Towards the noninvasive identification of pathologic responders to neoadjuvant pembrolizumab in muscle-invasive urothelial bladder cancer (MIBC).

Abstract

4540 Background: The possibility to predict the pathologic complete response (pT0) or pT≤1 after neoadjuvant immunotherapy may deeply impact the management of MIBC and orient clinical trials. The PURE01 study ( Necchi, JCO 2018) evaluated preoperative pembrolizumab before radical cystectomy (RC) in T2-3bN0MO patients (pts). The trial was amended to increase the sample size. Methods: Pts were assessed with bladder multiparametric magnetic resonance imaging (mpMRI: T2-weighted imaging, diffusion-weighted imaging, dynamic contrast enhancement [CE]) before and after treatment (3 cycles every 3 weeks) prior to RC. All mpMRI were made with bladder catheterization. We tested the following complete response (CR) definition: any measurable residual disease (T2W, morphology), but no residual restricted diffusion and no early CE signals. Exams were independently assessed by 2 senior radiologists and tested against the pT0/pT≤1 endpoints. Logistic regression models analyzed mpMRI and biomarker data on comprehensive genomic profiling (CGP) with FoundationONE CDx assay and PD-L1 combined positive score (CPS). Results: From 02/17 to 11/18, 84 pts were assessed before and after treatment (168 total mpMRI) and had the pathological response available from RC. Baseline CR (TURB effect) was observed in 14 (16.7%) pts. The mean overall percentage of agreement (OPA) across the mpMRI data was 98.2%. The sensitivity of post-therapy (PT) CR for pT0 response was 0.80 (95%CI: 0.61-0.92); the specificity for pT≤1 response was 0.94 (95%CI: 0.80-0.99). PT-CR was associated with both pT0/pT≤1 endpoints (p < 0.001) after adjusting for clinical T-stage and histology. Baseline CPS≥10 and TMB≥20 mut/Mb were strongly associated with pT0 response (OR: 4.46 and 14.9). mpMRI CR post-pembrolizumab significantly improved the c-index of the CPS/TMB models: up to 0.81 and 0.87 for pT0/pT≤1 endpoints. Conclusions: In post-pembrolizumab MIBC, the proposed definition of radiological CR through mpMRI is endowed with limited inter-observer variability and high accuracy in predicting pT≤1. CPS/TMB-selected pts who achieve CR with mpMRI can be reliably spared any surgical treatment. Clinical trial information: NCT02736266.