Up-and-down designs for phase I clinical trials

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Pharmacometrics

Background:

• Traditional "3 + 3" dose escalation designs face limitations in optimizing clinical trial efficiency and safety.
• Improved up-and-down designs exist but lack comprehensive guidance and practical assessment, hindering their adoption.
• Evaluating operating characteristics of dose-finding designs in realistic clinical settings is crucial.

Purpose of the Study:

• To review and comprehensively assess six up-and-down dose escalation designs.
• To compare the performance of these designs against a theoretical optimal bound.
• To provide guidance on selecting the most effective design for practical clinical settings.

Main Methods:

• Review of six distinct up-and-down designs: "3 + 3", accelerated titration, biased coin, k-in-a-row, group up-and-down, and cumulative group up-and-down.
• Extensive simulation studies to evaluate operating characteristics under various practical scenarios.
• Comparative analysis against the theoretical optimal bound for nonparametric designs.

Main Results:

• The cumulative group up-and-down design demonstrated superior performance across key metrics.
• This design excelled in accurately selecting the maximum tolerated dose (MTD).
• It also showed advantages in patient allocation to the MTD and overall patient safety.

Conclusions:

• The cumulative group up-and-down design is recommended for its robust performance in dose-finding studies.
• Its operating characteristics are generally close to the nonparametric optimal bound.
• While highly effective, further refinements may offer additional improvements in specific contexts.