Biology and performance of pre- and post-pembrolizumab (pembro) vesical imaging–reporting and data system (VI-RADS) to predict the pathological response in muscle-invasive urothelial bladder cancer (MIBC): Full data analysis from a clinical trials pipeline.

Abstract

4591 Background: The VI-RADS is a standardized reporting system that uses multiparametric magnetic resonance imaging (mpMRI) parameters to predict the probability of MIBC. It has been also used in studies using standard neoadjuvant chemotherapy. No studies have analyzed the performance of VI-RADS to predict the pT0 or pT≤1 response post-immunotherapy (IO) in MIBC. Methods: In PURE-01 study patients (pts) were staged with mpMRI before and after treatment (3 cycles of pembro) prior to radical cystectomy (RC). All mpMRI scans were centrally reviewed. Logistic regression models analyzed pre- and post-pembro VI-RADS against pT≤1 (primary endpoint) and pT0 (secondary endpoint). VI-RADS scores were dichotomized between 0-3 (0=no evidence of disease) and 4-5. Covariates included cT-stage, age, gender, PD-L1 combined positive score (CPS) and tumor mutational burden (TMB). Model performance was also tested in the ongoing NureCombo (nivolumab+nab-paclitaxel; NCT04876313) and SURE-01 (sacituzumab govitecan; NCT05226117) studies. Transcriptome-wide expression profiling was matched with VI-RADS scores. Results: In PURE-01, 115 pts had centrally-reviewed MRI scans (N=230 mpMRI), treated between 02/17 and 08/18. In total, 32 (27.8%) had pure/predominant non-urothelial carcinoma (UC) histology, 54 (47%) had a cT3-4N0 MIBC. Pre-pembro: 21 pts (18.3%) had no measurable disease (VI-RADS=0), 37 (32.2%) a VI-RADS 1-3 score, and 57 (49.5%) had a VI-RADS 4-5 score. 20 pts (17.4%) had a downstage from VI-RADS 4-5 to VIRADS 0-3 post-pembro. Pre-pembro VIRADS 4-5 were associated with higher angiogenesis and epithelial-mesenchymal transition (EMT) activity vs VI-RADS 0-3 (p=0.011 and 0.033). Both pre-pembro and post-pembro VI-RADS 0-3 scores were the only significant covariates against pT≤1 endpoint on multivariable analyses (MVA): the strongest effect was seen with post-pembro VI-RADS 0-3 against pT≤1 response (OR: 23.4, 95%CI: 7-95.3, p<0.0001). The AUC of this model was 0.90. Model-based decisions were supported in the range of threshold probabilities of 10-80%. Post-pembro VI-RADS 4-5 were characterized by higher stromal signature scores vs VI-RADS 0-3. When post-therapy VI-RADS were tested in evaluable pts from the combined NureCombo+SURE-01 cohort (N=17pts, 34 mpMRI): 90% of post-VI-RADS 0-3 revealed a pT≤1 stage at radical cystectomy. Conclusions: VI-RADS scores mirrored distinct biological features in MIBC, recapitulating tissue biomarkers of IO response. VI-RADS scores post-IO revealed a strong association with pathological downstaging and represented a robust feature based on which selecting pts for bladder-sparing strategies. Results seemed to be also applicable to other novel neoadjuvant therapies beyond IO monotherapy. Clinical trial information: NCT02736266 .