Golden Opportunity?
By Beth W. Orenstein  
Radiology Today    
Vol. 24 No. 3 P. 10

The Pros and Cons of Opportunistic Imaging  

More than 80 million CT scans are performed on patients for clinical indications each year in the United States, but these scans can provide significantly more information than is often utilized. By applying imaging technology or AI algorithms to scans, radiologists can often identify chronic or undiagnosed conditions other than the condition that was originally addressed. Harnessing additional information from imaging scans is known as opportunistic imaging, and interest in the practice is growing, says Perry J. Pickhardt, MD, a professor of radiology and chief of gastrointestinal imaging at the University of Wisconsin in Madison. 

Jessica H. Porembka, MD, FSBI, an associate professor of radiology in the division of breast imaging and vice chair of strategy and quality at the University of Texas Southwestern Medical Center in Dallas, offers this example of opportunistic imaging: A patient undergoes a CT to identify the cause of abdominal pain, and the radiologist also reports on the patient’s visceral fat, a risk factor for metabolic disease. Another example: a noncontrast CT to evaluate for renal stones is used to assess bone mineral density (BMD), which, when low, is a risk factor for osteoporosis.

Opportunistic imaging is different from incidental findings. If a physician orders a chest CT to determine why a patient is short of breath, and the CT shows a pulmonary embolism as the cause but also a small mass in the left breast, the breast mass is “an incidental finding,” Porembka says. One similarity between incidental findings and opportunistic imaging, however, is that both could be actionable, she says. With the incidental finding of a breast mass, physicians will want to know whether the patient has cancer and order more testing. With opportunistic imaging identifying low bone density, the patient’s physician may want to let the patient know of her risk for osteoporosis, which could lead to a future fracture, and recommend treatment that may prevent it, Porembka says.

Opportunistic imaging is not new. “People have been doing research on opportunistic imaging for a while,” Porembka says. Elizabeth Gausden, MD, MPH, an orthopedic surgeon at the Hospital for Special Surgery in New York, says she and colleagues have been looking at using CT opportunistically to assess bone quality for several years.

Increasing Interest
Although not new, opportunistic imaging is attracting increased interest, especially as more recent advances in machine learning and AI have made it possible to do more easily without additional costs or exposing patients to additional radiation, says Miriam A. Bredella, MD, MBA, FACR, a professor of radiology at Harvard Medical School, a musculoskeletal radiologist in the department of radiology at Massachusetts General Hospital in Boston, and lead author of a study on opportunistic imaging published in January 2023 in the American Journal of Radiology. Opportunistic imaging can be in play with any modality. However, the focus of the research in this area to date has been primarily on using CT to measure BMD, body composition, and vascular calcifications, Bredella says. The FDA has approved several fully automated commercially available AI algorithms that can be prospectively or retrospectively applied to routine CT exams.

Traditionally, the way to determine BMD is with a dual energy X-ray absorptiometry (DXA) scan. More than 10 million Americans have low BMD or osteoporosis, and more than 2 million Americans suffer fractures as a result every year. “The primary challenge in the management of osteoporosis is a lack of patient awareness,” Bredella wrote in the American Journal of Radiology study.

While a DXA scan can identify patients with low BMD who are at risk for osteoporosis, only a small number undergo screening. The number of minority patients who undergo screening is particularly low, Bredella notes. Early identification of risk is key to preventing osteoporosis- related fractures. Opportunistic imaging could more easily help identify patients who are at unsuspected risk for fractures and allow them to start treatment that can help prevent fractures in the future, she says. Bredella believes that opportunistic osteoporosis screening could be easily added to lung cancer screening CT or CT colonography. “Patients undergoing these examinations are typically at risk for bone loss, so opportunistic imaging increases the value of such screening examinations,” she says.

Before advances in machine learning and AI, quantifying BMD and body composition required laborious manual or semiautomated segmentation, both of which are time consuming. As a result, it was done on mostly small- and medium-sized patient samples. Automated methods have allowed quantification of BMD and body composition on CT exams that were performed for other purposes in large data sets. “They allow us to do large-scale population-based screening,” Bredella says.

More With Less
Another area where opportunistic imaging is gaining momentum is in measuring body composition, specifically abdominal fat. Visceral fat can be an indication of whether the patient is at risk for diabetes or metabolic syndrome, Porembka says. “You could potentially quantify the amount of visceral fat a patient has on an abdominal CT, give that to the referring provider, and say your patient is at risk for developing these diseases.”

One way to determine cardiometabolic risk is body mass index (BMI), but it is an imperfect measure, Bredella says in the study. “BMI does not explicitly reflect an increase in harmful fat depots or a decrease in muscle mass.” She says population studies have shown that the accumulation of visceral fat and the loss of muscle mass better predict adverse events in diverse populations, including patients with cancer or COVID- 19. In addition, opportunistic CT can quantify calcified plaques in abdominal and thoracic vessels, including the coronary arteries.

Cancer patients routinely undergo serial CT exams for staging, treatment response, and surveillance. Pickhardt, an author of a study published in October 2022 in the American Journal of Roentgenology, says now that body measurements can be made from these CT scans using fully automated AI-based segmentation and quantification tools, body composition measurements could be applied to provide more precise risk stratification as a component of personalized oncologic care.

Bredella sees yet another advantage of opportunistic imaging: It can help reduce health disparities. For example, screening for osteoporosis is especially low for ethnic and racial minorities and underserved populations. “Opportunistic osteoporosis screening could therefore help reduce the gap in screening disparities,” she says.

In addition, “using information from studies that have already been performed reduces costs and radiation exposure and ultimately increases the value of diagnostic imaging,” Bredella says.

Pickhardt agrees. In several of his studies on opportunistic imaging, he notes that opportunistic imaging can save cost “as it’s just using more information from the same study.”

Not Without Concerns
Opportunistic imaging does have drawbacks, however. One is that it could lead to information overload and, potentially, unnecessary treatments, Bredella says. “Only because we can measure something does not mean that we should.”

For example. “Some of the parameters that we can measure, such as [visceral adiposity], do not have reference values, so we do not know how much is too much and what to do with the added information,” she says. Also, although the FDA has approved some algorithms that can be applied prospectively or retrospectively, “there are often no clear guidelines on what to do with the additional information,” she adds. “This could lead to patient anxiety and unnecessary additional tests.”

Some have proposed opportunistic imaging as an additional screening exam, as a supplement to an annual physical exam, “such that we are not just taking information from a study already performed but performing another study,” Porembka says. One of the main concerns about opportunistic imaging used in this manner is the same as with any scan: incidental findings. What if, when using opportunistic imaging as a screening tool for visceral fat, it shows a renal mass that hadn’t been seen before? What if that mass is nothing? Can it be left alone? Opportunistic imaging could lead to more testing, some of which may be unnecessary, Porembka says. “Sometimes, there are things we see that don’t translate into clinical relevance and doing something about it could be worse than not.”

Gausden says while opportunistic imaging can provide orthopedic surgeons with information about their patients’ bone density that is useful, the thresholds for determining who is at risk have not yet been clearly defined. She notes many women in their 60s and 70s have undiagnosed osteoporosis and poor bone quality. Gausden’s patients undergo CT scans before their surgeries for planning purposes. Opportunistic imaging on those scans may tell her that the 62-year-old who is about to undergo a robotic hip replacement is at about a 10% risk of getting a fracture within the next five years. Does that mean Gausden should modify the way she plans to do the surgery? Should the patient be started on bone-building drugs? They are important questions that don’t yet have answers, she says.

Another question is who instigates opportunistic imaging: the radiologist or the physician requesting the CT? “The answer is, it depends,” Porembka says. Perhaps, she says, the physician who orders the scan should request opportunistic imaging at the same time.

“There is the possibility of actionable screening findings that need to be addressed, that need to be worked up,” Porembka says. “At some point, someone has to talk to the patient about what they see, especially if something is abnormal. If everything in the report is negative and nothing is abnormal, that’s one thing. But if there are these abnormalities, things that need further evaluation, or things that need intervention, who is talking to the patient about that and what the next steps are?”

Whose Views?
Gausden says the question of which physician specialty is responsible for managing low bone density and starting treatment for it is hotly debated—the woman’s primary care physician, endocrinologist, or orthopedic surgeon— and so who should order opportunistic studies to determine the woman’s bone density is debatable, as well.

Bredella says it could be the radiologist or the referring physician. “Opportunistic imaging, for example, for the assessment of bone mineral density, atherosclerotic vascular calcifications, or sarcopenia (muscle loss), could be incorporated routinely without a specific order, to increase the value of every CT, or it could be ordered specifically by the referring physician, if they are interested in this additional information,” she says.

Could opportunistic imaging become standard practice and adopted on a large scale? That remains to be seen. Porembka says a large part of the answer to that question is how it will be reimbursed. In addition, “it’s going to take extra work on both sides,” she says. “Even if it’s fully automated, you’re still going to be generating findings that go into reporting. Even if the reporting of all the findings is automated, radiologists still need to validate the findings before sending the report. It’s work for the radiologist, even with full automation. In addition to that, if the patient wants to discuss their findings, they may want to talk with the radiologist, and that’s potentially adding more work to radiologists who are already overworked.”

Porembka says, given the plusses and despite the minuses, she’s excited about opportunistic imaging. “I think it can make all the data that radiologists have access to more valuable, especially when it comes to population health,” she says. “It could lead to early intervention and prevention of some conditions.” However, she adds, “there are things that need to be worked through before there is large-scale, widespread adoption.”

Bredella also hopes that automated opportunistic assessment of CTs will become standard of care in the future. Opportunistic imaging clearly has a place in the care of “patients with cancer or chronic disease, who routinely undergo CT for staging or surveillance and who would benefit from the additional information,” she says.

Gausden believes that using diagnostic CT to infer region-specific osteoporosis could be extraordinarily valuable to orthopedic surgeons and primary care physicians. She says it certainly “merits further research.”  

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