Learning Objectives
• Understand the origin and assumptions of the LNT model of radiation injury
• Appreciate the statistical uncertainty underlying the concept of Lifetime Attributable Risk
• Recognize the fragility of estimates of radiation-induced cancers in a population

Abstract
Recently articles have appeared in the scientific literature and popular press describing thousands of cancers and cancer deaths induced in the US population annually by radiation from medical imaging procedures. The numbers of cancers and cancer deaths are presented without any of the substantial uncertainties underlying their generation. These uncertainties include the application of the linear no-threshold model of radiation injury to doses of a few mSv, adoption of the idea of effective dose for imaging procedures, use of the concept of Lifetime Attributable Risk with its broad statistical variability as a firm quantitative estimate of risk, and the practice of multiplying large population numbers by very small risk estimates to yield hypothetical numbers of individuals at risk in the population. Collectively, these uncertainties raise serious doubts about the validity of quantitative estimates of cancers and cancer deaths caused by medical imaging.