Accuracy of Biomarker Testing and Its Impact on Targeted Therapy Validation

• Activity Number: 445
• Type: Invited
• Date/Time: Wednesday, August 4, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Consulting
• Author(s): Meijuan Li*+ and Estelle Russek-Cohen and Rong Tang
• Companies: FDA and FDA and FDA
• Keywords: Biomarker test inaccuracy ; targeted therapy ; clinical trial ; personalized medicine

Abstract

Recent advances in technology have enabled examination of novel biomarkers and ignited interest in using biomarkers to guide treatments. Some have proposed evaluating drug and marker efficacy by assessing marker by treatment interaction. In these trials, the underlying tests are often assumed to have perfect diagnostic accuracy. This assumption rarely holds in reality and the inability of a test to accurately separate marker positive and negative populations may impact the success of a marker-treatment interaction trial. This talk aims to evaluate the impact of inaccuracy of biomarker test on such trials and simulate further discussion on this issue. We will propose a Bayesian statistical method for evaluating drug efficacy that accounts for error in biomarker test. Using simulation data, we will compare the accuracy and precision between the proposed model and the existing method.