Biomarker Assay Data: How Well Quantified Is the Limit of Quantification?

• Activity Number: 300
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Ying Guo*+
• Companies: Merck
• Keywords: Assay data ; Biomarker ; Limit of quantification ; Measurement error ; Bayesian method

Abstract

Laboratory assays to measure biomarkers are often subject to a limit of quantification (LOQ), which is a threshold below which estimates of biomarker have high coefficients of variation. In most LOQ problems, researchers are given an estimated value for the biomarker if this value is greater than the LOQ, and a missing value (< LOQ) otherwise. We question the logic of the LOQ approach for reporting values subject to measurement error. The method makes an implicit assumption that there is no measurement error for values above the LOQ, and unacceptable level of measurement error for values below the LOQ. A more plausible assumption is that there is measurement error throughout the measure's support. We propose a Bayesian method for data subject to the LOQ, which yields prediction intervals for the true biomarker value throughout the range of assay measurement, and allows for heteroscedasticity of measurement error. We illustrate the proposed method on assay data of eight fat-soluble vitamins. Our findings suggest that the current approach to transmit data from calibration assays is flawed, since it provides a distorted picture of the actual measurement error profile.