Accuracy of Vesical Imaging-Reporting and Data System for muscle-invasive bladder cancer detection from multiparametric magnetic resonance imaging

PURPOSE: The Vesical Imaging-Reporting and Data System (VI-RADS) was used to distinguish the invasive nature of bladder masses before surgery. These imaging criteria can be used to carefully select patients who are candidates for repeat transurethral resection of bladder tumor (Re-TUR-BT). One-third of patients are understage at the time of Re-TUR-BT. This study aimed to evaluate the discrimination accuracy of VI-RADS between non-muscle-invasive bladder cancer and muscle-invasive bladder cancer. MATERIALS AND METHODS: Patients with a bladder mass identified by cystoscopy who were assigned for TUR-BT were offered multiparametric magnetic resonance imaging (mpMRI) for VI-RADS. TUR-BT reports were compared with preoperative VI-RADS scores to evaluate the accuracy of discrimination of the muscle-invasive nature of the bladder mass. RESULTS: A total of 58 bladder tumor lesions were included, 13 with muscle-invasive bladder cancer and 45 with non-muscle-invasive bladder cancer. Sensitivity and specificity were 92.3% and 86.7%, respectively, when a VI-RADS cutoff of 4 or more was used to define muscle-invasive bladder cancer. Positive predictive value and negative predictive value were 66.7% and 97.5%, with an accuracy of 87.9%. The area under the receiver operating characteristic curve was 0.932 (95% confidence interval, 0.874–0.989), and the empirical optimal cutpoint from the Youden method was 3. CONCLUSIONS: VI-RADS is an accurate tool for correctly differentiating muscle-invasive bladder cancer from non-muscle-invasive bladder cancer. We found a cutpoint of VI-RADS 1–3 vs. 4–5 to have the highest specificity and accuracy for the discrimination of non-muscle-invasive from muscle-invasive bladder cancer.