Diagnostic value of CEUS-VI-RADS for predicting muscle invasion in bladder cancer: Validation of the scoring system and introduction of tumor-bladder wall contact length as a supplementary quantitative metric

PURPOSE: To validate the diagnostic performance of the Contrast-Enhanced Ultrasound-based Vesical Imaging-Reporting and Data System (CEUS-VI-RADS) and to explore the incremental value of the tumor-bladder wall contact length (TBCL) as a quantitative metric in predicting muscle invasion in bladder cancer. METHODS: This retrospective study analyzed 112 patients with bladder cancer who underwent preoperative CEUS between February 2021 and December 2025. Two radiologists, blinded to clinic-pathological results, independently assigned CEUS-VI-RADS scores to the index lesions and measured the TBCL on CEUS images. Logistic regression analysis was performed to identify independent predictors of muscle invasion. The diagnostic performance of each parameter was compared using receiver operating characteristic curve analysis. RESULTS: Inter-observer agreement was substantial for CEUS-VI-RADS scoring (kappa = 0.703) and excellent for TBCL measurements (intraclass correlation coefficient = 0.82-0.88). Quantitative analysis identified both the CEUS-VI-RADS score (odds ratio: 11.2) and TBCL (odds ratio: 1.02-1.07) as independent predictors of muscle invasion. The area under the curve for CEUS-VI-RADS in predicting muscle invasion was 0.86; using a score of 4 as the optimal cutoff, the diagnostic accuracy was 0.84. The area under the curve for TBCL was 0.75, and its diagnostic accuracy showed no significant difference compared to CEUS-VI-RADS (P > 0.01). CONCLUSION: CEUS-VI-RADS is a robust tool for predicting muscle invasion in bladder cancer. As an easily accessible and independent quantitative metric, TBCL serves as an effective supplement to CEUS-VI-RADS, collectively enhancing predictive accuracy.