On Balance
By Beth W. Orenstein
Radiology Today
Vol. 26 No. 6 P. 10

Study Raises Questions About Appropriate CT Use

More than 93 million CT scans were performed to detect tumors and diagnose many different diseases in the United States in 2023, according to a national survey of hospitals and imaging facilities. That’s up exponentially from just 3 million in 1980, according to the International Society for Computed Tomography. The increase in use can be attributed, at least in part, to CT’s increased availability and its ease of use.

“CT is available basically 24/7, almost anywhere in the country, and it gives you information very quickly,” says Rebecca Smith-Bindman, MD, a University of California, San Francisco, radiologist and professor of epidemiology and biostatistics and obstetrics, gynecology, and reproductive sciences.

Smith-Bindman is the lead author of a provocative study published in April in JAMA Internal Medicine that calls this continuing increase in CT scans rather alarming. The researchers concluded that the radiation emitted by this widely used imaging tool, even at low doses, could be potentially harmful. They projected that at the current usage pace, CT scans could eventually be responsible for 103,000 new cancers, which is roughly 5% of all cancers diagnosed in the United States in a single year.

The study and its conclusion were extensively covered in the popular media, with alarming headlines including NPR’s: “Study Highlights Cancer Risk From Millions of CT Scans Performed Annually.” CBS News also reported: “Radiation From CT Scans Could Lead to Thousands of Future Cancer Diagnoses, Study Finds.” People magazine wrote: “Just One Year of CT Scans Could Lead to Over 100,000 Cancer Diagnoses, Study Finds.”

Reaction from the imaging community was swift and reflects a deep concern. “Unfortunately, radiation and its effects are not taught in our school systems, and so most people are unfamiliar with the topic and believe attention-getting headlines,” says Cynthia McCollough, PhD, a CT imaging expert and past president of the American Association of Physicists in Medicine. “This saddens me because many patients are already worried about their health. The last thing they need is one more thing to worry about.”

McCollough, the director of Mayo Clinic’s CT Clinical Innovation Center, worries that some patients, seeing the headlines, will choose to forgo important preventive and diagnostic CT scans. Since the study’s publication, “I have had some patients decide not to have an exam— even in extreme cases where the scan was to be used for planning the patient’s upcoming surgery,” she says.

Like McCollough, Mark Supanich, PhD, director of medical physics at Rush University Medical Center in Chicago, is concerned about the messages the study’s conclusions send to patients. He, too, has seen “patients cancel diagnostic exams for disease follow-up and cite this study as a reason.”

Maxwell Amurao, PhD, MBA, chair of the ACR’s commission on medical physics, says patients shouldn’t hesitate to undergo CT imaging when the medical benefits outweigh the risks, including radiation risk, and that is something they should discuss with their physicians.

What Does the Evidence Say?
CT scans expose patients to ionizing radiation and, according to the National Cancer Institute, radiation can cause cell damage that can lead to cancer, even at low levels. What is debatable is whether CT scans themselves cause some cancers. In an interview with Radiology Today postpublication, Smith-Bindman says “numerous studies have provided direct evidence of cancer risks associated with CT imaging.”

However, Smith-Bindman says the JAMA Internal Medicine study is a modeling study. The paper is not intended to demonstrate that radiation will cause some cancers, she says. “This paper aims to quantify how many future cancers could result from the current use of CT scanning. That was our question, and we basically updated a previous 2009 publication using 2023 scan volume data and more accurate information about radiation doses and the use of multiphase scanning.”

While the authors raise valid concerns about cancer risk with their risk model, “there is no direct evidence linking low levels of ionizing radiation, like those used in CT scans, to cancer in the general population,” Supanich says. After the study was published, the ACR also issued a statement that said, “There are no published studies directly linking CT scans (even multiple CT scans) to cancer.” The ACR statement also pointed to a report in 2019 from the National Council on Radiation Protection and Measurement that said, “even with increased CT use, advances in technology and imaging protocol optimization have reduced population radiation burden (medical radiation dose per capita/person).”

Assessing Risk
Smith-Bindman disagrees that there is no peer-reviewed data to directly link CT radiation exposure to cancer. She cites two large EPICT studies published in highly ranked medical journals to back her statement that CTs can cause cancer. One was published in November 2023 in Nature Medicine, and another was published in January 2023 in Lancet Oncology. The studies, which pooled data from nine European countries, found increased risks of blood and brain tumors following CT imaging in children, teens, and young adults.

“These risks are comparable in magnitude to those observed in atomic bomb survivor studies and other cohorts that informed our risk estimates in our study, and indeed demonstrate risks even higher than we used in our modeling,” Smith-Bindman says.

For the recent study, the researchers analyzed 93 million exams from 61.5 million patients in the United States. They used a multicenter sample of CT examinations prospectively assembled between January 2018 and December 2020 from the UCSF International Dose Registry. Most of the scans were of adults, the largest number of scans being of those between 60 to 69 years old, while children accounted for about 4% of the scans. They excluded patients in the last year of their lives because CT imaging at this stage was unlikely to lead to cancer, and, in a sensitivity analysis, excluded patients in the last two years of life, the authors note.

The researchers found that adults 50 to 59 had the highest number of projected cancers: 10,400 cases for women; 9,300 for men. The most common adult cancers they project were lung, colon, leukemia, bladder, and breast. The most frequently projected cancers in children were thyroid, lung, and breast. The researchers wrote that the largest number of cancers in adults would come from CTs of the abdomen and pelvis, while in children, they would come from CTs of the head. They also project that those who underwent CT who were less than a year old were 10 times more likely to get cancer than others in the study.

The researchers projected future lifetime radiation-induced cancer risk using an updated version (4.3.1) of the National Cancer Institute’s Radiation Risk Assessment Tool (RadRAT) software. The software utilizes risk models from the National Academy of Sciences’ Biologic Effects of Ionizing Radiation (BEIR) VII report for 11 site-specific cancers (stomach, colon, liver, lung, breast, uterus, ovary, prostate, bladder, thyroid, and leukemia) and seven additional cancer sites (oral cavity or pharynx, esophagus, rectum, pancreas, kidney, and brain or central nervous system cancer plus a remainder category) using a more recent follow-up of the Japanese atomic bomb survivors and pooled analyses of other medically exposed cohorts. For a given cancer type, RadRAT estimates excess lifetime risk of cancer from the time of exposure based on user-supplied organ dose and US life table estimates of age and sex-specific baseline cancer rates.

Ongoing Research
Supanich believes the Smith-Bindman study’s reliance on data extrapolated from high-dose radiation exposure, such as those experienced by Japanese atomic bomb survivors, introduces uncertainty when applied to the medical imaging context. “This study does not establish a causal relationship between CT scans and cancer incidence,” he says.

Work on this topic continues in both the United States and abroad. “In fact, the National Academy of Sciences proposed a long-term strategy for research into low-dose radiation several years ago to try to answer the outstanding questions on the actual risk posed by these low radiation dose exams,” Supanich says. “Until we have a better understanding of the risk (if any) from diagnostic radiation exposure, claims of large numbers of induced cancers from CT scans should be viewed skeptically.”

McCollough also points out that the BEIR VII report from the National Academies of Science, which the authors used for their risk estimates, associates higher doses of radiation with higher cancer occurrences. However, she says, “the report focuses primarily on high doses and CT uses low doses (below 100 mGy). The risk data are for organ doses above 100 mGy, and CT delivers organ doses in only 10s of mGy.”

Smith-Bindman responds that many people died from the atomic bomb blasts, and many survivors were exposed to high radiation doses. Those patients have not been studied. The largest studies of atomic bomb survivors is the Life Span Study of 120,000 survivors, and this population has an elevated risk of nearly every cancer type.

“Their median dose was 40 mGy, doses we routinely use in medical imaging,” Smith-Bindman says. “The BEIR VII results show individuals with doses at even 10 mGy are at elevated risk of cancer. And, indeed, the EPICT studies are recent studies of children and young adults exposed to CT scans demonstrating an elevated risk of hematologic malignancy and brain cancer. We have a very clear understanding of risk, right now. Pretending that the BEIR VII report only focuses on high doses is untrue.”

Quality Improvement Efforts
Members of the radiology community also point out that they adhere to the golden rule of radiation safety known as ALARA (as low as reasonably achievable) for all imaging exams that require radiation. By doing so, they have decreased the average radiation dose per scan over time.

“Radiologists, medical physicists, and CT technologists engage in continuous quality improvement efforts to use only the amount of radiation necessary to ensure the diagnostic task for the exam is achieved,” Supanich says.

National efforts on reporting CT radiation, such as the ACR Dose Index Registry, aid in these efforts, as well, he adds. “All of these efforts should allay concerns about the theoretical risk of the scans and increase patient confidence that many reasonable efforts are undertaken to ensure that the image quality and radiation dose for the exam are optimized.”

Optimizing CT exam parameters, which include dose, “is a top priority in the radiology community,” McCollough agrees. “These efforts should indeed reassure patients and caregivers,” she says. She adds that any medical test can be used incorrectly, and it’s the medical community’s job to continue to develop clear guidelines regarding appropriate use and “to make sure these are well known outside the radiology department.”

Smith-Bindman says the JAMA Internal Medicine study is not advocating against CT scans when medically necessary. However, she says it should stand as a warning that physicians should only use CT scanning when it is necessary. “I don’t even understand how that can be a question,” she says. “Why would you use a test that’s not necessary?”

Radiation isn’t the only harm from CT, Smith-Bindman says. “There are false positives, contrast reactions, and anxiety related to waiting to get the test results. Our study is just one more piece of information to say that we should only use CT imaging when necessary.”

Checks and Balances
ACR’s Amurao agrees that, because CT has been widely available since the 1980s and is rather easy to use, it’s tempting to overuse it. To counter this temptation, Amurao says, the radiology community strongly encourages physicians who order imaging to adhere to the ACR’s appropriate use criteria. Those criteria are updated as needed and extremely helpful, he says.

Amurao also sees a role for smart programs or AI in helping to determine appropriate use and eliminate unnecessary scans. “Appropriate use criteria are not always used as much as we would want,” Amurao says, “but it’s really, really important.”

Equally as important, he says, is for imaging facilities to be certain that their equipment is working properly and that the CT scanner is not mistuned or improperly calibrated. He notes that checks and balances are in place to prevent that from happening.

While CT scans expose patients to radiation, there are risks associated with not receiving a CT scan, McCollough says. She cites a study of nearly 12,000 patients who had surgery to remove their appendix that was published in June 2019 in the American Journal of Surgery. The study found that the surgery was unnecessary for 10% of the group who had been diagnosed with ultrasound alone. Of those who had a CT scan prior to surgery, only 2.5% were misdiagnosed.

Lung cancer screening with CT is another example that demonstrates significant benefits from CT, McCollough says. “Mortality was decreased by 20% in the National Lung Cancer Screening trial in those who had CT,” she says. “If one uses the BEIR VII risk data, the increase in risk for the scan is less than 0.1%. Again, the benefits (20% observed decrease in death) are much bigger than the hypothesized risk (<0.1% increase in the risk of cancer occurrence).”

To Supanich, the bottom line is this: “The ability of CT scans to significantly alter and improve diagnostic accuracy and clinical management in emergency settings underscores their indispensable value. Such findings highlight the critical role CT scans play in delivering timely and effective patient care. In fact, use of diagnostic CT imaging can eliminate the need for surgeries or other procedures, providing better outcomes and less risk for patients.”

And to Smith-Bindman, the bottom line is this: “I am not advocating against CT scans at all. When medically necessary, the benefits clearly outweigh the risks. However, all clinicians, including radiologists, have a responsibility to use CT exams more wisely and ensure that those scans are done using the lowest effective doses for every patient in every setting. While the radiology community has identified dose optimization and reduction as a priority, in actual practice, profound variation in dose remains an urgent clinical need.”

—Beth W. Orenstein of Northampton, Pennsylvania, is a freelance medical writer and regular contributor to Radiology Today.