CHICAGO--A debate about radiation dose ended with a high five when an expert panel reached consensus about radiation dose, agreeing that CT scans should be limited to justified and optimized studies, during the Radiation Dose: Can it be Too Low session on Dec. 1 at the 96th annual scientific meeting of the Radiological Society of North America (RSNA).
Based on a debate format, the session asked four experts to argue the point—cancer induction should not be a consideration in ordering a medical imaging study.
Beyond mathematical models—back to basics
The medical community should stop arguing about numbers and risk estimates, argued Cynthia S. McCollough, PhD, professor of radiologic physics at Mayo Clinic in Rochester, Minn. She pointed out that it has not been scientifically demonstrated that any cancer risk exists below 100 mSv.
Existing risk models based on atomic bomb survivors are inconsistent with radiation biology. The heavily promulgated linear no threshold model on which most studies are based uses the assumption that response is proportional to dose and overlooks cellular repair mechanisms. Thus, it is not a sound platform for making risk estimates. “If a risk does exist [at low levels], it is too small to calculate,” she stated.
Radiologists and referring physicians have simple responsibilities to protect patients from excessive dose that do not include offering risk estimates or dose calculations, according to McCollough. The ordering physician needs to ensure that a scan is justified, and the radiologist needs to optimize the scan.
The flip side
“CT has revolutionized medical practice,” offered David A. Brenner, PhD, director of the Center for Radiological Research at Columbia University Medical Center in New York City. “We can’t say with certainty that there are any risks at all, but the evidence behooves us to behave as if there are. The issue is to optimize imaging decisions,” he continued.
Brenner argued that the relevant organ dose range is 5 to 100 mSv, and that existing studies based fairly directly on low-dose epidemiological evidence do show an increased cancer risk in that dose range. For example, atomic bomb survivors with an estimated radiation exposure in the 5 to 100 mSv range showed a small but statistically significant increase in cancer risk. And a study of 400,000 workers in the nuclear industry produced similar findings.
“Individual risks are small,” acknowledged Brenner, but given the prevalence of CT imaging, “population risks are not.”
Nonetheless, neither the atomic bomb nor nuclear industry study provides an ideal framework for estimating cancer risk from CT scans. Accurate risk estimates require direct epidemiological studies, which are just starting, said Brenner.
In the interim, as the medical community waits additional data, Brenner recommended that providers limit CT scanning to clinically necessary scans and appropriately lower dose. “For those CT scans that are clinically justified, the benefits far outweigh the individual risks. But there will be some radiation-dose induced cancers. Therefore, cancer induction should be a consideration in ordering and performing medical studies,” Brenner concluded.
McCollough poked holes in citations of risk, noting that studies of atomic bomb survivors in the 5 to 100 mSv range were averaged with most of the cohort falling in the higher ranges. The study of nuclear workers included questionable data from Canada. When those data are removed, the results are no longer statistically significant.
“We will always have error bars,” she explained. “The benefits of CT are real and known, and the risks are tiny and unknown.” Clinical studies may not provide the real answers about risk. That’s because accurate determination of risk represents a monumental undertaking. Finding the statistical risk of a 10 mSv CT requires matched cohorts of 5 million each, McCollough pointed out.
Instead of waiting for more data about risk, radiologists should temper the discussion and focus on the known benefits of CT. “The discussion shouldn’t be all about risk. The benefits need to be included when talking about risk,” stated John M. Boone, PhD, vice chair, research radiology, University of California, Davis.
Despite the consensus among panelists, an informal poll of the audience at the end of the session revealed that the issue was not resolved, with the audience evenly split on the questions of considering cancer risk in