| Objectives:1.To verify the predictive value of Vesical Imaging Reporting and Data System(Ⅵ-RADS)for muscular invasion of bladder urothelial carcinoma.2.To explore the efficacy of Ⅵ-RADS score combined with quantitative parameters in the diagnosis of muscular invasion in bladder urothelial carcinoma,and provide guidance and help for clinical treatment.Methods:The multi-parametric MRI and clinical data of 273 patients with pathologically confirmed bladder urothelial carcinoma were retrospectively analyzed.According to the number of tumor lesions,they were divided into single group(1 lesion)and multiple group(more than 1 lesion);According to the pathological results,they were divided into muscle-invasive bladder cancer(MIBC)group and non-muscle-invasive bladder cancer group(NMIBC).The Ⅵ-RADS score was performed on the MRI data(Selecting the lesion with the highest score and largest volume to analysis for multiple lesions),and the largest diameter,the ADC value,and the ADC value of the internal obturator muscle were measured,and the relative ADC value(rADC)was calculated.The difference of the Ⅵ-RADS score and the muscular invasion of the tumor was compared between 212 cases(137 cases of NMIBC,75 cases of MIBC)in the single group and 61 cases in the multiple group(34 cases of NMIBC,27 cases of MIBC).Receiver operating characteristic(ROC)curves were used to evaluate the predictive efficacy of Ⅵ-RADS scores in single and multiple groups.The sensitivity,specificity,positive predictive value,and negative predictive value of Ⅵ-RADS in the diagnosis of muscular invasion were obtained according to the best cut-off value.The coincidence rate between Ⅵ-RADS score and pathological results was calculated.The data of 273 patients were uniformly named as the overall group,and the gender,age,smoking history,presence or absence of hematuria,Ⅵ-RADS,maximum tumor diameter,ADC value,and rADC value of the overall NMIBC/MIBC group were compared between groups.The effective predictors of muscular invasion of bladder urothelial carcinoma were determined by binary logistic regression.The predictive performance of univariate and joint models were determined by ROC,and the difference of AUC value was compared by Delong test.Results:1.The Ⅵ-RADS scores in the single and multiple groups have statistical significance between the NMIBC/MIBC groups(P < 0.05).ROC curve analysis show that the Ⅵ-RADS of the single and multiple groups are > 3 as the best cut-off value,and the Yodon indexes are 0.8460 and 0.5566,respectively.The sensitivities for the diagnosis of muscular invasion are 85.33% and 70.37%,respectively.The specificities are 99.27% and 85.29%,respectively.Positive predictive values are98.5% and 79.2%,respectively.Negative predictive values are 92.5% and 78.4%,respectively.The coincidence rates with pathological results are 94.3% and 78.7%,respectively.The values of AUC are 0.959 and 0.821,respectively.The AUC value of single group is higher than that of multiple group,and the difference is statistically significant(P < 0.05).2.The Ⅵ-RADS score,ADC value,rADC value and maximum tumor diameter of the whole group are significantly different between the NMIBC/MIBC groups(all P values < 0.001).The results of multivariate binary logistic regression analysis show that Ⅵ-RADS score(OR=20.964,P < 0.001)and rADC value(OR=0.000,P < 0.001)are effective predictors of muscular invasion of bladder urothelial carcinoma.The risk of tumor muscular invasion will increase with the increase of Ⅵ-RADS score and decrease with the increase of rADC value.ROC curve analysis show that the Yodon indexes of combined model of Ⅵ-RADS and rADC,Ⅵ-RADS and rADC values are0.8729,0.7786,and 0.7444 at their optimal cut-off values(0.412,>3,0.922)respectively.The sensitivities for the diagnosis of muscular invasion of urothelial carcinoma of bladder are 93.14%,81.37% and 96.08%,respectively.The specificities are 94.15%,96.49% and 78.36%,respectively.Positive predictive values are 90.5%,93.3% and 72.6%,respectively.Negative predictive values are 95.8%,89.7% and97.1%,respectively.The AUC values are 0.982,0.938 and 0.931,respectively,and the AUC value of the combined model is the largest,with statistical significance(all P values < 0.001).Conclusions:1.The Ⅵ-RADS score has a high diagnostic value for muscular invasion of bladder urothelial carcinoma,and the diagnostic efficacy for single lesion is higher than that for multiple lesions.2.The diagnostic efficacy of Ⅵ-RADS score combined with rADC value is better than that of Ⅵ-RADS score alone. |