PURPOSE: To determine cure proportions in mixture-cure survival models using secondary endpoints or real-world data (RWD) when overall survival (OS) data is immature.

DESCRIPTION

: Mixture-cure models can reduce bias when assessing OS in disease areas where patients can be cured. However, limited follow up time in OS makes it difficult to estimate the cure proportion in the intervention arm of a randomized clinical trial (RCT).

We will illustrate two approaches for estimating cure proportions when only short-term OS data are available: 1) using progression-free survival (PFS), overall response rate (ORR), and duration of response (DOR) RCT data to estimate cure proportions in relapsed-refractory diffuse large B cell lymphoma (RR DLBCL) patients; and 2) incorporating long-term follow-up from RWD to estimate differences in cure proportions between a control and intervention arm:

The workshop agenda will consist of the following: 1) Federico Felizzi will discuss the rationale for using secondary endpoints to inform a mixture-cure approach and an overview of the methodology; 2) Per-Olof Thuresson will present an application to RR DLBCL for estimating cure proportions; 3) Danny Sheinson will present an example of mixture-cure modeling using long-term follow-up from RWD; and 4) Anirban Basu will conclude the presentations with a discussion of the robustness of the methods and under what conditions extrapolation to lifetime cure models can be performed.