The regression based defined approach for skin sensitization potency assessment based on NAM data from OECD accepted tests.

NAM for skin sensitization hazard assessment have been widely adopted and the research focus has moved to assess sensitization potency with these methods. Regression models for potency prediction with different combinations of input data and trained either vs. EC3 values from the Local Lymph Node Assay (LLNA) or vs. potency values combining EC3 data and human evidence were developed before. These models were combined into a fixed workflow to predict a single point-of-departure based on a given combination of available data. This Defined Approach (DA) is being validated by the OECD to be implemented in TG 497 in 2026. Here, this DA is described and predictivity is evaluated with data on 203 chemicals. In addition, data on reproducibility of the NAM-based potency predictions based on ring trial data are presented. Finally, special considerations for integrating and best using the kinetic DPRA in sensitization potency assessment are highlighted and solutions to technical limitations as provided in the updated standard operating procedure are discussed. This DA gives a fully transparent evaluation of the NAM data, and it makes best use of existing data, as it can directly use the data already generated for hazard identification and GSH subclassification.