Evaluation of surrogate endpoints in muscle-invasive bladder cancer (MIBC): A systematic review and meta-analysis.

Abstract

4580 Background: Overall survival (OS) is the gold-standard efficacy measure in oncology; however, it can take several years for OS data to mature, particularly in the clinically localized setting in patients undergoing curative intent treatment. To accelerate patient access to novel therapies, surrogate endpoints can be used to accelerate the assessment of new treatments when an early measure is reasonably likely or known to predict the clinical benefit for a target outcome such as OS. MIBC is a potentially curative disease with a complex and evolving treatment landscape involving radical cystectomy with or without systemic therapies, and bladder sparing strategies. Despite the need for surrogate endpoints in MIBC, there is limited research on their validity in this patient population. Methods: This study evaluated the trial-level surrogacy of event-free survival (EFS), progression-free survival (PFS), and disease-free survival (DFS) with respect to OS in MIBC. A systematic literature review (SLR) was conducted to identify randomized controlled trials (RCTs) that evaluated anti-cancer treatments (neoadjuvant, adjuvant, perioperative, and bladder sparing therapies) in MIBC and reported results for OS and ≥1 surrogate endpoint of interest. Studies published between Jan 1, 2000 and Jun 26, 2024 were identified by searching the MEDLINE, EMBASE, and CENTRAL databases. Grey-literature sources included recent conference proceedings and clinical trial registries. Study quality was assessed using the Cochrane Risk of Bias v2 tool. Data from studies with comparable outcome definitions for EFS, PFS, and DFS were combined into a broad composite outcome definition (cEFS). Trial-level surrogacy between the hazard ratio (HR) for cEFS and OS was evaluated. Analyses were conducted using a weighted linear regression (WLR) model and the bivariate Daniel & Hughes (D&H) model. Measures of surrogacy included the Pearson correlation coefficient (r) to measure the strength of association and the surrogate threshold effect (STE) to estimate the minimum HR for cEFS needed to reliably predict a HR for OS < 1. Results: 32 RCTs across 71 publications were included in the SLR; 14 were included in the cEFS analyses based on a feasibility assessment. Trials with a high-risk of bias (n = 1) or that evaluated the initiation of systemic therapy after disease progression (n = 4) were excluded. The HR for cEFS was strongly correlated with the HR for OS (r = 0.94; 95% CI: 0.72-0.99). Based on the STE, a HR <0.88 for cEFS would be needed to reliably predict a HR < 1 for OS. Results from the D&H model were consistent with these findings. Conclusions: These results suggest that at the trial level, the HR for cEFS is highly correlated with the HR for OS in MIBC across various treatment settings. cEFS may assist clinicians, regulatory agencies, and reimbursement bodies in contextualizing the benefits of novel treatment strategies in MIBC.