GC-MS based metabolomics strategy to distinguish three types of acute pancreatitis

Abstract

Acute pancreatitis (AP) is a progressive systemic inflammatory response with high morbidity and high mortality, which is mainly caused by alcohol, bulimia, gallstones and hyperlipidemia. The early diagnosis of different types of AP and further explore potential pathophysiological mechanism of each type of AP is beneficial for optimized treatment strategies and better patient's care. In this study, a metabolomics approach based on gas chromatography-mass spectrometry (GC-MS), and random forests algorithm was established to distinguish biliary acute pancreatitis (BAP), Hyperlipidemia acute pancreatitis (HLAP), and alcoholic acute pancreatitis (AAP), from healthy controls. The classification accuracies for BAP, HLAP, and AAP patients compared with healthy control, were 0.886, 0.906 and 0.857, respectively, by using 5-fold cross-validation method. And some special metabolites for each type of AP were discovered, such as l-Lactic acid, (R)-3-Hydroxybutyric acid, Phosphoric acid, Glycine, Erythronic acid, l-Phenylalanine, d-Galactose, l-Tyrosine, Arachidonic acid, Glycerol 1-hexadecanoate. Furthermore, associations between these metabolites with the metabolism of amino acids, fatty acids were identified. Our studies have illuminated the biomarkers and physiological mechanism of disease in a clinical setting, which suggested that metabolomics is a valuable tool for identifying the molecular mechanisms that are involved in the etiology of BAP, AAP, HLAP and thus novel therapeutic targets.

Keywords: Alcoholic acute pancreatitis; Biliary acute pancreatitis; Biomarker discovery; Hyperlipidemia acute pancreatitis; Metabolomics; Random forest.