Added value of synthetic MRI in refining biparametric MRI VI-RADS for predicting muscle-invasive bladder cancer.

PURPOSE: To evaluate the diagnostic value of synthetic MRI (SyMRI)-derived T1 and T2 values in predicting muscle-invasive bladder cancer (MIBC), and assess the feasibility of integrating them into the biparametric MRI Vesical Imaging-Reporting and Data System (bpVI-RADS).

METHODS

This prospective dual-center study included 42 patients (mean age: 71.90 ± 9.70 years, 90.5% male, 23.8% MIBC) with histologically confirmed urothelial carcinoma who underwent bladder MRI. Two radiologists independently assigned bpVI-RADS scores based on T2-weighted and diffusion-weighted images, blinded to pathological, with consensus reached through consultation with a senior radiologist. Tumor volumes were manually delineated on SyMRI-derived maps to extract T1 and T2 values. Two modified models (bpVI-RADS + T1 and bpVI-RADS + T2) were created by adjusting bpVI-RADS scores according to optimal T1 or T2 cutoff values determined via receiver operating characteristic (ROC) curve analysis. Diagnostic performance was assessed with ROC analysis, and AUCs were compared using the DeLong test.

RESULTS

MIBC lesions showed significantly lower T1 and T2 values than non-muscle-invasive bladder cancer (NMIBC). Optimal cutoff values for T1 and T2 to differentiate MIBC from NMIBC were 1370 msec and 115 msec, respectively. The bpVI-RADS + T2 achieved significantly higher AUC (0.994 vs. 0.881, P = 0.021) than bpVI-RADS at a cutoff score of ≥ 4. The bpVI-RADS + T1 showed slightly improved diagnostic performance than bpVI-RADS (0.913 vs. 0.881, P = 0.605).

CONCLUSION

SyMRI-derived T1 and T2 values could help distinguish muscle invasion in bladder cancer. Incorporating T2 values into the bpVI-RADS framework significantly enhances diagnostic performance for predicting MIBC.