Hepatocellular Carcinoma Epigenetic Patterns Correspond to Differences in Ethnoracial Status and Treatment Response in a Single-Center Retrospective Study.

ABSTRACT

PURPOSE: To correlate epigenetic patterns with ethnoracial status and locoregional therapy (LRT) response in patients with hepatocellular carcinoma (HCC). MATERIALS AND METHODS: DNA and RNA were extracted from 47 distinct formalin-fixed paraffin-embedded tumor samples from 42 patients with HCC (n = 14 Black, n = 19 White, n = 9 Hispanic). LRT response was determined using computed tomography (CT) or magnetic resonance (MR) imaging 3 months posttreatment of 35 tumors (n = 22 complete response, n = 13 retreatment candidates). RNA expression and DNA methylation were used to stratify patients by ethnoracial status and treatment response using partial least-squares discriminant analysis (PLS-DA). Results were validated using hierarchical clustering. Ingenuity pathway analysis was performed to identify upstream regulators and pathways. RESULTS: PLS-DA identified 100 genes and 12 methylated regions that differentiated tumors from Black from White/Hispanic patients. Hierarchical clustering clustered samples with the top 16 genes or the top 5 methylation regions. Dysregulated pathways included adrenomedullin pathway (P = .030), EIF2 signaling (P = .007), and several metabolic pathways. AGTR1 (log2fold = 1.59) and GSTM3 (log2fold = 2.53) represented potential differentially expressed therapeutic targets. PLS-DA identified 100 genes and 150 methylation regions that differentiated between complete responders and retreatment candidates. Hierarchical clustering clustered samples with the top 30 genes or the top 13 methylation regions. Dysregulated pathways included metabolic and DNA repair-related pathways. ASAP2 (log2fold = 0.29) and RAD50 (log2fold = 0.22) represented potential differentially expressed therapeutic targets. CONCLUSIONS: Variation in gene expression and DNA methylation patterns in patients with HCC corresponded to ethnoracial status and LRT response. These initial results suggest tumor profiling has the potential to close ethnoracial disparities and improve treatment stratification.