Statistical aspects of extrapolation of dichotomous dose-response data.

A mathematical model based on the hypothesis that carcinogenesis is a multistage process was proposed for the statistical problem of extrapolating the results of bioassay experiments performed at high dose levels of the carcinogenic risk at lower dose levels. This general multistage model had extrapolation characteristic similar to those of the more specific models that are commonly used; a graphic technique based on statistical likelihood procedures provided a measure of the uncertainty in this extrapolation problem. The examples used were 1) exposure to DES and mammary carcinoma and 2) exposure to DMN and liver carcinoma.