Metabolic Fingerprints of High-Risk Individuals for Pancreatic Cancer: A Multibiospecimen Investigation

Abstract

Metabolomics analyses of human samples have a high potential to identify new metabolic markers that are important for disease development and can be utilized for disease diagnostics. To identify these disease-specific metabolic variations that represent medically important signatures, these metabolites have to be distinguished from common metabolic changes at a baseline level. Herein, we report a comprehensive metabolomics analysis of human fecal, plasma, and urine samples of high-risk patients for the development of pancreatic cancer to explore early metabolic alterations associated with the presence of pancreatic cysts. This study provides detailed metabolic insights across multiple biospecimens using a combination of multimatrix sampling and cofactor-adjusted statistical analysis. Importantly, we have identified several potential candidate markers, including methylglutaric acid (area under the curve (AUC) = 0.88), hydroxyindoleacetic acid (AUC = 0.82), and hypoxanthine (AUC = 0.82) associated with the presence of pancreatic cysts in an exploratory receiver operating characteristics analysis. Furthermore, bile acids, alkaloids, and fatty acids were identified to be altered in the individuals with cyst formation. We also identified a series of metabolites relevant for age- and sex-dependent differences. While validation in larger, longitudinal cohorts remains necessary, our findings lay a strong foundation for future investigations into metabolic markers of pancreatic pathology.