Scientific Evidence Relevant to the Assessment of Solid Cancer Radiation Risk at Low Dose and Low Dose Rate

Abstract

The current System of Radiological Protection uses a dose and dose rate effectiveness factor (DDREF) with a numerical value of 2 when applying estimates of radiation risk derived from high doses and dose rates to settings involving low doses and/or low dose rates. The concept combines the low dose effectiveness factor (LDEF) when interpolating estimates of risk across dose levels, and the dose rate effectiveness factor (DREF) when extrapolating risk estimates from studies involving populations exposed to high dose rates to those exposed to low dose rates. In this report the current scientific evidence on the biological and health effects at those doses and dose rates is reviewed, with emphasis on human solid cancer incidence and mortality. Numerical evaluations of both DREF and LDEF are considered from studies of somatic cell mutation, cell transformation and cytogenetic endpoints. Life-shortening and all solid cancers combined are evaluated from historical studies on experimental animals (mice). A meta-analysis is described where risk estimates deduced from 29 human cohorts exposed to low dose rates were compared with those from the atomic bomb survivors (to address DREF), and a reanalysis of the curvature in the mortality data from the Japanese atomic bomb survivors on all solid cancers combined (to address LDEF) is presented. Finally, mechanistically-based ways to combine biological evidence with epidemiological data are considered. While considerable uncertainties remain, the ranges of LDEF and DREF values obtained here are narrower than those obtained in previous evaluations, and are largely consistent amongst the various sources of data reviewed. The overall conclusion of this report is that, based on current scientific evidence, an LDEF of much greater than 3 are not supported, and much less than 1 likewise. Similarly, it is concluded that a DREF value much larger than 3 or less than 1 is also unlikely.