5.1. Defining a biomarker’s boundary of utility
Background: We hypothesize that the predictive utility of the investigated biomarkers will be limited to specific drug classes. This class-specific nature of the biomarkers has direct policy implications for how to define and understand drug classes, and how to use biomarkers in making regulatory decisions for novel therapeutics. If we demonstrate that certain biomarkers cannot be assumed to be a universal predictors of clinical benefit (only of therapeutic response to a specific class of agents), then these results would support a more modest role for biomarkers as criteria of therapeutic innovation in research.
Aim: To identify the limitations of a biomarker’s utility, elucidate the patterns of research activities by which these limitations are determined, and provide policy recommendations for how lawmakers, research funders, and regulators ought to rely on biomarkers.
Approach: By pooling results from our four empirical projects, we will be able to measure the extent to which a biomarker’s utility is generalizable. How do the clinical benefit and cost-effectiveness associated with biomarker patient selection shift from one clinical indication to the next? Do certain kinds or uses of biomarkers generalize across drug classes or diseases better than others? Using our data as well as other databases available to the Division, we will be able to compare properties of biomarker research programs with more traditional drug development programs to answer questions such as to what extent biomarker-driven research programs lead to faster drug approvals.