Classification of Companion Diagnostics: A New Framework for Biomarker-Driven Patient Selection

Area of Science:

• Pharmacogenomics and Personalized Medicine
• Regulatory Science in Drug Development

Background:

• Personalized medicine utilizes companion diagnostics to stratify patients for targeted therapies based on differential benefit/risk.
• Demonstrating a positive benefit/risk balance is crucial for drug approval, even in biomarker-selected populations.
• Regulatory considerations include the validity of patient selection and benefit/risk assessments in biomarker-negative groups.

Purpose of the Study:

• To establish a systematic classification of biomarker-drug pairs to guide regulatory requirements and clinical development.
• To provide a framework for evaluating the molecular mechanism and clinical evidence supporting biomarker-drug associations.

Main Methods:

• A systematic classification of biomarker-drug pairs was developed.
• Biomarkers were categorized into five ascending levels based on increasing evidence of their predictive value for a specific drug.
• Classification considered comparative pharmacological and clinical evidence.

Main Results:

• A five-category classification system for biomarker-drug pairs was proposed.
• This system reflects increasing levels of evidence for a biomarker's predictive nature in relation to a specific drug.
• The classification is based on comprehensive pharmacological and clinical data.

Conclusions:

• The proposed classification facilitates regulatory decision-making and optimizes drug development strategies.
• It supports biomarker-related patient subgrouping during clinical programs and marketing authorization.
• The grade of evidence for a biomarker's differential predictive power indicates the utility of subgrouping.