Unique molecular underpinnings by genetic ancestry in localized clear cell renal cell carcinoma.

Abstract

4536 Background: Clinical studies evaluating the impact of race on outcomes in renal cell carcinoma (RCC) are inconsistent and conceivably influenced by both biological and socioeconomic factors. Self-reported African Americans (AA) are also largely underrepresented in clinical trials and molecular studies of clear cell (cc) RCC compared to their White (W) counterparts. Herein, we sought to comprehensively characterize molecular differences by genetic ancestry in localized ccRCC. Methods: We performed whole exome (WES) and transcriptome (RNAseq) sequencing on a case-matched cohort of AA and W patients (pts) who underwent nephrectomy for localized ccRCC at our institution. Germline single nucleotide polymorphisms were combined with reference panels from the 1000 Genomes Project, and ancestry was estimated using ADMIXTURE. Molecular subtypes were defined according to the IMmotion151 (IMm151) subtypes, and a random forest model was trained to predict subgroups. Results: From an initial selected cohort of 84 pts, WES and RNAseq data of adequate quality were available for 61 pts including 28 of African (AFR) and 33 of European (EUR) descent. We augmented our institutional cohort with a 3:1 EUR:AFR matched subset of the TCGA KIRC by stage, nuclear grade (NG), and gender. The final analytic cohort contained 75 pts of AFR and 177 pts of EUR descent. There were no significant differences in stage, NG, or gender between AFR and EUR groups. We observed that VHL [OR 0.22, CI [0.11-0.43], p <0.001) and PBRM1 (OR 0.4, CI [0.19-0.79], p = 0.005) mutations are enriched in pts with EUR ancestry and identified a unique association of KMT2C somatic mutations (OR 3.7, CI [1.1-13.5], p = 0.027) with AFR ancestry. Consistent with a lower frequency of VHL mutations, gene-set variation analysis revealed lower enrichment for angiogenesis (-0.23 vs 0.045; p < 0.001) in AFR vs EUR ancestry, respectively. AFR ancestry was associated with an increased proportion of the “Proliferative” IMm151 subtype (14.9% [n = 11/74] vs 3.4% [n = 6/177], p = 0.002). There was no association between overall survival (OS) and ancestry; however, IMm151 subtypes stratified OS with longest survival in the “Angiogenic/Stromal” group (median OS NR) and shortest OS in the T-Effector/Proliferative subtype (5.5yrs CI [2.3 - NR], p = 0.01). Conclusions: Our findings indicate that genetic ancestry is associated with distinct molecular subtypes of ccRCC. Biological drivers of ccRCC pathogenesis and therapeutic response may vary by ancestry, underscoring the need for racially diverse representation in future studies.