Plasma proteomic signatures improve risk stratification and personalized screening for gastric cancer

BACKGROUND: Accurate identification of individuals at high risk of gastric cancer (GC) remains a major challenge for effective screening. We aimed to identify plasma proteomic signatures and develop a risk prediction model for GC risk stratification. METHODS: Plasma proteomic profiling was performed using liquid chromatography-tandem mass spectrometry in a case-control discovery set (100 GC cases and 94 controls). Candidate proteins were evaluated in 52,552 UK Biobank participants with a median follow-up of 13.63 years, during which 92 incident GC cases were identified. Risk models integrating clinical, genetic, and proteomic factors were developed using LASSO-penalized Cox regression with stability selection and internally validated using bootstrap resampling. RESULTS: Among 2306 differentially expressed proteins in discovery, 25 were replicated in validation at nominal significance (P < 0.05) with consistent directions. Two proteins (CTSD and GGH) remained significant after false discovery rate correction. A primary proteomic model (clinical factors plus five proteins) improved discrimination versus clinical model (optimism-corrected C-index: 0.745 vs. 0.732, P = 0.046). Risk stratification revealed a clear GC risk gradient: hazard ratios were 6.08 (95% CI 2.15-17.20) for moderate-risk and 23.88 (95% CI 8.66-65.87) for high-risk groups. The risk score was also associated with GC risk as continuous variable (HR per standard deviation: 1.09, 95% CI 1.08-1.11). The 15-year cumulative incidence ranged from 0.02 to 0.56% across risk groups. Decision curve analysis indicated improved clinical utility. CONCLUSIONS: Plasma proteomic signatures may improve GC risk stratification beyond traditional clinical factors and could support more targeted screening strategies. Further validation is warranted.