NEJM: Do benefits outweigh radiation risk in medical imaging?
Reza Fazel, MD, from the division of cardiology at Emory University School of Medicine in Atlanta, and colleagues identified 952,420 subjects (mean age, 36 years old; 52.4 percent women) from UnitedHealthcare claims between Jan. 1, 2005 and Dec. 31, 2007. They focused on five healthcare markets: Arizona; Dallas; Orlando, Fla.; South Florida and Wisconsin, and examined claims from hospitals, outpatient facilities and physicians’ offices.
The researchers found 655,613 enrollees, or 68.8 percent, underwent at least one imaging procedure. CT and nuclear imaging accounted for 75.4 percent of the procedures, with 81.8 percent of these administered in outpatient settings.
Overall, Lauer questioned whether the benefits of medical imaging have been weighed against their potential “harms” -- cost and radiation exposure that could potentially lead to cancer. In an interview, Lauer, from the National Heart, Lung and Blood Institute, stopped short of saying that any diagnostic setting was overusing imaging.
“We don’t know if there is overuse or not, because the term ‘overuse’ implies that we know what the appropriate use is,” he said. “In order to understand the appropriate level of use, we need to know what value can be gained from the tests and, specifically, how much they improve health.”
Fazel and colleagues found that nearly 80 percent of the patients received an average of 2.4 mSv per year, which is equivalent to the natural background radiation that the average person absorbs in a year. Overall, moderate effective doses, defined as more than 3-20 mSv of radiation, were incurred by 19.38 percent of the enrollees per year. The researchers also reported that high (more than 20-50 mSv) and very high (more than 50 mSv) doses were incurred by 1.86 percent and .19 percent of the enrollees per year, respectively.
Although 80 percent of the study population received a low radiation dose, the remaining 20 percent is "disconcerting,” without proven benefits, he said.
John Lesser, MD, co-director of cardiac CT at the Minneapolis Heart Institute, said in an interview that these types of studies are helpful because they inform physicians about the radiation risks involved with the tests. He added that diagnostic techniques are “moving targets,” and therefore, difficult to assess as radiation doses are dropping with new technologies and techniques, especially in CT.
“We, as physicians, should lower the risk whenever possible, by lowering the radiation dose whenever we use these studies, and also assessing the individual patient based on their accumulated radiation dose,” Lesser stated.
Based on specific imaging procedures, Fazel and colleagues found that 20 procedures formed the largest contribution to the annual cumulative effective dose. The highest dose procedure was myocardial perfusion imaging, which accounted for more than 22 percent of the total effective dose.
Lauer noted in his perspective that myocardial perfusion scans increased by more than 6 percent between 1993 and 2001, with “no justification for their use.”
Though Lesser acknowledged that these high statistics may indicate an inappropriate ordering of nuclear imaging tests, he observed that physicians are only equipped with the best tools at their disposal to make the right decisions for their patients.
Fazel’s group also found that imaging procedures increased with advancing age and were more frequent for women than men. The authors noted that more than 40 percent of women in the study, compared with 30 percent of men, received doses exceeding 20 mSv. They wrote that due to the “related risks [that] accrue over a lifetime…cancer may be more likely to develop in women than in men after similar levels of exposure.”
In addition to this gender-specific consideration, Lauer said imaging protocols should be re-evaluated for all patient populations.
”The value of most imaging protocols in cardiovascular medicine is unknown. We don’t know if these tests save lives or prevent heart attacks. We need to conduct randomized, controlled trials to establish answers to these remaining questions,” he said.
Though mortality due to heart disease has been reduced by 34.3 percent between 1995 and 2005, according to American Heart Association statistics, Lauer said that there is no evidence to prove that this can be correlated to early detection through medical imaging tests.
However, he said that at least half of the mortality decrease can be ascribed to “standard prevention,” such as control of hypertension and reduction of high cholesterol. He added that there is “no evidence right now that imaging is playing a major role in the decline of coronary heart disease deaths.”
Lesser noted that such evidence is complicated to obtain. “When anyone attempts to prove that medical imaging improves longevity, it is very difficult because it is one step along the way in patient assessment and treatment, which is how it differs from procedures,” he noted. “However, imaging procedures allow the physician to make a diagnosis that may or may not necessarily improve mortality, such as the detection of coronary disease or the confirmation that coronary disease does not exist.”
Lesser concurred this type of proof can only emerge from a large, randomized trial that would require a great deal of time and money; however, it would effectively assess the use of diagnostic techniques leading to improved patient outcomes. In the meantime, he said, practitioners need to assess and treat their patients, and they often need imaging tests to bolster their diagnosis in real-life clinical practice.
“Because you can’t prove its benefits over risks through trials yet doesn’t mean physicians shouldn’t use imaging,” Lesser stated. “In various patient groups, I am forced to make a choice between lots of different diagnostic tests—not all radiation based.”
Lauer does not think that physicians should refer asymptomatic coronary artery disease patients for screening imaging tests, unless it is as part of a randomized trial. To support this, he cited the DIAD randomized trial of 1,000 asymptomatic patients with diabetes, which found that SPECT imaging did not reduce cardiac events. On the other hand, he stated that large, randomized trials have demonstrated that abdominal ultrasound can save lives.
Lauer cited the National Cancer Institute’s National Lung Screening Trial, funded by the National Institute of Health, as an example of such a randomized, controlled imaging trial. The investigators have randomized 50,000 asymptomatic, at-risk smokers to chest x-ray or CT screening scans, and are examining whether CT imaging can reduce deaths from lung cancer. This study is currently in its follow-up stage.
“We need the same type of trial for asymptomatic heart disease,” he stressed.
Lauer also cited mammography as an imaging procedure that has proven that its benefits justify its potential risks. A number of trials have clearly proven that mammography reduces breast cancer death rates, he said.
Until the emergence of large-scale randomized trial evidence is presented, Fazel and colleagues stated that strategies for “optimizing and ensuring appropriate use of these procedures in the general population should be developed.”