Regulatory acceptance of in silico approaches for the safety assessment of cosmetic-related substances

Abstract In silico approaches are becoming increasingly important tools in toxicology. This is especially the case for the cosmetics industry, where growing numbers of regions are implementing bans on the in vivo testing of ingredients and products, resulting in an increased need to be able to generate safety data in vitro and/or in silico. ‘In silico’ data include results obtained from (quantitative) structure activity relationship ((Q)SAR) models, chemical categories, grouping, read-across and physiologically-based (pharmaco)kinetic (PB(P)K) models, and ‘big data’ analysis. There are general and more specific factors to be considered when using different in silico methods, many of which are described in the multitude of guidance documents available to support their use for regulatory purposes. There are particular challenges associated with the use of in silico approaches for cosmetic safety assessment, but with proper consideration and justification as to how they have been addressed, none should be insurmountable. Whilst in silico methods are currently predominantly used for internal rather than regulatory decision making, as confidence grows in their applicability and predictivity, this is likely to change. Many of the most commonly cited issues with the use of in silico approaches relate to the need for better documentation and justification as to why the method used was appropriate. It is also frequently stated that they should be used within a weight of evidence (WoE) approach and with all available data, rather than meeting regulatory requirements as a stand-alone method. Whilst these considerations are important, it should be remembered that they are equally applicable to data generated using in vitro and in vivo methods.