A dual-energy CT radiomics model can better predict neoadjuvant chemotherapy response in locally advanced gastric cancer, impacting surgical decision-making.

• The dual-energy CT model achieved an AUC of 0.806 versus 0.729 with conventional CT in the training dataset (p = 0.041).
• High-risk patients identified by the dual-energy CT model had nearly double the overall survival and disease-free survival risk compared to low-risk patients (HR = 1.996, p = 0.012; HR = 1.873, p = 0.037).

This model offers surgeons a reliable tool for selecting patients who may benefit most from treatment.