You are here:
Comparing the strengths and limitations of concentration-response modeling pipelines for developmental neurotoxicity new approach methods (INA 2023)
Citation:
Carstens, K., J. Hsieh, A. Donmez, M. Scholze, K. Bartmann, T. Shafer, E. Fritsche, AND K. Paul-Friedman. Comparing the strengths and limitations of concentration-response modeling pipelines for developmental neurotoxicity new approach methods (INA 2023). International Neurotoxicology Association (INA) Meeting symposium May 2023, Durham, NC, May 21 - 25, 2023. https://doi.org/10.23645/epacomptox.23148260
Impact/Purpose:
Invited speaker presentation at the International Neurotoxicity Association (INA) conference in Durham in May 2023. The purpose of this presentation is to share and receive feedback on work comparing different concentration-response pipelines for developmental neurotoxicity new appraoch methods that are used by the US and European collaborators.
Description:
A battery of developmental neurotoxicity (DNT) new approach methodologies (NAMs) has been developed to evaluate key events in neurodevelopment, including proliferation, differentiation, apoptosis, neurite outgrowth, synaptogenesis, and neural network formation. The integration of in vitro assays, representing diverse neurodevelopmental processes, will be a critical next step for evaluating chemicals with DNT hazard potential. The DNT NAMs battery includes assays from different experimental platforms (e.g., microelectrode array and high-content imaging) or laboratories (e.g., US EPA and academic laboratories in the EU) that use different cell types and species. Additionally, assay developers rely on different dose-response modeling pipelines, such as the ToxCast Pipeline (tcpl) and CRStats, to estimate efficacy and potency. In this work, we evaluated the influence of different pipelines on model outcomes, including discrepancies in hit calling, cutoff thresholds, confidence intervals (CI), and benchmark concentrations (BMCs). For a set of 110 chemicals pipelined in tcpl and CRstats, hit calls were 95% concordant and BMCs were highly correlated (Pearson’s R=0.74) between the two pipelines. BMC CI widths were wider using CRstats compared to tcpl by a mean difference of 0.5 log10-µM, likely explained by differences in confidence levels (95% in CRstats versus 90% in tcpl). Moreover, different approaches to compute ‘selective’ bioactivity (activity below the threshold of cytotoxicity) were compared. The CRstats classification model was more stringent, identifying fewer selectively active hits compared to other approaches. Understanding strengths and limitations of dose-response modeling pipelines will be integral for informing biological interpretation of DNT NAMs and application for decision-making in the regulatory process for DNT. Disclaimer: This abstract does not necessarily reflect U.S. EPA policy.
URLs/Downloads:
DOI: Comparing the strengths and limitations of concentration-response modeling pipelines for developmental neurotoxicity new approach methods (INA 2023)
PIPELINE_COMPARISON_INA_FOR_CLEARANCE2.PDF (PDF, NA pp, 2266.501 KB, about PDF)