Dissecting genetic and immune drivers of heterogeneous responses to neoadjuvant immunochemotherapy in gastric cancer.

Although neoadjuvant immunochemotherapy (nICT) improves gastric cancer (GC) outcomes, resistance remains a challenge, highlighting the need for better patient selection and strategies to overcome resistance. Here, we analyze 110 patients with GC before and after nICT or chemotherapy (nCT) from the NEOSUMMIT-01 trial using multi-omic sequencing followed by functional validation. We identify five tumor microenvironment ecotypes (EC1-5) linked to therapy. nICT achieves response in EC1 (T cell activation), EC2 (tertiary lymphoid structures), and EC3 (vascular normalization), but not in EC4 (extracellular matrix organization) and EC5 (immunosuppressive macrophage enrichment). Notably, nICT resistance in EC5 is mediated by the interaction between APOA1 tumor cells and TREM2 macrophages. Additionally, we reveal multiple biomarkers associated with nICT efficacy, including SBS19, HLA-B15:02, FDXR expression, and FGFR pathway activity, and provide a multi-omic stratification model for treatment response-based patient stratification. This study provides mechanistic insights into nICT in GC, informs therapeutic decisions, and reveals potential targets.