Pancreatic cancer biomarkers among exosome-fractionated circulating miRNAs.

e15671 Background: Pancreatic cancer (PC) is one of the most deadly cancers. Existing tumor markers, such as carbohydrate antigen 19-9 and carcinoembryonic antigen, have relatively poor diagnostic sensitivity and specificity. Intraductal papillary mucinous neoplasm (IPMN) is related to pancreatic carcinogenesis and is treated as a precursor of PC. We analyzed the miRNA expression profiles of exosome-rich fractionated blood of IPMN, PC and control patients to identify biomarkers that would facilitate early diagnosis of PC. Methods: RNA from the exosome-rich fraction was extracted from eight PCs, eight IPMNs and eight non-pancreatic disease controls (NC). Extracted RNA was subjected to miRNA profiling using an Agilent miRNA microarray and Illumina next-generation sequencer (NGS). Results: IPMN and PC patients were discriminated from controls with 91.6% and 95.8% accuracy, respectively, using the expression profiles of 11 and 19 miRNAs determined by microarray and NGS analyses, respectively. Using NGS, 168 novel miRNA candidates were identified. Six candidate miRNAs were detected in patients with PC but not in those with IPMN or NC. One candidate miRNA was detected among all patients with PC but not those with IPMN. Conclusions: The results of our analysis of miRNAs associated with pancreatic disease were highly reproducible. We could identify pancreatic disease with a high probability using the expression profiles of existing miRNAs and novel candidate miRNAs. Therefore, miRNA expression profiles show promise for the early diagnosis of PC at the precancerous stage. (UMIN No.: 000017958)