Methodological Insights on Biomarker-Based Patient Selection: A Review of Scientific Advice Procedures at the
Cynthia Huber1, Joerg Zinserling2, Norbert Benda1,2
1Department of Medical Statistics, University Medical Center, Göttingen, Germany.
Abstract:
Biomarkers play a pivotal role in the selection and enrollment of trial participants. Particularly, predictive biomarkers help tailor medical care to individual patients; however, also prognostic biomarkers require consideration at the design stage. At the time of initiating a clinical trial, there may be uncertainty about whether a biomarker is predictive or prognostic, and the trial design may need to account for this. Relevant discussions between drug developers and regulators on the role of a biomarker in a specific drug development program are expected to take place during Scientific Advice (SA) procedures. SA procedures at the European Medicines Agency from January 1, 2018, to December 31, 2020, were systematically searched for methodological discussions around the use of predictive or prognostic biomarkers. The final analysis included 45 SA procedures for which key characteristics were summarized quantitatively. Selected methodological issues such as the cutoff selection of continuous biomarkers or study design considerations were elaborated in a qualitative summary. Our results identify commonly encountered points for discussion between drug developers and the European Medicines Agency for biomarker-informed patient selection and enrollment. Identified topics addressed during SA procedures include cutoff selection, study design, multiplicity control, and data-driven subgroup selection. The majority of the identified 45 SA procedures concern development programs in oncology. Addressing these issues upfront will allow for an improved dialogue between drug developers and regulators and support the drug development program and ultimately patient-centered access to medicines.
