Bayesian single-to-double arm transition design using both short-term and long-term endpoints.

Area of Science:

• Clinical Trials
• Biostatistics
• Oncology

Background:

• Phase II oncology trials often debate single-arm versus randomized double-arm designs.
• Transition designs offer a compromise but can prolong trial duration due to long endpoint evaluations.
• Delays in trial completion can postpone treatment for future eligible patients.

Purpose of the Study:

• To propose a Bayesian single-to-double arm design with short-term endpoints (BSDS) to accelerate phase II oncology trials.
• To utilize an intermediate short-term endpoint for early termination decisions, followed by long-term endpoint evaluation.
• To employ Bayesian posterior probabilities for decision-making and calibrate the design for frequentist operating characteristics.

Main Methods:

• Developed a Bayesian single-to-double arm design (BSDS) incorporating a short-term endpoint.
• Proposed design calibration steps to control frequentist operating characteristics and minimize expected sample size.
• Utilized Bayesian posterior probabilities for trial decision-making.

Main Results:

• Extensive simulations show the BSDS has comparable power and average sample size to conventional designs.
• The proposed BSDS significantly reduces overall trial duration.
• Applications demonstrated the design's utility in phase II oncology trials with binary endpoints.

Conclusions:

• The Bayesian single-to-double arm design with short-term endpoints (BSDS) effectively accelerates phase II oncology trials.
• BSDS provides a viable alternative to conventional designs, balancing efficacy with reduced trial time.
• This approach can expedite the delivery of novel cancer treatments to patients.