Exploring the Value of Dual-Source CT Dual-Energy Multiparametric Imaging for Predicting Lymph Node Metastasis of Short Axis ≥ 0.6 cm in Advanced Gastric Adenocarcinoma
Objective To predict short axis ≥0.6cm lymph nodemetastasis in advanced gastric adenocarcinoma based on clinical,traditional imaging and dual-energy features,and to explore the value of dual-energy imaging for qualitative diag-nosis of gastric adenocarcinoma lymph nodes.Methods Retrospective analysis of data of 172 patients with advanced gas-tric adenocarcinoma who underwent radical resectionin.To maximize the correspondence between imaging and pathological lymph nodes,the metastatic group lymph nodes were selected from 36 patients with pN3(n=114)and the non-metastatic group were selected from 26 patients with pN0(n=65),and the short axis of lymph nodes were at least 0.6 cm.The lymph nodes were randomly allocated into the training cohort(n=125)and validation cohort(n=54).The patients'gender,age,image features of primary tumor(location,borrmann classification,size,T-stage),image features of lymph nodes(size,loca-tion,shape,density,margin,degree of enhancement)and dual energy features of lymph nodes(λHU,IC,SIC,Overlay,DER,DEI,Rho,Zeff,fat fraction)were recorded by evaluating preoperative dual energy CT images.By using univariate and multi-variate Logistic regression analysis,traditional model was constructed based on clinical and traditional imaging features,and combined model was constructed based on clinical,traditional imaging and dual energy features.Receiver operating charac-teristics(ROC)curves and decision curve analysis(DC A)curves were used to evaluate the diagnostic performance of the two models.Results Primary tumor location,primary tumor maximum diameter ≥7.50 cm,and lymph node margin were independent risk factors of lymph nodes metastasis in the traditional model,with AUC of 0.854 and 0.793 in the training and validation cohort.Primary tumor location,primary tumor maximum diameter ≥7.50 cm,lymph nodes margin and fat fraction were independent risk factors of lymph node metastasis in the combined model,with AUC of 0.902 and 0.865 in the training and validation cohort,and the diagnostic efficiency was better than the traditional model(P＜0.05).The DC A curve of the combined model is mainly above the traditional model,with better clinical benefits.Conclusion The com-bined model has high diagnosis value for lymph node metastasis of short axis ≥0.6 cm in advanced gastric adenocarcino-ma.
NETL
NSTL
万方数据
