Designing Confirmatory Study Using Historical Data on Short-Term and Long-Term Endpoints

• Activity Number: 564
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Jiang Qian*+ and Qin Qin and Xuan Liu
• Companies: AbbVie and AbbVie and AbbVie
• Keywords: Short term endpoint ; long term endpoint ; sample size ; power ; historical data

Abstract

For some cancer types, there are opportunities for accelerated approval by FDA. Under those settings, usually a special protocol assessment (SPA) is desired and both the trial for accelerated approval and the confirmatory trial need to be part of the SPA. For other cancer types, there may be desire to design and start a Phase 3 study based on some exciting preliminary data on short term endpoint in an ongoing Phase 2 study. For the above situations, without much data observed on the long term endpoint, designing an appropriately powered Phase 3 (confirmatory) study is challenging. A statistical model is proposed utilizing historical data for the short term endpoint (e.g., pathological complete response rate for neoadjuvant breast cancer, rate of minimal residual disease for CLL) and the long term endpoint (e.g., event-free survival, progression-free survival) to enable the design of the confirmatory Phase 3 study, where the long term endpoint is the primary efficacy endpoint. A simulation study is presented to illustrate the proposed approach and its utility in the sample size and power calculation for a future confirmatory Phase 3 study.