Pathological outcomes and disease-free survival (DFS) in KEYNOTE-905: Neoadjuvant and adjuvant (neoadj-adj) enfortumab vedotin (EV) plus pembrolizumab (pembro) in participants (pts) with muscle-invasive bladder cancer (MIBC) who are cisplatin-ineligible.

Abstract

638 Background: The phase 3 KEYNOTE-905/EV-303 study (NCT03924895) showed significantly improved EFS, OS, and pCR rate with neoadj-adj EV + pembro added to radical cystectomy with pelvic lymph node dissection (RC + PLND) in pts with MIBC who were cisplatin-ineligible. We report pathological outcomes and DFS in the study. Methods: Pts with T2-T4aN0M0 or T1-T4aN1M0 MIBC ineligible for or declining cisplatin were randomized 1:1 to EV + pembro (3 cycles neoadj EV 1.25 mg/kg on d 1 and 8 + pembro 200 mg on d 1, RC + PLND, and adj 6 cycles EV + 14 cycles pembro) vs control (RC + PLND only). pCR (absence of viable tumor in tissue from RC + PLND [pT0N0]) rate was a key secondary endpoint; pathological downstaging (pDS; < pT2 [pT0, pTis, pTa, pT1] and N0) rate and DFS (time from post-surgery scan to local/distant recurrence or death) were secondary endpoints, all assessed by central pathology review or BICR. Pts were considered disease-free after RC + PLND for DFS analysis if they had complete resection (no gross residual disease) and no evidence of disease on a post-surgery scan. Results: As of June 6, 2025, median follow-up was 25.6 mo (range, 11.8–53.7). In the EV + pembro vs control arms, 149/170 pts (87.6%) and 156/174 (89.7%) underwent surgery; 147/149 (98.7%) and 149/156 (95.5%) had complete resection; and 138/149 (92.6%) and 123/156 (78.8%) had negative surgical margins. In addition to the previously reported pCR rate for all pts in the EV + pembro vs control arms (57.1% vs 8.6%; estimated difference 48.3%; 95% CI 39.5–56.5; P < .001), pDS rate was higher with EV + pembro vs control (65.9% vs 12.6%; estimated difference 53.1%; 95% CI 44.0–61.2). Table shows comprehensive pDS outcomes. Among pts evaluated for DFS (n = 135 in EV + pembro arm; n = 129 in control arm), median DFS was not reached vs 23.6 mo (HR 0.37; 95% CI 0.23–0.59). Conclusions: pCR, pDS, surgical outcomes, and DFS favored neoadj-adj EV + pembro and RC + PLND vs RC + PLND alone, supporting the primary results of KEYNOTE-905. These findings further establish neoadj-adj EV + pembro as a potential standard of care for pts with MIBC who are cisplatin-ineligible, addressing a key unmet clinical need. Clinical trial information: NCT03924895 . N; % (95% CI) EV + pembro, N=170 Control, N=174 pDS (<pT2N0) 112 65.9 (58.2–73.0) 22 12.6 (8.1–18.5) pT0N0 97 57.1 (49.3–64.6) 15 8.6 (4.9–13.8) pTisN0 7 4.1 (1.7–8.3) 2 1.1 (0.1–4.1) pTaN0 1 0.6 (0.0–3.2) 0 0.0 (0.0–2.1) pT1N0 7 4.1 (1.7–8.3) 5 2.9 (0.9–6.6) Non-pDS (≥pT2N0) 33 19.4 (13.8–26.2) 112 64.4 (56.8–71.5) Other* 2 1.2 (0.1–4.2) 15 8.6 (4.9–13.8) Incomplete resection 2 1.2 (0.1–4.2) 7 4.0 (1.6–8.1) Did not undergo surgery 21 12.4 (7.8–18.3) 18 10.3 (6.2–15.9) *Not evaluated centrally for pDS due to no available evaluable surgical sample, nonprotocol systemic therapy prior to RC + PLND or operational issues; considered non-pDS in analysis.