A feasibility study: Can information collected to classify for mutagenicity be informative in predicting carcinogenicity?

Abstract Carcinogenicityis a complex endpoint of high concern yet the rodent bioassay still used is costly to run in terms of time, money and animals. Therefore carcinogenicityhas been the subject of many different efforts to both develop short- term testsand non-testing approaches capable of predicting genotoxic carcinogenicpotential. In our previous publication (Mekenyan et al., 2012) we presented an in vitro – in vivo extrapolation workflow to help investigate the differences between in vitro and in vivo genotoxicitytests. The outcomes facilitated the development of new (Q)SAR models and for directing testing. Here we have refined this workflow by grouping specific tests together on the basis of their ability to detect DNA and/or protein damage at different levels of biological organization. This revised workflow, akin to an Integrated Approach to Testing and Assessment (IATA) informed by mechanistic understanding was helpful in rationalizing inconsistent study outcomes and categorizing a test set of carcinogenswith mutagenicity data on the basis of regulatory mutagenicity classifications. Rodent genotoxic carcinogenswere found to be correctly predicted with a high sensitivity (90–100%) and a low rate of false positives (3–10%). The insights derived are useful to consider when developing future (non-)testing approaches to address regulatory purposes.