673P - Project Optimus for dose optimization: Implementation strategies for your trial from the statistician’s standpoint

Background

The limitations of phase I dose-finding studies, aimed at identifying the Maximum Tolerated Dose (MTD), are well-known, especially in oncology where newer drugs exhibit different dose-toxicity relationships. This has shifted the focus from determining the MTD to finding the optimal dose. In 2021, the FDA launched Project Optimus, a framework designed to provide guidance for improving dose optimization during drug development. The framework outlines general requirements for dose optimization but has raised questions among sponsors designing or amending trials. This is largely because it emphasizes not only analyzing dose-limiting toxicities but also includes additional safety, pharmacokinetics, and efficacy endpoints. Our aim is to share our experience as statisticians in designing and conducting trials following the Project Optimus guidance. We hope to clarify common questions from sponsors, address perceived challenges, and accelerate the adoption of dose optimization.

Methods

Since the release of the Project Optimus draft guidance, we have supported the (re)design of several oncology clinical trials by offering statistical expertise in adaptive designs and innovative methodologies. We developed a new Bayesian method that incorporates multiple endpoints into dose optimization without alpha spending, providing a seamless inclusion in clinical trials.

Results

Drawing from our experience in designing phase 1/2 and phase 2/3 clinical trials and submitting protocols for regulatory approval, we present a statistical perspective on: - Practical implementation of Project Optimus in trials. - Statistical methodologies applicable to dose optimization. - Our approach to incorporating various safety, pharmacokinetics, and efficacy endpoints.

Conclusions

Dose optimization, rather than MTD identification, has become the gold standard in oncology drug development. While this shift presents design and analysis challenges, innovative statistical approaches can address them effectively. This presentation offers actionable information for sponsors from a statistician's viewpoint, aiming to simplify the implementation of dose optimization strategies.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.