基于增强CT静脉期影像组学模型术前预测进展期胃癌神经侵犯的价值OA

Objective:To investigate the value of an interpretable radiomics model based on venous phase contrast-en-hanced CT in the preoperative prediction of perineural invasion(PNI)in advanced gastric cancer.Methods:Imaging and clinical data were retrospectively collected from 110 patients with pathologically confirmed advanced gastric cancer admitted to Sichuan Cancer Hospital between January 2024 and December 2024,including 58 PNI-positive cases and 52 PNI-negative cases.A total of 1,223 radiomic features were extracted from venous phase contrast-enhanced CT images after target volume delineation.Patients were divided into a training cohort(77 cases)and a validation cohort(33 cases)at a ratio of 7:3.The least absolute shrinkage and selection operator was used for feature selection.Logistic regression was applied to estab-lish radiomics,clinical and combined models.The predictive performance of each model was evaluated by the area under the curve(AUC).A nomogram was used to visualize the predictive risk probability of the combined model,and decision curve analysis was performed to assess its clinical utility.SHAP(Shapley Additive Explanations)was conducted to evaluate the importance and contribution of clinical and radiomic features in the model.Results:For the prediction of PNI in advanced gastric cancer,the radiomics model achieved an AUC of 0.67(95%CI:0.55～0.80)in the training cohort and 0.72(95%CI:0.54～0.90)in the validation cohort;the clinical model yielded an AUC of 0.74(95%CI:0.63～0.84)and 0.70(95%CI:0.53～0.87),respectively;the combined model showed an AUC of 0.81(95%CI:0.72～0.91)and 0.80(95%CI:0.64～0.95),respectively.Conclusion:The contrast-enhanced CT-based combined model allows effec-tive preoperative prediction of PNI in advanced gastric cancer.