Advances in high-throughput and high-content (HT/HC) methods such as those used in the fields of toxicogenomics, bioinformatics, and computational toxicology have the potential to improve both the efficiency and effectiveness of toxicity evaluations and risk assessments. However, prior to use, scientific confidence in these methods should be formally established. Traditional validation approaches that define relevance, reliability, sensitivity, and specificity, such as those in use by ICCVAM, may not be readily applicable. HT/HC methods are not exact replacements for in vivo testing, and although run individually, these assays are likely to be integrated together for decision making. Many of these assays rely on robotics, which may be unique in each laboratory setting. Building on the frameworks developed in the 2010 Institute of Medicine Report on Biomarkers and the OECD 2007 Report on (Q)SAR Validation, we present constructs that can be adapted to address the validation challenges of HT/HC methods. These constructs are both flexible and transparent, though require explicit specification of context and purpose of use such that scientific confidence (validation) can be defined to meet different regulatory applications. Transparency in the prediction model algorithms is a necessity to establish scientific confidence. Prediction models, filters, decision points, validation/performance data sets, and assay results should be publicly available and subjected to independent scientific peer review. We discuss some of the shortcomings of recently published models and recommend improvements, including how anchoring the assays and their prediction models to Adverse Outcome Pathways (AOPs) could facilitate the interpretation of results and support scientifically defensible fit-for-purpose applications.

Approaches for establishing scientific confidence in 21st century methods for toxicity evaluations

246th National Meeting (2013)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.