OBJECTIVES: To comprehensively assess the diagnostic performance of Vesical Imaging-Reporting and Data System (VI-RADS) score for detecting the muscle invasion of bladder cancer. METHODS: PubMed, Web of Science, and Embase were searched up to November 20, 2019. QUADAS-2 tool assessed the quality of included studies. The diagnostic estimates including sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and the area under the curve (AUC) of hierarchical summary receiver operating characteristic (HSROC) were calculated. Further subgroup analysis, meta-regression and sensitivity analysis were conducted. RESULTS: Six studies with 1064 patients were finally included. The pooled sensitivity, specificity, and AUC value were 0.90 (95% CI 0.86-0.94), 0.86 (95% CI 0.71-0.94), and 0.93 (95% CI 0.91-0.95) for VI-RADS 3 as the cutoff value. The corresponding estimates were 0.77 (95% CI 0.65-0.86), 0.97 (95% CI 0.88-0.99), and 0.92 (95% CI 0.89-0.94) for VI-RADS 4 as the cutoff value. Meta-regression analysis revealed that study design (p value 0.01) and surgical pattern of reference standard (p value 0.02) were source of the heterogeneity of pooled sensitivity. No publication bias was observed. CONCLUSIONS: The VI-RADS score can provide a good predictive ability for detecting the muscle invasiveness of primary bladder cancer with VI-RADS 3 or VI-RADS 4 as the cutoff value. KEY POINTS: • VI-RADS score has high sensitivity and specificity for predicting muscle invasion. • The diagnostic efficiencies of VI-RADS 3 and VI-RADS 4 as the cutoff value are similar. • VI-RADS score could be used for detecting muscle invasion of bladder cancer in clinical practice.