Intratumoral and Peritumoral Radiomics of Contrast-Enhanced CT for Prediction of Disease-Free Survival and Chemotherapy Response in Stage II/III Gastric Cancer

Background We evaluated the ability of radiomics based on intratumoral and peritumoral regions on preoperative gastric cancer (GC) contrast-enhanced CT imaging to predict disease-free survival (DFS) and chemotherapy response in stage II/III GC. Methods This study enrolled of 739 consecutive stage II/III GC patients. Within the intratumoral and peritumoral regions of CT images, 584 total radiomic features were computed at the portal venous-phase. A radiomics signature (RS) was generated by using support vector machine (SVM) based methods. Univariate and multivariate Cox proportional hazards models and Kaplan-Meier analysis were used to determine the association of the RS and clinicopathological variables with DFS. A radiomics nomogram combining the radiomics signature and clinicopathological findings was constructed for individualized DFS estimation. Results The radiomics signature consisted of 26 features and was significantly associated with DFS in both the training and validation sets (both P<0.0001). Multivariate analysis demonstrated that the radiomics signature was an independent prognostic factor. The signature had higher predictive accuracy than TNM stage and single radiomics features and clinicopathological factors. Further analysis revealed that stage II/III GC patients with high scores were likely to benefit from adjuvant chemotherapy. Conclusion The newly developed radiomics signature is a powerful predictor of DFS, and it may predict which stage II and III GC patients benefit from chemotherapy.