The problem of thresholds in chemical carcinogenesis some views on theoretical and practical aspects.

I have tried to show that the threshold problem for carcinogens is still unsolved and that this unsatisfactory situation creates many problems for the scientific evaluation of the carcinogenesis problem as well as for its regulative aspects. In view of the extremely low acceptance of cancer risks in the general population, which is justified as well as understandable, carcinogenesis research must present new ideas to come nearer to a solution of such problems. More sound data on dose-time-effect curves in the low incidence range of 10-0.1% tumor incidences could indicated whether the established linearity in the higher incidence range might eventually by interrupted. Increase of animal numbers could also increase reliability of carcinogenicity experiments and allow the detection of low-dose effect; the disadvantages of this approach (“mega-mouse experiment”) however, in regard to cost and space necessary are evident. Increases of the mean life span by optimization of animal husbandry still would be another way of approach. However, no real breakthrough is to be expected here. The most promising approach seems to be the use of other “indicators” for carcinogenicity instead of tumor formation. Binding studies of carcinogens to biopolymers (Neumann 1980), especially DNA, as well as the determination of preneoplastic, enzyme-deficient islands (Kunz et al. 1978) have been shown to give dose-response curves parallel to the corresponding curves from carcinogenicity studies (Kunz) and extendable over 6 orders of magnitude of doses, without deviation from linearity (Neumann). Additional parameters might be found, which, altogether, might prove to be an important step toward further knowledge in this important field of carcinogenesis research.