Perspectives in low dose risk estimation

Currently, debate ranges from whether very low doses have disproportionately large consequences, to whether there is a practical threshold. Whilst epidemiology forlow doses can achieve somewhathigherpowerfrom conflation of results, the uncertainties are such that advances in radiobiology are essential. Conceptually, deviations from low-dose linearity can arise from changes in the cell, or in neighbouring cells, such that subsequent insults from irradiation or other carcinogens are more likely (sensitisation) or less likely (desensitisation) to result in cancer. Such feedback changes are common in living organisms (allergies and homeostatic stabilisation). Bystander effects indicate that neighbouring cells may be influenced. The effects of conditioning doses at high doses indicate that in some circumstances cells can retain a memory of earlier damage. What is measured in epidemiology is the net excess of cancers from irradiation above the much higher rates of cancer from environmental and spontaneous causes. Epidemiological dose-response curves are 'dressed'. Memory and group influences provide mechanisms which enable the 'naked' dose-response curve for the cell alone to be clothed. Junk DNA and anonymous carcinogens are also discussed.