Enhancing the flexibility and power of adaptive seamless phase 2/3 design with copula modeling between short-term and long-term endpoints.

AbstractAdaptive designs have garnered widespread attention in drug discovery studies as they are often more flexible, efficient, informative, and ethical. However, the planning of oncology trials with seamless phase 2/3 design is not as commonplace, especially when the primary endpoint involves time-to-event survival time and is expected to be immature at the interim for the purpose of treatment selection of the most-promising regimen. In this article, we establish a seamless phase 2/3 design with a review of the framework and adaptation to survival analysis, incorporation of treatment selection scheme guided by surrogate short-term endpoint at interim analysis, and integration of copula modeling for endpoints. We also explore parametric copula estimation using historical data, and investigate mis-specification among copulas families as well as against a straightforward family of the stratified model. For illustration, we provide extensive simulation with operating characteristics. The article concludes with a discussion of practical issues in the implementation of such designs. [ABSTRACT FROM AUTHOR]