Radiologists who have access to patient exposure histories will make recommendations for future imaging that take into account previously incurred imaging risks, even though models of radiation exposure and cancer risk indicate previous imaging should not affect decision making, according to a study published in the June issue of the American Journal of Roentgenology.

“Our findings raise concern that many radiologists have a limited understanding of how to address patient exposure histories when making prospective imaging decisions and underscore the need for related educational interventions,” wrote Pari V. Pandharipande, MD, MPH, of Massachusetts General Hospital in Boston, and colleagues.

At the center of the study was the radiologists’ acceptance of the linear no-threshold model, endorsed by the National Academy of Sciences’ Biological Effects of Ionizing Radiation committee. While counterintuitive, this model, when applied to cumulative exposures, implies that previously incurred radiation risks should not factor into prospective imaging decisions, explained the authors. This is because a new exposure confers the same risk regardless of whether the patient has had a prior exposure, and there’s no threshold that triggers a disproportionate increase in cancer risk.

“Therefore, when a physician considers ordering a CT for a patient with an exposure history, the physician’s risk-benefit analysis should include only risks from the CT under consideration; previously incurred risks cannot be mitigated and should not affect decision making,” wrote Pandharipande and colleagues.

The study assessed radiologists’ understanding and acceptance of this concept through a multicenter survey in which radiologists were asked to make recommendations for hypothetical patients with a history of multiple CT scans. The survey then questioned the factors that went into the decision making process. A total of 322 radiologists completed the survey.

More than 90 percent of respondents incorporated risks from the exposure history during decision making, according to the authors. This was despite the fact that 61 percent reported acceptance of the linear no-threshold model.

“Although [incorporating exposure history risks] is not consistent with the linear no-threshold model, most radiologists think that it is,” wrote Pandharipande and colleagues. They added that whether a radiologist accepted or rejected the model, only 36 percent made decisions that would be consistent with their beliefs.

Findings were consistent across institutions and training levels, according to the authors.

Pandharipande and colleagues emphasized that knowledge of prior imaging can be valuable, and gave the example of a patient who had undergone multiple negative CT studies within a month, indicating subsequent scanning would not have a favorable risk-to-benefit ratio.

“If radiologists, as well as the general medical community, operate using the most common decision-making processes evident in our study, an inappropriate shift toward less beneficial imaging tests may occur, creating the potential for substantial patient harms,” concluded the authors. “It is critically important that educational interventions in this area become a priority now, before the widespread availability of patient-level exposure histories.”