Based on studies of the irradiation of animals with neutrons a linear dose-response relationship was observed for the induction of most tumors at doses of 0.0 to 0.2 Gy; it was followed by a plateau or bending over of the curve at higher doses. Reducing the dose rate either had no effect on the dose-response relationship in the low-dose range or, in some instances, it increased the response per unit of dose. The differences in shape of the dose-response curve for cancer induction by gamma rays and neutrons resulted in the assignment of rather high relative-biologic-effectiveness (RBE) values for cancer induction at low doses. All the above data are consistent with biophysical models of radiation effects applicable to a variety of other end points, including radiation-induced cell-killing, induction of chromosomal aberrations, and radiation-induced mutation. These models predicted linear-quadratic dose-response relationships and reduced effectiveness per unit dose of low-LET radiation at low doses and low dose rates (Kellerer and Rossi 1972; Ullrich and Storer 1979).
Because of their consistency with projections from biophysical models of radiation effects, the combination of dose response and dose-rate data for tumor induction obtained from animal studies and data on various end points in animal and human cells provide substantial support for the application of a dose and dose-rate effectiveness factor (DDREF) in the estimation of cancer risks in human populations at low doses and low dose rates (UNSCEAR 1988; NRC 1990; ICRP 1991).
The high RBEs for neutrons at low doses (also predicted on the basis of biophysical models) observed in animal studies was important in the modification of quality factors used in risk estimates for neutrons (ICRP 1963). The neutron data are also likely to be important in the future analysis of data on atomic-bomb survivors, inasmuch as a portion of the dose that they received was from neutrons, but the contribution is still being evaluated.
After analysis of the results of long-term studies, it was recognized that understanding of radiation risks at low doses would not be improved by attempting to measure the effect at low doses on animals, but rather would require a better understanding of the underlying mechanisms. As a result, experimental studies of carcinogenesis since the last BEIR report have focused on mechanisms and on the cellular and molecular events involved in neoplasia. Over this time, the understanding of molecular events involved in the carcinogenesis process, in general, has increased dramatically. It is now clear that cancer development entails alterations in multiple genes that are involved in the regulation of progression through the cell cycle, cell growth and differentiation, and cell death, and in genes that are involved in the maintenance of genomic fidelity. A number of investigators have now demonstrated that alterations in genes that control genomic fidelity can play a major role in the early events leading to cancer by conferring a mutator phenotype on the affected cells (Loeb 1991, 1997). Cells with alterations in other critical genes later arise as a result of clonal selection.