A Rescue Approach in an Oncology Pivotal Study

• Activity Number: 172
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Guohui Liu*+ and Jianchang Lin
• Companies: Takeda and Takeda
• Keywords: Clinical Trials ; alpha spending function ; power and sample size ; time to event

Abstract

Recently, it has been seen that overall survival (OS), instead of traditional progression-free-survival (PFS), is used to determine sample size calculation even if PFS could be used for full approval with the purpose to show a more desirable risk-benefit profile for commercial reasons. In this case, the PFS significance might be obtained in the early stage of the study due to the fact that PFS is much shorter than OS and a large sample size. Therefore, an approach, with the rescue plan for PFS, is proposed. Instead of traditional alpha spending function, a large portion of alpha is allocated to the PFS analysis up to the projected time for agencies' approval, such as 0.045, no matter there is IA for PFS analysis or not. If the statistical significance has been achieved, the study will be claimed successful. If not, there still exists a small portion of alpha, say, 0.005 in this case, to have one more chance to test PFS again with much more accumulated number of events. In this talk, we show that the power is almost maintained while providing one more chance to claim success. This approach is particularly useful when the effect size assumption on PFS is quite aggressive. In additio