Abstract PR07: Transcriptome meta-analysis identifies new 5-gene classifier for early detection of pancreatic cancer

Abstract

Abstract Background: Pancreatic ductal adenocarcinoma (PDAC) is a truly devastating disease, primarily due to late diagnosis when curative resection is no longer possible and the lack of accurate early detection biomarkers. Therefore, increasing PDAC survival depends heavily on improving the accuracy and effectiveness of early PDAC detection. Methods:Key PDAC biomarkers were identified through an optimized meta-analysis strategy of raw data in 4 PDAC transcriptome datasets (training sets). Application of identical normalization and statistical testing procedures for each dataset generated a list of differentially expressed genes by empirical Bayes moderated t-statistic. The top ranking PDAC-specific genes from this list that were consistently differentially expressed in these training sets were used for developing a classifier for PDAC. The biomarker panel with the highest sensitivity and specificity in the training sets was validated for diagnostic performance in nine independent validation datasets. A subset of genes in the panel were evaluated by immunohistochemistry (IHC) in human PDAC tissue sections. Cell based assays assessed the effect of two genes on PDAC migration and invasion. Results and Conclusions: A 5-gene classifier panel was identified that disntinguished between PDAC and normal pancreas with high sensitivity, specificity and accuracy in the four training sets. Validation of this 5-gene predictor in 9 independent test sets confirmed the diagnostic accuracy to not only discriminate accurately between PDAC and healthy controls, but also between pancreatitis and PDAC. IHC analysis of two of the proteins encoded by the panel confirmed enhanced expression in PDAC cells. Cell-based experiments demonstrated that both proteins enhance migration and invasion of PDAC cells as well as soft agar growth, suggesting a causal link to PDAC development and progression. The immediate clinical utility upon further prospective validation of this PDAC biomarker panel will be accurate PDAC diagnosis in patients with pancreatic abnormalities on imaging, risk stratification and evaluation of the risk of malignancy in pancreatic cysts. This abstract is also presented as Poster B26. Citation Format: Manoj Bhasin, Octavian Bucur, Ken Ndebele, Jessica Plati, Andrea Bullock, Xuesong Gu, Eduardo Castan, Robert Najarian, Jen Jen Yeh, Channing Der, Jon Cody Haines, Karl Ruping, Rebecca Miksad, Roya Khosravi-Far, Towia Aron Libermann. Transcriptome meta-analysis identifies new 5-gene classifier for early detection of pancreatic cancer. [abstract]. In: Proceedings of the AACR Special Conference on Pancreatic Cancer: Innovations in Research and Treatment; May 18-21, 2014; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2015;75(13 Suppl):Abstract nr PR07.