Cancer risk assessment at the crossroads: the need to turn to a biological approach.

Mathematically based carcinogen risk assessment is based on a number of prudent default assumptions which are becoming progressively less tenable as new scientific evidence is adduced. For example, the assumptions that all rodent carcinogens will be carcinogenic in humans and that there is no safe dose of any carcinogen may, in specific examples, be shown to be untrue. The mechanisms by which carcinogens exert their effects, especially the induction of DNA lesions, DNA repair of these lesions, and cell proliferation, are considered; it is suggested that with recently developed experimental techniques they might be employed to develop a more biologically based approach to risk assessment and might avoid at least, some of the pitfalls associated with the present mathematically based carcinogen risk assessment models. They might lead to an improved appreciation of the shape of the carcinogen dose-response curve, at least at medium to high exposure levels.