 ± 0.86) and a coefficient of de-parameter optimization (15 decision trees), termination equal to (0.44 ± 0.22). Using the and a bootstrap 0.632 method was used for SHAP method, we determined the impact of performance evaluation. Our results showed features on model performance by identifying 79.4% Sensitivity (69.9-88.0?% confidence heart rate change, low signal frequencies, and interval), 65.0% Specificity (51.7-77.3 95% CI), systolic phase velocity as the most significant 77.8% Positive Predictive Value (70.3-86.0?% parameters. These findings improve our un-CI), and 66.5% Negative Predictive Value derstanding about the influence of age on the (58.0-76.2?% CI) in predicting OSA resolution PPG signal, offering potential insights for future after treatment. Moreover, oral anatomy clinical applications in cardiovascular risk pre- (Mallampati score), obesity (body mass in- vention. dex), age, and gasp and chokes presence during the night were highlighted by XAI as the PO-51 most relevant features for the model predic- Influence of fatigue in swimmers suffering tions, showing a minimum of 10% (gasp and from swimmer shoulder pain chokes) and a maximum of 18% (Mallampati Alessandra Raffini, Alex Buoite Stella, Miriam score) of relative importance. Accordingly, Martini, Laura Mazzari, Agostino Accardo we conclude that the information from the University of Trieste, Italy 139 The shoulder joint is susceptible to be dam-stents are sized to the distal main vessel, and a aged in sports with overhead actions, often proximal optimization technique (POT) is per-leading to swimmer shoulder pathology. Fa- formed with a larger diameter balloon to en-tigue can also worsen and in-crease the risk of sure that there is no malapposition at the prox-overuse injuries. Evaluating shoulder kinemat- imal area of the stent. Intracoronary ima- ics during swimming is crucial to identify injury- ging can be used to determine the proximal related movement patterns and to provide a and distal diameters of the vessels of interest; correct physiotherapy treatment. To meas- however, today this advanced imaging is only ure kinematics, inertial and magnetic meas- used in approximately 13% of PCI procedures. urement systems (IMMSs) offer a very versat- Therefore, it is important to understand com-ile approach with respect to traditional video- mon human coronary anatomies and relative based systems. This preliminary study fo- presentations of such tapering to avoid over-cuses on the effects of fatigue on shoulder expanding the given vessel, which can cause joint kinematics in swimmer with swimmer dissections or rupture. The Visible Heart Labor-shoulder compared to healthy swimmers, by atories® receives human heart donations, and using IMMS. 11 young swimmers (5 patholo-these hearts are fixed in an end-diastolic state, gical, 7 male) took part in the study. Each such that the coronaries remain patent. Fol- participant executed 40 seconds of dry front lowing fixation, the left and right coronary os-crawl followed by a fatiguing protocol and by tia were cannulated with venogram catheters, other 40 seconds of dry front crawl. We ana-and contrast was injected into the coronaries. lyze the arm movement relative to the thorax Each heart was then micro-CT scanned (˜80-examining the differences of the movement micron resolution), and the 3D scans were re-amplitudes between healthy and pathological constructed. The coronary arteries were seg- subjects and before and after fatigue exer- mented within Mimics software, and center-cise according to the three rotations: Flex- lines were generated for each artery includ-ion/Extension, Abduction/Adduction, and In- ing the left main, left anterior descending ternal/External rotation. Some slight non- (LAD), circumflex, and right coronary artery significant differences were found after exer- (RCA) and their associated branches. Next, cise compared to before in all the three ro- an automated program was used to meas-tations while a significant difference between ure and record the cross-sectional diameters healthy and pathological subjects was found and corresponding locations of all the points in Flexion/Extension rotation both before and along each centerline, and to calculate the after fatigue exercise. The use of IMMS al- rates of tapering (slope of the linear regression lowed to verify the repeatability of the kin- between the diameter and distance) of each ematic movement and to quantify the rota- branch. To date, this semi-automated ap- tion angles identifying which component of proach was successfully used to quantify the the movement is most affected by the swim- coronary artery tapering of 14 human coron-mer shoulder pathology. However, a larger ary artery branches. Within this sample set, the number of subjects is necessary in order to vessels had maximum diameters between 1.5 confirm the results. and 3.5 mm and tapered at an average rate of 0.018 mm per 1 mm of vessel. This rate in- PO-53 dicates that coronary arteries taper relatively Semi-Automated Approach to Analyze slowly, which suggests that sizing up a POT bal-the Tapering of Coronary Vessels from loon by 0.5 mm may not be necessary. Al-Contrast-Enhanced Micro-CT Scans of ternatively, using a higher pressure with same Human Coronary Artery Anatomy size balloon may be sufficient to account for Amanda DeVos, Paul A. Iaizzo the artery tapering. The described method- University of Minnesota, United States ologies can also be used to build a dataset Coronary artery anatomy must be taken of tapering rates of coronary arteries, as well into consideration when sizing percutaneous as, the cross-sectional areas and diameters of coronary intervention (PCI) devices, includ-healthy coronary arteries, as future means to ing stents and dilation balloons. It is well train programs to identify diseased vessels. documented that all coronary vessels taper; however, the locations and variability of such tapering has not been reported. Typically, 140 PO-55 SDD (1.35). Despite some technical challenges Reliability of the SPIMON, a portable 3-D encountered, sensor placement and measure-spinal range of movement measurement ment processes were efficient and straight-prototype device forward, highlighting areas for further refine- Aris Dermitzakis 1, Athanasios Lampropoulos1, ment and potential enhancement of SPIMON’s Ioannis Kalamaras1, Emil Valchinov2, Vasileios design and functionality. Syrimpeis3, Dimitra-Tania Papanikolaou1, Evdokia Billis1 PO-57 1University of Patras, Greece Surface electromyographic (sEMG) 2AEON P.C., Patras Science Park, Greece biofeedback device as an outcome meas- 3General Hospital of Patras, Greece ure for an individualized pelvic floor muscle Measuring the range of movement (ROM) (PFM) training program for urinary incon-in the spinal column is crucial for diagnos- tinent women: preliminary results ing and rehabilitating spinal pathologies. Tra-Dimitra-Tania Papanikolaou1, Aris Dermitzakis 1, ditional methods for assessing spinal move- Emil Valchinov2, Konstantinos Giannitsas1, Evan-ments, such as flexion and extension, rely gelos Dermatas1, Epaminondas Mitronikas1, Evdokia Billis1 on sophisticated and expensive equipment 1University of Patras, Greece that often falls short in usability, portabil- 2AEON P.C., Patras Science Park, Greece ity, cost-effectiveness, and time efficiency. This study introduces SPIMON, an innovative, Urinary incontinence (UI), defined as the wearable, and portable electronic device de- involuntary loss of urine, significantly im- signed to monitor spinal ROM. SPIMON is af- pacts approximately one-third of adult wo-fixed directly to the spine and communic- men globally, encompassing groups such as ates wirelessly with a tablet to record move- young athletes, pregnant women, and the ment data. Our research primarily investig- elderly, thereby severely affecting their qual-ates the inter-examiner and test-retest reliab- ity of life. Recognizing pelvic floor muscle ility of SPIMON in measuring the spinal ROM (PFM) training as the primary therapeutic ap-of healthy subjects, with a focus on its ap- proach for UI, this study assesses the effic- plication by novice examiners. Twenty asymp- acy of a surface electromyographic (sEMG) tomatic participants (11 men, 9 women) with biofeedback device used as an evaluative an average age of 22.74±1.4 years and a body tool in a tailored, 12-week PFM training pro- mass index (BMI) of 22.12±2.5 were enrolled gram. Ten women, aged 60.2 ± 9.8 years in the study. Measurements were conduc- with a mean body mass index (BMI) of 28.8 ted by two final-year physiotherapy students ± 5.9, participated in this intervention, which who had undergone a brief training session was meticulously crafted based on individual on the device’s operation, including sensor PFM performance assessments facilitated by placement and data collection. Reliability as- the NeuroTrac sEMG device. Administered sessments involved test-retest measurements by specialized women’s health physiotherap-by each examiner and inter-examiner com- ists, the program’s structure included an ini- parisons, focusing on spinal flexion, exten- tial baseline assessment, a concluding eval-sion, and lateral flexion movements. Data uation at 12 weeks, and intermediate weekly analysis was performed using intraclass cor- supervised sessions. Primary outcome meas-relation coefficients (ICC2,1), standard error ures focused on several PFM parameters: en-of measurement (SEM), and smallest detect- durance, measured by the average time parable differences (SDD) through SPSS. The find- ticipants could maintain a contraction at over ings revealed excellent test-retest reliability 50% of their maximal voluntary contraction for flexion, extension, and right-side flex- (MVC); strength, quantified through MVC in ion (ICC>0.91, SEM=0.87-2.37, SDD=0.86-1.68) microvolts via sEMG; frequency of sustained and moderate reliability for left-side flexion MVCs; and the number of fast PFM contrac- (ICC=0.55 & 0.68, SEM=0.95 & 0.68, SDD=0.92 tions, maintaining greater than 80% of peak and 1.10). Inter-examiner reliability was very MVC values for one second. The results indic- good across all spinal ROM measurements ated statistically significant improvements in (ICC=0.85-0.95, SEM=1.73-2.06, SDD=1.12-1.96). three out of the four primary PFM paramet-However, deviations were noted in left-side ers. Endurance saw an increase from an av-flexion within a subgroup, indicating moder- erage of 6.1 ± 1.6 seconds to 7.3 ± 1.2 seconds ate ICC (0.40) with acceptable SEM (2.09) and (p=0.009), the number of sustained MVCs im- 141 proved from 5.8 ± 1.1 to 7.3 ± 1.1 (p=0.002), and signal. A mathematical algorithm indexes the the frequency of fast contractions escalated trigger waves according to m-sequence for from 6 ± 1.4 to 7.5 ± 1.3 (p=0.002). Although averaging. It extracts the VEP response for an increase in PFM strength was observed each sector by the difference between the av- (from 48.3 ± 38.2 µ V to 53.4 ± 41.6 µ V), it did not eraged signal where a specified sector shifts reach statistical significance (p=0.55). Sec- and the signal where it does not. ondary outcomes, assessed through the Greek Twenty subjects will be recruited to test our version of the International Consultation on method and device. The inclusion criteria are Incontinence Questionnaire-Urinary Incontin- unilateral and local retinal lesions. The mfVEP ence Short Form (ICIQ-UI SF) and the “1-hour between the eyes will be compared, inferring pad test,” also demonstrated significant im- the accuracy of the system. provements in UI severity and urine loss, re- Results: The mfVEP performed by TC is spe-spectively, enhancing from 11 ± 5.6 to 4.8 ± 3.2 cific and sensitive to detect and topograph- (p=0.003) and from 4.3 ± 3.4g to 0.5 ± 0.4g ically map the functional impact of the retinal (p=0.008). These findings affirm the sEMG lesions, compared with the healthy contralat- biofeedback device’s utility as a potent dia- eral eye. gnostic and monitoring tool in PFM training re- Conclusion: It is possible to record func-gimes tailored to combat UI, advocating for its tional changes in small regions of the ret- broader application in clinical practices tar- ina and visual pathways in a non-invasive geting similar therapeutic outcomes. and non-expensive way, which does not oc- cur with electroretinography (ERG) or conven- PO-59 tional VEPs. As TC easily allows the dissemina- Tempo Certo Project - Adapting EEG sys- tion of access to evoked potentials using EEG tems to perform Multifocal Visual Evoked systems, our method will optimize the func-Potentials for clinical care and research tional diagnosis and follow-up of several dis- Igor Telles1, Arthur Conceição1, Paulo eases (e.g., glaucoma, diabetes mellitus, mul- Galhanone2, Mauricio Cagy1, Dimitri Abramov2, tiple sclerosis, and other acute processes) with Antonio Miranda de Sa 1 high precision and sensitivity. 1Federal University of Rio de Janeiro, Brazil Although neurodiagnosis has advanced sub- 2Oswaldo Cruz Foundation, Brazil stantially with the advent of neuroimaging, Objective: Development of protocols for lower accessibility has become a significant the acquisition of multifocal evoked poten- problem worldwide due to the high cost of tials (mfVEP) applied to the Tempo Certo imaging. On the other hand, mfVEPs are very (TC) project, a portable analog device that sensitive and constitute a technique that al- adapts conventional electroencephalogram lows assessing the functionality of visual neural (EEG) equipment to perform evoked potential systems, being an alternative for screening examinations with low cost, greater flexibility, and follow-up of patients, in accordance with and absolute precision, which makes it applic- policies for universalizing health systems. able both in research and in medical practice. Methodology: The mfVEP provides local visual PO-61 field VEP responses, simultaneously allowing Microstates analysis for dry and gel-based the topological study of small parts from the multichannel electroencephalography retina to the visual cortex. The stimulus is Hannes Oppermann, Patrique Fiedler, Jens a circular shift pattern, shaped like a log- Haueisen arithm checkboard centered on the fovea Technische Universität Ilmenau, Germany (radius=20–25o), divided into 60 sectors, with Spatial analysis of EEG data, e.g. using nine black/white squares each. Half of all sec- short-term stable microstates, is increasingly tors shift at a pseudorandom combination (m- used in neuroscience and clinical applica-sequence), ranging from 15 to 30Hz. The stim- tions. ulation is provided through a media running in At the same time, scenarios involving mobility, a suitable player. TC synchronizes the events sports, and home-based activities are be- (pattern shifts) by an eccentric probe in the coming more prevalent in EEG studies. screen under the TC photocell. A trigger sig- For this purpose, dry EEG electrodes are more nal is sent to the EEG recorder for each event. and more commonly used. So, multiple individual VEP responses are gen- Thus, our objective was a comparison of erated simultaneously from the occipital EEG microstates analysis between dry and gel-142 based EEGs. 256-channel EEGs were recorded chick chorioallantoic membrane for the devel-from 30 volunteers using dry and gel-based opment of tumor organoids, no significant dif- electrodes during resting state eyes closed ference was documented in the effect of 2.5 and eyes open. Microstates were extracted and 1 �g/ml bleomycin. Furthermore, we ex-for each measurement and time-domain amined the potentiation of radiosensitivity fol- parameters were calculated. lowing ECT in 3D UM spheroids. We found a high degree of consistency 2D and 3D UM cell cultures were electro-between the microstate maps extracted from porated with 8 pulses (100 �s pulse duration, dry and gel-based measurements for both 5 Hz repetition frequency) of a 1000 V/cm eyes closed and eyes open conditions. The pulse strength alone or in combination with topographic similarities between the average 0.11 mg/mL, 0.28 mg/mL, 0.55 mg/mL or 1.11 maps for dry and gel-based recordings were mg/mL calcium chloride or 1.0 �g/mL or 2.5 above 81.5% for each of the seven extracted �g/mL bleomycin. In addition, in vivo UM cell maps. line-xenografts as well as patient derived tu- We conclude that topographic microstate mor samples were treated. The results suggest analyses are feasible using multichannel EEG a dose-dependent ATP depletion with a wide setups with new dry EEG electrodes. range of sensitivity among the tested UM cell lines, as CaEP and bleomycin electroporation PO-63 significantly reduce cell viability at similar ap- The potential use of electroporation in com- plied voltage settings. bination with chemotherapeutics or cal- Further investigations and the development of cium chloride as a therapeutic option for suitable electrodes for the eye are necessary uveal melanoma patients: first in vitro and for the establishment of EP as an adjuvant op-in vivo results tion for the advanced UM. Miltiadis Fiorentzis University Hospital Essen, Department of Ophthal- PO-65 mology, University Duisburg-Essen, Germany Acoustic pressure analysis due to high- Uveal Melanoma (UM) represents the most frequency electric field on the biological cell common primary intraocular malignant tumor during reversible electroporation in adults. Despite its high metastatic and mor- Mayank Kumar, Rochish Thaokar tality rate, a standardized therapeutic regi- Indian Institute Of Technology Bombay, India men has not yet been established for the con-The biological cells when exposed to an solidation of a protocol for the surgical treat- electric field undergo electrodeformation ment of UM. parallel to the direction of the applied elec- In pursuance of novel treatment modalities, tric field. This electrodeformation produces we investigated the effect of electroporation acoustic pressure waves in the extracellular (EP) in combination with various chemothera- fluid which could be measured by the pressure peutic drugs or with calcium chloride predom- transducers. The multiphysics investigation inantly on ocular melanoma 2D and 3D cell on this deformation and acoustic pressure cultures as well as in vivo in the CAM assay. produced during reversible electroporation is Specifically four uveal melanoma (UM) cell proposed using different cell contours. The lines were tested for the efficacy of EP in con- transmembrane potential, total elastic strain junction with bleomycin and cisplatin, where energy, Maxwell stress tensor(MST) and Helm-a higher resistance of all cell lines was noted holtz wave equation in the cell are compared to cisplatin. Bleomycin showed similar results using multiphysics for action under unipolar at lower concentrations with electroporation and bipolar nano-pulse electric fields. The conditions of 750 V/cm and 20 pulses, whereas polarization of cervical cells is expressed using the same effect was achieved for only two cell Debye expression which occurs as the cells lines with 1000V/cm and 8 pulses with 2.5�g/ml are dispersive at high frequencies. Our model cisplatin. In five uveal cell lines of primary and explains the temporal evaluation of electro- metastatic origin, used for the formation of poration and the acoustic wave produced in 3D spheroids, the most significant reduction the fluid due to the bi-lipid membrane under of size was observed after EP with 750 V/cm constant stress which produces electrode-and 2.5 �g/ml bleomycin. In the first in vivo res- formation. This method can find application in ults after implantation of cell pellets on the acoustic imaging where the back-projection 143 method is applied to assess whether the sponsible for X-ray generation, with its per-cell has undergone apoptosis or the mem- formance directly affecting imaging resolu- brane of the cell has lost its integrity during tion and diagnosis accuracy. The tube will electroporation. gradually deteriorate due to usage, aging, and improper maintenance, which has a negat- PO-67 ive impact on the performance of CT, and Texture Analysis of H-scan Ultrasound Im- even leads to costly maintenance and casu-ages for the Characterization of Breast Tu- alties. Existing research mostly focuses on op-mors timizing its design to improve the inherent re- Zhanjie Zhang 1, Sio Hang Pun1, Peng Un Mak1, liability of tube. Prognostic and health man-Hung Chun Li2, Kung Jui Hou2, Mang I. Vai1 agement during clinical application are in- 1University of Macau, Macao adequate for bolstering operational reliabil- 2Lingyange Semiconductor Inc., China ity and furnishing maintenance guidance to Texture analysis of breast ultrasound im- mitigate hospital downtime losses. This re- ages is widely used in the tissue character- search propose a prognosis and uncertainty ization of breast tumors, as the texture fea- management framework for filament degrad-tures are closely connected to the scatter- ation of X-ray CT tubes with time series data. ing patterns and microstructures of the tu- First, time series data of X-ray tube in clinical mors. H-scan imaging is a novel ultrasound application is procured via the Medical Inter-imaging method to differentiate the backs- net of Things (IoMT). Second, a degradation in-cattering patterns inside different tissue struc- dex is constructed based on the filament de-tures. In this work, the feasibility of H-scan im- gradation mechanism. Based on setting para- age texture analysis for the characterization of meter fixed and data smoothing techniques, breast tumors is assessed. Five texture features the stable filament degradation data is ex- (Contrast, Energy, Entropy, Homogeneity, and tracted by mitigating interference from envir- Correlation) are obtained from the H-scan onmental noise and scanning positions. Third, images based on the Open Access Series of a long-time prediction model based on one-Breast Ultrasonic Data (OASBUD) involving 100 dimensional convolutional neural network is breast tumors of 78 female patients. Texture devised to learn the filament degradation pro-analysis results demonstrated significant dif- cess, while the Monte Carlo dropout method ferences (p<0.05) between benign and malig-is employed to implement uncertainty man- nant breast tumors for the selected features agement throughout the degradation pro-derived from H-scan images. Moreover, sev- cess. Finally, the temporal IoMT data of CT eral textural features (Energy, Entropy, Correl- and tubes in West China Hospital, Sichuan Uni-ation) from H-scan images began to show sig- versity are utilized to demonstrate the effect- nificant differences between certain BI-RADS iveness of the proposed framework. This work category levels (level 4a with level 5, level 4b aims to fill the gap in operational reliability with level 5). In contrast, no significant dif- prognosis of X-ray tube in clinical application. ference was in the features extracted from B- Further, the predictive and uncertainty man- scan images of the aforementioned BI-RADS agement results have the potential to be used category groups. Overall, this study demon-for predictive maintenance of CT in hospit- strates that texture analysis of H-scan images als, thereby reducing maintenance costs and is helpful for characterizing breast tumors. avoiding safety incidents. PO-69 Prognosis and Uncertainty Management for Poster session Filament Degradation of X-ray CT Tubes with IoMT time series data Wednesday morning Jie Zhong, Heng Zhang, Qiang Miao, Qilin Liu, Jin poster session Huang Jun 12, 10:00 - 10:30 Sichuan University, China X-ray Computed Tomography (CT) repres- PO-02 ents a pivotal medical apparatus utilized in BIM operations and clinical engineering in hospitals for disease diagnosis and treatment. the training of the biomedical engineer X-ray tube is a critical component in CT re-Luis Vilcahuaman Pontifical Catholic University of Peru, Peru 144 The use of computer tools for the design teen sex-matched healthy lean participants and construction of hospitals under the BIM (C group, 14f, 13.7 – 25.4 years; median age: (Building Modeling Information) method 16.9 years) were studied. We non-invasively allows integrative and collaborative work recorded beat-to-beat values of mean blood between different specialists in architecture pressure (MBP) by the photoplethysmographic and engineering. However, after construction, volume-clamp method (Finometer Pro, FMS, it is necessary to guarantee the operation Netherlands), RR intervals by ECG (CardioFax of the hospital by complying with the as- ECG-9620, NihonKohden, Japan), and car-pects of functionality, efficiency, safety and diac output (CO) by impedance cardiography cost control. The traditional approach of (CardioScreen® 2000, Medis, Germany) during the clinical engineer is aimed at addressing supine rest (15 min) and head-up tilt (HUT, 8 biomedical equipment, without emphasizing min). Next, beat-to-beat values of peripheral hospital systems and the environment. Both vascular resistance (PVR) were calculated as require integrative and collaborative work, so the ratio of MBP and CO. Spectral coupling it is necessary to improve the traditional ap- (a measure of causal coupling strength in the proach to guarantee the operation of current frequency domain) and spectral gain (barore-hospitals. This requires improving the training flex sensitivity, a measure of the response of of the biomedical engineer, and in particular the RR and PVR to the unit change in MBP) in the training of the clinical engineer, who, the cardiac chronotropic (from MBP to RR) beyond medical equipment, must contribute and vascular resistance baroreflex (from MBP to guaranteeing the functionality and safety to PVR) arms were evaluated using the partial of the entire technological environment of a spectral decomposition method in the low clinical service. As a result, it is expected to frequency band (0.04 – 0.15 Hz). recover existing hospitals and to better plan Significantly lower spectral gain values were the operation phase in a modern hospital. observed in the vascular resistance baroreflex arm in the O group during the supine rest PO-04 phase (P = 0.014). During HUT, the spectral Baroreflex function in young patients with gain in both baroreflex arms was significantly obesity: analysis of the cardiac chrono- lower in the O group (P ≤ 0.002). No significant tropic and vascular resistance baroreflex differences were observed in the spectral arms coupling values. Jana Čerňanová Krohová 1, Barbora Obesity-related impairment of baroreflex Czippelova1, Zuzana Turianiková1, Miriam Kuricova1, function, affecting both heart rate and vaso- Dana Cernochova1, Luca Faes2, Michal Javorka1 motion control, could be one of the factors 1Comenius University and University Hospital, Slov- contributing to the future development of hy-akia pertension in this high-risk associated group. 2University of Palermo, Italy Supported by grant VEGA no. 1/0283/21. Impairment of baroreflex function is one PO-06 of the factors that could contribute to the Artificial Intelligence Advancements in Fetal development of hypertension, through an Monitoring: Enhancing Prenatal Care imbalance in the sympathetic-vagal out- Dragoș-Daniel Țarălungă, Ionut Manea, flow to the heart and blood vessels. Obesity Rares-Marin Preoteasa, Bogdan Cristian Florea, significantly increases the risk of arterial hy- Georgeta-Mihaela Neagu pertension. We assume that the preclinical National University of Science and Technology Po-stage of hypertension development in young litehnica Bucharest, Romania patients with obesity could precede arterial hypertension for years. The main objective Pregnancy and the delivery of a healthy of this study was to investigate potential baby rank among the crucial milestones in changes in baroreflex function as one of the human life cycle. Obstetrical science the mechanisms potentially contributing to is devoted to ensuring these events unfold future hypertension development in young as smoothly as possible, promoting the well-normotensive patients with obesity. being of both the infant and the mother. The Nineteen normotensive adolescents and primary challenge in achieving this objective young adults (O group, 14f, age range: 13.7 – is the occurrence of fetal deaths within the 25.4 years; median age: 16.8 years) and nine- uterus, which represents a significant obstacle to the overall goal of a healthy outcome for 145 the baby and the mother. Fetal monitoring ers. is an essential aspect of prenatal care, aim- ing to assess the health and well-being of PO-10 the developing fetus during pregnancy. There The Study of Non-Invasive Blood Informa-are various approaches for fetal monitoring tion Measurement and Monitoring Method used in clinical practice: cardiotocography via Wearable AWPPG Device (CTG) and Doppler ultrasound, fetal electro-Chung ChiFu 1, Lu PoWen2, Cheng Chun Chang1 cardiography, fetal ecography etc. However 1National Taipei University of Technology, Taiwan there are crucial limitations: inter- and intra- 2Taipei Medical University, Taiwan observer variability, invasivity, low signal to In recent years, with the proliferation of noise ratio (SNR) etc. The aim of the present wearable devices, wearable watches and study is to evaluate the current contribution of wristbands capable of measuring physiolo-artificial intelligence (AI) advancements in of- gical information have become indispensable fering innovative solutions to enhance accur- tools for many in modern society. These acy, early detection, and overall efficiency in wearable devices allow individuals to monitor assessing fetal health. real-time physiological data, such as heart rate, blood pressure, and blood oxygen con- PO-08 centration, to maintain their health. However, Using localized electroporation for trans- there is currently no non-invasive method fecting cardiac cells available on the market for measuring blood- Jernej Jurič, Vid Jan, Tina Turk, Lea Rems related information within the human body. University of Ljubljana, Slovenia Therefore, this study aims to investigate The permeabilization of the cell membrane whether wearable devices can provide a non-and concurrent transfer of chemical or bio- invasive means of acquiring blood-related logical agents into cells can be facilitated information through related signal processing through diverse methodologies. Among these, algorithms and neural network models, utiliz- electroporation, allowing enhanced delivery ing an All-Wavelength Photoplethysmography of exogenous molecules into the cell interior (AWPPG) approach combined with neural net-after exposure to pulsed electric fields, is re- work models for accurate measurement of garded as a viable alternative to viral, lipid, blood information. The research is divided into and other nanoparticle-based delivery sys- three main components: the development of tems. Traditionally, bulk electroporation (BEP) a non-invasive wearable device, the collection has been utilized for most transfections in vitro. of blood information in a clinical setting, and In BEP, cells are placed between two elec- the exploration of one-dimensional signal trodes and subjected to an approximately ho-algorithm models. mogeneous electric field. The disadvantage Currently, we have collected blood-related of BEP is that it increases membrane per- data at the hospital using our self-developed meability over a large membrane surface area AWPPG measurement device. We have also and is thus associated with quite high cellular trained artificial intelligence models for vari- damage. By localizing the electric field over ous blood parameters. The accuracy rate a small membrane area using a nanoporous for blood glucose is seventy percent, while substrate, it is possible to improve transfer effi- for hemoglobin, platelet count, sodium ion ciency while providing greater cell survival and concentration, calcium ion concentration, better control over gene expression. In this and total protein content, the accuracy rate study, we investigated the potential of local- is approximately sixty to seventy percent. ized electroporation for transfecting rat car- diac cell line H9c2 and isolated rat ventricular PO-12 cardiomyocytes, by using polyethylene tere- Development of Neurorehabilitation Bicycle phthalate (PET) membranes with 0.4 µ m pore with Posture Control Detection diameter, which are commercially available in Ya-Ju Chang 1, Hsiao-Lung Chan1, Jr-kai Yu1, the form of cell culture inserts. We identified Miao-Ju Hsu2, Rou-Shayn Chen1 1 appropriate process parameters and success- Chang Gung Memorial Hospital and University, fully electroporated the cells, employing either Taiwan 2 propidium iodide or a plasmid encoding the Kaohsiung Medical University, Taiwan green fluorescent protein as validation mark- Patients with central nervous system injur- 146 ies such as stroke, spinal cord injury, Parkinson’s (ET) with the ventral intermediate nucleus disease, etc., often suffer from impaired motor of the thalamus (Vim) and the zona incerta functions, affecting their quality of life. Among (Zi) are common targets. Implantation in these impairments, balance and gait disturb- these structures usually results in a reduction ances are crucial rehabilitation issues, and ad-of tremor, providing efficient therapy while justments in posture balance are necessary avoiding side effects is a larger challenge. during walking to prevent falls and improve This work aims at creating and validating a walking efficiency. In neurorehabilitation, bi-method for mapping side effects in Zi-DBS for cycle training is believed to establish a central ET. pattern generator (CPG) for controlling walk- Methods: Stimulation parameters and pre-ing and is commonly used. However, tradi- and postoperative imaging from 71 ET patients tional neurorehabilitation bicycles lack simu-were included in this study. The amplitude and taneous posture balance control training or contact that generated side effects were re- detection, potentially reducing the translation trieved during a monopolar screening where of training effects into walking ability. There- the effect was evaluated in each contact on fore, the purpose of this study is to develop a quadripolar lead with increasing amplitude a rehabilitation bicycle with posture and bal-in steps of 0.2-0.3 V up to 5V. Based on the ance detection and training functions and to registered side effects, 8 groups were identi-conduct a human study. fied (speech affection, visual affection, dizzi- In the first part of the research, we developed ness, paresthesia (all), paresthesia arm/hand, a saddle force plate for bicycle use, utilizing ataxia arm, muscular contraction, sweating) three transducers to replace the traditional for further analysis. For each stimulation set-four transducers, calculating the COP (Cen- ting, the electric field around the lead was sim-ter of Pressure) data of the saddle. In the ulated (n=852) with the finite element method second part of the research, we recruited 30 and transformed to an MRI template nor- subjects to study the correlation between COP malized from the same cohort. A probabil-displacement and trunk inclination angle dur- istic map (PSM) was computed for each side-ing non-riding and riding conditions. effect based on a voxel-wise Wilcoxon signed The research findings revealed that both static rank test, with a permutation test for correc-and riding states showed a significant correl- tion of family-wise error. To validate the maps, ation between COP displacement area and a support vector machine (SVM) classification body inclination angle, but a high linear cor- model was created based on four predictors relation was observed only in static conditions. extracted from the overlap between the PSM Additionally, during riding, the COP displace- and the electric field. The validation was per- ment area varied significantly with different formed in a leave-one-out fashion, where the inclination angles and riding speeds. PSM and SVM were computed with all but one The results of this study confirm the feasib- patient and evaluated using the left-out pa-ility of simultaneously detecting and training tient. The process was repeated until all pa-body posture control during cycling, although tients had been left out. the actual body inclination angle requires ad- Result: PSMs were computed for all side ef- ditional calibration. The equipment and train- fects, but sweating did not result in any cluster ing modes developed in this study could be surviving the analysis. Generally, a trend of applied to the rehabilitation of patients with lower p-values for side effect PSMs with higher neurological disorders in the future. occurrence, up to side effects occurring in around 100 stimulations. This trend was also PO-14 seen in the validation, where accuracy and Mapping side effects for deep brain stimu- true positive rate increased with increasing lation in essential tremor sample size. The true negative rate was higher Teresa Nordin 1, Erik Österlund2, Dorian Vogel3, than the true positive rate for all side effects Anders Fytagoridis2, Simone Hemm3, Karin Wårdell1 except paresthesia. 1Linköping University, Sweden Conclusion: A workflow for predicting side ef- 2Karolinska Institute, Sweden fect occurrence in deep brain stimulation was 3University of Applied Sciences and Arts Northwest- developed. For a stable result, the number of ern Switzerland, Switzerland included simulations is important and should Introduction: Deep brain stimulation (DBS) be systematically investigated in the future. is an established therapy for Essential tremor Future studies will be conducted to combine 147 improvement and side effects to predict the The statistical analyses on the complete data-optimal DBS setting for a wide therapeutic set revealed sweet spots of varying sizes: 28.88 window. mm3 for the t-test and 58.25 mm3 for the Wil- coxon test. The Dice coefficient between the PO-16 two was 66%. The significant volume obtained Exploring the impact of data and statistical with the t-test was entirely enclosed in the methods for defining sweet spots in Deep volume extracted with the Wilcoxon test. For Brain Stimulation the best configurations sweet spot volumes Vittoria Bucciarelli 1, Dorian Vogel1, Teresa were 567.63 mm3 for t-test and 560.88 mm3 Nordin2, Karin Wårdell2, Jérôme Coste3, Jean- for Wilcoxon, with a Dice coefficient of 99.3%. Jacques Lemaire3, Raphael Guzman4, Simone Stimulations activating the sweet spot were Hemm1 correlated to improvement with similar coeffi- 1FHNW University of Applied Sciences and Arts cients between methods but higher when us-Northwestern Switzerland, Switzerland ing sweet spots derived from the complete 2Linköping University, Sweden dataset. 3Université Clermont Auvergne, France Users should be conscious of the strong in- 4University Hospital Basel, Switzerland fluence that the dataset and chosen stat- The precise location of the sweet spot (op- istical method exert on the extracted sweet timal target), for Deep Brain Stimulation (DBS) spot in group analysis. Restricting analysis to is still under investigation. Nevertheless, its optimal parameters leads to inflated volumes identification holds significant importance in undermining the ability to identify high im- guiding algorithms for the automatic pro- provement areas. Despite the difference in ex-gramming of DBS parameters. Previous stud- tracted volumes between the methods, both ies have predominantly used t-tests and Wil-provided a significant correlation with clinical coxon tests to generate probabilistic stimu- improvement. Thus, further criteria to inform lation maps. A comprehensive evaluation of the method’s selection are needed. both the dataset and the chosen statistical method’s impact on the extracted sweet spot PO-18 from group analysis is lacking. This study aims REDCap and SQLite: a powerful combina- to compare the outcomes of these statistical tion for streamlining metadata capture in tests applied to two DBS datasets with distinct Deep Brain Stimulation research compositions. Marc Stawiski1, Vittoria Bucciarelli 1, Dorian The data used in this study was obtained Vogel1, Jérôme Coste2, Jean-Jacques Lemaire2, from intra-operative stimulation tests on 6 Es- Siegward Elsas3, Ute Gschwandtner4, Peter Fuhr4, sential Tremor (ET) patients with implants in Ethan Taub4, Simone Hemm1 1 the ventral-intermediate nucleus (VIM). Stim- FHNW University of Applied Sciences and Arts ulation parameters were used to generate Northwestern Switzerland, Switzerland 2 patient-specific electric field simulations, sub- Université Clermont Auvergne, France 3 sequently transformed to a group-specific MRI Arlesheim Hospital, Switzerland 4 template. Voxel-wise one-tailed one-sample University Hospital Basel, Switzerland t-tests and Wilcoxon signed-rank tests were The field of Neurosciences research faces performed to identify voxels significantly as- significant challenges in managing hetero-sociated with tremor improvement exceed- geneous data types. While the Brain Ima- ing 60% (p< 0.05). False Discovery Rate (FDR) ging Data Structure (BIDS) offers a standard-correction was applied to mitigate false pos- ized format for brain imaging data organiza-itives. The significant clusters obtained with tion, its limitations in capturing detailed clin- the two methods were compared visually and ical metadata, such as longitudinal clinical by calculating their total volume, intersection scores or demographic information become volume, and Dice coefficient. Additionally, apparent in data analysis pipelines. To ad-validation was conducted by correlating the dress this critical limitation, this work proposes overlap between the electric field and sweet a framework for enhanced metadata capture spot with the improvement observed in the based on the example of Deep Brain Stimu- clinic. The same tests were then repeated on a lation (DBS). It leverages BIDS with Research subset of the data, including only the 50 best Electronic Data Capture (REDCap), a secure configurations per patient weighted by amp-web-based platform designed for validated litude. data collection, and SQLite, a lightweight re- 148 lational database management system cap-PO-20 able of storing structured metadata. BIDS of- Numerical Study of Hemodynamic Charac-fers organization of data in neuroscience re- teristics on the Surgery of Pulmonary Artery search, and RedCap user-friendly clinical data Banding entry for clinicians. Jinlong Liu 1, Weiru Luo2, Jiwen Xiong1, Yanjun Patient-related metadata (e.g. demograph- Sun1 ics, medication, neurological assessments) is 1Shanghai Jiao Tong University, China collected via structured forms within REDCap. 2Waseda University, Japan Medical image data, originating from Brain-Pulmonary artery banding (PAB) is a com- lab Elements neurosurgical planning station, monly used procedure for the restriction of is stored and transferred in the anonymized pulmonary blood flow by surgically creating a DICOMDIR format. To ensure BIDS compli- stenosis in the main pulmonary artery (MPA). ance, a custom Python script extracts relevant Two types of PAB are available in clinic. One information from the DICOMDIR data. This ex- is called the extraluminal pulmonary artery tracted data is converted into the Neuroima- banding (E-PAB), reducing the lumen of the ging Informatics Technology Initiative (NIfTI) MPA by placing a band outside the MPA. format, and the resulting files are stored in a The other is the intraluminal pulmonary artery BIDS-compliant directory structure. Addition- banding (I-PAB) with suturing a perforated ally, image references are established within patch on the inner wall of the MPA to re- the SQLite database, which also serves as re- strict the pulmonary flow. There is still contro-pository for metadata extracted from RED- versy over which surgical method to choose. Cap. Here, we reported our numerical studies on the To validate the proposed framework, data was hemodynamic characteristics of different PAB collected and analyzed from two medical in- methods based on a three-dimensional model stitutions: the University Hospitals Basel and of pulmonary artery, thus to help surgical Clermont-Ferrand. Our framework has suc- planning. The model of pulmonary artery was cessfully captured data for 107 patients, with reconstructed according to patient-specific an average of 35 imaging files including MRI computed tomography images. Computer- and CT scans and labeled anatomical struc- aided design (CAD) was used to perform vir-tures. This data is stored and managed within tual surgeries of E-PAB and I-PAB, respectively. the patient-centric SQLite database. The Pulsatile simulations and hemodynamic ana- database schema comprises 28 tables, 230 lysis were conducted to capture the differ- data fields (e.g. patient ID, age, implanted po- ences of postoperative hemodynamics using sition, image file path) and 33 established re- the method of computational fluid dynamics lationships between these tables. The advant- (CFD). The local hemodynamic features in Eages of using such a system can be particularly PAB and I-PAB models were illustrated with the appreciated in the use-case of group-level pressure drop, velocity streamlines, wall shear analysis. The system design facilitates efficient stress (WSS) and energy loss (EL). The results retrieval of structured clinical data. Further- showed the flow field closed to the banding more, the framework effectively handles and site in the I-PAB model was presented with stores files generated during post-processing more flow turbulence due to the block of the steps, ensuring data integrity and traceability. patch. The pressure drop and WSS between This work demonstrates the successful integ- the upstream and downstream of the band-ration of a translational tool for streamlin- ing site were higher in the I-PAB model, coming data collection and organization between pared to that of E-PAB model. The EL of I-PAB clinics and research institutes. Our framework model was more than that of E-PAB model. captures both standardized imaging data The study indicated that the E-PAB was as-and comprehensive patient-metadata within sociated with better hemodynamic perform- a unified system. This enables researchers not ance in controlling pulmonary artery pressure only in the field of DBS but from the Neur- and reducing the cardiac burden. The numer-osciences in general to leverage the richness ical simulation of CFD combined with CAD is of both types of data, potentially leading to an effective approach for the investigation of improved clinical decision-making and ulti-preoperative surgical planning and postoper- mately, better patient outcomes. ative hemodynamic prediction. 149 PO-22 tion demonstration system and clinical eval- Monte Carlo Simulated Photoplethysmo- uation system suitable for different centers graphy Signals for the Validation of an In and domestic advanced minimally invasive Vitro Wrist Phantom surgical instruments. Methods: To construct Raghda Al-Halawani, Meha Qassem, James the survey questionnaires, a robust method- May, Panicos A. Kyriacou ological approach was employed, combining City, University of London, United Kingdom expert consultations, the Delphi method, and Photoplethysmography (PPG) is an optical an extensive review of existing literature to en-technique used for the continuous monitoring sure the inclusion of relevant and comprehens-of blood volume changes, and is integrated ive content. These questionnaires were then in a variety of medical devices and consumer electronically distributed to various levels of wearables. Meanwhile, in silico and in vitro medical institutions across the nation, lever-experiments have shown to facilitate our un- aging the expertise of specialists from multiple derstanding of light-tissue interactions to as- disciplines. The responses collected were sub-sess the feasibility and accuracy of PPG-based jected to rigorous statistical analysis to extract technologies in acquiring PPG signals from the meaningful insights. This systematic approach human anatomy. Therefore, this study demon- facilitated a detailed evaluation of the cur- strates an approach to validate an in vitro rent state of high-end minimally invasive sur-study of a developed wrist phantom through gical instruments in these institutions, focusing a Monte Carlo (MC) simulation to compare on the availability, deployment, and training PPG signals generated from experimental and capabilities associated with advanced med- computational models. The MC model sim- ical equipment. Results: The configuration of ulates a PPG signal using MATLAB, and ac- high-end endoscopic systems and minimally counts for the optical and mechanical prop- invasive surgical instruments in medical insti-erties of the phantom and vessel in synchrony tutions at all levels is generally low, with signi-with changes in internal blood pressure. Key ficant room for increase in the proportion of features of the PPG signal, such as onsets, and domestically produced equipment. Further- systolic and diastolic slopes are visually ap- more, the frequency and capability of training parent. Additionally, the use of in-built MAT- related to endoscopic systems are also slightly LAB functions is recommended to improve lacking. Conclusion: Equipment and training the morphology of the PPG signal to optimise are crucial foundations for the promotion of computational resources. Overall, the current advanced minimally invasive surgery, largely study shows the applicability and potential of determining whether medical institutions are Monte Carlo simulations to validate PPG data capable of performing high-end minimally in-acquired from in vitro models. vasive operations. Currently, there are is- sues with insufficient configurations of high- PO-24 end endoscopic systems and minimally invas- Identification of Barriers in the Adoption ive surgical instruments, inadequate training, and Promotion of Advanced Endoscopic and a low proportion of domestically pro-Systems and Minimally Invasive Surgical duced equipment. There is an urgent need for Equipment in China: Evidence from a Survey homogeneous and standardized research and of 320 Healthcare Facilities application of domestic advanced minimally Zhichen Wang, Fei Sun, Qian Zhang, Jing Sun, invasive surgical instruments and techniques Jingyi Feng nationwide to effectively address these exist- Department of Clinical Engineering and Material ing problems. Supplies, the First Affiliated Hospital of Zhejiang University, China PO-26 Co-transfection of anti-fibrotic microRNAs Objective: To thoroughly understand the as a treatment for oral submucous fibrosis current status of high-end endoscopic sys- Kai-Chiang Yang 1, Chih-Yuan Fang1, Hsing-Yu tems and minimally invasive surgical equip- Chen2, Yung-Hsin Cheng2 ment configurations in medical institutions 1Taipei Medical University, Taiwan across China’s six major regions, as well as the 2National Taiwan University of Science and Techno- situation regarding endoscopy-related edu- logy, Taiwan cation and training. This aims to provide a data foundation for constructing an applica- Oral submucous fibrosis (OSF) is a disease characterized by excessive extracellular mat- 150 rixes (ECM) deposition (especially collagen) fortable exam that uses ionizing radiation. Ul-and is highly associated with the areca quid trasound is used as an auxiliary exam for dia- chewing habit of patients. Arecoline, a nat- gnosis, but studies show the potential of Ultra-ural alkaloid of the betel nut, may activate the sound Transmission Tomography (UTT) to re- transforming growth factor-beta (TGF- β) sig- place mammography. UTT began in the 1970’s naling pathway and subsequently switch on with pioneering studies by Greenleaf et al. and downstream genes related to fibrosis and ECM Carson et al., imaging the speed of sound and production. TGF- β/Smads cascade is known attenuation coefficient. Nowadays there are as the major inducer of myofibroblast differ- numerous methods of tomography acquisi-entiation in OSF. Several studies have found tion, two of these methods can be categor-that the dysregulation of microRNA (miR) is ized into linear-array transducers and ring- involved in the progression of various fibrotic array transducers. Our objective is to com-diseases. Therefore, we surveyed 16 fibrotic- pare simulation and reconstruction systems to related miRs and proposed delivering exogen- determine the most efficient reconstruction ous miR mimic or inhibit simultaneously as a method. Full-Waveform Inversion (FWI) is a treatment for OSF. Human oral submucosal seismic technique renowned for its high res- fibroblasts were subjected to varying concen- olution and adaptability to ultrasound tomo- trations of arecoline (5, 10, 20, 40, or 100 graphy applications. In our study, we integ- µ g/mL) for either 24 or 48 hr, and cell viabrated open-source implementations for linear ility, death, mRNA expression of TGF- β and arrays developed by Rehman Ali (2022) and alpha-smooth muscle actin ( α-SMA), and pro- utilized the ring array provided by the Stride tein production of STRO-1, type I collagen, vi-framework (Cueto et al,. 2022) for compar- mentin, and α-SMA were evaluated by West- ative analysis, both use FWI as reconstruction ern blot. Subsequently, based on the res- method.The phantom utilized was adapted ults of these analyses, the condition involving and modified from a cone-beam breast CT 20 µ g/mL of arecoline and 24 hr of expos- image (J. Uhlig et al., 2017; Rehman Ali, 2022). ure was selected. The stimulated OSF fibro- For the linear array, we utilized k-wave to simu-blasts were transfected with 16 miRs individu- late a pair of transducers with 192 elements, at ally by using cationic lipid transfection re- a central frequency of 1 MHz, 0.6 mm pitch, and agents. Based on the ability to downregu- a spatial grid with 0.6 mm. For the ring array, late TGF- β and α-SMA mRNA expressions, miR- we used the toolkit offered by Stride to simu-29a mimic, miR-150 inhibitor, miR-196a mimic, late 128 elements, with a central frequency of and miR-509-5p mimic were further selected. 1 MHz, and a spatial grid of 0.5mm. Our simu-Co-transfection of the above 4 miRs further lations using both tools revealed that the lin- downregulated the mRNA levels of COL1A1 ear transducer yielded a mean error in speed and matrix metalloproteinase-7 in myofibro- of sound of 1.98 m/s, with a standard devi-blasts. Furthermore, transfected cells’ wound ation of 2.98 m/s. For the ring array, we ob-closure, cell migration, and collagen gel con- tained a mean error of 10.1 m/s and a deviation traction abilities were inhibited. In conclusion, of 6.8 m/s. In the analysis, the linear trans- co-transfection of anti-fibrotic miRs can be a ducer demonstrated a better estimation. To treatment for OSF. determine the best methodology, we need to observe the differences between frameworks; PO-28 for example, the differences in the simulation Analysis of Open-Source Softwares for algorithm can create discrepancies. Our next Ultrasound Tomography Based on Full- step is to eliminate these differences to have a Waveform Inversion better understanding of the requirements ne- Lucas da Costa, Nilton Assugeni Neto, João cessary to achieve a quality parameter for lin- Henrique Uliana, Théo Pavan, Antonio Adilton ear and ring arrays. Carneiro Department of Physics – Faculty of Philosophy, Sci- PO-30 ences and Letters at Ribeirão Preto, University of Improving diagnostic performance of an São Paulo, Brazil automated melanoma diagnostic system Breast cancer is the second-leading cause using fast style transfer data augmentation of cancer-related deaths in women (Siegel et Takashi Nagaoka, Mitsutaka Nemoto, Yuichi al,. 2024). The gold standard for diagnosing Kimura breast cancer is mammography, an uncom- Kindai University, Japan 151 Melanoma is a highly malignant skin tumor stride length by leveraging tri-axial accelera-that is difficult to distinguish from benign le- tions and tri-axial angular velocities gathered sions. Quantitative methods are required to from the lateral side of shoes. Our pro- support early detection of melanoma, and AI- posed model incorporates a U-Net architec- based melanoma diagnosis system is becom- ture designed for extracting crucial features, ing increasingly popular. Deep learning re- which was trained using randomly selected quires a large number of images for training, 1-second data segments from 10 individuals but it is difficult to obtain a large amount of with Parkinson’s disease. Subsequently, the U-medical information such as dermoscopic im- Net’s encoder path outputs serve as features ages, and some form of data augmentation is for predicting stride length through a mul- needed. In this study, we used fast style trans- tilayer regression model. Utilizing leave-one-fer (FST), as a data augmentation method out cross-validation between the predicted and verified the change in diagnostic accur- stride length and the ground truth acquired acy. In this study, 1000 melanoma and non-from the GAITRite electronic walkway, a re- melanoma cases are randomly selected from markable R-squared correlation of 0.88 and a dataset called HAM10000 (”Human Against an error percentage of 8.385% were achieved Machine with 10000 training images”). Us- when employing the 1-second segments pre- ing the dataset, 999 fake images are gener- ceding toe-off. We conducted additional ated from a single image using the FST in ad- tests to assess the impact of inaccurate de- vance. The melanoma discrimination model is tection of the toe-off event. This was simu- built using automated machine learning. Au- lated by randomly shifting time offsets ran-toKeras 1.1.0 is used to build an EfficientNet- ging from 0.05, 0.1, 0.2, to 0.5 seconds in both based deep learning model. The model ini-negative and positive directions. It’s worth tially used common augmentation methods noting that the evaluation specifically focused such as rotation and flipping. During training, on the 0.5-second offset shifting, eliminating our model randomly replaces some of the im-the necessity for stride identification. These ages with fake images pre-generated by the configurations yielded R-squared correlations FST. Diagnostic accuracy is calculated using a of 0.898, 0.876, 0.87, and 0.81, along with er- 5-fold cross-validation. The diagnostic accur- ror percentages of 7.96%, 8.63%, 8.74%, and acy of our model with and without augment- 10.80%, respectively. For future work, recruiting ation by the FST was 83.0+2.0% and 81.1+2.2%, a larger number of individuals with PD would respectively. be beneficial to encompass a broader spec- trum of gait variations within the PD popula- PO-32 tion. Estimating stride length through a deep learning model utilizing foot inertial data in PO-34 individuals with Parkinson’s disease A Convolutional Neural Network (CNN) based Electrocardiogram (ECG) Pattern Hsiao-Lung Chan, Ya-Ju Chang, Ai-Tung Hsieh, Rou-Shayn Chen Analysis under Cycling Movement with the Chang Gung Memorial Hospital and Chang Gung Applications into the Development of an University, Taiwan Exhaustion Stratification System Szi-Wen Chen 1, Yu-Hsuan Huang1, Rou-Shayn Gait disorders present a significant chal- Chen2 lenge in Parkinson’s disease, often manifest- 1Chang Gung University, Taiwan ing as a reduction in stride length, particularly 2Chang Gung Memorial Hospital, Taiwan in off-medication situations. Utilizing an iner- tial measurement unit on the lower extremity This research introduces a new approach provides a convenient method for measuring utilizing electrocardiogram (ECG) data and stride length, irrespective of space constraints. deep learning techniques to effectively strat-While recent advancements in deep learn- ify and quantify exercise-induced fatigue us- ing have proven effective in estimating stride ing a novel indicator, dubbed the exhaustive length from the inertial data, these models status index (ESI). The study synchronizes ECG predominantly rely on data from healthy sub- recordings with cycling movements, facilitat- jects, with limited representation from indi- ing real-time monitoring of physiological re- viduals with gait disorders. In this study, we sponses during exercise. To capture nuanced developed a deep learning model to predict patterns indicative of fatigue, ECG signals are meticulously segmented into short segments, 152 optimizing data granularity. These segmented PO-36 ECG signals serve as two-dimensional (2D) in- The Prediction of Sleep Quality using Heart puts to a convolutional neural network (CNN), Rate Varia-bility Modulations during Wake-a powerful tool for extracting complex fea- fulness tures and patterns from input data. Lever-Andrea Di Credico, David Perpetuini, Pascal Iz-aging the CNN architecture, the study cal- zicupo, Giulia Gaggi, Nicola Mammarella, Alberto culates the ESI, a metric designed to reflect Di Domenico, Rocco Palumbo, Pasquale La Malva, the subject’s exhaustive status accurately. By Daniela Cardone, Arcangelo Merla, Barbara Gh-analyzing ECG patterns captured during ex- inassi, Angela Di Baldassarre ercise, the ESI provides valuable insights into University “G. d’Annunzio” of Chieti-Pescara, Italy the individual’s physiological response to exer-Sleep quality is a vital component of one’s tion. During the experimental protocol, exer- overall health and well-being. In-adequate cise cessation criteria are standardized, with sleep quality is linked to various adverse con- termination triggered when the Borg scale sequences, including cognitive decline, mood value reaches 20, maintaining constant res- disruptions, and an elevated susceptibility to istance and a cycling cadence of 60 revolu- non-communicable diseases. Hence, it is cru-tions per minute (RPM). This stringent con- cial to precisely evaluate the quality of sleep, in trol ensures consistency across participants order to identify individuals who are at risk and and facilitates accurate assessment of fa- to develop suc-cessful interventions. Import-tigue levels. Signal segmentation methodo- antly, it has been shown that sleep quality can logy adopts a blind approach, dividing ECG impact physiological processes even when a signals into 2-second segments with a 50% person is awake, leading to changes in heart overlap. This segmentation strategy optimizes rate variability (HRV). From this standpoint, the data utilization while preserving temporal in-utilization of wearables and contactless tech- formation critical for fatigue classification. In nologies that can measure HRV without caus- order for training process, the training data- ing any discomfort is extremely well-suited for set is divided into two distinct categories: less- evaluating sleep quality. Nevertheless, there exhaustive and near-exhaustive, comprising is a dearth of studies that analyze the correl- ECG signals recorded from the vicinity of two ation between HRV and sleep quality during extremes of fatigue status. For each subject, waking. The aim of this study is to create a each category comprises 40 ECG signals util- machine-(ML) learning model that uses HRV ized for model training, ensuring robustness data to estimate sleep quality, as evaluated by and generalizability of the developed classi-the Pittsburgh Sleep Quality Index (PSQI). The fication system. To evaluate model perform- measurement of HRV was conducted using a ance, our study employs well established deep wearable photoplethysmography (PPG) sen- learning architectures, including GoogleNet, sor positioned on the fingertip. Subsequently, ResNet-18, and ResNet-50. Numerical results models were created to classi-fy sleep qual-showcase the efficacy of these CNN models in ity based on the PSQI score. By employing the accurately stratifying exhaustive status, high- current ap-proach, a classification good ac- lighting their potential as valuable tools for curacy of 76.7% was achieved. In summary, fatigue assessment in exercise physiology re- this study has the potential to facilitate the use search. Furthermore, regression analysis ex- of wearable and contactless technology for amines the relationship between the calcu-monitoring sleep quality in ergonomic applic- lated ESI and subjective Borg scale values, re- ations. vealing a positive correlation. This finding un- derscores the utility of the ESI as a reliable PO-38 indicator of exercise-induced fatigue, com- Comparative Analysis of Depression Detec- plementing subjective perception metrics with tion using EEG Signals objective physiological assessments. Overall, Stefana Duta, Georgeta-Mihaela Neagu, Alina- this study contributes to advancing our un- Elena Sultana derstanding of fatigue dynamics during exer- National University of Science and Technology Po-cise and exploits the potential of ECG-based litehnica Bucharest, Romania deep learning approaches in enhancing per-This paper presents a comprehensive ana- formance monitoring and athlete manage- lysis of Multilayer Perceptron (MLP) models for ment strategies. the classification of EEG signals in the con- text of depression state detection. Exper- 153 iments were conducted using two separate ac-tresses repeating 200 words in seven dis-databases: the De-pression Rest Database tinct emotions. Furthermore, we in-troduced and the MDD vs. Control Database. For an age-related analysis to gain insights into the Depression Rest Database, the MLP model how age might impact the human capacity reached an accuracy of 84.65% on the training to express emotions through vocal expression. set but faced challenges with validation, plat- Results obtained showed an accuracy of 99.6% eauing at 68.79%. Conversely, the MLP model in the emotional assessment for vo-cal re- excelled in the MDD vs. Control Database, cordings of both actresses employing GTCCs. achieving an accuracy of 89.99% on the train- The age-related studies presented a 100?curing data and 88.97% on the validation data. acy in the emotional assessment restricted to It displayed high precision and recall values the vo-cal samples from the younger actress, for both healthy and depressed classes, indic- compared to the 98.6% obtained in the one ating a balanced performance. Additionally, restricted to the older actress. These findings feature selection was explored on a combined suggest that the age could influence the emo- database, yielding promising results but with tional expression. room for further optimizations. The novelty of this study lies in its investigation into whether PO-42 the com-bination of two datasets, both ori- A Machine Learning Framework for Gait and ented toward the common objective of de- EMG Analysis for Post-Stroke Motor Dys- pression, demonstrates superior performance functions Assessment compared to the individual analyses conduc-Francesco Romano, David Perpetuini, Daniela ted on each dataset. Cardone, Arcangelo Merla University D. D’Annunzio of Chieti-Pescara, Italy PO-40 A stroke is a notable medical disorder char- Unveiling Age-Related Patterns in Vocal Ex- acterized by the abrupt cessation of blood cir-pression of Emotions: A Machine Learning culation to the brain, causing a deficiency of Approach with Mel and Gammatone Fre- oxygen and nutri-ents to brain tissue, which quency Cepstral Coefficients leads to cellular damage or demise. The ef- Michele Giuseppe Di Cesare, David Perpetuini, fects of stroke on individuals may range sig-Daniela Cardone, Arcangelo Merla nificantly, from modest impairments to pro- University D. D’Annunzio of Chieti-Pescara, Depart- found disability. Stroke treatment entails ex- ment of Engineering and Geology, Italy tended therapy, often empha-sizing gait re- The significance of emotional assessment habilitation. Gait rehabilitation programs fo-has gained increasing recognition across di- cus on enhancing gait symmetry, velocity, and verse fields, such as psychology, healthcare, independence, since these qualities have a education, and social sciences. It is deemed substan-tial impact on the likelihood of pa- crucial for comprehending and addressing tients reintegrating into their pre-existing sura broad spectrum of outcomes, including roundings. Significantly, the assessment of mental health, academic performance, pa- muscle activation patterns and neuromus- tient experiences, and social acceptance. A cular control using Electromyography (EMG) noteworthy aspect of human emotions lies measurement during walking is of utmost im- in their expression through a diverse range portance for stroke patients. Machine learning of vocal sounds, which can be interpreted meth-ods have the capacity to provide useful and understood by the listener. From this insights on the patterns of walking and muscle perspec-tive, the application of Mel Fre- activation, functional results, and strategies quency Cepstral Coefficients (MFCCs) and for rehabilitating stroke patients. The object- Gammatone Frequency Cepstral Coefficients ive of this research is to use a Support Vec- (GTCCs) emerges as a brilliant approach to tor Ma-chine classifier to categorize individu-portraying the concealed vocal components als with Stroke from healthy controls by ana-that express emo-tional states. In this study, lyzing gait and EMG measures, both alone and we employed Machine Learning (ML) tech- in combination. The best performances were niques to leverage MFCCs and GTCCs, aiming obtained employing both gait and EMG fea- to construct a classifier capable of assessing tures, reaching a test accuracy of 91.4%. The different emotions. We utilized the freely avail- findings can foster the employment of ML apable Toronto Emo-tional Speech Set (TESS), a proaches to help clinicians to diagnose the dataset comprising vocal recordings of two stroke severity. 154 PO-44 plex set of educational challenges. Address- Comet Assay in the Digital Era: A Review of ing these needs, this paper presents the de-the use of Artificial Intelligence for the Ana- velopment and implementation of an innov-lysis of DNA damage based on the results of ative virtual reality (VR) simulator designed to the Comet Assay enhance the competency of anesthesiologists Adna Softić 1, Nejra Merdović1, Velid Dlakić1, Em-in performing TEE within the specific proced- ina Mrđanović1, Lejla Mahmutović2, Daria Ler3, Lejla ural context of Amplatzer device deployment Gurbeta Pokvić1 for ASD closure. This advanced simulator lever- 1Verlab Research Institute for biomedical engineer- ages high-resolution three-dimensional mod-ing, medical devices and artificial intelligence, Bosels, meticulously crafted from de-identified nia and Herzegovina contrast-enhanced computed tomography 2International University of Sarajevo, Bosnia and (CT) and micro-computed tomography (�CT) Herzegovina scans, to forge a highly realistic and interact- 3ASA Institute, Bosnia and Herzegovina ive training milieu. The rapid advancement of genetic test- The TEE simulation module, endowed with auing techniques has revolutionized our under- thentic TEE controls, provides a one-to-one standing of organism health and opened the correspondence in terms of realism and in- way for personalized approaches to care and teractivity, aiming to markedly elevate the prevention. In light of this progress, this re- anesthesiologists’ proficiency in the naviga-view article explores the intersection of ge- tion and interpretation of echocardiographic netic testing, particularly through the Comet images throughout the surgical procedure. assay, and artificial intelligence (AI) method- While the simulator automates the Amplatzer ologies for analyzing DNA damage and pre- device deployment phase to concentrate the dicting organismal health status. Focusing on learning experience on mastering TEE skills, it research published within the last five years, also imparts an indispensable comprehension this paper synthesizes the current landscape of the device deployment process, thereby of studies leveraging Comet assay and AI, with rounding out the educational experience. In- specific attention to artificial neural networks tegrating an interactive TEE simulation with (ANN), convolutional neural networks (CNN), an automated process for the deployment of Fuzzy Inference System CNN (FIS CNN), Faster the Amplatzer device, the simulator allows for Region based CNN (Faster R-CNN), and Mask the continuous imaging of the catheter and Region based CNN (Mask R-CNN). Through device via echocardiography, offering a coma comprehensive examination of recent liter- prehensive procedural visualization. This edu-ature, this review elucidates the potential of cational innovation represents a significant integrating advanced genetic testing and AI leap forward in the domain of anesthesiology techniques to enhance our understanding of training, offering a cost-efficient, accessible DNA damage and its implications for organ- solution that necessitates nothing more than ism health, thus opening the way for more pre- a VR headset and a computer of adequate cise and personalized approaches to health-capacity. The introduction of this simulator care and disease prevention. into the training regimen of anesthesiologists is anticipated to substantially improve the un- PO-46 derstanding, execution, and efficiency of TEE- Advancing Pediatric Cardiology Train- guided procedures, which are critical for the ing with a VR-Based TEE and Amplatzer successful closure of ASDs. Preliminary feed-Deployment Simulator back from the use of this simulator has been David Buyck 1, Enrique Vergara-Escudero1, overwhelmingly positive, highlighting its effic- Susana Arango2, Varun Aggarwal1, Benjamin acy in enhancing essential procedural skills Gorbaty1, Paul A. Iaizzo1 among anesthesiologists. Future research en- 1University of Minnesota, United States deavors will aim to provide a quantitative as- 2Harvard University, United States sessment of the simulator’s impact on educa- In the specialized field of pediatric cardi- tional outcomes and explore its potential in-ology, the imperative to train anesthesiologists tegration into broader training curricula. Such in the proficient execution of transesophageal research will pave the way for improved pro-echocardiography (TEE) during the closure of cedural success rates and patient outcomes atrial septal defects (ASD) introduces a comin the field of pediatric cardiology, thereby underscoring the significant role of innovative 155 educational tools in advancing medical prac- (LDA) were trained for: (i) EEG features, (ii) clin-tice and patient care. ical and sociodemographic variables, and (iii) the combination of them. These models were PO-48 trained independently to predict short-term (7 Automatic Prediction of Pediatric Postoper- days) and medium-term (28 days) PBD. The LR ative Behavioral Disturbance from Neuronal model using both EEG and sociodemographic Activity During Anesthesia variables achieved the highest performance in Javier Gomez-Pilar 1, Laura Gutiérrez-de Pablo2, terms of accuracy in the test group (Acc=0.92, Carlota Gordaliza3, Estefanía Gómez-Pesquera3, AUROC=0.97) for short-term PBD, as well as for Jesús Poza1, Eduardo Tamayo4, Roberto Hornero1 medium-term (Acc=0.82, AUROC=0.74). These 1Universidad de Valladolid and Centro de Investiga- results highlight the importance of EEG fea-ción Biomédica en Red en Bioingeniería, Biomater- tures during anesthesia induction to enhance iales y Nanomedicina (CIBER-BBN), Spain the robustness of models trained by conven- 2Bioengineering Department, University Carlos III of tional sociodemographic variables to predict Madrid, Spain the emergence of PBD. These findings open 3Faculty of Medicine, University of Valladolid, Spain new doors in PBD prediction, enabling tailored 4Biomedical Research Networking Center in Infec- treatment even before the first symptoms aptious Diseases (CIBERINFEC), Carlos III Institute of pear. Health, Spain This research was funded by “MI- The process of undergoing general anes- CIU/AEI/10.13039/501100011033” and by thesia can be distressing for children and has “ERDF A way of making Europe” through the potential to induce adverse behavioral the project PID2022-138286NB-I00; by GRS changes. This alteration, named Postoper- 2636/A/22 funded by ‘Gerencia Regional de ative Behavioral Disturbance (PBD), manifests Salud de Castilla y León’; and by ‘CIBER en not only during the induction of anesthesia but Bioingeniería, Biomateriales y Nanomedi- also upon awakening from it, as well as after cina (CIBER-BBN)’ through ‘Instituto de Salud discharge home. While previous studies have Carlos III’. predominantly assessed behavioral dynam- ics during anesthesia induction and emer- PO-50 gence stages, post-hospitalization alterations Investigation of the Use of Gait Analysis for remain underexplored. Using electroenceph- Prosthetic Patients: Evaluating the Rela-alographic (EEG) recordings obtained during tionship Between Gait Outcome Measures a surgery procedure to monitor neural activity and Function (K-level) involving anesthesia induction, we hypothes- Amy Edwards, Thomas Stone ize the feasibility of predicting children who are Cambridge University Hospitals, United Kingdom at a higher risk of experiencing PBD. There-Background: In the UK, there are 14 ac- fore, this study aims to develop an automatic credited gait laboratories however, only three predictive model for identifying pediatric pa-state prosthetics as a main clinical activity. tients who will experience PBD, based on EEG, Observational gait analysis is used to de-clinical, and sociodemographic variables, en- termine K-level classification. K-levels reflect abling personalized treatment and minimiz- functional and ambulation ability and are ing adverse consequences. For this, 100 pe-classified from K0 (lowest) to K4 (highest). diatric EEG recordings (patients aged 2 to K-levels are used to prescribe prosthetic com-12 years) from SedLine equipment (4 frontal ponentry and track rehabilitation. However, electrodes) were used. EEG signals were pre- the literature is unclear on which outcome processed and segmented into three states measures are best to assess K-level. There is no (awake, anesthesia, and surgery). Spectral gold standard measurement of K-level. K-level (relative power, median frequency, spectral determination is observational consisting of entropy, and spectral asymmetry) and non-patient-perceived scores, questionnaires, and linear features (sample entropy and Lempel- timed-walking tests. Patients can be poor Ziv complexity) were extracted from the three at self-reporting and observed measures are EEG states. Fast Correlation-Based Filter was subjective. Quantitative gait measures can used to perform automatic feature selection overcome these issues. in the training subset (50% of the total pop- Objectives: This study investigates the re- ulation). Then, models based on logistic re- lationship between K-level and clinically gression (LR) and linear discriminant analysis 156 relevant quantitative gait outcome measures as ventilator-induced diaphragmatic dys-obtained by instrumented gait analysis. New function. Stimulation of the phrenic nerve can quantitative clinical measurements of move- keep the diaphragm active and overcome ment fluency have also been investigated. this complication. For non-invasive electrical Methods: 6 K2 and 6 K3 transfemoral am- stimulation, the anatomy and especially the putees performed walking, sit-to-stand and thickness of the subcutaneous fat layer are sit-to-walk activities. Outcome measures important in determining the stimulation investigated included temporal-spatial para- current required to activate the nerve. meters, knee angles, hip range of motion Three different volume conductor models of (ROM), symmetry indices, hesitation, and the neck are built to represent different neck smoothness. circumferences (35 cm, 41 cm, and 50 cm) Results: K3 participants walked significantly including different subcutaneous fat tissue faster than K2 participants and with signi-layers. The subcutaneous fat tissue thick- ficantly greater step lengths, stride lengths nesses with a variation from 1 mm to 23 mm and reduced periods of double support. No represents the variability of the German male difference was observed in maximum knee population. The volume conductor model of angle in loading response. K3 users have the neck has a high level of anatomical detail significantly increased maximum swing knee and contains skin, subcutaneous fat tissue, angles on both sides compared to K2 users. muscles, bones, intervertebral discs, cartil-K3 users also have significantly increased and age, thyroid, trachea, esophagus, internal air, more symmetrical hip ROM. No other signi- blood vessels, and nerves. The phrenic nerve ficant differences in symmetry were present includes three fascicles, including Epineurium, between the groups. In sit-to-stand, K3 users Perineurium, and Endoneurium. Electrical are smoother. In sit-to-walk, K3 users are stimulation is incorporated via two square smoother and less hesitant. Cluster analysis electrodes with an edge length of 10 mm and suggests that parameters investigated may a distance between them of 25 mm (edge be used to objectively classify patients and to edge) placed on the neck surface at the that some individuals fall between K-levels. posterior border of the sternocleidomastoid Conclusions: There are biomechanical dif- muscle. We apply a monophasic rectangular ferences between K2 and K3 users that can stimulation pulse with a pulse width of 150 be objectively measured. Cluster analysis has �s. Finite element simulation of the electrical been used to develop an objective tool to potential distribution is done under steady-classify patients based on a combination of state conditions using COMSOL Multiphysics relevant gait parameters. This would reduce with a tetrahedral mesh of about 90 million subjectivity in classifying patients. elements. Electric potential values in the Implications: The results of this study will be phrenic nerve elements are coupled with a used to inform physiotherapists on which biophysiological model of the phrenic nerve objective outcome measures are most rel- to calculate the activation threshold. The bio-evant to ensure new prosthetic users reach physiological nerve model for the evaluation their functional potential. The findings of this of the action potential generation is based study could be used clinically in guiding per- on the McIntyre-Richardson-Grill axon model. sonalised prosthetic prescription, targeting The modeled phrenic nerve axons have a interventions and optimising rehabilitation diameter ranging from 5.7 �m to 14 �m. protocols. The activation threshold depends on the neck circumference and the axon diameter. For the PO-52 smallest neck circumference, the activation Influence of the Subcutaneous Fat Layer current is between 11.6 mA and 81 mA, for Thickness on the Activation of the Phrenic the middle neck circumference between 20.3 Nerve via Non-Invasive Electrical Stimula- mA and 139.3 mA, and for the largest neck tion circumference between 29.3 mA and 225.6 mA Laureen Wegert, Alexander Hunold, Marek Zi- with the minimum activation threshold for the olkowski, Irene Lange, Tim Kalla, Jens Haueisen largest axon of 14 �m and the maximum value Technical University of Ilmenau, Germany for the smallest axon of 5.7 �m. Mechanical ventilation is a life-saving We conclude that neck circumference is an technique used every day, but it is of- essential parameter to be taken into account ten associated with complications such for the design of individualized phrenic nerve 157 stimulation setups. ern Switzerland, Switzerland 2Université Clermont Auvergne, France PO-54 Use of electrical household appliances and Structural brain templates are the found- risk of all types of tumours: A case-control ational element for group analysis in neur-study oscience, providing an anatomical reference when analyzing data from different patients. Muhammad Waseem Khan, Shabana Noori When creating such templates, the topic of Balochistan University of Information Technology the appropriate number of brains to obtain a Engineering and Management Sciences, Pakistan stable anatomy must be addressed. The goal The use of electrical appliances using ex- of this study was to estimate the number of tremely low frequency (ELF) magnetic fields patients required to reach convergence in the (MF) has increased in the past few years. These creation of a cohort-specific anatomical tem- ELF MF are reported to be linked to several ad- plate through exemplary calculations for data verse health effects including (but not limited (Ptolemee Electrophysiologie project: IRB 5921, to) neurodegenerative diseases, reproductive CE-CIC-GREN-18-03) from Clermont-Ferrand health effects and certain types of tumours. university hospital (France). Previous studies conducted on the topic have Preoperative imaging data from a group of 47 reported mixed effects. Based on the findings patients (Parkinson’s: 30, essential tremor: 17) of previous studies, the international agency who received deep brain stimulation was used for research on cancer has classified ELF MFs in an iterative, non-linear, mixed-modality, as a possible human carcinogen. We studied unbiased anatomical normalization pipeline the use of common household electrical appli-published previously. It consists of iterative ances and suspected risk of tumours in a multi- non-linear normalization of all original images hospital-based case-control study. Only a to an anatomical template updated after couple of studies have been conducted on the each iteration and implemented to use both T1 association between risk of tumours and use of and WAIR (white matter attenuated inversion electronic devices using low frequency (LF) MF. recovery, a modality specially designed to enThe study was conducted in four tertiary care hance grey matter contrast). During the pre-hospitals (regional cancer referral centers) in operative planning, up to 35 deep brain struc- district Quetta. The study participants com- tures were manually labeled by a single expert. pleted questionnaires on fourteen (14) com- The normalization process was repeated, in- monly used household electrical appliances. In creasing the number of included patients, res- total, 316 patients (158 tumour cases and 158 ulting in 5 different templates. The perform- control cases) were included in the final ana- ance of the normalization was quantified us- lysis. The study results showed increased risk ing the pairwise overlap between anatomical of tumour only for couple of devices whereas structures across patients. A logistic function the risk of tumour for all other studied devices was then fitted on the median values of that was below unity. An increased risk of tumour score for each template to estimate the num- was observed for computer screen use OR: 2.77 ber of patients necessary to obtain a variation (95% CI: 0.90 – 8.55) and use of microwave lower than 5%. oven OR: 2.65 (95% CI: 0.73 – 9.61). The current In this study we estimated the number of study serves as a pilot study of primary data patient images required to obtain a stable and will be helpful in future epidemiological re- group-specific anatomical template. Manual search studies on the topic in the region. Fur- segmentation of deep brain structures was ther research on the risk of electronic devices used to benchmark templates with increas-using LF MF is necessary, involving diverse par- ing number of patients included. Results might ticipants from various parts of the country. differ depending on the specific MR sequences used. PO-56 Acknowledgements: This work was suppor- How Many is Enough? The Influence of pa- ted financially by the Swiss National Science tient count on structural normative tem- Foundation (205320_207491) plate quality Dorian Vogel 1, Jérôme Coste2, Jean-Jacques Lemaire2, Simone Hemm1 1University of Applied Sciences and Arts Northwest- 158 PO-58 mental trials utilizing both PPG and GSR fea- Assessing stress level by utilizing heart rate tures in comparison with classifications based variability and galvanic skin response fea- on solely PPG or GSR features. tures The obtained results suggest that the chosen Sandro Radan 1, Jana Bratanić2, Ana Sušac1 metrics can reliably be used in assessing and 1University of Zagreb, Croatia multilevel classification of stress during differ- 2University of Split, Croatia ent cognitive tasks. This work provides a new approach to ana- lyzing and assessing stress level by utilizing two PO-60 Sensitivity Analysis of Reconstruction Fiber physiological signals: photoplethysmography Tracking Parameters from Diffusion Tensor (PPG) and galvanic skin response (GSR), re- Magnetic Resonance Imaging of Reanim-corded with the wearable wrist-mounted sys- ated Ex Vivo Swine Hearts tem Shimmer3 GSR+. Participants were sub-Neal Duong, Paul A. Iaizzo jected to varying levels of stress induced by University of Minnesota, United States the Stroop color-word test and the Trier Social Stress Test (TSST). This study investigates the optimiza- The PPG signal was firstly bandpass filtered in tion of cardiac fiber orientation mapping the range 0.05-8 Hz by applying a second or- of formalin-fixed swine hearts through the der Butterworth filter. Afterwards, peaks and applications of diffusion tensor magnetic res- their positions, which we later used for extract- onance imaging (DT-MRI) and advanced fiber ing the time-, frequency- and non-linear fea- tracking techniques. DT-MRI, a modality that tures, were extracted by applying the Event gauges the movement of water molecules Related Moving Average (ERMA) algorithm. in soft tissue, is a powerful tool for under- We extracted a total of 17 features (6 time- standing the relative fiber orientations within based, 5 frequency-based and 6 non-linear cardiac structures. Utilizing the Dipy fiber features) in accordance with the overview of tracking software and employing customized heart rate variability (HRV) norms and metrics. denoising methods, our research focuses The GSR signal was preprocessed and de- on unraveling the sensitivity of parameter composed using the Ledalab software leav- selections in reconstructing fiber orientations ing us with continuous tonic and phasic activ- within both the left and right ventricles. ity components. We extracted the peaks and The analysis framework involves a multi-step their positions of the tonic component fol- process, commencing with the reanimation lowed by identifying the skin conductance re- of hearts, followed by fixation in formalin, and sponse (SCR) events. Additionally, the skin the strategic application of agarose solutions conductance level (SCL) was determined by to reduce imaging artifacts. Subsequent taking the average of the GSR signal in the scanning using a Siemens 3T MRI machine, time window. In total, 10 features (8 time- coupled with image file conversions to a NIFTY based and 2 frequency-based features) were format for compatibility with Dipy; enabled extracted following the guide for analysing a comprehensive analysis of cardiac fiber electrodermal activity. The baseline features orientations. Manual segmentation utilizing were used to normalize the features extracted ITK-SNAP further refines each dataset to solely during the experiment by dividing those fea- myocardial structures, minimizing noise and tures by the corresponding baseline feature. facilitating focused computational analyses The extracted features were fed to differ- over traditional thresholding methods. ent machine learning algorithms (k-nearest Parameters critical to fiber tracking, such as neighbours (kNN), random forest, naïve Bayes, step size, stopping criteria based on fractional decision tree, etc.) to obtain stress level clas- anisotropy, and maximum angle deviation, sification during different stages of the exper- were systematically adjusted to explore their iment. Each classifier was trained and valid- sensitivity for optimal ventricular reconstruc-ated using different subsets of selected fea- tions. The study carefully evaluated these tures, whereby multifold cross-validation was parameters across a spectrum, ranging from performed to prevent overfitting. step sizes of 0.025 to 0.3, stopping criteria We found increased classification accuracy from 0.025 to 0.3, and max angle thresholds and F1 scores of multilevel classification of spanning from 15 to 75 degrees. Visual as- stress during different stages of the experi- sessments, supplemented by user-assigned 159 ranks, aided in evaluating representation ferrite nanoparticles (MnFe2O4@CIT) in three fidelities by comparing reconstructed images concentrations (1%, 2% and 3%) within gelatin with established research models. Optimized phantoms. Then, we varied systematically fiber tracking techniques identified crucial the distance from the magnetizing coil and parameters; notably a stopping threshold the amplitude and duration of the excitation below 0.1 and an angle limit between 45 and pulse. Next, by using an ultrasound platform 60 degrees. (Verasonics Vantage System) we acquired The present findings have promising implic- the RF data and reconstructed the echo ations for clinical applications, not only in maps, being able to additionally assess the enhancing diagnostic accuracies but also nanoparticles localization and viscoelasti- perhaps in guiding electrophysiology inter- city parameters of the phantom samples. ventions. We envision applications such as Results highlight the significant decrease of determining optimal left bundle branch pa- the nanoparticles displacement when dis- cing depths, visualizing ventricular ablations, tance from the coil is increased by just a few and characterizing myocardial infarctions’ millimeters or the inclusion concentration is morphologies and location. Furthermore, our reduced. Furthermore, changing the pulse research contributes valuable physiological duration resulted in different spectral content context to ventricular function and electrical of the acquired signals, which is useful for behaviors, thereby providing critical inputs for viscoelasticity analysis via the shear wave computational modeling. In the future, our velocity maps. In conclusion, standardiza- group aims to extrapolate these optimized tion of such parameters can provide more techniques to human heart specimens within accurate information of the effects of nano- our Visible Heart� human specimen library, particles within tissues and the corresponding fostering additional translational models properties gathered from these studies. for understanding human cardiac fiber ori- entations. The potential impact of these PO-64 findings extends beyond swine models, and Unveiling the phenolics of chestnut honey- may provide promising advancements in based propolis and in silico phytoestrogen the comprehension and application of car- activity diac imaging in both preclinical and clinical Idris Arslan settings. Zonguldak Bülent Ecevit University Department of Biomedical Engineering Zonguldak Turkey, Turkey PO-62 An estrogen deficiency at menopause in- Insights into Magnetomotive Ultrasound: duces climacteric symptoms in humans in-Evaluating Key Parameters for Enhanced cluding hot flashes, diaphoresis, sleep disturb-Imaging ances and progressive bone loss. A phytoes- Ariane Sanches, Nicholas Zufelato, David trogen can be defined as a plant-derived xen- Collazos-Burbano, João Henrique Uliana, Théo oestrogen molecule with estrogenic activity Pavan, Antonio Adilton Carneiro and not associated with the endocrine sys- University of Sao Paulo, Brazil tem. The present study aims to investigate the Magnetomotive ultrasound is a recent chemical constituents and their potency as a technique that combines magnetism and phytoestrogen of chestnut-honey based pro- ultrasound imaging to localize magnetic nan- polis from Çarşamba district (Sam-sun, Turkey) oparticles within tissues and determine their by in silico molecular docking simulation. The viscoelastic properties based on the shear present study evidenced the phenolic con- wave velocity. For the optimal operation stituents from chestnut-honey based propolis of this technique, key factors include the possess phytoestrogen activity. intensity and gradient of the magnetic field, the transient characteristics and width of PO-66 the applied pulse, as well as the magnetiza- Enhancing Cross-Domain Adaptability of tion and concentration of the nanoparticle, Existing Computer-Aided Endoscopic Le- among others. The primary objective of this sion Detection Using Plug-and-Play Tracker study is to evaluate the influence of these Yijie Ku, Hui Ding, Guangzhi Wang parameters on misleading results when not School of Biomedical Engineering, Tsinghua Univer-considered in data collection. For this pur- sity, China pose, we applied citrate-coated manganese 160 Computer-aided detection (CADe) for en-fectively mitigat-ing the disadvantages of the doscopy can help physicians to locate and Ponseti method. The developed orthosis can identify lesions better, but there are still many be manufactured globally using the 3D print-false positives (FP) when processing cross- ing process, with a manufacturing cost of domain data. This paper proposes SE-SORT, a around €150, excluding assembly costs. In plug-and-play tracker, designed to be seam- summary, a new solution was proposed within lessly integrated as a post-processing plu- the same treatment method, effectively elim- gin into existing CADe systems to enhance inating skin complications, and enabling cost- the adaptability to cross-domain data. The effective manufacturability on a global scale. proposed tracker adds trajectory initializa- tion thresholds into the tracking association PO-70 strategy, reducing the impact of high confid- A complex spinal surgery lifting system for ence FPs on the matching process. Experi-prone positioning ments show that the modified tracker effect- Lutong Li 1, Stuart Watson2, Glyn Smurthwaite2, ively reduces the impact without significantly John Large2, Andrew Weightman1, Glen Cooper1 1 affecting the processing speed. This allows the The University of Manchester, United Kingdom 2 detector of CADe system to tolerate lower de- Salford Royal NHS Foundation Trust, United King- tection confidence thresholds, thus improving dom the overall accuracy on cross-domain data. Spinal surgery is one of the fastest-growing Compared to existing SORT trackers, the pro- surgical procedures and improving prone pos-posed tracker exhibits better accuracy and ition is always a big challenge, specifically for higher efficiency in endoscopic lesion detec- patients with ad-vanced deformity. Current tion and tracking. This work will help to improve spinal surgery tables (e.g. Allen Table, Hill-the generalization and expanding the clinical rom LTD, Newbury, Berks, UK) can meet the application scope of related works on endo- simple positioning requirement, but they lack scopic real-time CADe. the necessary flexibility and customisation to adequately address the needs of patients with PO-68 more complex requirements, such as patients Design and Analysis of a Biomedical Or- with ad-vanced deformities. This study iden-thosis for Clubfoot Corrective Device tified the clinical challenge of prone posi- Shahrol Mohamaddan tioning through a literature review and focus Shibaura Institute of Technology, Japan group with six professionals (one clinical engin- Clubfoot is a common congenital foot de- eer, two spinal anaesthetists and three mech- formity that leads to constant pain and sig- anical engineers) and proposes a solution to nificant limitations if left untreated or not address the problem. A two-stage study was treated adequately. The most used method conducted. A detailed design specification of for treating clubfoot is the Ponseti method. It the smart lifting system was developed based involves a correction phase where about five on the literature review and focus group res-plaster casts are applied and changed weekly. ults, which includes four cat-egories, ‘Hard-This treatment lasting about 2 to 3 months, ware’, ‘Software’, ‘Safety’ and ‘User’s prefer-is the most chosen method due to its high ence’ requirements in the first stage. The success rate. However, treated babies of- second stage is the design and evaluation of ten experi-ence skin complications caused by the smart lifting system. According to the stiff and tight casts. Previous research showed design requirements, a smart lifting system is that viable solutions already exist including or- designed for spinal surgery with 170 kg loading thoses. In this re-search, a developed method capability and 250 mm lifting height. Multiple known as VDI 2221 was applied and the printa- iterations of focus groups involving profession-ble orthosis using 3D printer was selected as als were under-taken to assess the smart lifting an alternative to Ponseti meth-od. Calcula- system, and it successfully satisfies all identi- tions and finite element method (FEM) analysis fied design requirements. This study proposes demonstrated that the orthosis made of PA6- a potential solution that can be used to ad-CF provides sufficient stiffness and strength, dress the difficulty of positioning patients with assuming the weight force of the foot is 10 advanced deformities and reduce the need to N. The selected design was de-veloped based use manual handling of patients during sur- on requirements and functional analysis, ef- gery. 161 Authors’ Index A Ballarin, Virgina Laura 110, 59, Cardenas Caceres, Pablo 121, Abdullah, Amira Raudhah 70, Balling, Susann 91, Cardone, Daniela 153, 154, 154, 135, Banderet, Grégoire 64, Caricato, Marco 127, Abramov, Dimitri 142, Barrio Cortes, Jaime 60, Carneiro, Antonio Adilton 151, Abu Bakar, Nur Nabilah 135, Bassi, Francesco 64, 97, 128, 160, Accardo, Agostino 64, 128, 139, Batista Napotnik, Tina 74, Cartocci, Alessandra 115, 97, 89, 42, 53, Baumgarten, Daniel 100, Cartocci, Alessandra 44, Adoberg, Annika 81, Bećirović, Faruk 137, Carvajal, Diego 71, Adolf, Jindřich 106, Beiramvand, Matin 52, Casan-Pastor, Nieves 47, Adorno, Bruno 40, Bigaki, Maria 91, Castaldo, Rossana 69, Aggarwal, Varun 155, Billis, Evdokia 141, 141, Castaldo, Rossana 115, Ahluwalia, Arti 113, Bini, Fabiano 41, Castella, Lorenzo 118, Ajcevic, Milos 89, 53, 97, 128, 42, Biscontin, Alessandro 97, Castillo, Andrés 60, Aki, Fumitaka 136, 129, 127, 130, Björkman, Mats 94, Cea, Gloria 66, 71, Akulauskas, Mykolas 130, Bliznakov, Zhivko 104, 39, Cerar, Andraž 57, Al-Halawani, Raghda 150, Bliznakova, Kristina 104, Cernochova, Dana 145, Alabdah, Fahad 84, 117, Bliznakova, Kristina 39, Cevenini, Gabriele 44, Aldana Palomino, Ana Cristina Blystad, Ida 131, Chae, Minseong 135, 121, Boadu, George 114, Chan, Hsiao-Lung 146, 152, Alshammari, Adel 84, 117, Bocchi, Leonardo 139, Chang, Ya-Ju 146, 152, Amato, Francesco 118, 67, Boccia, Rosa 116, Chang, Cheng Chun 146, Anagnostopoulos, Spyridon Bodenstorfer, Mirjam 68, Chaudhary, Ansh 116, 91, Boscarino, Tamara 83, Checa Rifa, Pedro 66, 132, Andellini, Martina 40, 88, Botezatu, Radu 127, Chen, Chien-An 110, 107, Angelone, Francesca 118, 67, Botis, George 117, Chen, Szi-Wen 152, Arango, Susana 155, Botta, Alessio 117, Chen, Jia-Jin 107, 110, Arienzo, Arianna 118, Bouda, Sabrina 91, Chen, Rou-Shayn 152, 152, Arpag-McIntosh, Sezgi 56, Boukany, Pouyan 74, Chen, Mei-Fen 120, Arredondo, Maria Teresa 71, Boye, Carly 56, Chen, Lei 138, 60, 72, Brandão, Mariana 120, Chen, Hsing-Yu 150, Arroyo, Peña 72, 60, Bratanić, Jana 159, Chen, Pao-Huan 110, Arroyo, Nubia 76, Bromis, Konstantinos 117, Chen, Rou-Shayn 146, Arslan, Idris 160, Bruña, Ricardo 80, 80, Cheng, Yung-Hsin 150, Arth, Kevin 101, Bucciarelli, Vittoria 148, 148, Chern, Jen-Suh 133, Artioli, Rebecca 73, Buchner, Teodor 65, ChiFu, Chung 146, Arund, Jürgen 81, Budil, Martin 90, 104, Chikweto, Francis 46, Assugeni Neto, Nilton 151, Bulysheva, Anna 56, Chikweto, Francis 128, Auškalnis, Liudas 130, Chiu, Wen-Tai 49, Buoite Stella, Alex 139, Ayinde, Babajide 91, Cibulka, Jaroslav 39, Buyck, David 155, Azarbarzin, Samaneh 108, Cieza Huané, Leslie Yessenia C Á 121, Cabrera, Maria Fernanda 72, Cifrek, Mario 85, 86, Álvarez, Daniel 138, 82, 71, 71, 60, Cimperman, Tina 74, Árnason, Jón Andri 105, Cagy, Mauricio 142, Ciobanu, Valentina 98, B Cai, Kui 138, Cipollina, Andrea 90, Babic, Ankica 66, Calcagno, Andrea 118, Collazos-Burbano, David 160, Babić Jordamović, Naida 137, Calderón Bocanegra, Fran- Collesi, Chiara 73, Bacco, Luca 43, cisco Carlos 119, Comani, Silvia 80, 78, 80, Bachmann, Maie 53, Cancilla, Nunzio 90, Conceição, Arthur 142, Balciuniene, Neringa 91, Cannilla, Rossella 73, Cooper, Glen 40, 117, 84, 40, 161, Ballarin, Virginia 76, Cappuccio, Francesco 115, Cordella, Fulvio 118, 163 Coro, Florinda 113, D’Urso, Giordano 117, Giuffrè, Mauro 42, 89, Coste, Jérôme 158, d Goh, James 45, 50, Coste, Jérôme 148, 148, da Costa, Lucas 151, Gómez-Pesquera, Estefanía Crapanzano, Fabio 42, 156, de Brito Martins, Ayana 118, Croce, Lory 89, 42, del Campo, Félix 138, Gomez-Pilar, Javier 156, 138, Cruces Chancahuaña, Andre Gorbaty, Benjamin 155, Jesus 121, E Gordaliza, Carlota 156, Cubides, Mauricio 76, Edwards, Amy 156, Goretti, Francesco 115, Culjak, Ivana 85, Ehlich, Jiří 47, Gozal, David 138, 82, Curti, Filis 99, Elsas, Siegward 148, Grandone, Elvira 91, Cybulski, Gerard 87, F Grassi, Roberto 67, Cybulski, Gerard 65, Faes, Luca 145, Gschwandtner, Ute 148, Czippelova, Barbora 62, 145, Fang, Chih-Yuan 150, Gu, Xiaoyue 138, Č Feng, Jingyi 150, Guillen, Sergio 72, Čerňanová Krohová, Jana 62, Ferreira-Santos, Daniela 138, Gurbeta Pokvić, Lejla 137, 155, 145, Fico, Giuseppe 72, 70, 66, 60, Gutiérrez-de Pablo, Laura 156, D 72, 71, Gutiérrez-Tobal, Gonzalo C. Fiedler, Patrique 80, 79, 80, 78, 82, 115, D’Ambrosio, Antonio 83, 78, 142, Gutiérrez-Tobal, Gonzalo C. D’Antoni, Federico 93, Filho, Edson 80, 138, Daino, Giovanni Luca 121, Filipovic, Nenad 104, 94, 136, Dal Ferro, Matteo 73, Guzman, Raphael 148, Filippi, Simonetta 106, Davalos, Rafael 48, 55, Guzmán Rodríguez, Natalia Finti, Alessia 41, Davalos, Rafael V. 34, 86, Fiorentzis, Miltiadis 143, Dawes, Helen 40, Guzmán-Alarcón, Marta 86, Florea, Bogdan Cristian 145, De Batlle, Jordi 66, Guðmundsdóttir-Korchai, Flores-Luna, Rosa 76, De Fano, Antonio 80, 80, Ragnhildur 105, Folkvord, Frans 91, De Maria, Carmelo 113, Fournier, Guillaume 118, H Dermatas, Evangelos 141, Fox, Daniel 69, Haberl Meglič, Saša 75, Dermitzakis, Aris 141, 141, Francia, Piergiorgio 139, Haenni, Etienne 64, Dermitzakis, Aris 123, 122, Francis, Michael 56, Haleem, Salman 115, DeVos, Amanda 140, Franzò, Michela 41, Hallbeck, Martin 131, Di Baldassarre, Angela 153, Fridolin, Ivo 63, 81, Hána, Karel 41, Di Cesare, Michele Giuseppe Frljak, Sabina 57, Haraguchi, Ryo 45, 154, Fuhr, Peter 148, Haritou, Maria 117, Di Credico, Andrea 153, Fukaya, Aoi 46, Haron, Athia 40, Di Domenico, Alberto 153, Fytagoridis, Anders 147, Hasugian, Jimmy 60, Di Pasquale, Elisa 73, Haueisen, Jens 78, 78, 142, 79, Dianderas Jorge, Angel G 157, Eduardo 121, Gaggi, Giulia 153, Hauffe, Ralf 78, Didaskalou, Stylianos 91, Galhanone, Paulo 142, Havlík, Jan 62, Dinescu, Sorina 98, 99, Gallego, Alba 66, Ding, Hui 160, Gao, Yueming 81, 86, 85, Hay, Alana 48, Dionisio, Pietro 91, Gao, Hong 88, Hayashi, Hirosane 134, Djordjevic, Srdjan 64, García, María 115, 63, Helgason, Þórður 105, Dlakić, Velid 155, García, Constantino A. 86, Heller, Richard 56, Dobranici, Alexandra-Elena Garcia, Renato 120, Hemm, Simone 148, 147, 148, 34, 99, Garibizzo, Gianluca 69, 158, Dobrisan, Raluca 98, Gautier, Maxime 91, Henriques, Jorge 91, Doksanský, Marek 41, Geng, Xiangnan 88, Hensley, Jacob 56, Doležal, Jaromír 106, Ghinassi, Barbara 153, Herrero-Tudela, María 115, 63, Drosatos, George 91, Ghisalberti, Gianpaolo 121, Hidalgo-Bastida, Araida 117, Drougka, Ioanna 91, Giacca, Mauro 33, 73, Hinrikus, Hiie 53, Dukov, Nikolay 39, 104, Giannitsas, Konstantinos 141, Ho, Mary 48, Duong, Neal 159, 92, Gil, Paulo 91, Hoad, Damon 69, 67, Duta, Stefana 153, Ginstman, Fredrik 108, Holmar, Jana 81, 164 Hornero, Roberto 63, 82, 115, Kaneuji, Ayumi 134, Lhotska, Lenka 119, 39, 138, 156, Karwat, Piotr 61, Li, Lutong 161, Hosono, Junto 137, Keraka, Margaret 113, Li, Guobin 138, Hospodková, Petra 104, 90, Khan, Muhammad Waseem Li, Hung Chun 144, 85, Hou, Kung Jui 144, 158, Li, Dongming 85, Hren, Rok 121, 101, Khazaei, Mohammad 80, 80, Li, Xin 88, Hsieh, Ai-Tung 152, Kijima, Tashihiko 46, Liebert, Adam 83, 109, Hsu, Miao-Ju 146, Kim, Joohyeong 106, Lim, Chwee Teck 44, Huang, Chih-Chung 129, Kim, Yu-Chan 135, Lin, Kang-Ping 60, 45, 49, 120, Huang, Yi-Jing 110, Kimura, Tatuhiro 136, Lin, Wen-Chi 120, Huang, Yu-Hsuan 152, Kimura, Tatsuhiro 127, Lin, BO-TSEN 110, Huang, Yu-Jui 110, Kimura, Yuichi 151, Lin, Wen-Chen 120, Huang, Jin 144, Kircher, Marco 91, Lindén, Maria 94, Hundur, Madžida 137, Klaić, Luka 81, Lipping, Tarmo 52, Hunold, Alexander 157, Kliment, Radim 41, Lipšová, Vladimíra 106, Hyttinen, Jari 92, Klint, Elisabeth 131, Lithgow, Brian 109, I Klubarská, Irina 104, Liu, Jinlong 149, 82, Kojic, Milos 136, Liu, Qilin 144, Iadanza, Ernesto 42, 69, 68, 121, Kolková, Zuzana 62, Liu, Jinfen 82, 44, 122, 115, Komosar, Milana 79, Liu, Tse-Ying 49, Iaizzo, Paul A. 159, 140, 155, 92, Koo, Dongkyu 135, Lo, Chao Chen 107, Ibrahim, Nahimiya Husen 132, Kos, Monika 73, Lombardi, Sara 139, Ilić, Đorđe 94, Kos, Bor 55, Lombroni, Ivana 71, Ingebretsen, Jim 60, 72, Koss, Peter 100, Longo, Umile Giuseppe 84, Iršič, Jernej 54, Kouris, Ioannis 117, Longo, Isabel 93, Iscra, Katerina 128, 53, Krašna, Simon 64, López, María I. 115, Ishiyama, Kazushi 46, Kresevic, Simone 42, 97, 89, López, María I. 63, Izzicupo, Pascal 153, Krishnan, Shankar 58, 77, 33, 76, Lopez-Perez, Laura 72, J Ku, Yijie 160, Loureiro, João 91, Jacobs, Edward 55, Kumar, Mayank 143, Lu, Shao-Hung 120, Jagličić, Zvonko 83, Kuricova, Miriam 145, Lučev Vasić, Željka 81, 85, 86, Jan, Matevž 54, Kuroda, Tomohiro 45, Luman, Merike 81, Jan, Vid 146, 73, Kwak, Dai-Soon 133, Lungu, Adriana 99, 98, Jarm, Tomaž 102, Kyriacou, Panicos A. 150, Luo, Weiru 149, Jarmolowska, Joanna 128, L Lupiañez, Francisco 60, 72, Jaroszeski, Mark 56, Luschi, Alessio 44, 121, 42, 122, La Malva, Pasquale 153, Jaufenthaler, Aaron 100, Labyt, Etienne 101, M Javorka, Michal 62, 145, Lai, Ying-Hui 49, Ma, Wenjing 138, Jazbinšek, Vojko 109, 101, 83, Lampropoulos, Athanasios 141, Macas, Andrius 91, Jegelevičius, Darius 130, Lange, Irene 157, Maccaro, Alessia 66, 72, Jeon, Hojeong 135, Large, John 161, Maček Lebar, Alenka 56, Jiang, Ruixin 138, Lasabová, Zora 62, Macku, David 103, Jin, Wei 88, Lass, Jaanus 53, Madete, June 113, Jodko-Władzińska, Anna 99, Le Prado, Matthieu 101, Maeda, Sumiko 134, Jung, Hwachul 135, Lee, Kang-Sik 135, Magjarević, Ratko 102, 76, 83, Jurcic, Krunoslav 83, 85, Lee, Dong-Ho 135, 75, Jurčová, Kateřina 106, Lee, Jessica 92, Mahmutović, Lejla 155, Jurič, Jernej 73, 146, Leelasiri, Puwadej 129, 127, 130, Mahnič-Kalamiza, Samo 48, K Legros, Mathieu 91, Maierová, Lenka 62, Kačerová, Ilona 106, Leis, Liisi 81, Mak, Peng Un 144, 85, Kakkos, Ioannis 117, Lemaire, Jean-Jacques 158, Malbrain, Manu 118, Kalamaras, Ioannis 141, 148, 148, Maljević, Dubravka 34, Kaldoudi, Eleni 91, Ler, Daria 155, Mallo, Irene 71, Kalla, Tim 157, Lewandowski, Marcin 61, 95, Mammarella, Nicola 153, Kampekete, Muyanga 128, Lhotská, Lenka 106, 103, Manco, Melania 40, 165 Manea, Ionut 145, Miric, Ana 104, Pagnano, Maria Elisabetta 127, Mangano, Valentina 113, Mitronikas, Epaminondas 141, 116, Mansoubi, Maedeh 40, Mitryukovskiy, Sergey 101, Palacios-Laloy, Agustin 101, Mao, Qing 138, Mlamba, Bazil 113, Palermo, Andrea 83, Marhl, Urban 83, 101, 109, Mohamad, Mohd Riduan 135, Pallikarakis, Nicolas 123, 75, 122, Marin, Minodora 98, Mohamaddan, Shahrol 161, 102, Marinozzi, Franco 41, Mohd Effendy, Nadia 135, Palumbo, Rocco 153, Marozas, Vaidotas 91, Moltani, Lara Alessia 91, Papadelis, Christos 106, Martín Gómez del Moral Her- Moreno Elescano, Sergio En- Papadopoulou, Stella-Lida 91, ran, Rodrigo 72, rique 121, Papanikolaou, Dimitra-Tania Martín-Montero, Adrián 82, Mou, Pedro Antonio 85, 141, 141, Martinez-Licona, Fabiola 89, Moussavi, Zahra 109, 108, Paredes, Simão 91, 111, Moustakidis, Pavlos 91, Parrish, Amy 118, Martini, Miriam 139, Mrđanović, Emina 155, Passerini, Michele 90, Martynenko, Alexander 52, 52, Mubanga, Boyd 128, Pastor, Xavier 52, 52, Marusic, Uros 128, Muehlsteff, Jens 65, Pastore, Juan 76, Mascianà, Gianluca 127, Munaretto, Laura 53, Patwardhan, Manali 48, Mašek, Matyáš 104, Munsanje, Foster 128, Pavan, Théo 151, 160, Matarrese, Margherita Anna Musa, Intan Maslina 70, 135, Pałko, Tadeusz 65, 63, 87, Grazia 84, 106, 116, 127, Mužík, Jan 41, Pecchia, Leandro 36, 83, 67, Matjačić, Zlatko 35, Mwaura, Salome 111, 132, 69, 103, 127, 116, 93, 59, 84, Matsopoulos, George 117, 97, 106, 43, Matsumoto, Takeo 45, N Pecchia, Leandro 115, 69, 72, Mattila, Henna 87, Nagaoka, Takashi 151, 66, Matušková, Lenka 62, Nakajyou, Yuuta 134, Pecchia, Leandro 88, Mauritzon, Stina 108, Nathaniel, Emmanuel 69, Pęczalski, Kazimierz 63, 87, 65, May, James 150, Neagu, Georgeta-Mihaela Pereira-Rodrigues, Pedro 138, Mayr, Winfried 95, 153, 127, 145, Pérez-Tirador, Pablo 86, Mazzari, Laura 139, Nemoto, Mitsutaka 151, Perpetuini, David 154, 154, 153, Mbakah, Symon 112, 123, Nicolae, Carmen-Valentina Perše, Martina 73, McGrath, Brendan 40, 99, Petrella, Lorena 91, Medrano, Alejandro 71, Nicosia, Alessia 90, Petrosino, Lorenzo 93, Mejbar, Mohammed-Reda 78, Nikolic, Milica 104, Piaggio, Davide 69, 67, 132, 69, Melillo, Paolo 116, Noori, Shabana 158, 66, Merdović, Nejra 155, Nordin, Teresa 131, 148, 147, 34, Piccolo, Salvatore 118, Mergl, Vitezslav 39, Novak, Petr 39, Piemonte, Vincenzo 93, 132, 83, Merino, Beatriz 72, 60, 72, 71, Novikov, Dmitry 91, Pilt, Kristjan 63, 81, Merla, Arcangelo 153, 154, 154, Nydal, Anton 66, Pinos, Eduardo 76, Merlo, Marco 53, Pletkus, Justinas 130, Merone, Mario 93, 43, O Poglajen, Gregor 57, Mezzatesta, Vincenzo 121, Ohshima, Hiroshi 127, 136, 129, Põld, Toomas 53, Miao, Qiang 144, 130, Polsi, Katia 69, Miklavčič, Damijan 77, 53, 48, Oi, Tomoya 127, 129, Ponsiglione, Alfonso Maria 67, 54, 73, Oksala, Niku 91, 118, Miladinovic, Aleksandar 97, Olăreț, Elena 99, Potoupnis, Michail 91, Miladinovic, Aleksandar 128, Olimzoda, Abdumumin 78, Pousset, Nicolas 91, 53, Ono, Yumie 45, Poveda, Alvaro 76, Milanič, Matija 121, Oppermann, Hannes 80, 142, PoWen, Lu 146, Milev, Minko 104, Otero, Abraham 86, Poza, Jesús 156, Milicevic, Bogdan 136, Otten, Alex 56, Preoteasa, Rares-Marin 145, Milivojevic Dimitrijevic, Ö Prinz, Thorsten 91, Nevena 104, Österlund, Erik 147, Prolič Kalinšek, Tine 54, Milosevic, Miljan 136, Pun, Sio Hang 85, 144, Miloulis, Stavros-Theofanis 117, P Minev, Teodor 39, Paats, Joosep 81, Q Miranda de Sa, Antonio 142, Päeske, Laura 53, Qassem, Meha 150, 166 Qian, Ying 138, 88, Shung, Dennis 42, Thairi, Cecilia 73, R Siino, Marco 90, Thaokar, Rochish 143, Sikwese, Jimmy 128, Tichá, Marie 41, Radan, Sandro 159, Simončič, Urban 121, Tinnirello, Ilenia 90, Raeisi Nafchi, Khadijeh 80, 80, Simonelli, Francesca 116, Tisell, Anders 131, Raffini, Alessandra 139, 128, Sinagra, Gianfranco 73, Todorov, Todor 104, Rahman, S. M. Musfequr 87, Sivarasu, Sudesh 114, 113, 96, Tognetti, Linda 44, Ramirez Martinez, Santiago Sládková, Petra 41, Trotovšek, Blaž 64, Humberto 119, Smajlhodžić-Deljo, Merima Tsai, Yi-Ching 133, Rasser, Ronja 100, 137, Tsai, Cheng-Lun 120, Raumonen, Pasi 87, Smrčka, Pavel 41, Tuohy, Joanne 48, Reberšek, Matej 73, Smurthwaite, Glyn 161, Turianiková, Zuzana 145, 62, Rems, Lea 146, 73, 74, Sobiech, Judyta 65, Turk, Tina 146, 73, Richard, Patrick 64, Softić, Adna 155, Richter, Johan 131, Ț Song, Lipeng 84, Ritrovato, Matteo 40, 88, Țarălungă, Dragoș-Daniel 127, Sołtysiński, Tomasz 60, Rizzi, Jacopo Giulio 53, 145, Sparaco, Rossella 118, Roca-Umbert, Ana 60, Stancu, Izabela-Cristina 99, U Rocha, Teresa 91, 98, Uliana, João Henrique 151, 160, Roglić, Matija 81, Stawiski, Marc 148, Um, Seung-Hoon 135, Romain, Rudy 101, Stergar, Jošt 121, Romano, Francesco 154, V Stiens, Johan 118, Romano, Maria 118, Vai, Mang I. 144, 85, Stokes, Katy 115, Romero-Oraá, Roberto 115, 63, Valchinov, Emil 141, 141, Stone, Thomas 156, Rotbei, Sayna 117, Vaquerizo-Villar, Fernando 82, Stručić, Marko 73, Rubegni, Pietro 44, Vasile, Bogdan Stefan 99, Su, Wanzi 69, 67, Rujas, Miguel 72, 60, 72, Vasile, Titus Mihai 127, Sultana, Alina-Elena 153, 127, Rutkowski, Jaroslaw 101, Vehkaoja, Antti 91, Sun, Fei 150, Rutkūnas, Vygandas 130, Vergara-Escudero, Enrique Sun, Qi 82, Rytis, Jurkonis 91, 155, Sun, Jing 150, S Verma, Atul 33, Sun, Yanjun 149, Vidović, Domagoj 85, Sakai, Hitomi 134, Sun, Wei-Cheng 107, Vihmaru, Merilin 53, Sakalauskas, Andrius 91, Sunday, Conix 128, Vilcahuaman, Luis 144, Sakuma, Ichiro 51, 45, Sušac, Ana 159, Villimar Rodriguez, Ana Isabel Salameh, Zaid 48, Suzuki, Hanako 46, 60, Salman, Ali 43, 115, Syrimpeis, Vasileios 141, Virkki, Johanna 87, Sanches, Ariane 160, Š Vitale, Vincenzo Norman 118, Sander, Tilmann H. 101, 83, 109, Šeketa, Goran 83, Vitale, Jacopo 97, 116, 127, 99, Šmerc, Rok 48, Vitale, Jacopo 69, Sani, Lorenzo 42, Štublar, Jernej 54, Vogel, Dorian 34, 148, 158, 148, Sansone, Mario 67, 147, Santos, Pedro 55, T Voigt, Jens 100, Sassi, Martina 84, 97, Takahashi, Reon 130, 129, 127, Vrtovec, Bojan 57, Satta, Francesca 42, Takano, Noriyuki 134, 134, Vrtovec, Bojan 57, Sau, Francesca 90, Tamayo, Eduardo 82, 156, Vulović, Aleksandra 94, Sawosz, Piotr 109, 83, Tamburro, Gabriella 80, 80, Vyapooree, Alvin 69, Schiaffini, Riccardo 88, Tan, Ka Liong 135, Schloetelburg, Cord 91, Tang, Yuan 40, W Schnabel, Allard 100, Tanner, Risto 81, Wacker, Josias 64, 66, Serša, Gregor 121, Tapper, Sofie 108, Wallace, James 132, Shahbakhti, Mohammad 52, Tarousi, Marilena 117, Wang, Dongan 98, Shaikh, Asif 87, Taub, Ethan 148, Wang, Jiamei 86, 85, Shin, Seonjin 133, Tavernise, Federica 139, Wang, Guangzhi 160, Shiraishi, Yasuyuki 46, Tayebi, Salar 118, Wang, Zhichen 150, Shiraishi, Yasuyuki 45, Telles, Igor 142, Wang, Han 86, Shuang, Yang 86, Testa, Francesco 116, Wang, Ming Chen 49, 167 Wårdell, Karin 131, 147, 34, 148, Yamazaki, Kiyoyuki 130, 136, Zemljič, Gregor 57, 108, 127, 129, Zequera, Martha L. 76, 119, Warsito, Indhika 78, Yambe, Tomoyuki 46, Zhang, Shuheng 138, Waswa, Augustine 113, Yang, Kai-Chiang 150, Zhang, Heng 144, Watanabe, Hayato 106, Yang, Chi-Hung 129, Zhang, Taomin 138, Watson, Stuart 161, Yang, Jiejie 86, Zhang, Qian 150, Wegert, Laureen 157, Yang, Yuzhi 88, Zhang, Zhanjie 144, Wei, Ziliang 81, 86, Yang, Yueqi 88, Zheng, Hairong 50, 58, Weightman, Andrew 40, 40, Yang, Yu 133, Zheng, Junan 138, 161, Yasrab, Robail 116, Zhong, Jie 144, Zhong, Hui 88, Wojtkiewicz, Stanislaw 83, 109, Yeow, Raye Chen-Hua 46, Ziolkowski, Marek 157, Wu, Shi-Yi 120, Yeow, Raye 50, Zisimopoulos, Spilios 123, 122, Władziński, Michal 99, Yilmaz, Gürkan 64, Yin, Yadong 85, Zivanovic, Marko 104, X Yokosawa, Koichi 106, Zsigmond, Peter 108, Xiong, Jiwen 149, 82, Yu, Jr-kai 146, Zufelato, Nicholas 160, Xu, Yingxin 88, Zupan, Blaž 34, Z Xu, Xiang 88, Ž Zajki-Zechmeister, Tibor 97, Xu, Zhimeng 86, Žumer, Jan 64, Zamudio, Luis 76, Y Zappasodi, Filippo 80, 80, Ł Yamauchi, Yasushi 137, Zarghami, Ashkan 118, Ładyżyński, Piotr 76, 168 Document Outline P R O G R A M M E P L E N A R Y L E C T U R E S ' A B S T R A C T S O R A L P R E S E N T A T I O N S ' A B S T R A C T S P O S T E R P R E S E N T A T I O N S ' A B S T R A C T S Authors' Index