Beyond PD-L1 and HER2: emerging predictive biomarkers in gastroesophageal tumors

Gastroesophageal adenocarcinoma (GEA) remains a major cause of cancer morbidity and mortality, with substantial geographic variation and marked inter- and intratumoral heterogeneity. The implementation of predictive biomarkers, including human epidermal growth factor 2 (HER2) overexpression/amplification, programmed death ligand 1 (PD-L1) expression, deficient mismatch repair/high microsatellite instability (dMMR/MSI-H), and claudin 18.2 (CLDN18.2), has redefined first-line treatment and enabled biomarker-directed therapeutic selection. Nevertheless, most patients eventually progress, reflecting imperfect biomarker performance, spatial and temporal heterogeneity, coexpression patterns, and complex tumor microenvironment (TME) interactions. Beyond PD-L1, emerging biomarkers such as Epstein–Barr virus (EBV) status, tumor mutational burden (TMB), and transcriptomic immune signatures provide deeper insight into tumor immunogenicity, although their predictive value remains inconsistent. To overcome resistance, novel therapeutic strategies are targeting alternative inhibitory immune checkpoints [cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), T cell immunoreceptor with Ig and ITIM domain (TIGIT), lymphocyte-activation gene 3 (LAG-3), and T cell immunoglobulin and mucin-domain containing protein 3 (TIM-3)] and costimulatory pathways (e.g. CD137/TNF receptor superfamily member 9; normally named 4-1BB), as well as emerging targets such as B7-H3 and immunoregulatory axes including C–C motif chemokine receptor 8 (CCR8), Dickkopf-related protein 1 (DKK1), transforming growth factor beta (TGF-β), and prostaglandin E2 receptor 4 (EP4). Concurrently, emerging targetable alterations such as fibroblast growth factor receptor 2b (FGFR2b) and mesenchymal–epithelial transition factor (MET) are reshaping precision treatment, whereas next-generation platforms, including bispecific antibodies, antibody–drug conjugates, and adoptive cellular therapies, are expanding therapeutic opportunities beyond conventional targeted approaches. Notably, these advances have been achieved mainly in the advanced disease setting, whereas approaches for localized GEA remain a critical unmet need. Future progress will require a deeper understanding of the tumor–TME interactions, through integrated, longitudinal, and multidimensional biomarker frameworks that capture genomic, transcriptomic, and immune contexture, alongside adaptive trial designs and equitable access to molecular testing and targeted therapies, to refine precision oncology and improve outcomes in this biologically complex disease.