Disease-free survival as a surrogate endpoint for overall survival in adults with early-stage pancreatic cancer: A trial-level surrogacy analysis.

Abstract

708 Background: In early-stage cancer, overall survival (OS) is the gold-standard endpoint for assessing therapeutic efficacy. However, long follow-up periods make OS challenging and costly to measure in clinical trials. Disease-free survival (DFS), which captures the time to recurrence or death, has been proposed as a surrogate endpoint for OS in early-stage pancreatic cancer (PC). This surrogacy relationship could allow for a more rapid assessment of treatment benefits, but its validity requires robust analysis across clinical trials. Understanding the trial-level correlation between DFS and OS is crucial to optimizing trial design and grant accelerated drug approval. Methods: We conducted a systematic search of Embase, PubMed, and Cochrane databases through August 2024, identifying randomized controlled trials (RCTs) that evaluated adjuvant or neoadjuvant systemic therapies in adult patients with PC. Included studies reported both OS and DFS outcomes. Hazard ratios for DFS (HR_DFS) were analyzed as surrogate endpoints for OS (HR_OS). Statistical analyses utilized linear regression models, weighted by trial sample sizes, to determine the strength of the correlation. Surrogacy strength was categorized by the coefficient of determination (R²), with R² ≥ 0.7 indicating a strong correlation, 0.69–0.5 a moderate correlation, and values < 0.5 indicating weak correlation. Results: Out of 3,048 studies screened, 30 RCTs met the inclusion criteria, comprising 5,275 patients in adjuvant therapy trials and 557 patients in neoadjuvant therapy trials. The median follow-up duration ranged from 16 to 53.5 months, with 12 phase II trials included. Overall, DFS showed strong surrogacy for OS in adjuvant trials (R² = 0.70; 95% CI, 0.50–0.91). Subgroup analysis indicated strong correlation in phase III trials (R² = 0.71; 95% CI, 0.46–0.96) and moderate correlation in phase II trials (R² = 0.67; 95% CI, 0.21–1.00). The strongest surrogacy was observed in studies incorporating adjuvant radiation therapy (R² = 0.81; 95% CI, 0.52–1.00). In the analysis of neoadjuvant trials, HR-DFS also presented strong correlation with OS (R² = 0.90; 95% CI, 0.56–1.00). Conclusions: There is a strong trial-level correlation between DFS and OS in both adjuvant and neoadjuvant therapy trials for early-stage pancreatic cancer, particularly in phase III studies and trials incorporating adjuvant radiation therapy. However, the strength of this surrogacy varies depending on the treatment setting and trial design.