|
Abstract:
|
In oncology clinical trials, longitudinal data are often collected for monitoring tumor reduction along with a time-to-event outcome, which is usually employed to evaluate the primary hypothesis on the treatment effect. The existence of longitudinal measurements is directly related to the status of the time-to-event outcome; more sophisticated modelling techniques of jointly modeling of longitudinal and survival outcomes, needs to be considered for providing valid inference on treatment effect. In multiple myeloma, with the advent of transformative agents, overall survival has improved dramatically. As a result, the time-to-event endpoint will take much longer time to mature and it becomes challenging to conduct clinical investigations in a reasonable time frame. We propose a new design framework to evaluate the potential efficiency gain obtained by leveraging information from the longitudinal outcome for making correct decision at the end of Phase 2 with the consideration of the correlations between endpoints. Simulation studies show that our design substantially outperforms some existing designs.
|
