Cancer Risk in Animals Exposed to Low Doses of Ionising Radiation

For the purpose of a global analysis of the effects of low radiation doses in mammals, the International Centre for Low Dose Radiation Research (ICLDRR) is collecting, in one database, all accessible information on cancer induction in animals following exposure to low doses and dose rates of ionizing radiation. A large number of these studies are identified and succinctly described in the International Radiobiology Archives of LongTerm Animal Studies (IRA), Vol. 1, Jul. 1996, Document US-DOE-RL-96-72. Data from studies published in the open literature since the publication of the IRA are being collected and incorporated into the database. A description of the database, of the methods used to analyse the data, and preliminary observations are presented, along with basic statistical analyses and dose-response model fitting conducted on the data accumulated at this time. In some studies, a high natural cancer incidence in controls makes it difficult to detect an increased risk of cancer in exposed animals, whereas protective effects cannot be detected in animals with a very low natural incidence of cancer. In the ensemble of data collected to date, the general observations are as follows. The Linear No-Threshold hypothesis (LNT) seems to overestimate the risk of cancer in animals for internally deposited alpha and beta emitters. The linear, probit, and logit models do not, generally, provide a good description of the relationship between dose and risk. In animals exposed to low doses and dose rates of ionising radiation, the three possible types of dose responses are observed: unambiguous reductions in cancer risk, clear risk increase, and insensitivity of cancer incidence to low dose radiation exposures. However, the apparently strong protective effect of low radiation doses seen in some of the experiments calls for further investigation and explanation.