Discussing Radiation Risks Associated With CT Scans With Patients

Technological advancements in diagnostic and therapeutic imaging have greatly increased physicians' ability to care for patients. As a result, use of medical imaging has increased exponentially. For example, Prokop¹ estimated that since the 1980s, use of computed tomography (CT) has doubled almost every 2 years. The increase in utilization has led to a parallel increase in concern regarding radiation risks. Almost half (48%) of the total dose of ionizing radiation exposure for individuals in the United States (including background radiation) has been attributed to medical tests and procedures.¹ It is estimated that use of CT may be associated with 1.5% to 2% of all cancers in the United States in the future.²

There is evidence suggesting that too many imaging examinations are being performed. Potential causes of overuse include fear of litigation, miscommunication, and self-referral among non–radiologist-owned imaging clinics.³ Also, there is a general lack of awareness regarding radiation risks among both health care workers and patients.⁴ For example, one study suggested that approximately 92% to 95% of patients are not informed of any radiation risks prior to their CT scan.⁴

A number of models can be used to estimate the risks of ionizing radiation. The dominant model was created by the Biological Effects of Ionizing Radiation (BEIR) subcommittee of the US National Institute of Science.⁵ Although much of the data on which the evolving model is based originate from atomic bomb data, information and data from other sources also have been included. According to the BEIR subcommittee, the most accurate current model is the linear and nonthreshold model, which predicts that the risk of outcomes (primarily cancer) is directly proportional to the radiation dose received, is additive, and does not respect a minimum risk threshold.

Smith-Bindman et al⁶ measured the radiation doses associated with 11 types of CT studies performed in 1119 adults in the San Francisco, California, area. Using data from the BEIR VII report, the authors derived cancer risk estimates from these CT exposures. For example, a 20-year-old woman undergoing a CT to rule out pulmonary embolism would have an estimated risk of developing cancer of approximately 1 in 330, compared with 1 in 880 in a 20-year-old man.⁶ A routine abdominal CT with contrast in a 40-year-old woman is associated with an estimated risk of developing cancer of approximately 1 in 870, compared with 1 in 942 in a 40-year-old man.⁶ The risk decreases with lower dose, older age, and male sex.

Several initiatives have been developed to reduce radiation risk. These include the ALARA (as low as reasonably achievable) principle, the Image Gently Alliance campaign, dose-reduction protocols, the evidence-based appropriateness criteria and the National Radiology Data Registry (both established by the American College of Radiology), and the Dose Check initiative of the Medical Imaging and Technology Alliance. However, a critical component missing from this effort is the mandatory dissemination of radiation risk information to patients.

Courts have declared that any risk from an investigation or treatment that a reasonable person would want to know to determine whether to undergo a proposed procedure must be communicated.⁷ This is called the “prudent person standard.”⁸ Another definition for the disclosure threshold is a risk that is higher than the background risk of carrying out normal daily activities.

Because the estimated risk of radiation-induced cancer from a typical abdominal CT scan is approximately 1 in 300 to 1 in 2000 depending on dose, age, sex, and body part scanned, it is clear that this information must be provided to patients.

Who should be responsible for providing this information to patients? On one hand, the clinician requesting the study presumably would have the most complete clinical information, may know how many previous CT scans the patient has undergone, and, therefore, in the majority of circumstances, would be most appropriate for obtaining consent. On the other hand, the radiologist has the most information about the specific radiation risks and, therefore, may be most appropriate for obtaining consent. However, radiologists may object because they currently feel vulnerable in that they generally cannot affect the performance of imaging examinations once patients' clinicians have ordered them.

Another concern is the potential for patients to misinterpret radiation risks. However, the concept of “therapeutic privilege,” wherein physicians were assumed to know best and make health decisions for their patients, was discredited decades ago. Physicians must provide patients with the necessary information for a given treatment, including benefits, risks, and alternatives, so patients can make an informed decision.

A corollary is the concern that some patients may refuse a potentially lifesaving examination because of fear of radiation. The counterargument is that physicians must explain the potential benefits of the imaging test to patients and explain why they believe the test is indicated. If a patient still refuses the examination at that point, it is ultimately his or her choice. This is no different from disclosure of the risks of anesthesia or surgery. Evidence exists that explaining the risk of radiation will not dissuade patients from undergoing the examination, even when the risk is explained to parents of the most radiation-sensitive age group—children.⁹

Others will point out that these risks are only estimates and great uncertainty remains. However, many radiologists may experience cognitive dissonance when confronted with these numbers. There is justifiable reason to be skeptical whether radiation at low doses increases the risk of cancer. However, the BEIR VII subcommittee had access to all competing models and its model is arguably the best available at this time; therefore, it should be used.

Twenty years from now, it will not be a valid argument that patients were not informed of the risks from radiation because physicians were operating under a world of uncertainty. The final counterargument will be one of the time and resources required to obtain informed consent. Clearly, that argument (ie, “I didn't tell you the risks because it would take too much time.”) will not prevail.

The medical profession cannot rely on industry, government, or the legal system to solve the dilemma of whether and how to discuss radiation risk with patients. A program of mandatory informed consent is needed for imaging examinations and procedures associated with relatively high levels of ionizing radiation.

The threshold for obtaining consent is debatable but, at the least, should include CT scans, many nuclear medicine examinations, and procedures involving a substantial amount of exposure from fluoroscopy, including many interventional radiology and cardiology procedures.

In addition to satisfying physicians' duty to warn patients of risks, this initiative may also have the benefit of helping to reduce the potentially substantial number of examinations and procedures that are not clinically indicated and to accelerate efforts to reduce radiation doses for these procedures.