Early completion based on adjacent dose information for model-assisted designs to accelerate maximum tolerated dose...

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Pharmacology

Background:

• Phase I clinical trials are crucial for determining the maximum tolerated dose (MTD).
• Traditional MTD identification can be time-consuming, delaying subsequent trial phases.
• Model-assisted designs offer a framework for optimizing dose escalation strategies.

Purpose of the Study:

• To present and evaluate an early completion method for accelerating MTD identification in Phase I trials.
• To assess the accuracy and impact of this method on trial outcomes using simulation and real clinical trial data.
• To demonstrate the method's compatibility with model-assisted dose-finding designs.

Main Methods:

• Developed an early completion method utilizing toxicity data from adjacent dose levels.
• Calculated dose-assignment probabilities based on multiple dosage information.
• Evaluated the method's performance using simulations and data from an actual clinical trial, assessing MTD selection accuracy and trial efficiency.
• Compared the proposed method against standard approaches for MTD identification.

Main Results:

• The early completion method accurately identifies the MTD with minimal reduction in accuracy compared to standard methods.
• In some scenarios, the proposed method even shows improved MTD selection accuracy.
• The method accelerates MTD identification, enabling quicker progression to later trial phases.
• The early completion approach integrates seamlessly with model-assisted designs.

Conclusions:

• The proposed early completion method effectively accelerates MTD identification in Phase I trials.
• This method maintains high accuracy and can even enhance MTD selection precision.
• It poses no issues when applied to model-assisted designs, offering a valuable tool for clinical trial optimization.