Abstract B74: Identification of prognostic molecular subtypes of ovarian serous cystadenocarcinoma by isoform-level gene expression analysis

Abstract Characterization of molecular aberrations in ovarian serous cystadenocarcinoma (OVSC) is highly crucial for developing precision medicine approaches. While gene-level analysis of gene expression profiles of The Cancer Genome Atlas (TCGA) OVSC samples revealed four molecular subgroups (differentiated, proliferative, mesenchymal, and immunoreactive), the association of the derived subtypes with the clinical and pathologic features, such as stage, cytoreduction, and chemotherapy, is not statistically significant. Considering the fact that the majority of human genes produce multiple transcript-variants that could be involved in different functional pathways, we reanalyzed the TCGA exon-array expression profiles at the transcript variant or isoform level by following the novel informatics methodology we developed and successfully applied for glioblastoma (Pal et al., Nucleic Acids Res 2014). Non-negative matrix factorization (NMF) clustering of TCGA ovarian serous adenocarcinoma samples, based on isoform-level gene-expression profiles, recaptured the four known molecular subgroups but switched the subtype for 26% of the samples, resulting in significant (p=0.009) survival differences among the refined subgroups. The survival analysis was further extended to multivariate Cox proportional hazard model considering the different subgroups, and the clinical parameters as the predictor variables. Patients classified under mesenchymal group (HR = 1.44; 95% CI, 0.99-2.08; P < 0.05) and in stage III (HR = 2.39, 95%of CI, 1.25-4.56, P < 0.05), IV (HR= 3.16; 95% of CI, 1.58-6.32, P < 0.05) were identified to be at high risk or decreased survival, while those receiving intraperitoneal (IP) chemotherapy (HR, 0.63; 95% of CI, 0.38-1.03, P < 0.1) were associated with longer survival. We are in the process of translating the isoform-level gene signature into clinically adaptable molecular subtyping assay and validating on an independent patient cohort, with potential implications for developing robust diagnostic assays for OVSC patient stratification. Citation Format: Arunima Shilpi, Brandon L. Seagle, Manoj Kandpal, Shohreh Shahabi, Ramana V. Davuluri. Identification of prognostic molecular subtypes of ovarian serous cystadenocarcinoma by isoform-level gene expression analysis. [abstract]. In: Proceedings of the AACR Conference: Addressing Critical Questions in Ovarian Cancer Research and Treatment; Oct 1-4, 2017; Pittsburgh, PA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(15_Suppl):Abstract nr B74.