Learning To Speak Clearly About Computed Tomography Safety

The enemy is among us, and it's sneaky. With every "unexpected" death due to heart attack, patients and physicians become more desperate for a "no fail" detector for the atherosclerotic disease that is the dominate killer in our communities. Into this hue and cry come amazing new imaging technologies, which promise better (if not fail-safe) early diagnosis. Beautiful pictures, no big needles--what could be better?

The problem, as we all know, is that screening asymptomatic patients with radiation-based imaging technology has not been proven to decrease cardiovascular mortality or to extend life. Given the recent Centers for Medicare & Medicaid Services' decision to pay for these examinations, it is unclear when we can expect definitive data from adequately sized, carefully performed, prospective studies to further clarify the value of cardiac computed tomography (CT) modalities and their most appropriate role among alternatives. In the meantime, direct-to-consumer advertising and aggressive marketing to physicians continue to push the use of these techniques into clinical gray zones where scientific evidence is sorely lacking. More and more patients will receive CT scans for a broader spectrum of indications and with variable impact on their care. All of this is clearly accompanied by a sobering truth: We are creating an increase in radiation exposure among the population and, with it, an expanding risk for malignancy.

Radiation safety issues have been important in our field for a long time, given our reliance on nuclear and angiographic methods. The amount of radiation exposure attributable to CT procedures for cardiac disease is rapidly increasing. A survey of ordering and performing physicians demonstrated very low rates of awareness of CT's radiation exposure and risk.[1] There have been many professional, governmental, and lay publications highlighting the degree of radiation exposure from CT scans of different territories and with different protocols [2,3], and patients and insurers are asking pointed questions about risk-benefit ratios. While insurers and professional organizations wrangle over utility and reimbursement, the individual care provider, faced with his/her own patient, finds his or herself with a difficult task: How do we put imaging's diagnostic role and radiation dose in context for our patients? How do we clearly, and without bias, communicate its risk?

In the Science Advisory, Ionizing Radiation in Cardiac Imaging, Dr. Thomas Gerber and his colleagues, representing the American Heart Association's Council on Clinical Cardiology and Council on Radiology and Intervention, take on the task of explaining how radiation exposure is measured and of explaining approaches to understanding the excess radiation to which we expose our patients, putting it into context in terms of long-term health. Their synopsis of the variety of methods that have been used to estimate the individual patient's exposure, serves to explain why these values have been a moving target. These measures have grown from parameters that were developed to reflect population-wide impact from major acute radiation exposure, which is dramatically different from the far lesser, but potentially repetitive, exposure that patients may receive from diagnostic imaging. The measures are dismayingly confusing and poorly standardized with respect to the individual and have been subject to significant modification in the past 20 years. Even when using the most current definitions for exposure, it remains problematic to predict the biological impact on the patient. It is humbling to review how limited our knowledge of the biology of the dose/risk continuum remains: Is it a linear function that can be extrapolated downwards from devastatingly high exposures to the much lower and parceled exposures experienced by most adults today?[4] Or, are there thresholds below which the risk is very small (the linear-quadratic relationship)? Other organizations such as the National Research Council of the National Academies [5] and the National Commission for Radiation Protection [6] have opted to refer to the linear model, as it is the most conservative position with respect to safety. Assuming that there is no level of inconsequential exposure, the burden is on all of us to make sure that exposure is kept as low as possible.

The argument has been made that the premalignant effects of radiation are likely to be delayed by decades, and therefore, patients of 60 years or older may not face the same lifetime increase in malignancy as younger patients.[7] Exposure of children and young adults is clearly worrisome, due both to their increased susceptibility to radiation effects and the long lives ahead of them. Not all malignancies will have a 30-year delay; however, there is a spectrum of time course of cancer versus exposure, with leukemia incidence increasing after as little as 5 to 10 years after an exposure. Our population is aging and living longer; many older adults may live long enough to experience the consequences of incremental risk of malignancy.

This advisory does an outstanding job of educating the non-expert on just how murky the waters remain with respect to the safety and value of the expanding use of these techniques. They call for better standardization and point, hopefully, to likely improvements in the near future. At the same time, the clinical role and reliability of these tests need to continue to be held up to high standards for both safety and clinical performance. We need to understand how best to use this information and how it complements or supplants other diagnostic methods. CT methods are not the only tool in the armamentarium and must be used only after consideration of other options and prior test results as well as risk scores from well-validated predictive models. Scans prompted by direct-to-consumer advertising or self-referral are less likely to take this other information into account. In all cases, there is potential for repetitive studies or additional radiation exposure, especially if the patient's care is not managed with appropriate oversight and follow-up.

In the midst of all of these unknowns, physicians, strongly motivated not to miss an opportunity to detect and treat coronary disease and lured by the optimism of this compelling technology, need to find a way to put the potential risks of these studies into a meaningful context for themselves and their patients. The authors provide comparisons to other events with attributable risk, as have other groups. For my part, I do not find them compelling nor can I imagine using them in conversation with my patients. Is it relevant that a family history of breast cancer likely conveys a much greater lifetime risk of malignancy than a single coronary CT? The family history cannot be avoided; the examination can. Discussing the risks of bicycle riding or swimming sounds like dissuasion from exercise to me. The estimates relating exposure to passive smoking may be the most helpful of those on the list, since this is both remediable and helps to endorse better health behavior. Perhaps the most reasonable approach is to give two examples, pairing reassuring with sobering comparisons (the CT angiogram may expose you to as much radiation as 1,000 chest X-rays, but the risk is much lower than inhaling someone else's cigarette smoke on a chronic basis). The most important follow-up question that the physician should anticipate, and indeed should have already carefully asked themselves is as follows: What are the other options for getting the information that we need?

There are many other questions that are critical to this puzzle, which at its core is about the risk/benefit of these procedures from the individual patient's perspective. No radiology-based imaging should be offered without consideration and discussion of risks and options with the patient. This is particularly critical when the benefits are unproven, or the exposure is significant, as with CT methods. We must be honest with ourselves and our patients in that we still await proof of improved outcomes and the incremental benefit of these tests compared to other approaches. The clinician sitting across from a patient will need reliable information about doses, as provided in this advisory, and relevant comparisons to help the patient understand the choices. Hopefully the time spent discussing the testing is matched by an effort to educate about well-proven interventions such as smoking cessation, lifestyle modification, and compliance with measures to control blood pressure and lipid levels. These conversations take time and are inadequately reimbursed, but in the end, analysis may have more influence on the patient's health than many other exciting and cutting edge modalities that we can provide.