Discovery and validation of biomarkers for Epstein-Barr virus associated gastric cancer

Epstein-Barr virus-associated gastric cancer (EBVaGC) displays unique clinicopathological hallmarks, yet serology-based tools for its detection are still limited. Here, we develop a functional EBV proteome microarray covering 72 viral proteins and apply it to profile antibody responses in 62 gastric cancer patients. The resulting landscape reveals an IgG-skewed humoral signature specific to EBVaGC and identifies 34 EBV antigens exhibiting differential reactivities. Multivariable logistic regression integrates complementary markers into an optimal five-analyte panel (LF2_IgG, BBLF2_IgG, BLRF2_IgG, BPLF1-2_IgA, and BGLF4_IgA) that achieves outstanding discrimination performance (AUC = 0.93) in an independent validation set ( n = 316). The panel's performance is further validated in a community-based screening cohort ( n = 474), where it achieves 87.3% sensitivity and 88.3% specificity for distinguishing EBVaGC from non-malignant gastric conditions spanning gastritis to dysplasia (AUC = 0.94). Together, these results establish a serological framework for EBVaGC diagnosis and provide a scalable strategy for population-level screening that could materially improve the management of this virus-driven malignancy.