Risk vs. Benefit Calculations

In normal circumstances, exposure to medical radiation does no harm and provides a tremendous diagnostic advantage for treating physicians and thus patients. Imaging allows the physician to examine a patient internally without using an invasive procedure, such as surgery. This suggests physicians and radiologists must engage in a risk vs. benefit calculation when patients are exposed to substantial doses of ionizing radiation during treatment.

In the 1950 and 1960s, pregnant women were often exposed to abdominal radiologic imaging procedures to help the obstetrician gauge whether a normal delivery was possible (reviewed by Linet et al., 2012). A survey was eventually conducted that found an increased risk of pediatric cancer in the children born to these women [Risk Ratio (RR) 1.39, 95% CI, 1.31-1.47]. Since this finding was based on a survey of parents, the results languished under a cloud of skepticism until the findings were confirmed independently using medical records. It was estimated that infants that had undergone radiologic imaging during gestation experienced a 5.4-fold increased risk for pediatric cancer. Current estimates suggest the risk for this procedure has declined to 1.5- to 2.2-fold, due to lower radiation levels being used. In addition, the prevalence of in utero radiologic imaging has declined significantly due to the development and widespread use of ultrasonography.

For children and adolescents, multiple research studies have produced mixed results when trying to assess the risk of pediatric cancer (reviewed by Linet et al., 2012). Although it is generally believed that there is a small increase in pediatric cancer risk for children exposed to diagnostic radiation, few studies have been able to reach statistical significance. If there is a cancer risk, it would be for leukemia. In terms of the lifetime risk of cancer due to pediatric exposure to diagnostic radiation, what evidence that exists suggests the risk is real. For example, young women monitored for scoliosis and tuberculosis during childhood have an increased risk of developing breast cancer later in life (RR, 2.86, p = 0.058), but not lung cancer or leukemia.

Adults also suffer from an increased risk of cancer due to diagnostic radiation exposure (reviewed in Linet et al., 2012). The RRs for 3-month, 2-year, and 5-year follow-ups were 1.17 (95% CI, 0.8-1.8), 1.42 (95% CI, 0.9-2.2), and 1.04 (95% CI, 0.6-1.8) for leukemias other than chronic lymphocytic leukemia. These findings were generated from HMO medical records. Another study estimated the level of radiation received by the bone marrow and predicted a 1.4-fold increased risk of acute myeloid leukemia within 3 to 20 years following exposure. With respect to chronic myeloid leukemia, a number of studies have found a small increased risk during the 20-year period subsequent to radiation exposure.

Scientists have searched for a link between diagnostic radiation exposure and the emergence of solid tumors in adults, but only a few have reported statistically significant results (reviewed by Linet et al., 2012). Meningiomas and parotid tumors were linked to the number of full-mouth dental X-rays conducted before 1945 or prior to the age of 20. Adult tuberculosis patients undergoing repeated radiological examinations of their chest have an increased risk of breast cancer. In addition, elevated levels of chromosomal translocations were found in the white blood cells from radiology technicians.

These findings reveal real risks, but also a dearth of well-controlled studies. Linet and colleagues (2012) reported that such studies are underway, studies which hope to provide a clearer picture of the cancer risk caused by diagnostic radiation. Such studies should be more robust, not only because the science has improved, but because patients are being exposed to greater levels of ionizing radiation more frequently. In the meantime, radiologists and physicians are relying on the linear, no-threshold model for guidance.

Current risk estimates suggest the 70 million CT scans performed in 2007 in the U.S. could eventually cause 29,000 cancers (15,000-45,000) (reviewed by Linet et al., 2012). Half of this risk was attributed to abdominal and pelvic examinations. The most common cancer to result is expected to be lung cancer, followed by colon cancer and then leukemia. The cancer risk for specific procedures was also examined and whole-body CT scans between the ages of 45 to 70 years could cause cancer in 1 in 53 patients, mammograms in 1 in 1111 women, lung CTs in 1 in 435 males and 118 females, and CTs for coronary artery calcification in 1 in 2500 males and 1667 females.

These risk estimates can be used to calculate risk vs. benefit ratios, which in turn can be used to justify using or not using a radiologic diagnostic procedure. Of all the patients who enter a hospital emergency room, trauma patients will likely be exposed to significant amounts of diagnostic radiation. Laack and colleagues (2011) examined the medical records of 642 level II trauma patients in Rochester, Minnesota for treatment outcomes. The median radiation dose received was 24.7 mSv (interquartile range = 6.2-26.6), which did not differ substantially among the four age groups studied (15-19, 20-40, 41-60, and over 60). The primary radiologic examination was CT scans, including whole-body CTs. The largest doses were produced by CT scans of the abdomen and pelvis (14.1 mSv), thoracic spine (15.7 mSv), and whole-body (20.9 mSv). The maximum dose received among all patients was 54.75 mSv. The estimated increased risk of death from cancer due to CT scans was estimated to be 0.195% (0.053-0.218) for patients under 20 years of age, 0.137% (0.046-0.167) for patients between 20 and 40 years of age, 0.113% (0.030-0.126) for patients between 41 and 60 years of age, and 0.050% (0.022-0.085) for patients over the age of 60. Death due to trauma among these trauma patients was 0.6%. The patients who died had a median age of 90 and all died of intracranial injuries. No patients under 80 years of age died of trauma-related injuries. By comparison, the overall risk of death due to CT scans was just 0.1%.

The study by Laack and colleagues (2011) reveals how an evidence-based risk/benefit calculation can be conducted. All such studies need to consider age as a primary determining factor of risk. For example, the oldest patients in this study had the highest mortality risk and the lowest risk of CT scan-related cancer. For these patients, the risk/benefit calculation clearly indicates that CT scans should be done to enhance the quality of care that elderly trauma patients receive. The same conclusion cannot be reached for younger patients. Level II trauma patients between the ages of 15 and 19 are more likely to survive their injuries because of their age, but the estimated cumulative cancer risk is four times higher than for patients over the age of 60. Laack and colleagues (2012) suggest that efforts to reduce radiation exposure should therefore be focused on younger patients. The limitations to their study included an inability to determine if CT scans had saved any lives and the authors admit that if a single life had been saved in the youngest age group then the risk/benefit calculation would swing dramatically in favor of performing CT scans on young trauma patients.

Reconsidering the Linear, No-Threshold Model

Studies that examined the health of atomic blast survivors in Japan, suggests that exposures below 200 mSv did not cause an adverse effect over the course of their lifetime (reviewed by Jargin, 2012). Epidemiological studies examining other demographics have revealed the lower limit for adverse effects of ionizing radiation exposure is between 100-200 mSv. Jargin (2012) argues that there is substantial evidence to support the conclusion that low doses of radiation may not be a threat to human health and may even prove beneficial (hormesis). He cites a study that revealed cancer rates are lower for people living at higher elevations, where radiation from the sun and other interstellar sources is less diminished by the atmosphere. With respect to animal studies, exposure to 70-140 mGy per year of ionizing radiation increased the lifespan of mice. The transition zone for irradiated mice seems to be around 100 mGy per year, with cancer decreasing below this limit and increasing above it. If low doses of ionizing radiation are indeed protective against cancer, the mechanism involved is likely the induction of cellular DNA repair machinery by radiation-induced DNA damage. However, above a certain threshold, such a mechanism would predict that the rate of DNA damage exceeds the capacity of the repair machinery and mutations begin to accrue. If this theory is correct, then the linear, no-threshold model is inaccurate for low dose exposures typically received during medical imaging procedures.

Hendee and O'Connor (2012) would agree with Jargin's (2012) argument. They cite scientific evidence that suggests workers in the nuclear industry benefit from increased exposure to ionizing radiation, thus producing the so-called "healthy worker effect." They also argue that using data from the survivors of the nuclear bomb blasts in Japan to predict health outcomes from medical imaging is inappropriate. The health of the two groups is…