Prognostic Implications of MRI Melanin Quantification and Cytogenetic Abnormalities in Liver Metastases of Uveal Melanoma

Simple Summary Melanin content in uveal melanoma is suspected to influence tumoral microenvironment and antitumoral response and is related to higher risk of metastasis and death in the primary disease. However, the prognostic impact of melanin content in liver metastases of uveal melanoma (LMUM) remains unexplored. The aim of our retrospective study was to evaluate the prognostic implications of melanin quantification assessed by MRI with clinical, pathological, and genetic features of LMUM. We found in a population of 63 patients eligible for margin-free resection of LMUM that MRI melanin quantification was an independent prognostic factor associated with overall survival. Liver metastases with high MRI melanin content and high genetic risk "M3/8g" were associated with lower overall survival compared with that of liver metastases with low MRI melanin content and/or low/intermediate genetic risk. The level of pigmentation in "M3/8g" LMUM identified two subsets that were correlated with distinct clinical outcomes. To evaluate the prognostic implications of melanin quantification assessed by magnetic resonance imaging (MRI) with respect to the clinical, pathological, and genetic features of liver metastases of uveal melanoma (LMUM). This single-center retrospective cohort study included 63 patients eligible for margin-free resection of LMUM between 2007 and 2018. Comparative genomic hybridization of resected liver metastases on microarrays was performed for genetic risk classification. Metastases exhibiting monosomy 3 with any type of gain of chromosome 8 (M3/8g) were considered high-genetic-risk. MRI melanin quantification using the mean T1 signal (mT1s) in liver metastases was assessed quantitatively on preoperative imaging examination and compared to that of gross pathological evaluation. The association between MRI melanin quantification and overall survival (OS) was assessed by multivariate analysis using the Cox proportional hazards model. Gross pathological assessment of melanin content and MRI melanin quantification were strongly correlated (r = 0.8, p < 0.001). Independent prognostic factors associated with OS were disease-free interval <= 24 months (HR = 3.1; 95% CI, 1.6-6.0; p < 0.001), high-genetic-risk (HR = 2.2; 95% CI, 1.1-4.8; p = 0.04), mT1s > 1.1 (HR = 2.3; 95% CI, 1.2-4.7; p = 0.019), and complete hepatic resection (HR = 0.3; 95% CI, 0.2-0.7; p = 0.004). In patients with high-genetic-risk, mT1s showed a significant association with OS (HR = 3.7; 95% CI, 1.5-9.3; p = 0.006). The median OS was 17.5 months vs. 27 months for >1.1 and <= 1.1 mT1s tumors, respectively (p = 0.003). We showed that the level of pigmentation in M3/8g LMUM identified two subsets that were correlated with distinct clinical outcomes.