Metabolic signatures for gastric cancer diagnosis and mechanistic insights: a multicenter study.

ABSTRACT: Early detection of gastric cancer (GC) is critical for improving prognosis, yet conventional biomarkers lack sensitivity and specificity, necessitating non-invasive, high-performance diagnostic tools. This study integrated untargeted metabolomics and machine learning to develop a plasma metabolite panel for GC diagnosis and mechanistic insights. Plasma and tissue samples from two cohorts (n = 597) were analyzed using ultra-performance liquid chromatography-mass spectrometry (UPLC-MS). A six-metabolite panel was identified and validated, demonstrating excellent diagnostic performance (area under the curve: 0.947-0.982 in discovery; 0.920-0.951 in validation) and superior sensitivity (0.900-0.940) compared to conventional markers (0.020-0.240). Isovalerylcarnitine (C5), a key component, was

SUBMITTER: Zhu J 

PROVIDER: S-EPMC12686445 | biostudies-literature | 2025 Dec

REPOSITORIES: biostudies-literature