Recent Development in Biomarker and Subgroup Identification for Tailored Therapeutics

• Activity Number: 60
• Type: Topic Contributed
• Date/Time: Sunday, August 4, 2013 : 4:00 PM to 5:50 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Russell Wolfinger*+ and Richard C. Zink
• Companies: SAS Institute Inc and JMP Life Sciences, SAS Institute, Inc.
• Keywords: Subgroup Identification ; Biomarkers ; Personalized Medicine

Abstract

Multiple biomarkers-and an even larger number of subgroups defined by these biomarkers-are often assessed for their impact on efficacy or safety response using clinical trial data, with the objective of developing tailored therapeutics. A central question in these efforts is whether sufficient confidence can be obtained in a biomarker/subgroup with a clinically meaningful effect to enable key decisions that range from designing the next trial around this subgroup, to investing in further development of this compound, and to determining the indicated population(s) in regulatory approval. In this presentation, I will compare a number of biomarker/subgroup identification methods which have recently been proposed to make this possible. Emphasis will be given to the approaches used to account for multiplicity, both for the testing of hypotheses and estimation of biomarker/subgroup effects.