Evaluating the ability of defined approaches to predict the human skin sensitisation potential of chemicals previously untested in new approach methodologies

Much progress has been made in the development and validation of New Approach Methodologies (NAMs) for assessing skin sensitisation, as part of the global move away from animal testing. While there are now numerous in silico models and in chemico/in vitro assays available, none are currently thought to be a one-for-one replacement for the animal tests, but rather several NAMs are combined within a Defined Approach (DA), such as those described in OECD guideline 497. In this study, 22 chemicals were chosen which have known human sensitisation potential, but which have not previously been tested in NAMs, to the best of the authors’ knowledge. New in chemico/in vitro data were generated for each chemical in three assays, and this was combined with in silico predictions from two models to generate predictions from four DAs. The data was used to assess the performance of the individual NAMs and DAs within a less well understood area of chemical space, and to learn more about their applicability domains. The newly generated data are made available herein in the expectation that they will be useful to others who are developing and/or validating DAs which assess the risk of chemicals causing human skin sensitisation.