Pancreatic ductal adenocarcinoma and chronic pancreatitis may be diagnosed by exhaled-breath profiles: a multicenter pilot study.

Abstract

BackgroundThe diagnosis of pancreatic adenocarcinoma and chronic pancreatitis often rely on expensive and invasive diagnostic approaches, which are not always discriminative since patients with chronic pancreatitis and pancreatic adenocarcinoma may present with similar symptoms. Volatile organic compounds (VOCs) in expired breath, could be used as a non-invasive diagnostic biological marker for detection of pancreatic pathology. Detection and discrimination of pancreatic pathology with an electronic nose has not yet been reported.PurposeThe objective of this pilot study was to determine the diagnostic potential of an electronic nose to identify pancreatic adenocarcinoma and chronic pancreatitis by analyzing volatile organic compoundg (VOC) profiles in exhaled air.Patients and methodsIn a multicenter study, the exhaled air of 56 chronic pancreatitis patients, 29 pancreatic adenocarcinoma patients, and 74 disease controls were analyzed using an electronic nose based on 3 metal oxide sensors (MOS). The measurements were evaluated utilizing an artificial neural network.ResultsVOC profiles of chronic pancreatitis patients could be discriminated from disease controls with an accuracy of 0.87 (AUC 0.95, sensitivity 80%, specificity 92%). Also, VOC profiles of patients with pancreatic adenocarcinoma differed from disease controls with an accuracy of 0.83 (AUC 0.87, sensitivity 83%, specificity 82%). Discrimination between chronic pancreatitis and pancreatic adenocarcinoma showed an accuracy of 0.75 (AUC 0.83, sensitivity 83%, specificity 71%).ConclusionAn electronic nose may be a valuable diagnostic tool in diagnosis of pancreatic adenocarcinoma and chronic pancreatitis. The current study shows the potential of an electronic nose for discriminating between chronic pancreatitis, pancreatic adenocarcinoma and healthy controls. The results from this proof-of-concept study warrant external validation in larger cohorts.