Abstract 2383: Multiregional RNA sequencing identifies intratumor transcriptomic heterogeneity in a subset of early-stage hepatocellular carcinoma

BACKGROUND: Hepatocellular carcinoma (HCC) is known to have histological intra-tumor heterogeneity. Little is known about the underlying molecular contributions to this observation, or its potential impact on resistance to therapies. AIMS: 1) To evaluate intra-tumoral molecular heterogeneity of primary HCCs using multi-regional RNA-seq; 2) to correlate molecular alterations with histological features at distinct areas. METHODS: We analyzed 55 fresh-frozen tissues from 10 patients with BCLC-A HCC treated with surgical resection. Multi-regional sampling included 38 HCCs and 16 adjacent non-tumoral tissues (average 5.4 samples per patient). We performed H&E staining and RNA-seq on all samples. Each sequenced section was evaluated for a panel of 11 histological features. Data analyses included clustering (consensus and multidimensional scaling (MDS)), sample ordering via linear graph modeling of expression, pathogen quantitation, gene set enrichment analysis (GSEA), and nearest template method (NTP) for gene signature prediction. RESULTS: Patients enrolled were mostly males (70%), with an average age of 59, all with single-nodule HCC, and a median tumor size of 5.3 cm (range 3-16 cm). Background liver disease was HBV in 50% of cases without histological cirrhosis in 80%. MDS and consensus clustering identified heterogeneity in 40% (4/10) of patients as defined by at least one sample not clustering with their counterparts. Sample ordering uncovered dominant tumor heterogeneity evolution paths, including HBV specific ones. GSEA revealed differential enrichment (FDR<0.05) of Cancer Hallmark Gene Sets in the 4 heterogeneous tumors. Indeed, these 4 tumors showed concordant gene set enrichment (FDR<.05) in all regions only 19% of the time compared to 83% in homogenous tumors on average. NTP further confirmed heterogeneity in molecular classification predictions in 2/4 heterogeneous tumors (e.g., Proliferation class). There were no significant differences in size or regions sequenced between heterogeneous and homogenous tumors. Molecular outliers show distinct histological features such as high mitotic rate and poorer differentiation. CONCLUSION: Transcriptomic intra-tumor heterogeneity is frequent in a subset of surgically resected HCC, and it correlates with histological features. Integrative analysis of CNVs, expressed mutations, allele specific expression, and gene fusions is ongoing. Citation Format: Amanda Craig, Mehmet E. Ahsen, Ismail Labgaa, Ashley Stueck, Delia D’Avola, Stephen C. Ward, Maria Isabel Fiel, Ganesh Gunasekaran, Josep Llovet, Swan Thung, Myron Schwartz, Bojan Losic, Gustavo Stolovitzky, Augusto Villanueva. Multiregional RNA sequencing identifies intratumor transcriptomic heterogeneity in a subset of early-stage hepatocellular carcinoma. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2383.