Abstract 762: A non-invasive miRNA signature for the identification of pancreatic ductal adenocarcinoma (PDAC) patients with poor molecular subtypes and worse prognosis

Abstract

Abstract Purpose: Pancreatic ductal adenocarcinoma (PDAC) is a highly aggressive disease with dismal prognosis. Recent advances in treatment paradigms have been encouraging; however, identification of appropriate patient populations and a better understanding of the underlying tumor heterogeneity could further improve patient outcomes. Recently, four distinct molecular subtypes of PDAC were established, and in particular, squamous subtype was attributed with worse prognosis. Nonetheless, the clinical application of genome-wide transcriptome-based subtyping is cumbersome. Therefore, in this study, we identified miRNA networks regulating the expression of various genes in each PDAC subtype, with a goal to develop tissue and liquid biopsy-based miRNA biomarkers that can be translated into the clinic. Experimental Design: We used a mRNA-miRNA network-based approach to unravel miRNA networks regulating various PDAC subtypes. Using logistic regression models and area under the curve (AUC) analysis, we evaluated the performance of a miRNA-signature for identifying poor molecular subtypes of PDAC. Using qRT-PCR assays, we also analyzed the prognostic significance of the miRNA panel in 433 PDAC patients, including TCGA (n=183), multiple in-house retrospective, tissue cohorts (n=199) and a prospective, preoperative serum cohort (n=51). Univariate and multivariate Cox-proportional hazard models were used for data analysis, and Youden’s index derived cut-off thresholds were used to plot the Kaplan Meier (KM) curves for overall survival. Results: We identified a panel of 9-miRNAs that were significantly dysregulated (upregulated: miR-205-5p and -934; downregulated: miR-192-5p, 194-5p, 194-3p, 215-5p, 375-3p, 552-3p and 1251-5p) in squamous subtype; and was highly accurate in identifying all poor molecular PDAC subtypes proposed previously (Squamous AUC=0.90, Basal AUC=0.89 and Quasimesenchymal AUC=0.83). Intriguingly, expression level of all these 9-miRNAs clearly stratified overall survival (OS) in patient subgroups, and patients with high-risk scores exhibited significantly worse OS (Hazard ratio (HR): 2.48, p<0.0001). Intriguingly, our miRNA signature was equally robust even in pre-operative serum specimens for identifying patients with poor OS (HR: 2.85, p=0.02), highlighting its significance for treatment decision-making in the clinic. Conclusions: We for the first time report a unique miRNA signature for identifying patients with poor molecular subtype PDAC who are at risk for worse overall survival. Our 9-miRNA signature can be easily adapted into the clinical setting using simple qRT-PCR based assays. The discrimination of high-risk patients even in the preoperative setting could improve the selection of PDAC patients for neoadjuvant treatment, especially in those with unresectable disease. Citation Format: Raju Kandimalla, Tadanobu Shimura, Saurav Mallik, Susan Tsai, Douglas B. Evans, Hideo Baba, Yasuhiro Kodera, Xi Chen, Ajay Goel. A non-invasive miRNA signature for the identification of pancreatic ductal adenocarcinoma (PDAC) patients with poor molecular subtypes and worse prognosis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 762.