Multistage, Bayesian Adaptive Design for Non-Small Cell Lung Cancer: A Step Toward Choosing Right Subpopulation

• Activity Number: 633
• Type: Topic Contributed
• Date/Time: Thursday, August 13, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Serap Sankoh* and Zhaowei Hua and Hongliang Shi and Mingxiu Hu
• Companies: Takeda Pharmaceuticals and Takeda Pharmaceuticals and Takeda Pharmaceuticals and Takeda Pharmaceuticals
• Keywords: Clinical trials ; Bayesian analysis ; Survival analysis ; Adaptive design ; Posterior probability ; Markov chain Monte Carlo (MCMC) methods

Abstract

A multistage, Bayesian adaptive design is used in a Phase 2 non small cell lung cancer (NSCLC) study to monitor progression free survival (PFS). This study consists of up to 3 stages, each evaluating active treatment versus control in a distinct population with NSCLC. Three criteria are set to determine whether to proceed to the next stage. 1: If there is at least a 72% posterior probability of the PFS hazard ratio (HR) exceeding a pre-defined boundary, the study will stop for efficacy. 2: If there is at least a 15% posterior probability of the PFS HR falling below a pre-defined boundary, the study will be terminated for futility. 3: If criteria 1 or 2 not met, an adaptation rule will be employed to either define a new subpopulation or continue enrolling additional patients in the current population. We will discuss the operating characteristics of the design parameters and how they influence power and type I error.