All Times EDT
Keywords: biomarker, precision medicine, companion diagnostic, diagnostics
The promise of precision medicine largely depends on biomarkers that identify right patients for right treatments. When the pathway of a disease and mechanism of action (MOA) of a novel target drug for treating the disease are clear, and the compound is used as monotherapy, molecules that are directly involved in onset of the disease and/or associated with MOA would correctly pick out patients who are likely to respond to the new treatment. However, in reality, disease pathway and/or drug MOA are not always known, and combination therapies for patients who no longer respond to monotherapies may be necessary. Under these circumstances, finding likely responders based on disease biology and MOA is not an easy task. One of common strategies for addressing this challenge is to discover and validate predictive biomarkers in a form of biomarker panel and / or algorithm of multiplex biomarkers. It is worthy to note that developing such complex biomarkers, panels and/or algorithms usually is an evolving and iterative process of multi-cycles of hypothesis generation and testing in the clinical stage of drug development. On the other hand, in accordance with the draft FDA guidance for drug-diagnostic co-development, biomarker definition and their cutoffs should be locked down, and companion diagnostic assays (CDx) should be analytically validated before the initiation of pivotal clinical trials in which CDx under development will be clinically validated. Clearly there is a gap between drug development and validation of complex biomarkers under the current co-development guidance. In this talk, we will review real-world drug-diagnostic co-development process, identify key technical and statistical issues and propose a few practical considerations. Our thought process calls for closer collaboration between regulation makers and clinicians, translational scientists, assay developers and statisticians.