Risks of radiation-induced cancer at high doses and dose rates

Abstract

The Armitage-Doll model of carcinogenesis is fitted to the Japanese bomb survivors with the DS86 dosimetry. It is found to provide an adequate description of solid cancer incidence (but not that of leukaemia) as a function of radiation dose when up to two radiation-affected stages are assumed; the optimal model for solid cancers is one in which there are two radiation-affected stages separated by two additional stages. Leukaemia is best modelled by assuming a single radiation-affected stage. The fit of the Armitage-Doll model is contrasted with that of a less biologically based (relative risk) model which is fitted jointly to the leukaemia data in the Japanese cohort and in the UK spondylitics, as well as to solid cancer data in the Japanese. The literature on exposure to high doses at high dose-rates from therapeutic irradiation is surveyed with a view to determining patterns for variation in risk by age at exposure, time since exposure and evidence for curvilinearity of dose-response. Reductions in relative risk with increasing age at exposure are documented for all cancer types.