A Multi-Arm Multi-Stage Group Sequential Phase 2/3 Design with Dose Selection for Oncology Trials

Seamless phase 2/3 design with dose selection has generated increasing interest in clinical trial design, aiming to expedite drug development with fewer participants. The existing designs consist of two stages, with multiple doses evaluated in the first stage (phase 2) and the selected dose (s) tested in the second stage (phase 3). In this article, we extend the two-stage design to a multi-stage design that selects a single dose and tests it based on a time-to-event endpoint. The proposed design controls the Type-I error rate by using the combination test along with the group sequential method and the closed testing procedure. The simulation results confirm the control of Type-I error rate and indicate that the proposed design offers attractive operating characteristics compared to the traditional sequential approach with separate phase 2 and 3 trials.