Background: Largely due to late diagnosis when curative resection is no longer possible, Pancreatic Adenocarcinoma (PDAC) is a devastating disease marked by a mortality rate nearly equivalent to its incidence rate. While serum CA19-9 is associated with PDAC, it lacks the sensitivity and specificity needed for early and accurate diagnosis. Thus, early diagnosis of PDAC is critical to improving survival for PDAC patients and for accurate PDAC risk stratification.

Methods: We have identified key PDAC biomarkers through an optimized meta-analysis strategy of raw data (training sets). From this analysis, the top ranking genes that distinguish PDAC from normal pancreas were selected through statistical significance comparison of the consistency in differential PDAC gene expression across studies. The biomarker panel with the highest sensitivity and specificity in the training sets was chosen for assessment of predictive power in nine independent validation datasets. For a subset of genes in the panel, immunohistochemistry (IHC) analysis was performed in human PDAC and pancreatitis to validate our panel.

Results and Conclusions: We have identified a 5-gene classifier panel with specificity of 88.9% and sensitivity of 100% with an overall accuracy of 93.3% and AUC of 0.94 for discriminating PDAC from non-malignant (normal or pancreatitis) tissue. IHC analysis of two of the proteins 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 indicating that they may be causative in PDAC. The 5-gene predictor discriminated PDAC from healthy controls, as well as pancreatitis, with highest accuracy in nine independent validation sets. Our identified PDAC panel may facilitate early diagnosis and improve risk-stratification.

Citation Format: Manoj Bhasin, Andrea Bullock, Xuesong Gu, Octavian Bucur, Robert Najarian, Jon Cody Haines, Karl Ruping, Rebecca Miksad, Towia Libermann, Roya Khosravi-Far. Early detection of pancreatic cancer using a new 5-gene classifier. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr LB-173. doi:10.1158/1538-7445.AM2014-LB-173