PS24: Dose Optimization Statistical and Design Considerations

The FDA's Project Optimus guidelines have signalled a shift from dose-finding to dose optimization, requiring different requirements for dose optimization. Dose-finding in oncology has traditionally been based on a paradigm of cytotoxic drugs where trials have been designed around determining the maximum tolerated dose (MTD). However, with modern oncology therapies, there are examples where the MTD may not exist or where efficacy begins to plateau and higher doses do not result in improved efficacy. The MTD may no longer be the optimal dose and approaches for dose optimization in this context are needed. A framework that includes a robust evaluation of the totality of information across efficacy, safety, biomarkers, and other endpoints that balances benefit and risk to determine an optimal biologic dose (OBD) is necessary. Further consideration regarding design and comparing doses needs to be developed as well. Backfills are becoming more popular and included in dose-escalation trials. An approach that has gained popularity is the use of "backfill" patients on dose levels already deemed safe by a dose-escalation algorithm. Enrolling further patients to backfill cohorts provides an opportunity for further investigation of dose response and exposure response relationships, which is a key area of focus for Project Optimus. Dose randomization is another key area of Project Optimus and with this being a new area is open to innovative approaches to tackle dose comparison and show sufficient evidence for the OBD. With the addition of the dose optimization phase there has been a shift in rethinking design strategy. Seamless designs are being proposed that tie various stages together. To be more efficient and make informed decisions a novel seamless phase II/III design with dose optimization (SDDO framework) has been proposed to address the design need for Project Optimus. In this session we will have perspectives from industry, academia, and regulatory regarding how to address the backfill, dose comparison decision framework, and seamless designs for dose optimization in this new landscape. Case studies will highlight design and analysis features which facilitated dosage selection. Speaker 1 will discuss the role of backfill in dose optimization, focusing on its statistical and design considerations. Speaker 2 will discuss a seamless phase II/III design with dose optimization for oncology drug development. We will close with speaker 3 on a regulatory perspective on dose optimization and the latest views on design and approaches. They will present case studies on how considerations were explored in practice regarding comprehensive evaluations of both the risks and benefits associated with treatment, rather than solely concentrating on DLTs and the MTD.