Biomarker Discovery with Highly Left-Censored Multiplex Immunoassay Data

• Activity Number: 7
• Type: Invited
• Date/Time: Sunday, August 9, 2015 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Elizabeth Hill* and Elizabeth Slate
• Companies: Medical University of South Carolina and Florida State University
• Keywords: Left-censoring ; Bayesian hierarchical model ; Biomarker discovery ; Multiplex immunoassay

Abstract

Multiplex immunoassays (MIAs) are moderate-throughput platforms for simultaneous quantitation of a panel of analytes, and increasingly are popular for targeted biomarker identification. Often little is known about analytes' expected concentrations in samples derived from the target population. As a consequence, MIA data can be plagued by high proportions of concentrations flagged either as 'out-of-range' - samples for which the observed response falls below the lower asymptote of a 5-parameter logistic calibration curve - or as extrapolated beyond the smallest standard concentration. We present a unified analysis approach in the context of a Bayesian hierarchical model that accounts for calibration error induced by background estimation and standard curve fitting, and that models subjects' observed fluorescence values as a function of unobserved (latent) analyte concentrations, with accommodation for left-censoring via variance function specification. We compare our approach to conventional analysis methods and apply our method to a cytokine and chemokine array analysis of serum specimens from HPV-associated and non-associated head and neck cancer patients.