Epistemological problems in assessing cancer risks at low radiation doses.

Abstract

Historically, biology has not been subjected to any epistemological analysis as has been the case with mathematics and physics. Our knowledge of the effects in biological systems of various stimuli proves to be dualistic in a complementary (although not mutually exclusive) way, which bears resemblance to the knowledge of phenomena in quantum physics. The dualistic "limbs" of biological knowledge are the action of stimuli and the response of the exposed, biological system. With regard to radiogenic cancer, this corresponds to the action of the ionizations and the response of the exposed mammal to that action, respectively. The following conclusions can be drawn from the present analysis: Predictions as to radiogenic cancer seem often if not always to have neglected the response variability (variations in radiosensitivity) in individuals or among individuals in populations, i.e. the predictions have been based exclusively on radiation doses and exposure conditions. The exposed individual or population, however, must be considered an "open statistical system", i.e. a system in which predictions as to the effect of an agent are only conditionally possible. The knowledge is inverse to the size of the dose or concentration of the active agent. On epistemological grounds, we can not gain knowledge about the carcinogenic capacity of very low (non-dominant) radiation doses. Based on the same principle, we can not predict cancer risks at very low (non-dominant) radiation doses merely on the basis of models, or otherwise interpolated or extrapolated high-dose effects, observed under special exposure conditions.