The overrated role of 'promotion' in mechanistic modelling of radiation carcinogenesis.

It is argued that the 'cellular' dose-response relationships of model parameters cannot be determined from mechanistic model fits to experimental or epidemiologic cancer data. Baseline population cancer incidence data show conclusions about intermediate cell kinetics to be especially questionable. Here we recommend that mechanistic models should be applied solely in one direction: by starting from known cellular dose-response relationships the models aim at providing a biologically motivated consistent description of the development of radiation-induced cancer for different exposures, which is very important for low-dose risk estimation.