ABSTRACT
Risks from radiation of inducing cancer are derived from epidemiological studies of Japanese citizens who survived the Hiroshima and Nagasaki atomic bombs. These show higher incidences of cancer in certain tissues and among younger individuals. Once adjustments have been made for sex and age at exposure, epidemiological results show linear relationships between cancer incidence and organ dose between 0.1–0.2 and 4 Sv, but do not satisfy statistical criteria for such a relationship at lower doses. Two models have been developed to extrapolate risks to lower doses, one based on excess absolute risk (EAR) for each tissue and the other excess relative risk (ERR) with a modification to take prevalence of cancer incidence in the target population into account. Predictions of cancer incidence from calculations by ICRP 103, BEIR VII, and RadRAT are compared. These use slightly different epidemiological data, apply alternative weightings for EAR and ERR, derive results for different populations, and use different factors to correct results for application to low doses. Cancer incidences derived for individual organs differ by factors of 2–4, but the overall risks from whole-body exposures are within a factor of 2. Since uncertainties are large, results of calculations must be used with caution.
