|
Abstract:
|
The main goal of co-develop targeted therapies and companion diagnostics is to identify a biomarker subpopulation that are more likely to respond to a drug, as it improves clinical program success probability by increasing efficacy and limiting safety concerns. In an ideal framework, for a continuous biomarker, a cutoff point is first determined to select biomarker-positive patients using data from Phase I or II studies, and the selected threshold, companion diagnostic and treatment effect are then confirmed in pivotal studies. While cutoff point selection is straightforward for biomarkers with bimodal distribution, it is quite challenging to do so and dichotomize biomarkers with continuous distribution. Moreover, challenges in Phase I/II studies, such small samples size, assay analytical performance and different patient population or comparative arms, also make it difficult to select a good biomarker cutoff point prior to initiation of pivotal studies. We discuss these practical challenges and some alternative strategies for cutoff point selection.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.