Simultaneously optimizing dose and schedule of a new cytotoxic agent.

Area of Science:

• Clinical trial design
• Pharmacology
• Biostatistics

Background:

• Traditional phase I trials use a single treatment course, varying only dose per administration.
• Actual medical practice involves complex treatment schedules with dose adjustments and delays.
• Patient toxicity risk is influenced by both dose and the overall treatment schedule.

Purpose of the Study:

• To develop a practical phase I clinical trial design that mirrors clinical practice.
• To account for both dose per administration and the patient's treatment schedule.
• To optimize trial designs for patient safety and treatment efficacy.

Main Methods:

• An outcome-adaptive Bayesian design was developed.
• The design simultaneously optimizes dose and schedule based on time-to-toxicity.
• Computer simulations were used to calibrate design parameters.

Main Results:

• A phase I trial in allogeneic bone marrow transplantation was designed using this method.
• Simulations show superior performance in identifying optimal (dose, schedule) combinations.
• The new method assigned more patients to optimal treatment combinations compared to dose-only methods.

Conclusions:

• This is the first phase I trial design flexible enough to vary dose, timing, and number of administrations.
• The design requires larger sample sizes due to examining more treatment combinations.
• Assumptions include independent effects of administrations and a constant toxicity hazard.