An efficient single-arm Bayesian adaptive trial algorithm to evaluate de-intensified oncologic treatment

Area of Science:

• Oncology
• Biostatistics
• Clinical Trial Design

Background:

• Evaluating de-intensified oncologic treatments can reduce patient toxicity and improve quality of life.
• Challenges exist in rare cancers due to long trial durations needed for sufficient event data.
• Traditional randomized trials may be impractical for uncommon cancers requiring de-intensified strategies.

Purpose of the Study:

• Introduce a Bayesian adaptive method for single-arm trial design.
• Enable efficient analysis of survival data in challenging clinical trial settings.
• Facilitate robust estimations and predictions for trial design and sample size determination.

Main Methods:

• Utilizes a Bayesian adaptive approach incorporating prior knowledge and historical control arms.
• Employs an R package, "BayesAT," for flexible modeling and multi-stage interim analyses.
• Designed for efficient survival data analysis, particularly for de-intensified oncologic treatments.

Main Results:

• The method's validity is confirmed through extensive simulation studies and sensitivity analyses.
• Successfully applied to a pediatric Hodgkin lymphoma trial.
• Demonstrates effective use of prior information and interim analyses to accelerate conclusions.

Conclusions:

• The Bayesian adaptive method offers an efficient approach for clinical trials with survival data constraints.
• The "BayesAT" package provides a practical tool for implementing this method.
• This approach expedites conclusions on treatment efficacy, especially for de-intensified strategies in rare cancers.