Development of a biomarker decision support system.
Savage-RE Jr.; Maier-A; Haber-L; Hack-E; Lotz-WG; Schulte-P; Fowler-B
Proceedings of the American Association for Cancer Research (AACR) 96th Annual Meeting, April 16-20, 2005, Anaheim/Orange County, CA, Abstract 2209. Philadelphia, PA: American Association for Cancer Research, 2005 Apr; 46:189
For over two decades, scientists have been touting the importance, and ultimate application of biomarkers in reducing disease and protecting individuals from the harmful effects of exposure to occupational and/or environmental chemicals. However, few scientists apply stringent or recommended criteria to biological end points before proclaiming them biomarkers. While established guidelines for biomarker validation exist, methods for their implementation and case studies testing the methods are rare. This pilot study seeks to develop and demonstrate the use of a framework for integrating complex and multifaceted data, validating biomarkers, and incorporation of the biomarkers into an occupational risk assessment. The objectives are 1) to identify an occupationally relevant case-study chemical; 2) to develop a structure for a biomarker database that can be used to organize the diverse types of data; 3) to use Bayesian analysis and regression techniques to test and validate (or discount) biomarkers along the entire exposure-disease continuum and 4) use the biomarker database information to develop a risk assessment and compare that with a risk assessment developed using historical, traditional methodologies. A survey was developed and disseminated to 59 occupational safety and health professionals to help identify an occupationally relevant, data rich, case study compound. The Biomarker framework includes multiple steps. The data are binned into one of the exposure-disease categories (e.g., exposure, internal dose, effective dose, early effects, mild/moderate effects, or severe effects). The data are ranked based on relevance, study quality, and continuity (e.g., exposure and gene expression and disease data), and the highest ranking data are reformatted and entered into a database. A Bayesian belief network is used to validate the biomarkers by analyzing the strength of the dependencies between exposure, the potential biomarkers, and disease. Regression analysis is used to generate dose-biomarker relationships that are examined to confirm or reject biomarkers. The framework lays out an approach to consider a variety of biomarkers from the exposure-disease continuum for the enhancement of occupational risk assessment.
Biomarkers; Risk-analysis; Statistical-analysis; Questionnaires
Conference/Symposia Proceedings; Abstract
Proceedings of the American Association for Cancer Research (AACR) 96th Annual Meeting, April 16-20, 2005, Anaheim/Orange County, CA