ASNC INFORMATION STATEMENT Variability in radiation dose estimates from nuclear and computed tomography diagnostic imaging Edward P. Ficaro, PhD, a Pat Zanzonico, PhD, b Michael G. Stabin, PhD, c Gilbert L. Raff, MD, d Randall C. Thompson, MD, e Andrew J. Einstein, MD, PhD, f Milena J. Henzlova, MD, g Matthew J. Budoff, MD, h Vasken Dilsizian, MD, i Warren K. Laskey, MD, j Joao Lima, MD, k Joseph L. Roti Roti, PhD, l and Timothy M. Bateman, MD e INTRODUCTION Radiation risk to both the patient and the imaging personnel from medical imaging procedures involving the use of ionizing radiation is a point of concern and controversy in the diagnostic imaging community. With the increased utilization of computed tomography (CT) imaging both standalone and in conjunction with nuclear tomographic imaging (positron emission tomography [PET] and single-photon emission computed tomogra- phy [SPECT]), radiation risk has received significant attention both scientifically and in the public media. Several studies are reporting significant increased risk to the patient receiving low ionizing radiation from CT and nuclear studies. 1-3 These presentations lead to many questions by the imaging community on what is best for their patients, and some imaging centers are using tab- ulated dosimetry values to help determine the imaging protocol without a clear understanding of relative cer- tainty in the tabulated values. The presentation of risk data in ‘‘cancer units’’ has generated concern and questions from patients, referring physicians, and diagnostic imaging professionals. The expression of radiation dose in units of cancer risk is misleading for a number of reasons. First, radiation dosimetry values for radiopharmaceuticals are in general subject to considerable uncertainty, 4,5 as they are based on limited biokinetic data (taken from a handful of patients or from animal data) and are derived for ‘refer- ence’ (median) individuals (adult males, females, etc.). Doses to individual patients have significant variability due to the stochastic nature of radiation dosimetry. Sec- ond, while average dosimetry or cancer risk values can be expressed for a population, the absolute, incremental cancer risk value for any given individual from a given radiation exposure (e.g., due to a myocardial perfusion SPECT study) may not be derived from these population models. Furthermore, the confidence limits for the dose and risk estimates are not defined, and thus individual risk values should not be reported as a deterministic value with known confidence limits. Also of importance is the uncertainty in converting radiation dosimetry to cancer risk. The extrapolation of radiation dosimetry to cancer risk is based on the linear, no-threshold (LNT) hypothesis. This hypothesis is the subject of considerable controversy as there are no data available at the dose levels experienced in diagnostic medical imaging. The data for diagnostic imaging dose level are extrapolated from observations of cancer incidence from survivors of the atomic bomb detona- tions and other populations exposed to high doses and dose rates. The LNT works well in defining conservative guidelines for the safe use of radiation and other forms of policy setting. However, its use to predict cancer incidence or deaths at low doses and dose rates remains a matter of considerable scientific discussion and con- troversy. Reporting of numerical estimates of risk associated with individual dose levels associated with particular nuclear medicine and/or CT exams represents a use of the LNT model that is inappropriate given our current understanding of the risks associated with low Unless reaffirmed, retired, or amended by express action of the Board of Directors of the American Society of Nuclear Cardiology, this Information Statement shall expire as of January 2014. From the University of Michigan, a Ann Arbor, MI; Memorial Sloan- Kettering Cancer Center, b New York, NY; Vanderbilt University, c Nashville, TN; William Beaumont Hospital, d Royal Oak, MI; Mid America Heart Institute, e Kansas City, MO; Columbia University Medical Center, f New York, NY; Mount Sinai Medical Center, g New York, NY; Los Angeles Biomedical Research Institute at Harbor-UCLA, h Torrance, CA; University of Maryland Medical Center, i Baltimore, MD; University of New Mexico, j Albuquerque, NM; Johns Hopkins Hospital, k Baltimore, MD; Washington Uni- versity School of Medicine, l St. Louis, MO. Reprint requests: Edward P. Ficaro, PhD, University of Michigan, Ann Arbor, MI. 1071-3581/$34.00 Copyright Ó 2008 by the American Society of Nuclear Cardiology. doi:10.1007/s12350-008-9026-0