A Bayesian Response-Based Adaptive Randomization Design for Phase II Trials of Targeted Therapy

• Activity Number: 298
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 8:30 AM to 10:30 AM
• Sponsor: International Chinese Statistical Association
• Author(s): Yiyi Chen*+ and Zunqiu Chen and Motomi Mori and Byung Park and Jeffrey W Tyner and Marc M Loriaux and Tibor Kovacsovics and Brian J Druker
• Companies: Oregon Health and Science University and Oregon Health and Science University and Oregon Health and Science University and Oregon Health and Science University and Oregon Health and Science University and Oregon Health and Science University and Oregon Health and Science University and Oregon Health and Science University
• Keywords: Bayesian design ; Adaptive design ; Clinical trials ; Allocation ratio ; Targeted therapy

Abstract

We have developed a novel in-vitro screening technique to evaluate the sensitivity of primary leukemia cells against a panel of small-molecule kinase inhibitors. To evaluate the clinical utility of drug selection based on the in-vitro drug sensitivity assay, we propose a novel Bayesian response-based adaptive randomized phase II trial design to evaluate whether assigning assay-matched treatment brings therapeutic benefits compared to assay mismatched assignments of the treatment. The primary endpoint of the trial is the objective response. A hierarchical Bayesian model is used for evaluating the assay-matched effects for the J drugs included in the trial. The operating characteristics of the design are assessed through a simulation study for six potential allocation rules. We find that the modified Rosenberger, Sallard, Ivanova, Haper and Ricks (RSIHR) allocation ratio has the most desirable performance for a variety of scenarios we consider in the simulation study.