The validity, reliability, and reviewer acceptance of VI-RADS in assessing muscle invasion by bladder cancer: a multicenter prospective study

Objective To assess diagnostic validity and reliability of VI-RADS in predicting muscle invasion by bladder cancer (BCa) and evaluate reviewer acceptance of VI-RADS for clinical routine. Methods A prospective multicenter study enrolled 331 patients with suspected/untreated BCa who underwent preoperative multiparametric MRI examination (mp-MRI) of the urinary bladder. Four experienced radiologists independently evaluated all mp-MRI using VI-RADS. The diagnostic validity of VI-RADS for predicting muscle invasion by BCa was calculated using histopathology of the first transurethral resection bladder tumor (TURBT) and second TURBT as the reference standards. The kappa statistics (kappa) were applied to assess the interreader agreement (IRA). Reviewer acceptance was evaluated with questionnaires. Results The risk of muscle invasion in VI-RADS 2, 3, 4, and 5 after the first and second TURBT was 21.8%, 45.8%, 69.6%, and 96.4% and 24.4%, 58.3%, 87%, and 99.2%, respectively. The overall diagnostic validity of VI-RADS was high. The optimal cut-off value for predicting muscle invasion after first TURBT was > VI-RADS 3 (sensitivity = 84.1% and specificity = 92.3%), and after second TURBT was > VI-RADS 2 (sensitivity = 89.9% and specificity = 90.1%). VI-RADS categorization showed a very good IRA (kappa = 0.93). Reviewers fully agreed with the statement, "The application of structured reporting of bladder tumor should be encouraged" (score = 20). Conclusions VI-RADS showed high diagnostic validity and reliability for predicting muscle invasion by BCa, especially VI-RADS 4 and 5. However, VI-RADS 2 and 3 require further modifications to enhance their diagnostic validity. VI-RADS is highly encouraged to be used in daily practice.