Abstract
The tumorigenic dose 50 (TD(50)) is a widely used measure of carcinogenic potency which has historically been used to determine acceptable intake limits for carcinogenic compounds. Although broadly used, the TD(50) model was not designed to account for important biological factors such as DNA repair and cell compensatory mechanisms, changes in absorption, etc., leading to the development of benchmark dose (BMD) approaches, which use more flexible dose-response models that are better able to account for these processes. Using a nitrosamine dataset as a case study, we compare the impact of moving to a BMD-based limit as opposed to a TD(50)-based limit. Although there are differences in individual potency estimates between the two approaches for some compounds, we show that the key metrics such as the 5(th) percentile of the respective potency distributions, used when calculating class-specific default acceptable intakes, are not greatly affected. Furthermore, potency estimates for nitrosamine compounds relevant to read-across do not vary by more than a factor of 3, which is little in the context of the inherent variability in a biological response, in an overall landscape wherein potencies can vary by four orders of magnitude. This suggests a move to BMD-based limits is achievable without significant disruption to existing limits while utilising a more robust methodology.