Abstract:
|
Companion diagnostics; the foundation of personalized medicine, has become a rapidly-evolving area within the past several years. One of the keys to implementing a successful companion diagnostic lies in the ability to successfully identify the correct cutoff point for a continuous biomarker; or simultaneously for multiple biomarkers. A method for deriving a robust biomarker signature for prognostic or predictive biomarkers will be presented. The method to be presented is a tree-based ensemble statistical algorithm, called BATTing (Bootstrapping and Aggregating Thresholds from Trees). Important validation procedures will also be discussed.
|