Validity of event-free survival as a surrogate endpoint in haematological malignancy: Review of the literature and health technology assessments

Abstract

Objective: Event-free survival (EFS) is increasingly used as a primary endpoint in trials of haematological malignancies (HMs). A key consideration is whether EFS can reliably predict survival.

Methods: We conducted a review of the scientific literature and health technology assessments to evaluate evidence for EFS-OS surrogacy in HMs and acceptability of EFS by payers.

Results: Evidence of surrogacy varies by indication and line of therapy. In first-line AML, EFS is highly correlated with OS at the trial-level supporting its use as an early endpoint for traditional approval of treatments with curative intent. Surrogacy was also demonstrated in first-line DLBCL but remains unexplored in relapsed/refractory setting where post-transplant EFS24 was not prognostic of survival. In first-line FL, PTCL, T-LBL, and MCL, EFS24 is prognostic of survival but trial-level surrogacy has not yet been evaluated.

Conclusion: Strong EFS-OS correlation required for surrogacy may only be achievable in HMs with treatments characterised by high rates of durable remissions. Nevertheless, EFS24 is associated with favourable outcomes and remains a clinically meaningful endpoint in HMs.

Keywords: Correlation study; Event-free survival; Health technology assessment; Outcome prediction; Overall survival; Predictor variable; Surrogate endpoint.