Group sequential multi-arm multi-stage survival trial design with treatment selection

Area of Science:

• Clinical Trials
• Biostatistics
• Survival Analysis

Background:

• Multi-arm multi-stage (MAMS) trials are gaining traction for evaluating multiple treatments simultaneously.
• A significant barrier to routine MAMS adoption is the computational complexity of determining stopping boundaries.
• Existing methods, like Jaki and Magirr's, require intensive computation for time-to-event endpoints.

Purpose of the Study:

• To develop an improved group sequential MAMS survival trial design.
• To overcome the computational challenges associated with traditional MAMS boundary calculations.
• To offer a more efficient and accurate approach for designing and analyzing MAMS survival studies.

Main Methods:

• The study proposes a novel group sequential MAMS survival trial design.
• The design is based on the sequential conditional probability ratio test.
• It provides explicit solutions for futility and efficacy boundaries for any number of stages and arms.

Main Results:

• The new method eliminates complex computational efforts for trial design.
• It accurately determines the required number of events for both fixed-sample and group sequential designs.
• An innovative interim analysis procedure is introduced that maintains study power.

Conclusions:

• The proposed method simplifies the design of group sequential MAMS survival trials.
• It offers computational advantages over existing methods, enhancing routine adoption.
• The design provides accurate event numbers and preserves statistical power during interim analyses.