Simplified Integration of ADC Quantification Enhances Biparametric and Multiparametric VI-RADS for Detrusor Muscle Invasion Prediction in Bladder Cancer

Objectives To assess the diagnostic performance of an adjusted VI-RADS (adj-VI-RADS), combining ADC measurements with biparametric(bp)/multiparametric (mp) magnetic resonance imaging (MRI), for predicting detrusor muscle invasion in bladder cancer. Methods In this retrospective study, 184 patients underwent bladder mpMRI before TUR-B/cystectomy. Two radiologists independently evaluated the images. T2-Weighted imaging, difussion weighted- imaging, and dynamic contrast enhanced images were scored according to the VI-RADS criteria. Whole-tumor (wADC) and normalized ADC (nADC) values were calculated. An optimal nADC cutoff for muscle invasion was determined by ROC analysis. Scores were upgraded if nADC was ≤ this cutoff, deriving the adj-VI-RADS. Results A total of 140 patients had NMIBC, while 42 had MIBC. Both readers found significantly lower mean wADC and nADC values in MIBC compared to NMIBC (e.g., Reader 1 nADC: 0.317 vs. 0.49; Reader 2 nADC: 0.39 vs. 0.524; all p < 0.001). The optimal nADC cutoff for predicting muscle invasion was 0.403. In ROC analysis, Reader 1 obtained the following area under the curve (AUC) values: baseline bp-VI-RADS: 0.84; baseline mp-VI-RADS: 0.92; adj-bp-VI-RADS: 0.89; and adj-mp-VI-RADS: 0.94. For Reader 2, the corresponding values were as follows: baseline bp-VI-RADS: 0.82; baseline mp-VI-RADS: 0.85; adj-bp-VI-RADS: 0.89; and adj-mp-VI-RADS: 0.90. For all models, a VI-RADS category ≥ 4 was the optimal cutoff for predicting muscle invasion. Conclusion Integrating nADC quantification via a standardized upgrade rule significantly enhances VI-RADS diagnostic precision for detecting muscle invasion in BC. Our contrast-free adj-bp-VI-RADS protocol achieves accuracy comparable to that of contrast-enhanced mp-VI-RADS, offering a safer and more cost-efficient alternative.