Genomic and transcriptomic landscape of oral pre-cancers (OPCs) and risk of oral cancer (OC)

6009 Background: The molecular landscape of OPCs and its association with neoplastic progression is largely unknown. We report the results of high throughput DNA/RNA profiling of OPCs from pts in the Erlotinib Prevention of Oral Cancer trial (EPOC), with long-term prospective follow-up. Methods: We performed next generation sequencing of 201 cancer genes (MD Anderson T200 platform) in 170 OPCs from EPOC, and RNA profiling using HTG EdgeSeq Oncology Biomarker Panel containing 2,560 transcripts in a subset of 141 OPCs. 73 pts developed invasive OC during a median follow up of 7.3 years, from whom 23 paired OCs were profiled to characterize the evolutionary trajectory from OPCs to OCs. OPC molecular features were correlated with OC-free survival. Results were compared with TCGA invasive OC DNA/RNA profiles and an independent set of 86 OPCs with RNA data. Results: Similar to TCGA, C > T was the predominant substitution. The top mutated genes in OPCs were TP53 (29%), CDKN2A (15%), NOTCH1 (11%) and PIK3CA (7%), which were also frequently mutated (albeit at higher rates) in OCs from EPOC or TCGA. There was a progressive increase of tumor mutation burden (TMB, P < 0.05) and frequency of high-risk TP53 mutations (P = 0.02) from hyperplasia, to dysplasia, to invasive OCs (P < 0.05). Median TMB was higher in OPCs from pts who developed OC (2.45 mut/Mb) vs those who did not (1.22 mut/Mb) (P < 0.01). Pts with TP53 mutated OPCs had shorter OC-free survival compared to TP53 wild-type (HR 1.81, 95% CI 1.13-2.90, P = 0.01). A prognostic score was derived from a Cox regression model which identified 12 mRNA transcripts associated with OC risk (HR 4.72, 95% CI 2.51-8.86, P < 0.01), and which was validated in the independent set of 86 OPCs (HR 2.68, P < 0.01). This score was also associated with shorter overall survival when applied to invasive OCs from TCGA pts (HR 2.72, P < 0.01). Conclusions: This is the first large-scale cohort of OPC pts with long-term, prospective follow up and comprehensive RNA/DNA profiling. We demonstrated an association between TMB, TP53 mutations, a 12-gene RNA signature score in OPCs, and OC risk. This study may provide a framework for similar efforts of pre-cancer molecular profiling in the oral cavity and other sites, such as the PreCancer Atlas of the NCI.