Surrogacy of Intermediate Clinical Endpoints for Overall Survival in Patients With Localized Muscle-Invasive Bladder Cancer

BACKGROUND: Although intermediate clinical endpoints (ICEs) may expedite completion of randomized controlled trials (RCTs) evaluating perioperative systemic treatments for localized muscle-invasive bladder cancer (MIBC), no validated surrogate for overall survival (OS) has been established. We aimed to assess the surrogacy of pathologic complete response (pCR), pathologic objective response (pOR), and disease-free survival (DFS) for OS. PATIENTS AND METHODS: We analyzed 4,828 patients with MIBC (cT2-T4N0M0) who underwent radical cystectomy (RC) with or without neoadjuvant chemotherapy (NAC) across 29 European centers (2001-2024). The inverse probability of treatment weighting (IPTW) approach was used to adjust for confounding between NAC and RC-only groups. Surrogacy was evaluated using: (1) adapted Prentice criteria to test whether each ICE remained a significant predictor of OS while the treatment effect disappeared in IPTW-adjusted multivariable Cox models; (2) the proportion of treatment effect explained (PTE); and (3) an emulated 2-stage meta-analytic framework to estimate the pseudo-trial-level R2 between treatment effects on each ICE and OS across 1,000 replicates of 5 random clusters. The surrogate threshold effect (STE) was calculated for ICEs demonstrating strong surrogacy (R2≥0.7). RESULTS: Overall, 1,288 (26.7%) patients received NAC followed by RC and 3,540 (73.3%) underwent RC alone. In IPTW-adjusted Cox regression analyses including NAC and each ICE separately, pCR (hazard ratio HR, 0.32; 95% CI, 0.24-0.41; P<.001), pOR (HR, 0.26; 95% CI, 0.21-0.31; P<.001), and DFS (HR, 5.17; 95% CI, 4.56-5.86; P<.001) were independent predictors of OS. The PTE was 0.42 (95% CI, 0.22-0.54), 0.48 (95% CI, 0.24-0.58), and 0.84 (95% CI, 0.66-0.96) for pCR, pOR, and DFS, respectively. At the pseudo-trial level, the R2 was 0.22 (95% CI, 0.20-0.25), 0.33 (95% CI, 0.31-0.36), and 0.83 (95% CI, 0.81-0.84) for the correlation between treatment effects on pCR, pOR, and DFS and OS, respectively. The STE was 0.82 (95% CI, 0.81-0.84) for DFS. CONCLUSIONS: We observed uncertainty regarding the surrogacy of pCR and pOR in patients undergoing RC with or without NAC for localized MIBC. Only DFS consistently mediated the treatment effect on OS, supporting its use as a surrogate for RCT dimensioning when a recurrence or death risk reduction of ≥18% is expected.