Update on linear non-threshold dose-response model and implications for diagnostic radiology procedures.

Cancer risk estimates are used in the setting of radiation protection standards by international and national organizations, and for this purpose need to be developed for low doses of radiation. The approach has involved extrapolation from cancer mortality and incidence values at higher doses to predict the low-dose estimates. Such an extrapolation has generally involved the use of the linear non-threshold (LNT) theory. Recent reports from the National Research Council (BEIR VII) and the International Commission on Radiological Protection (ICRP) have considered the appropriateness of the use of LNT for the purposes of radiation protection standard setting. The overall conclusion from both committees was that current scientific evidence remains consistent with the LNT hypothesis, while appreciating that this might not rule out the possibility that other extrapolation models might well be valid but require further evaluation and additional research to establish their validity. The dose and dose-rate effectiveness factor (DDREF) is used for adjustment in the extrapolation from high to low doses and from high to low dose rates. The BEIR VII committee proposed a new Bayesian approach for estimating DDREF and concluded that a value of 1.5 best fit the data. This is a departure from the previously used value of 2, which is still proposed by ICRP in its most recent recommendations. The current cancer risk estimation process as utilized by ICRP and BEIR VII is used here to assess the potential risks from annual whole-body computed tomography (CT) screens using information and an approach published by Brenner and Ellington. The major conclusion is that potential radiation risks need to be considered along with the pros and cons of the detection limits of the procedure and the impact of false positives.