Evaluating the contribution of circulating tumor DNA (ctDNA) towards magnetic resonance imaging (MRI) in clinical staging and response assessment of patients with muscle-invasive bladder cancer (MIBC).

Abstract

798 Background: In patients (pts) with MIBC, optimizing clinical staging (cT-stage) and response assessment after neoadjuvant therapies has become critical in order to envision next-generation bladder-sparing strategies. We have previously reported on the performance of bladder MRI to stage and predict the pathological response to pembrolizumab with the use of modified vesical-imaging reporting and data system (VI-RADS, PMID: 37803523). Using ctDNA yields the potential to increase the prediction of pathological stage, and improve the accuracy of the clinical complete response (cCR) definition. Methods: We report interim results from a prospective biomarker study assembling data from a platform of neoadjuvant trials testing novel perioperative therapies and radical cystectomy (RC) in pts with MIBC (NCT06341478). Pts were staged before and after neoadjuvant therapy with pelvic MRI; modified VI-RADS score consisted of the addition of VI-RADS 0 category to indicate no evidence of residual tumor. MRI images were assessed by 2 independent radiologists. ctDNA monitoring was assessed before and after neoadjuvant therapy and centralized using Signatera ctDNA Assay (Natera Inc., San Carlos, CA). Mean tumor molecules (MTM)/mL were reported. Fisher’s exact test was used for comparing groups. Results: From 07/23 to 09/24, 29 pts with available information on both MRI and ctDNA at baseline and/or after treatment, before RC, were collected. Nineteen (65.5%) had a cT2-stage, 4 received neoadjuvant nivolumab+abraxane (NCT04876313), 15 sacituzumab govitecan (SG) alone (NCT05226117) and 10 SG+pembrolizumab (NCT05535218). At baseline, 0/10 pts with VI-RADS 0-3 had detectable ctDNA vs 9/13 (69%) of VI-RADS 4-5 pts (p = 0.001). Mean baseline ctDNA count was 2.31 MTM/mL. Only 17.6% of cT2 pts revealed detectable ctDNA (mean: 2.51 MTM/mL) vs 100% cT3-4 (mean: 2.39 MTM/mL) (p < 0.001). Likewise, post-neoadjuvant therapy, 0/10 pts with VI-RADS 0-3 had detectable ctDNA vs 4/7 (57%) of VI-RADS 4-5 pts with a mean count of 294.8 MTM/mL (p = 0.010). 2/3 pts with clearance of ctDNA post-therapy had VI-RADS 4-5 scores post-therapy. After neoadjuvant treatment 11/14 (78.6%) pts with VI-RADS 0-3 achieved a pathologic complete response (pCR) vs 1/9 (11.1%) with VI-RADS 4-5 (p < 0.001). However, the only patient with pCR and post-neoadjuvant VI-RADS 3-4 relapsed. Post-neoadjuvant ctDNA was negative in 8/11 (72.7%) pts with pCR and in 0/5 without pCR (p = 0.023). Conclusions: Initial findings from a prospective biomarker study revealed a suboptimal association between local and systemic tumor assessment in pts with MIBC. In the effort of optimizing the definition of cCR to systemic therapy, ctDNA revealed critical limitations that would limit its use in bladder-sparing strategies. More mature data will be presented at the meeting. Clinical trial information: NCT06341478 .