A novel Bayesian seamless phase I/II design.

Area of Science:

• Clinical Trial Design
• Biostatistics
• Pharmacology

Background:

• Optimizing dose selection in early-phase clinical trials is crucial for drug development.
• Existing designs may not efficiently balance dose escalation with patient allocation to promising doses.

Purpose of the Study:

• To propose a novel Bayesian Phase I/II clinical trial design.
• To integrate a hybrid mTPI for MTD targeting and adaptive randomization for dose expansion.

Main Methods:

• Employs a hybrid mTPI method for dose escalation in Phase I.
• Utilizes a randomization allocation schema for adaptive patient assignment in Phase II.
• Inherits a mechanism for simultaneous dose escalation and expansion from prior literature.

Main Results:

• Extensive simulations demonstrate significant sample size reduction.
• The design efficiently allocates more patients to optimal doses.
• Outperforms two competing designs in simulations.

Conclusions:

• The proposed Bayesian Phase I/II design offers improved efficiency in dose-finding studies.
• This approach optimizes patient allocation and sample size utilization.
• Provides a valuable alternative for early-phase oncology trial design.