Bayesian Optimal Interval (BOIN) Design

The optimal interval design is a novel Bayesian phase I clinical trial design for finding the maximum tolerated dose (MTD) for single-agent and drug-combination trials. It strictly adheres to ethical considerations and optimizes the dose assignment for each patient enrolled into the trial.

Description

The software (i.e., an R package called "BOIN") implements the Bayesian optimal interval (BOIN) design of Liu and Yuan [1]. The BOIN design is motivated by the top priority and concern of clinicians, which is to effectively treat patients and minimize the chance of exposing them to subtherapeutic or overly toxic doses. The prominent advantage of the BOIN design is that it achieves simplicity and superior performance at the same time. The BOIN design is algorithm-based and can be implemented in a simple way similar to the traditional 3+3 design. The BOIN design yields average performance comparable to the continual reassessment method (CRM) in terms of selecting the MTD, but has a substantially lower risk of assigning patients to subtherapeutic or overly toxic doses.

Features

• Written in R
• Very easy to use
• Extensively documented
• Handles both single-agent and drug-combination phase I trials
• Enables the user to generate operating characteristics for protocol preparation

Why use the BOIN Design?

• Very simple to implement  Similar to the 3+3 design, the dose escalation/deescalation rule is prespecified and can be tabulated, thus conducting a trial design by this method does not require additional software. During the trial conduct, clinicians can simply count the number of patients who experience toxicity and compare the observed toxicity rate with the prespecified dose escalation/deescalation boundaries to determine dose assignment until the trial is completed.

• Highly Ethical  The design optimizes the dose assignment for each patient in the sense that it minimizes the chance of assigning a patient to either a subtherapeutic dose or an overly toxic dose. This is consistent with the clinician's point of view for conducting clinical trials, i.e., maximizing the patient's treatment benefit.

• Superior performance  Both theoretically and numerically, it has been shown that the optimal interval design has desirable statistical properties and superior operating characteristics --- better than commonly used phase I designs.

• Accommodates both single-agent and drug-combination trials The BOIN design can be used to design both single-agent and drug-combination phase I trials. The resulting designs are easy to implement.

• Solid statistical justifications The design optimizes a sensible statistical criterion. It has a desirable finite-sample property (i.e., is coherent) and a desirable large-sample property (i.e., converges to the target dose). See the reference for more details.

Documentation

An extensive statistical tutorial, R manual, and protocol template are provided with the software. This tutorial covers the statistical basis of the method, and has guidelines for how to use the method and software.

Credits

Ying Yuan and Suyu Liu implemented the numerical algorithms using R.

References

[1] Suyu Liu and Ying Yuan (2015) Bayesian Optimal Interval Designs for Phase I Clinical Trials, Journal of the Royal Statistical Society: Series C, 64, 507-523.