Long-term effects of low-level radiation exposure, experimental studies

Important considerations in the quantitative estimate of risk of carcinogenesis in the human being are the related factors of the shape of the dose-effect curve, the risk per rad at very low doses (about 10 rad or less) vs. higher doses, and the effect per rad of higher doses delivered at very low dose rates. Data on the human being, though extensive and adequate for quantitative risk estimation at high doses and dose rates, do not allow definitive conclusions on these factors. Data on eukaryotic cells are represented well and in detail by a modified quadratic relationship with a definite linear term, and indicate that both the simple linear and the pure quadratic relationships are rare or essentially nonexistent. A dose rate effect is ubiquitous, but the extent varies widely among different tumors and species. Extensive data on carcinogenesis in mammals are in general represented well by the linear-quadratic model, although simple linearity and the pure quadratic (or threshold) models cannot be excluded in a few tumor systems. Considerations and uncertainties in addition to curve shape and dose rate may tend to offset any credit that might be afforded by dose rate. Hence, while it is almost certain that dose rate does at least to some degree affect (reduce) the current risk estimated based on linearity, it also seems clear that, at least for some time, current carcinogenesis risk estimates based on linearity probably will continue to be used as an approximation of the degree of risk associated with radiation exposure at very low doses and dose rates