Time to Treatment Failure as an End Point in Pragmatic Oncology Trials: Association With Progression-Free Survival Across Randomized Studies.

Abstract

Purpose: Time to treatment failure (TTF), defined as the time from random assignment to treatment cessation for any reason, is an end point readily obtainable from routine clinical and real-world data sources. Despite its growing use, evidence describing its relationship with progression-free survival (PFS) remains limited. We evaluated the association between TTF and PFS across tumor types and treatment modalities to assess the validity of TTF as an end point for pragmatic oncology trials.

Methods: A systematic search identified randomized trials of advanced solid tumors reporting TTF and PFS. Correlations between hazard ratios (HRs) for TTF and PFS were assessed using inverse variance-weighted linear regression, stratified by prognosis (median overall survival <12 v ≥12 months). Subgroup analyses were performed according to therapy type and line of treatment. Individual patient data (IPD) were reconstructed from published Kaplan-Meier curves to enable arm-level comparisons.

Results: Fifty-eight analysis units encompassing 19,099 patients across 13 tumor types were included. TTF and PFS were strongly correlated in both poor- (r = 0.81; 95% CI, 0.59 to 0.94) and good-prognosis (r = 0.79; 95% CI, 0.60 to 0.91) groups. In the intervention arm, median TTF and PFS were comparable in poor-prognosis trials (9.6 v 11.0 months, P = .12) but diverged in good-prognosis settings (16.8 v 23.0 months, P < .001). Correlation was the highest in chemotherapy-based (r = 0.93-0.96) and first-line (r = 0.88-0.90) trials and weaker in targeted-therapy (r = 0.35-0.55) and later-line settings (r = 0.35-0.82).

Conclusion: TTF offers a practical, patient-centered end point, reflecting both treatment efficacy and tolerability. Its strong correlation with PFS supports its role as a clinically meaningful end point for pragmatic oncology trials in advanced cancers.