Radiation risks of medical imaging: separating fact from fantasy.

During the past few years, several articles have appeared in the scientific literature that predict thousands of cancers and cancer deaths per year in the U.S. population caused by medical imaging procedures that use ionizing radiation. These predictions are computed by multiplying small and highly speculative risk factors by large populations of patients to yield impressive numbers of "cancer victims." The risk factors are acquired from the Biological Effects of Ionizing Radiation (BEIR) VII report without attention to the caveats about their use presented in the BEIR VII report. The principal data source for the risk factors is the ongoing study of survivors of the Japanese atomic explosions, a population of individuals that is greatly different from patients undergoing imaging procedures. For the purpose of risk estimation, doses to patients are converted to effective doses, even though the International Commission on Radiological Protection warns against the use of effective dose for epidemiologic studies or for estimation of individual risks. To extrapolate cancer incidence to doses of a few millisieverts from data greater than 100 mSv, a linear no-threshold model is used, even though substantial radiobiological and human exposure data imply that it is not an appropriate model. The predictions of cancers and cancer deaths are sensationalized in electronic and print public media, resulting in anxiety and fear about medical imaging among patients and parents. Not infrequently, patients are anxious about a scheduled imaging procedure because of articles they have read in the public media. In some cases, medical imaging examinations may be delayed or deferred as a consequence, resulting in a much greater risk to patients than that associated with imaging examinations. © RSNA, 2012.