Joint Analysis of Progression and Survival with Missing Data from a Cancer Clinical Trial

• Activity Number: 468
• Type: Topic Contributed
• Date/Time: Wednesday, August 7, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Dianne Finkelstein*+ and David A Schoenfeld
• Companies: MGH and Harvard University and MGH and Harvard University
• Keywords: Progression free survival ; CEST ; interval censoring ; Clinical trial ; cancer

Abstract

Clinical Trials often assess efficacy by comparing treatments on the basis of two or more event-time outcomes. In the case of cancer trials, the comparison of treatments is often summarized by the analysis of progression-free survival (PFS), which is the minimum of the time to progression or death. This analysis has several issues. Disease progression is only known to have occurred in the interval (which can be of varied length) between examinations (interval censored). Also, if progression is not detected before death from the disease, then even an interval-censored analysis must be modified to handle the fact that the censoring is not an independent process. Finally, PFS does not use all information available on each outcome. In this paper, we consider a joint model for progression and death, assuming that the effect of covariates on the hazard of progression, and of death after progression obey proportional hazards models. We then develop a test based on the joint distribution of progression and survival using the very simple and intuitive approach of CEST. We will focus on the application of these methods in cancer research.