Abstract
Gastric cancer (GC) often presents a poor prognosis due to its asymptomatic phenotype at early stages. Upper endoscopy, which is the current gold standard to diagnose GC, is invasive with limited sensitivity for detecting gastric preneoplasia. Non-invasive biomarkers, such as blood circulating proteins, offer a promising alternative for the early detection of gastric lesions. In this prospective study, we identified plasma protein biomarkers for gastric preneoplasia and cancer using mass spectrometry-based proteomics in an exploratory cohort (n = 39). Fifteen promising protein candidates emerged to distinguish patient categories and were further confirmed by enzyme-linked immunosorbent assays (ELISA) in plasma samples from a validation cohort of 138 participants. Our predictive models demonstrated high classification performance with a minimal set of biomarkers. A four-protein panel (ARG1, CA2, F13A1, S100A12) achieved 94.1-98.2% AUROC (95% CI) for distinguishing cancer from non-cancer cases, while a five-protein panel (ARG1, CA2, HPT, MAN2A1, LBP) reached 97.3-99.5% AUROC (95% CI) for distinguishing cancer or preneoplasia from healthy or non-atrophic gastritis cases on the full cohort. Leveraging simple blood sampling, this strategy holds promise to detect high-risk gastric lesions, even at asymptomatic stages. Such an approach could significantly improve early detection and clinical management of GC, offering direct benefit for patients.