Radiol Oncol 2024; 58(4): 509-516. doi: 10.2478/raon-2024-0063 509 research article Liver volumetry improves evaluation of treatment response to hepatic artery infusion chemotherapy in uveal melanoma patients with liver metastases Sebastian Zensen1, Hannah L Steinberg-Vorhoff1, Aleksandar Milosevic1, Heike Richly2, Jens T Siveke3,4, Marcel Opitz1, Johannes Haubold1, Yan Li1, Michael Forsting1, Benedikt Michael Schaarschmidt1 1 Institute of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Germany 2 Department of Medical Oncology, West German Cancer Center, University of Duisburg-Essen, Essen, Germany 3 Bridge Institute of Experimental Tumor Therapy, West German Cancer Center, University Medicine Essen, Essen, Germany 4 Division of Solid Tumor Translational Oncology, German Cancer Research Center (DKFZ) and German Cancer Consortium (DKTK), partner site Essen, Heidelberg, Germany Radiol Oncol 2024; 58(4): 509-516. Received 2 August 2024 Accepted 9 October 2024 Correspondence to: Sebastian Zensen, Ph.D., M.D., Institute of Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Hufelandstraße 55, 45147 Essen, Germany. E-mail: sebastian.zensen@uk-essen.de Disclosure: No potential conflicts of interest were disclosed. This is an open access article distributed under the terms of the CC-BY license (https://creativecommons.org/licenses/by/4.0/). Background. In uveal melanoma patients, short-term evaluation of treatment response to hepatic artery infusion chemotherapy (HAIC) using the Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 criteria is challenging due to the diffuse metastatic spread. As liver enlargement can frequently be observed, this study aims to compare RECIST 1.1 and liver volumetry (LV) for the evaluation of HAIC treatment response. Patients and methods. Treatment response was evaluated in 143 patients (mean age 65.1 ± 10.9 years, 54% fe- male) treated by HAIC by RECIST 1.1 and LV on CT imaging performed before and after HAIC. In LV, different increases in liver volume were evaluated to set an effective threshold to distinguish between stable disease (SD) and progressive disease (PD). Overall survival (OS) was calculated as the time from first HAIC to patient death using Kaplan-Meier test and multivariate analysis was performed for RECIST 1.1 and LV. Results. In the overall population, median OS (mOS) was 13.5 months (95% CI 11.2–15.8 months). In LV, a threshold of 10% increase in liver volume was suited to identify patients with significantly reduced OS (SD: 103/143 patients, mOS 15.9 months; PD: 40/143 patients, 6.6 months; p < 0.001). Compared to RECIST 1.1, LV was the only significant prognos- tic factor that was able to identify a decreased OS. Conclusions. In uveal melanoma patients with liver metastases, LV with a threshold for liver volume increase of 10% was suitable to evaluate treatment response and would be able to be used as a valuable add-on or even alterna- tive to RECIST 1.1. Key words: uveal melanoma; computed tomography; liver volumetry; staging Introduction Uveal melanoma (UM) is the most frequent prima- ry malignancy of the eye and accounts for around 5% of all melanomas.1,2 Over the course of the dis- ease, 50% of all patients develop metastases, with the liver being the most common site in 70–90% of cases.3-5 If liver metastases occur, the prognosis Radiol Oncol 2024; 58(4): 509-516. Zensen S et al. / Liver volumetry in uveal melanoma liver metastases510 worsens considerately with a 1-year survival rate of about 13% and a median overall survival (mOS) of 2–5 months.6,7 Due to their diffuse and infiltra- tive growth pattern, liver metastases rapidly lead to fatal liver failure. Hence, even in the presence of extrahepatic metastases, aggressive local tu- mor treatment is key to improve survival.8 Due to diffuse metastatic spread, therapies targeting the whole organ such as transarterial chemoemboli- zation (TACE), radioembolization (RE) or hepatic arterial infusion chemotherapy (HAIC) are pos- sible treatment options.9,10 Here, especially HAIC plays an important treatment option due to its low rate of side effects that has been shown to prolong progression-free survival with less severe hemato- logic side effects.11 As liver metastases in UM patients often show rapid progression demanding immediate chang- es in the therapeutic regimen, short-term stag- ing is necessary to evaluate treatment response. However, established tumor response assess- ment of UM liver metastases using the Response Evaluation Criteria in Solid Tumors (RECIST) 1.1 and its derivatives is challenging. As it is dif- ficult to define a single lesion due to the diffuse liver involvement, a high interobserver measure- ment variability and thus an inconsistent assess- ment of response to treatment can be observed.12-14 However, we observed that considerate liver en- largement occurs in the later stages of the disease. Furthermore, liver volume would be a parameter that could be easily and (potentially automatically) monitored by liver volumetry (LV) over the course of the disease.15 To validate if changes in liver volume can also be observed in the earlier stages of the disease, the aim of this study is to compare RECIST 1.1 and LV for the evaluation of treatment response to HAIC in UM patients with liver metastases. Patients and methods Patient cohort In this retrospective observational study de- sign, all UM patients who underwent first HAIC for treatment of unresectable UM liver metas- tases in our department between October 2013 and December 2020 were identified using the Radiology Information System (RIS). Inclusion cri- teria were: 1) HAIC as only liver directed therapy of liver metastases; 2) no prior surgical therapy of liver metastases; 3) no additional interventions in addition to or during the first HAIC, such as coil embolization of hepatic arteries or use of degrada- ble starch microspheres (DSM); 4) abdominal CT imaging performed no more than 5 days before and at least 5 weeks after first HAIC but before second HAIC. Patients without CT imaging before or after first HAIC were excluded. Ethical approval for this retrospective single-center study was granted by the local ethics committee and the requirement to obtain informed consent was waived (19-8703-BO). Hepatic artery infusion chemotherapy HAIC was performed as described by our research group before via a transfemoral access.10,16 Then, a microcatheter was placed either into the proper hepatic artery or selectively into the left and the right hepatic artery and a starting dose of 40 mg melphalan was infused via an automated injector. In our department, HAIC was repeated every 6 to 8 weeks for local tumor control, as this time interval is considered safe and feasible based on pharma- cokinetic data from intravenous administration of the chemotherapeutic agent.17,18 Before each HAIC, a contrast-enhanced CT scan was performed to assess tumor response and intensify local tumor treatment in case of disease progression. Evaluation of treatment response by RECIST 1.1 and liver volumetry A CT scan was performed one to three days before the first HAIC. The next CT scan was performed 6–8 weeks after the first HAIC without intermedi- ate further local therapy of liver metastases, usu- ally on the day before the second HAIC. All CT scans were acquired in arterial phase of the liver and in venous phase of the whole abdomen. Then, CT images acquired before and after first HAIC were evaluated by LV and RECIST 1.1. using syn- go.via (Siemens Healthineers, Erlangen, Germany). LV was performed software-based manually in consensus by two radiologists blinded to outcomes using CT images of the venous phase. RECIST 1.1 evaluation was restricted to the liver. In accord- ance with RECIST 1.1, the maximum diameter of up to two lesions were analyzed. To correct for perfusion differences, we aimed to assess one le- sion in each liver lobe.19 To assess the impact of treatment induced changes detectable by RECIST 1.1 and liver volumetry, OS was calculated as the time from first HAIC to patient death. No separate analysis was performed for patients with extrahe- patic metastases, as their presence is known not to affect survival.8 Radiol Oncol 2024; 58(4): 509-516. Zensen S et al. / Liver volumetry in uveal melanoma liver metastases 511 Statistics and data analysis Statistical analysis was performed using GraphPad Prism 5.01 (GraphPad Software, San Diego, USA) and SPSS Statistics 28 (IBM, New York, USA). To determine normal distribution, D’Agostino- Pearson test was applied. Normally distributed data are reported as mean ± standard deviation (SD), non-normally distributed data as median and interquartile range (IQR). Wilcoxon matched pairs test was used to analyze target lesion param- eters and liver volumes. Interrater concordance of LV and RECIST 1.1 was assessed by Cohen’s κ-coefficient. Concordance was classified as pub- lished by Landis and Koch as no agreement (κ < 0), slight (κ:0.00–0.20), fair (κ:0.21–0.40), moderate (κ:0.41–0.60), substantial (κ:0.61–0.80) or almost perfect (κ:0.81–1.00) agreement.20 Overall survival between different groups of liver volume changes, RECIST 1.1 evaluation and combined assessment were compared using Kaplan-Meier curves and log-rank (Mantel-Cox) test. Cox proportional haz- ards regression model was used to determine haz- ard ratios (HRs) and the corresponding 95% con- fidence intervals (CI) of RECIST 1.1 and LV evalu- ation. A p-value lower than 0.05 was considered statistically significant. Results Patient cohort characteristics Between October 2013 and December 2020, 239 pa- tients underwent their first HAIC for the treatment of UM liver metastases, of which 96 patients did not meet the inclusion criteria and were therefore excluded. A total of 143 patients could be included in the analysis (Figure 1). Treatment-based exclusion criteria were: prior surgical therapy for liver metastases (22%, 21/96), additional coil embolization of hepatic arteries (18%, 17/96), first HAIC limited to one liver lobe (2%, 2/96) and additional use of degradable starch microspheres (DSM) (1%, 1/96). Imaging-based ex- clusion criteria were: no CT scan before or after first HAIC (38%, 36/96), no current CT scan prior to the intervention (17%, 16/96), no appropriate target lesion for RECIST 1.1 evaluation (2%, 2/96) and CT scan not evaluable due to accompanying liver he- matoma (1%, 1/96). Mean patient age at first HAIC was 65.1 years (SD 10.9, range 28–85) and 54% (77/143) of patients were female. A median number of five HAICs were performed (IQR 3–9, range 1–26). At the time point of data collection (December 2021), a total of 86% (123/143) were deceased, 9% (12/143) were alive and 6% (8/143) were lost to follow-up with a median follow-up time of 1.8 months (IQR 1.6–2.0). Median time period between CT scans before and after first HAIC was 48 days (IQR 44–53). mOS of all patients was 13.5 months (95% CI 11.2–15.8). Feasibility of liver volumetry for evaluation of treatment response In the entire study population, liver volume be- fore the first HAIC was 1735 ml (IQR 1431–2189 ml, range 889–7116 ml) and after the first HAIC was 1780 ml (IQR 1461–2329 ml, 827–7078 ml, p < 0.0001). The change in liver volume was a median increase of 4% (IQR -2.6% - +11.1%) ranging from a decrease of 20.4% to an increase of 37.6%. First, we performed an explorative data analysis to assess the impact of different changes in liver volume on overall survival using Kaplan-Meier curves (Figure 2). mOS was comparable for decreasing liver vol- ume (53/143 patients, mOS 15.9 months) and a small increase in liver volume up to 10% (50/143 patients, mOS 15.4 months, p = 0.7852, Figure 2). In con- trast, both an increase in liver volume of 10–20% (25/143 patients, mOS 7.9 months, 95% CI 3.6–12.2 months) and more than 20% (15/143 patients, mOS 5.7 months, 95% CI 4.8–6.6 months) were associ- ated with significantly decreased mOS compared to both decreasing or up to 10% increasing liver volume (p < 0.001, Figure 2, Table 1). Accordingly, FIGURE 1. Flowchart of analyzed study population with exclusion criteria. HAIC = hepatic arterial infusion chemotherapy Radiol Oncol 2024; 58(4): 509-516. Zensen S et al. / Liver volumetry in uveal melanoma liver metastases512 Treatment response evaluation by liver volumetry with a threshold of 10% increase in liver volume Liver volume measurements used to evaluate treatment response to HAIC according to LV with a threshold of 10% increase in liver volume are shown in Table 2. In LV, mOS was significantly shorter in patients with PD (6.6 months, 95% CI 4.4–8.8 months, 40/143 patients) than with SD (15.9 months, 95% CI 12.7–19.1 months, 103/143 patients, Chi-square = 39.28, p < 0.001) (Figure 3B). Initial liv- er volumes prior to the initial HAIC between pa- tients with PD (1903 ml, IQR 1481–2529 ml) and SD (1678 ml, IQR 1426–2176 ml) were not significantly different (p = 0.2007, Table 2). Combined treatment response evaluation by RECIST 1.1 and liver volumetry Two image examples of concordant evaluations by RECIST 1.1 and LV are shown in Figure 4. The agreement of LV with RECIST 1.1 was only considered as fair according to the inter-rater reli- ability analysis with about a quarter (35/143) of dis- cordant evaluations (κ=0.289, 95% CI: 0.118-0.461, Table 3). Therefore, we further compared the discord- ant evaluations with both RECIST 1.1 and LV ap- plied in combination. Here, in patients with SD ac- cording to RECIST 1.1, mOS was still significantly shorter if changes in LV were > 10% and there- fore considered as PD according to LV (RECIST 1.1 SD / LV PD: mOS 6.6 months, 27/143 patients) compared to LV SD (RECIST 1.1 SD / LV SD: mOS 16.6 months, 95/143 patients, Chi-square=28.45, p< 0.001) (Table 3, Figure 5). In contrast, for all cases with changes < 10% in LV, mOS was not significantly different regard- less the results of the RECIST 1.1 assessment: (RECIST 1.1 PD / LV SD: mOS 12.8 months, 8/143 patients, RECIST 1.1 SD / LV SD: mOS 16.6 months, 95/143 patients, Chi-square=1.84, p = 0.175, Table 3, FIGURE 2. Kaplan-Meier curves of overall survival differentiated by change in liver volume of patients with uveal melanoma with liver metastases after first hepatic artery infusion chemotherapy. TABLE 1. Comparison of median overall survival (mOS) differentiated by different changes in liver volume before and after first hepatic artery infusion chemotherapy Liver volume change Log-rank (Mantel-Cox) test Increase 0–10% Increase 10–20% Increase > 20% N mOS [months] Chi-square p Chi-square p Chi-square p Decrease 53 15.9 0.1 0.7852 14.2 0.0002 32.3 < 0.0001 Increase 0–10% 50 15.4 15.8 < 0.0001 33.4 < 0.0001 Increase 10–20% 25 7.9 1.9 0.162 Increase > 20% 15 5.7 an increase in liver volume of more than 10% was chosen as the threshold to classify patients as PD by LV, as this was associated with significantly decreased mOS. In contrast, a decrease of liver volume or an increase up to 10% was considered SD. For LV, no patient was evaluated as partial response if a RECIST 1.1 analogue threshold of a 30% decrease in volume was chosen. A complete response for LV is not applicable. Treatment response evaluation by RECIST 1.1 Measurements of liver target lesions used to eval- uate treatment response to HAIC according to RECIST 1.1 are shown in Table 2. In RECIST 1.1, mOS was significantly shorter in patients with PD (8.5 months, 95% CI 5.5–11.5 months, 22/143 pa- tients) than with SD (14.6 months, 95% CI 11.9–17.3 months, 121/143 patients, Chi-square = 9.302, p = 0.0023, Figure 3A). No patient was classified as complete response or partial response according to RECIST 1.1 criteria. Radiol Oncol 2024; 58(4): 509-516. Zensen S et al. / Liver volumetry in uveal melanoma liver metastases 513 Figure 5). Univariate Cox hazard regression analy- sis indicated that both RECIST 1.1 and LV showed high prognostic value, whereas in the subsequent multivariate analysis only LV remained an inde- pendent prognostic factor (Table 4). Discussion HAIC is an important and valuable palliative treatment option for liver metastases in patients with UM.1,18,21 Short-term assessment of treatment response can be difficult using the commonly used RECIST 1.1 criteria because lesion delinea- tion is challenging due to diffuse organ involve- ment, leading to increased, reader-dependent measurement variability and inconsistent treat- ment response evaluation.12-14 The results of our study can be subsumed in three key points. First, when selecting 10% increase in liver volume as the threshold for PD in LV, more patients with signifi- cantly lower OS are identified than by RECIST 1.1. Second, LV and RECIST 1.1 show only fair agree- ment in the evaluation of treatment response to HAIC. Third, even patients with RECIST 1.1 SD have significantly lower OS when an increase in liver volume of 10% or more is observed in LV. Tumors that involve the liver often show an asymmetrical and heterogeneous necrosis pattern, which complicates a precise evaluation of treat- ment response in follow-up imaging.12 Therefore, in patients with disseminated liver metastasis, as in UM, measurements of target lesions are often not reliable, making accurate assessment of treat- ment response difficult.22 However, growing liver metastases lead in parallel to an enlargement of A B TABLE 2. Results of evaluation of treatment response to hepatic artery infusion chemotherapy (HAIC) by RECIST 1.1 and liver volumetry Before first HAIC After first HAIC p-value RECIST 1.1a (Sum of) longest diameter(s) of target lesion(s) [mm] Total study cohort (n = 143) 48.6 (IQR 36.3–69.2) 50.8 (IQR 35.3–76.5) 0.0008 SD (n = 122) 47.7 (IQR 36.2–70.5) 48.0 (IQR 34.5–71.1) 0.3485 PD (n = 21) 55.2 (IQR 34.3–68.9) 73.0 (IQR 46.5–85.3) < 0.0001 Liver volumetry Total liver volume [ml] Total study cohort (n = 143) 1735 (IQR 1431–2189) 1780 (IQR 1461–2329) < 0.0001 SD (liver volume decreases or increases up to max. 10%) (n=103) 1678 (IQR 1426–2176) 1714 (IQR 1430–2151) 0.6691 PD (liver volume increases more than 10%) (n=40) 1903 (IQR 1481–2529) 2203 (IQR 1692–2946) <0.0001 a RECIST 1.1 criteria as published.19 Values are given as median and interquartile range (IQR). SD = stable disease; PD = progressive disease FIGURE 3. Overall survival of evaluation of treatment response by RECIST 1.1 (A) and liver volumetry with a threshold of 10% (B) of liver volume increase of uveal melanoma patients with liver metastases treated by hepatic artery infusion chemotherapy. Kaplan-Meier curves show overall survival separately for patients evaluated as stable disease (SD) and progressive disease (PD). In liver volumetry, patients with an increase in liver volume more than 10% were classified as PD and with decrease or increase below 10% as SD. RECIST 1.1 criteria as published.19 Radiol Oncol 2024; 58(4): 509-516. Zensen S et al. / Liver volumetry in uveal melanoma liver metastases514 the liver volume, which can be easily assessed by LV.22,23 Our results show that increases in liver volume of up to 10% are not associated with significant- ly reduced OS compared to decreasing liver vol- ume. However, a threshold of 10% liver volume increase in LV is well suited to identify patients with significantly reduced OS. Therefore, for the clinical application of LV to evaluate treatment re- sponse, we propose a volume increase of 10% as the threshold to distinguish between PD and SD in UM patients with liver metastases. For liver metastases in CRC, LV was also shown to be use- ful for evaluating treatment response, and the threshold for differentiating between SD and PD was 9.5% liver volume gain, which was very simi- lar to our finding.22 Our results show that LV as well as RECIST 1.1 are suitable for evaluation of treatment response but show only moderate inter- rater reliability with about a quarter of discordant cases. Here, LV can identify more patients than RECIST 1.1 whose life expectancy is significantly decreased. Our data show that even if patients are evaluated as PD by RECIST 1.1, they do not have a significantly decreased OS if their liver volume does not increase by more than 10%. However, if patients are considered SD by RECIST 1.1, but their liver volume increases by more than 10%, their OS is still significantly decreased. These findings are underlined by Cox regression analysis. Here, both LV and RECIST 1.1 have a high prognostic value in assessing treatment response after HAIC. However, after subsequential multivariate analy- sis, only LV remained an independent prognostic factor. Hence, LV might be a helpful tool to identify non-responders to HAIC that might profit from treatment escalation or potentially other treatment approaches such as RE, radiotherapy or surgery, which are established concepts in other hepatic malignancies apart from hepatocellular carcino- mas.9,24-26 Additionally, when local treatment op- tions are no longer feasible, systemic therapies such as immunotherapy or targeted therapies may offer further treatment possibilities. Furthermore, liver volumes between SD and PD evaluated patients were not significantly different before the first HAIC, so initial liver volume was TABLE 3. Median overall survival and accordance of treatment response evaluation by RECIST 1.1 and liver volumetry with a threshold of 10% increase in liver volume Liver volumetry SD (Liver volume decreases or increases up to max. 10%) PD (Liver volume increases more than 10%) Total RECIST 1.1 criteria SD 16.6 months (n = 95) 6.6 months (n = 27) 14.6 months (n = 122) PD 12.8 months (n = 8) 7.7 months (n = 13) 8.5 months (n = 21) Total 15.9 months (n = 103) 6.6 months (n = 40) 12.6 months (n = 143) Cohen’s κ = 0.289, 95% CI = 0.118–0.461 SD = stable disease; PD = progressive disease FIGURE 4. Image examples of CT examinations in portal venous phase before and after first hepatic artery infusion chemotherapy (HAIC) for two patients with evaluations according to RECIST 1.1 and liver volumetry (LV) as stable disease (SD) and progressive disease (PD). Liver metastases are marked with arrows. Radiol Oncol 2024; 58(4): 509-516. Zensen S et al. / Liver volumetry in uveal melanoma liver metastases 515 not a predictor of significant liver volume change in this study. Although RECIST 1.1 is the widely used and standardized method for assessing response to treatment in oncologic, LV not only offers addi- tional information but also has methodological advantages compared to RECIST 1.1: LV is a robust method that can be performed as part of the usual CT imaging performed for staging. As the whole organ is assessed, common problems in RECIST 1.1 evaluations leading to inaccurate therapy response evaluation such as varying contrast or poorly de- lineated lesions due to diffuse organ involvement as well as inter- and intrareader variability can be circumvented by LV.27-29 In addition, intrareader variability, which is a frequent problem in RECIST 1.1 measurements, might be reduced and thus im- prove patient response assessment.30 Here, espe- cially advancing developments in software and artificial intelligence might transform LV into an automatically acquired datapoint.31-33 This would allow LV to be easily included as an additional parameter in staging and clinical practice. Despite these promising initial results, these approaches are nevertheless so far experimental and are there- fore neither established in clinical routine nor ready for clinical use. The limitations of our study are its retrospective and single-center study design. Evaluation by LV and RECIST 1.1 was performed by the same radi- ologist for each of the examinations to avoid inter- observer variability. Therefore, these data should be confirmed in prospective studies once auto- mated software solutions for liver volumetry are commercially available. Furthermore, evaluation of treatment response was assessed only after the first HAIC, so follow-up studies should confirm applicability to later time periods in treatment and course of the disease. In conclusion, in UM patients with liver me- tastases, LV might be a suitable and in the future robust method to evaluate treatment response by a reliable identification of non-responders to HAIC and a consecutively shortened life expec- tancy. Hence, it can be used as a valuable add-on or even alternative to RECIST 1.1 to evaluate treat- ment response in this patient cohort. A threshold for liver volume increase of 10% was effective in distinguishing PD from SD in UM patients with liver metastases. FIGURE 5. Kaplan-Meier curves for the combined RECIST 1.1 and liver volumetry (LV) evaluation of treatment response with a threshold of 10% increase in liver volume in uveal melanoma patients with liver metastases treated by hepatic artery infusion chemotherapy. PD = progressive disease; SD = stable disease . TABLE 4. Univariate and multivariate Cox proportional hazards regression model of evaluation of treatment response to hepatic artery infusion chemotherapy by RECIST 1.1 and liver volumetry Analysis Univariate Multivariate Covariate Category n Median OS(95% CI) HR (95% CI) p HR (95% CI) p RECIST 1.1a SD 122 14.6(11.9–17.3) Reference Reference PD 21 8.5(5.5–11.5) 2.11 1.29–3.45 0.003 1.19 0.92–1.55 0.184 Liver volumetry SD (liver volume decreases or increases up to max. 10%) 103 15.9(12.7–19.1) Reference Reference PD (liver volume increases more than 10%) 40 6.6(4.4–8.8) 1.84 1.51–2.26 <0.001 1.77 1.43–2.19 < 0.001 a RECIST 1.1 criteria as published.19 HR = hazard ratio; OS = overall survival; PD = progressive disease; SD = stable disease Radiol Oncol 2024; 58(4): 509-516. Zensen S et al. / Liver volumetry in uveal melanoma liver metastases516 Acknowledgments We acknowledge support by the Open Access Publication Fund of the University of Duisburg- Essen. References 1 Rowcroft A, Loveday BPT, Thomson BNJ, Banting S, Knowles B. Systematic review of liver directed therapy for uveal melanoma hepatic metastases. HPB (Oxford) 2020; 22: 497-505. doi: 10.1016/j.hpb.2019.11.002 2 Singh AD, Turell ME, Topham AK. Uveal melanoma: trends in incidence, treatment, and survival. Ophthalmology 2011; 118: 1881-5. doi: 10.1016/j. ophtha.2011.01.040 3 Rietschel P, Panageas KS, Hanlon C, Patel A, Abramson DH, Chapman PB. Variates of survival in metastatic uveal melanoma. J Clin Oncol 2005; 23: 8076-80. doi: 10.1200/JCO.2005.02.6534 4 Damato BE, Coupland SE. Differences in uveal melanomas between men and women from the British Isles. Eye 2012; 26: 292-9. doi: 10.1038/ eye.2011.272 5 Diener-West M, Reynolds SM, Agugliaro DJ, Caldwell R, Cumming K, Earle JD, et al. Development of metastatic disease after enrolment in the COMS trials for treatment of choroidal melanoma: Collaborative Ocular Melanoma Study Group Report No. 26. Arch Ophthalmol 2005; 123: 1639-43. doi: 10.1001/archopht.123.12.1639 6 Gragoudas ES, Egan KM, Seddon JM, Glynn RJ, Walsh SM, Finn SM, et al. Survival of patents with metastases from uveal melanoma. Ophthalmology 1991; 98: 383-90. doi: 10.1016/s0161-6420(91)32285-1 7 Chew AL, Spilsbury K, Isaacs TW. Survival from uveal melanoma in Western Australia 1981-2005. Clin Exp Ophthalmol 2015; 43: 422-8. doi: 10.1111/ ceo.12490 8 Bedikian AY, Legha SS, Mavligit G, Carrasco CH, Khorana S, Plager C, et al. Treatment of uveal melanoma metastatic to the liver. A review of the M. D. Anderson cancer center experience and prognostic factors. Cancer 1995; 76: 1665-70. doi: 10.1002/1097-0142(19951101)76:9<1665::aid- cncr2820760925>3.0.co;2-j 9 Schelhorn J, Richly H, Ruhlmann M, Lauenstein TC, Theysohn JM. A single- center experience in radioembolization as salvage therapy of hepatic metas- tases of uveal melanoma. Acta Radiol Open 2015; 4: 2047981615570417. doi: 10.1177/2047981615570417 10 Heusner TA, Antoch G, Wittkowski-Sterczewski A, Ladd SC, Forsting M, Verhagen R, et al. Transarterial hepatic chemoperfusion of uveal melanoma metastases: survival and response to treatment. RoFo 2011; 183: 1151-60. doi: 10.1055/s-0031-1281743 11 Leyvraz S, Piperno-Neumann S, Suciu S, Baurain JF, Zdzienicki M, Testori A, et al. Hepatic intra-arterial versus intravenous fotemustine in patients with liver metastases from uveal melanoma (EORTC 18021): a multicentric randomized trial. Ann Oncol 2014; 25: 742-6. doi: 10.1093/annonc/mdt585 12 Duran R, Chapiro J, Frangakis C, Lin M, Schlachter TR, Schernthaner RE, et al. Uveal melanoma metastatic to the liver: the role of quantitative volumetric contrast-enhanced MR imaging in the assessment of early tumor response after transarterial chemoembolization. Transl Oncol 2014; 7: 447-55. doi: 10.1016/j.tranon.2014.05.004 13 Erasmus JJ, Gladish GW, Broemeling L, Sabloff BS, Truong MT, Herbst RS, et al. Interobserver and intraobserver variability in measurement of non-small- cell carcinoma lung lesions: implications for assessment of tumor response. J Clin Oncol 2003; 21: 2574-82. doi: 10.1200/JCO.2003.01.144 14 Thiesse P, Ollivier L, Di Stefano-Louineau D, Négrier S, Savary J, Pignard K, et al. Response rate accuracy in oncology trials: reasons for interobserver variability. Groupe Français d’Immunothérapie of the Fédération Nationale des Centres de Lutte Contre le Cancer. J Clin Oncol 1997; 15: 3507-14. doi: 10.1200/JCO.1997.15.12.3507 15 Koitka S, Gudlin P, Theysohn JM, Oezcelik A, Hoyer DP, Dayangac M, et al. Fully automated preoperative liver volumetry incorporating the anatomical location of the central hepatic vein. Sci Rep 2022; 12: 16479. doi: 10.1038/ s41598-022-20778-4 16 Zensen S, Opitz MK, Ludwig JM, Haubold J, Richly H, Siveke JT, et al. Radiation dose aspects of hepatic artery infusion chemotherapy in uveal melanoma patients with liver metastases. Cardiovasc Intervent Radiol 2022; 45: 841-5. doi: 10.1007/s00270-022-03130-1 17 Pinguet F, Culine S, Bressolle F, Astre C, Serre MP, Chevillard C, et al. A phase I and pharmacokinetic study of melphalan using a 24-hour continuous infu- sion in patients with advanced malignancies. Clin Cancer Res 2000; 6: 57-63. PMID: 10656432 18 Boone BA, Perkins S, Bandi R, Santos E, McCluskey K, Bartlett DL, et al. Hepatic artery infusion of melphalan in patients with liver metastases from ocular melanoma. J Surg Oncol 2018; 117: 940-6. doi: 10.1002/jso.24984 19 Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009; 45: 228-47. doi: 10.1016/j.ejca.2008.10.026 20 Landis JR, Koch GG. The measurement of observer agreement for categori- cal data. Biometrics 1977; 33: 159-74. PMID: 843571 21 Karydis I, Gangi A, Wheater MJ, Choi J, Wilson I, Thomas K, et al. Percutaneous hepatic perfusion with melphalan in uveal melanoma: a safe and effective treatment modality in an orphan disease. J Surg Oncol 2018; 117: 1170-8. doi: 10.1002/jso.24956 22 Kalkmann J, Forsting M, Stattaus J. Liver volume variations as a parameter to assess therapy response in advanced metastatic liver disease. Onkologie 2011; 34: 30-4. doi: 10.1159/000323373 23 Purkiss SF, Williams NS. Growth rate and percentage hepatic replacement of colorectal liver metastases. Br J Surg 1993; 80: 1036-8. doi: 10.1002/ bjs.1800800838 24 Opitz M, Zensen S, Ludwig JM, Weber M, Alatzides G, Seifert R, et al. Radiation dose aspects and establishment of diagnostic reference levels for90Y radioembolisation during angiographic procedure. J Radiol Prot 2022; 42(3). doi: 10.1088/1361-6498/ac8f9e 25 Schaarschmidt BM, Wildgruber M, Kloeckner R, Nie J, Steinle V, Braat AJAT, et al. 90Y Radioembolization in the treatment of neuroendocrine neo- plasms: results of an international multicenter retrospective study. J Nucl Med 2022; 63: 679-85. doi: 10.2967/jnumed.121.262561 26 Schaarschmidt BM, Kloeckner R, Dertnig T, Demircioglu A, Müller L, Auer TA, et al. Real-life experience in the treatment of intrahepatic cholangiocar- cinoma by 90Y radioembolization: a multicenter retrospective study. J Nucl Med 2023; 64: 529-35. doi: 10.2967/jnumed.122.264598 27 Krasovitsky M, Lee YC, Sim H-W, Chawla T, Moore H, Moses D, et al. Interobserver and intraobserver variability of RECIST assessment in ovarian cancer. Int J Gynecol Cancer 2022; 32: 656-61. doi: 10.1136/ ijgc-2021-003319 28 Suzuki C, Torkzad MR, Jacobsson H, Aström G, Sundin A, Hatschek T, et al. Interobserver and intraobserver variability in the response evaluation of cancer therapy according to RECIST and WHO-criteria. Acta Oncol 2010; 49: 509-14. doi: 10.3109/02841861003705794 29 Kuhl CK, Alparslan Y, Schmoee J, Sequeira B, Keulers A, Brümmendorf TH, et al. Validity of RECIST Version 1.1 for response assessment in metastatic cancer: a prospective, multireader study. Radiology 2019; 290: 349-56. doi: 10.1148/radiol.2018180648 30 Gafita A, Rauscher I, Fendler WP, Murthy V, Hui W, Armstrong WR, et al. Measuring response in metastatic castration-resistant prostate cancer using PSMA PET/CT: comparison of RECIST 1.1, aPCWG3, aPERCIST, PPP, and RECIP 1.0 criteria. Eur J Nucl Med Mol Imaging 2022; 49: 4271-81. doi: 10.1007/s00259-022-05882-x 31 Hagen F, Mair A, Bitzer M, Bösmüller H, Horger M. Fully automated whole- liver volume quantification on CT-image data: comparison with manual volumetry using enhanced and unenhanced images as well as two different radiation dose levels and two reconstruction kernels. PloS one 2021; 16: e0255374. doi: 10.1371/journal.pone.0255374 32 Cai W, He B, Fan Y, Fang C, Jia F. Comparison of liver volumetry on contrast- enhanced CT images: one semiautomatic and two automatic approaches. J Appl Clin Med Phys 2016; 17: 118-27. doi: 10.1120/jacmp.v17i6.6485 33 Suzuki K, Epstein ML, Kohlbrenner R, Garg S, Hori M, Oto A, et al. Quantitative radiology: automated CT liver volumetry compared with inter- active volumetry and manual volumetry. AJR Am J Roentgenol 2011; 197: W706-12. doi: 10.2214/AJR.10.5958