Abstract:The Bayesian optimal interval (BOIN) design is a novel phase I trial design for finding the maximum tolerated dose (MTD). It can be used to conduct both single-agent and drug-combination trials. 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 most important advantage of the BOIN design is that it is easy to implement in practice and has superior operating characteristics. The BOIN design is algorithm-based and thus can be implemented in a way similar to the traditional “3+3” design. The performance of the BOIN design is comparable to that of the continual reassessment method (CRM, one of the best performing model-based phase I trial designs) in terms of selecting the MTD, but has a lower risk of assigning patients to subtherapeutic or overly toxic doses (i.e., better patient ethics). The idea behind the BOIN design is straightforward. The conduct of phase I trials can be viewed as a sequence of decision-making steps of dose assignment for patients who are sequentially enrolled into a trial. At each moment of decision making, based on the observed data, we take one of three actions: escalate, deescalate or retain the current dose. Under the standard assumption that efficacy monotonically increases with toxicity (for cytotoxic agents), an ideal trial design would escalate the dose when the current dose is below the MTD in order to avoid treating a patient at a subtherapeutic dose level; deescalate the dose when the current dose is above the MTD in order to avoid exposing a patient to an overly toxic dose; and retain the same dose level when the current dose is equal (or close) to the MTD. However, such an ideal design is not available in practice because we do not know whether the current dose is actually below, above or equal (or close) to the MTD, and need to

Influence of the slope in percentile estimation through binary regression for dose-finding experiments

Adaptive designs for identifying optimal biological dose for molecularly targeted agents.

A Bayesian Dose-Finding Design for Phase I/II Clinical Trials with Nonignorable Dropouts.

A Bayesian Quasi-Likelihood Design for Identifying the Minimum Effective Dose and Maximum Utility Dose in Dose-Ranging Studies

Bayesian hybrid dose-finding design in phase I oncology clinical trials.

Bayesian Dose Finding for Combined Drugs with Discrete and Continuous Doses

Dose-Response Curve Estimation: A Semiparametric Mixture Approach

A Bayesian Dose Finding Design for Oncology Clinical Trials of Combinational Biological Agents

Sequential Continual Reassessment Method for Two-Dimensional Dose Finding.

Gboin: a Unified Model-Assisted Phase I Trial Design Accounting for Toxicity Grades, and Binary or Continuous End Points

Bayesian Dose Finding by Jointly Modelling Toxicity and Efficacy As Time-To-Event Outcomes

Dose Finding in Oncology Trials Guided by Ordinal Toxicity Grades Using Continuous Dose Levels

Sample size calculation in dose-response research

Simultaneous confidence bounds in monotone dose–response studies

On the Relative Efficiency of Model-Assisted Designs: a Conditional Approach.

Phase I Trial Design for Drug Combinations with Bayesian Model Averaging.

AIDE: Adaptive intrapatient dose escalation designs to accelerate Phase I clinical trials

An investigation to improve a nonlinear mixed-effects approach for EC50 estimation based on multi-donor dose–response data

Statistical Tutorial for Using Bayesian Optimal Interval (BOIN) Design for Phase I Single-Agent and Drug-Combination Trials

Seamless Phase II/III Design: a Useful Strategy to Reduce the Sample Size for Dose Optimization

Bayesian model for early dose-finding in phase I trials with multiple treatment courses