Circulating tumor DNA (ctDNA) to guide response-adapted bladder preservation in muscle invasive bladder cancer (MIBC): Integrated analysis of the RETAIN trials.

LBA632 Background: The phase II RETAIN 1 and 2 trials evaluated a response-adapted approach to identify pts with MIBC who may safely undergo cystectomy-sparing active surveillance (AS) after neoadjuvant therapy (tx). RETAIN-1 evaluated AMVAC and RETAIN-2 combined nivolumab with AMVAC. Although ctDNA is prognostic in pts treated with cystectomy, its role in selecting or monitoring pts on AS is unknown. We report updated RETAIN-2 outcomes and an integrated ctDNA analysis from RETAIN-1/2 to evaluate the role of ctDNA in risk stratification for bladder preservation. Methods: RETAIN-2 enrolled pts with cT2-T3N0M0 MIBC treated with AMVAC plus nivolumab. TURBT samples were sequenced for ATM , ERCC2 or RB1 mutations. Pts with >1 mutation and clinical complete response (restaging TUR, urine cytology and CT), entered AS; others received bladder-directed tx. The primary endpoint was 2-yr metastasis-free survival (MFS) in the ITT population. Plasma from both trials was analyzed with Signatera at baseline and post-tx; RETAIN-2 included 3- and 6- mo draws. Results: Among 71 evaluable RETAIN-2 pts (median age 68; 77% male; 42% cT3), 57 (80.3%) remained metastasis-free at 29.2-mo follow-up. The estimated 2-yr MFS was 79.8% (95% CI 70–90%) in the ITT population and 80% (95% CI 64–100%) in AS pts. Among 22 AS pts, 8 (36%) had bladder recurrence, 4 (18%) developed metastases, 4 needed salvage cystectomy, 3 salvage chemoradiation; 16 (73%) remained metastasis-free with an intact bladder. Across both trials (RETAIN-1/2), 274 ctDNA timepoints from 111 pts were analyzed. Baseline and post-tx ctDNA positivity were 42.2% and 13.6%, respectively. Among baseline positives, 72.7% (32/44) cleared ctDNA. Pts who were ctDNA-negative post-tx or cleared ctDNA (from baseline positive) had markedly lower recurrence risk than those who with post-tx ctDNA-positive or non-clearers (recurrence rates: 34.8% and 43.8% for negative/clearance vs 85.7% and 91.7% for positivity/non-clearance; p < 0.001 and p < 0.01, respectively). Among AS pts who were ctDNA-negative post-tx, the 12-/24-mo MFS was 97.1%/82.4%, indicating strong prediction of metastatic control. However, the 12-/24-mo recurrence-free survival (RFS) was 62.9%/50.7%, driven largely by local recurrences. Of 22 AS pts who recurred, 19 were ctDNA negative post-tx. Only 3 AS pts were ctDNA-positive. Conclusions: RETAIN-2 is expected to meet its primary endpoint of 2-yr MFS. While post-tx ctDNA negativity predicted metastatic control, it did not reliably predict local-only recurrences among AS pts. This limitation may reflect the field effect characteristic of bladder cancer or simply the poor sensitivity of plasma ctDNA for detecting minimal residual disease confined to the bladder. Future bladder-sparing strategies may benefit from integrating response-adapted selection with serial ctDNA and urine tumor DNA monitoring. Clinical trial information: NCT04506554 .