Digital PCR-based plasma cell-free DNA mutation analysis for early-stage pancreatic tumor diagnosis and surveillance.

Abstract

Cell-free DNA (cfDNA) shed from tumors into the circulation offers a tool for cancer detection. Here, we evaluated the feasibility of cfDNA measurement and utility of digital PCR (dPCR)-based assays, which reduce subsampling error, for diagnosing pancreatic ductal adenocarcinoma (PDA) and surveillance of intraductal papillary mucinous neoplasm (IPMN). We collected plasma from seven institutions for cfDNA measurements. Hot-spot mutations in KRAS and GNAS in the cfDNA from patients with PDA (n = 96), undergoing surveillance for IPMN (n = 112), and normal controls (n = 76) were evaluated using pre-amplification dPCR. Upon Qubit measurement and copy number assessment of hemoglobin-subunit (HBB) and mitochondrially encoded NADH:ubiquinone oxidoreductase core subunit 1 (MT-ND1) in plasma cfDNA, HBB offered the best resolution between patients with PDA relative to healthy subjects [area under the curve (AUC) 0.862], whereas MT-ND1 revealed significant differences between IPMN and controls (AUC 0.851). DPCR utilizing pre-amplification cfDNA afforded accurate tumor-derived mutant KRAS detection in plasma in resectable PDA (AUC 0.861-0.876) and improved post-resection recurrence prediction [hazard ratio (HR) 3.179, 95% confidence interval (CI) 1.025-9.859] over that for the marker CA19-9 (HR 1.464; 95% CI 0.674-3.181). Capturing KRAS and GNAS could also provide genetic evidence in patients with IPMN-associated PDA and undergoing pancreatic surveillance. Plasma cfDNA quantification by distinct measurements is useful to predict tumor burden. Through appropriate methods, dPCR-mediated mutation detection in patients with localized PDA and IPMN likely to progress to invasive carcinoma is feasible and complements conventional biomarkers.