1.Multiparametric CT features for prediction of the risk classification of gastric stromal tumor
Chengyao XIE ; Zhiqi YANG ; Xiaofeng CHEN ; Zhiqiang ZHAGN ; Sheng ZHANG ; Xiong ZHANG ; Wenfeng LING
Journal of Practical Radiology 2024;40(3):394-397
Objective To investigate the value of multiparametric CT features for predicting the risk classification of gastric stro-mal tumor(GST).Methods The clinical data from 139 patients with GST were retrospectively collected.According to the patho-logical risk results,the patients were divided into two groups:a low-risk GST group(including very low-and low-risk)with 75 patients and a high-risk GST group(including medium and high-risk)with 64 patients.The CT features between low-risk GST group and high-risk GST group were compared using chi-squared test or t-test.The risk factors of high-risk GST were identified by univariate analysis.The prediction models were built by multivariate logistic regression.The performance of models were evaluated by receiver oper-ating characteristic(ROC)curve.Results There were significant differences in the maximum tumor diameter,minimum tumor diameter,arterial phase enhancement degree,venous phase enhancement rate,arterial phase enhancement degree rate,venous phase enhance-ment degree rate,cystic,and necrosis between low-risk GST group and high-risk GST group,which were associated with the risk classification of GST.The area under the curve(AUC)of the quantitative features-based model that combined maximum tumor diam-eter,minimum tumor diameter,arterial phase enhancement degree,venous phase enhancement rate,arterial phase enhancement degree rate and venous phase enhancement degree rate,showed a significantly higher performance than the qualitative features-based model that incorporated cystic and necrosis(0.981 vs 0.850,P<0.001).Conclusion Maximum tumor diameter,minimum tumor diameter,arterial phase enhancement degree,venous phase enhancement rate,arterial phase enhancement degree rate,venous phase enhance-ment degree rate,as well as cystic and necrosis,are associated with the risk classification of GST and can predict the high-risk GST.