Incidental findings are a growing problem, amid wide variation in follow-up recommendations
In the everyday world, it is commonplace to use the adjectives incidental and ancillary to describe something or someone insignificant; minor even constitutes one definition for incidental. In the world of radiology, though, incidental and ancillary findings identified in imaging studies administered to assess the presence, absence, or treatment response of another condition are far from insignificant.
Rather, they have assumed great importance. Not only is there heightened interest among radiologists in making recommendations for follow-up care after these findings (as the specialty considers and reconsiders its role in the management of population health), but the question of how and when to respond also remains a topic of great debate. The incidental or ancillary finding—sometimes called an incidentaloma—is an increasing phenomenon in radiology.Preload: Preview
- Increased use of cross-sectional imaging and technological leaps in CT have resulted in an abundance of incidental findings in radiology—particularly among the Medicare crowd.
- Recommendations for follow-up care for incidental findings, however, remain highly variable and greatly individualized.
- The implications are anything but incidental: Patient care aside, complicating factors include medicolegal, cost-efficiency, and technology issues.
- Accountable care could drive adoption of standardized ways to address incidental findings; in the meantime, ACR® white papers, software, and individual efforts fill the gap.
Leonard Berlin, MD, FACR, serves in the department of radiology at NorthShore University HealthSystem Skokie Hospital in Illinois and is a professor of radiology at Rush University and the University of Illinois at Chicago. He says that such findings surface more frequently than ever before, with CT and MRI exams—especially the former—yielding the majority of them. The catalysts have been a sharp rise in the number of CT exams performed in the United States annually, along with marked improvements in spatial and contrast resolution gained using newer CT equipment.
“In the early 1980s,” Berlin notes, “we were performing just a few million CT scans per year. For the past few years, however, we have been at nearly 80 million CT scans annually. With the equipment that’s available today, we can see tiny areas of density—in the abdomen, bones, and elsewhere—as never before.”
In general, studies of the abdomen and pelvis produce the most troublesome incidental findings. Lincoln L. Berland, MD, FACR, is professor emeritus and vice chair for quality improvement and patient safety, as well as chief of body CT and the 3D laboratory in the department of radiology at the University of Alabama at Birmingham. He states that by default, any imaging exam of a large region of the body is a screening for conditions other than the one that the procedure it is intended to identify or track, and the more extensive the region, the greater the number of possible incidental findings.
As would be expected, Berland adds, incidental findings uncovered during exams of the abdomen and pelvis become more troublesome as patients age. He says, “As an example, in an 80–year-old patient with probable kidney cancer and other already-diagnosed diseases, there are bound to be multiple incidental findings. This is not necessarily so in a young, healthy patient, even with a large-region scan.”
Results of a study1 published in Archives of Internal Medicine lend credence to these assertions and add fuel to the fire. Specifically, of abdominal/pelvic CT and thoracic CT studies reviewed by the researchers, a respective 61% and 55% yielded an incidental finding. In contrast, nuclear-medicine exams showed an incidental finding in just 4% of studies scrutinized.
CT scans of the abdomen and pelvis uncovered the greatest proportion (9.2%) of incidental findings that were subsequently followed up; of the 1,376 study participants (1,426 exams), nine had ovarian or adnexal masses, and five had indeterminate lung nodules. For every decade of age, people evaluated in the study were determined to be 1.5 times more likely to have an incidental finding.
By some estimates, the likelihood that an incidental finding is (or will become) malignant is quite small. The process of determining which findings warrant follow-up and which should be ignored, however, as well as what that follow-up should entail, remains troublesome to radiologists. Such difficulties remain, despite inroads made by the ACR.
In 2010, the ACR’s Incidental Findings Committee2 published a white paper in the Journal of the American College of Radiology: JACR. The paper offered recommendations for managing incidental findings in the kidneys, liver, adrenal glands, and pancreas. Recommendations contained within it were based on available literature and on committee members’ personal experiences, including whether their actions had spurred increased or decreased utilization of medical procedures.
Three years later, Berland wrote an introduction3 to four new white papers covering incidental-findings management and written by the ACR’s Incidental Findings Committee II. The papers provide recommendations for managing incidental adnexal, splenic, nodal, biliary, vascular, and cholecystic CT and MRI findings .
Flowcharts within the papers feature algorithms for steps involving information that has an impact on incidental-findings management: categorization, demographics, patient history, and study results. Also covered are action steps (performing a study, providing follow-up care, or intervening through biopsy or surgery) and indications that no further action is required because the lesion can be deemed benign.
Marta E. Heilbrun, MD, MS, is an assistant professor in the department of radiology at the University of Utah; she practices at University Radiology Associates (Salt Lake City, Utah). Heilbrun designates for follow-up care any incidental finding with a size or complexity that places it in the category of concern.
She says that many incidental findings uncovered in her work are too small to fit into the category of concern, but seem to merit investigation, for one reason or another. She adds, “I will look to support from guidelines: ACR appropriateness criteria or National Comprehensive Cancer Network guidelines. The challenge comes when different guidelines conflict, which can be the case.”
Berland uses a similar approach—and like Heilbrun, he continues to see high levels of complexity. “The ACR white papers and other guidelines have taken what used to be a Wild West scenario, when it came to follow-up, and added some logic to it,” he explains. “We have some basic guidance about how to handle a specific incidental finding, with specific features, in a specific situation—and that’s good, but there are still huge variations in how to address some findings. It remains the $64,000 question.”
The task can be even more complicated in a screening setting, according to Caroline Chiles, MD, professor of radiology in the Comprehensive Care Center at Wake Forest Baptist Medical Center (Winston-Salem, North Carolina) and a principal investigator for the National Lung Screening Trial. For physicians working in an environment of this type, there exists the temptation to order a follow-up study whenever there is the slightest risk that an incidental finding might be relevant.
“In deciding on relevance, we look at the clinical setting and the patient history plus the risk, and there are guidelines for when further evaluation is needed,” Chiles states. “At what point do we stop? We don’t want every exam to prompt another. That would be a cost-efficiency nightmare.”
Ascertaining adherence to recommendations for follow-up exams also presents its share of challenges. Some radiology services have yet to figure out a viable system for handling this step, and many do so in an informal manner. Some vendors and institutions reportedly are devising systems designed to indicate when specific recommendations are not being followed, but integrating these with existing software used by radiology services has been a somewhat tricky endeavor, Berland notes.
According to Heilbrun, radiology (as a whole) will become more diligent about tracking follow-up care when larger health-care institutions formulate the requisite measurements. The spread of accountable-care organizations will move things forward, too, she says, as these organizations demand justification for follow-up visits and procedures.
Medicolegal Implications and More
Imminent and less-imminent changes notwithstanding, other issues cloud the picture, and medicolegal implications top the list. On one hand, the low likelihood that an incidental finding will prove to be malignant—or the absence of other risk factors—often spurs reluctance, among radiologists, to report that finding to the referring physician and recommend further testing, as well as hesitation to inform the patient. “On the other hand,” Berlin says, “there is the chance that if a radiologist doesn’t say anything about a finding, and a malignancy is found two years later, a lawsuit will ensue.”
Although it is difficult to predict the standard of care that would be applied in court in each case, one lawsuit hints at what could become the norm, Berlin says. In this instance, a 47–year-old man went to an emergency department with flank pain. A radiologist, interpreting a CT exam ordered by the emergency-department physician, termed the exam normal, with the exception of a hypodense, well-circumscribed mass in the left kidney: “most likely a cyst,” the radiologist reported. The patient had a diagnosis of renal-cell carcinoma, which subsequently led to his death. A lawsuit, filed by his family, was won by the plaintiffs.
Berlin’s preferred course of action is to let the referring physician know, for example, that there is a small area of density with a remote likelihood of a malignancy, leaving the rest up to that referring physician and the patient. It’s just as significant, though, that follow-up work for an incidental finding has the potential to be medically detrimental. Berland has seen some patients suffer surgical complications or renal failure as a result of follow-up procedures and has witnessed one death attributable to treating an incidental finding. “It’s a very delicate balance,” he says.
Then, there are the financial implications. David Rosman, MD, MBA, is associate director of business development at Massachusetts General Hospital (MGH) in Boston and is medical director, Mass General Imaging Worcester. Rosman believes that a general lack of consistency, among radiologists, in issuing recommendations for incidental-findings follow-up has as much of an impact on the cost of patient care as it does on the quality of that care.
He says, “We are remarkably inconsistent, as a specialty, at making the same recommendation, for the same finding, every time. Most of the time, thanks to a variety of documents like the ACR’s white papers on incidental findings, there is a correct answer—and we still demonstrate inconsistency.” This, in turn, raises doubts in ordering physicians’ minds, causing them to question radiologists’ recommendations and order expensive follow-up procedures that might not be warranted.
Rosman notes that the timing of acting (or not acting) on ancillary findings also has a bearing on cost. He cites, as an example, the short- and long-term expenditures incurred in following a pulmonary nodule.
Although it is true that the short-term cost of tracking such a nodule will always exceed the short-term cost ($0) of opting against doing so, he explains, a small wedge resection to remove a malignancy would prove less expensive than the more extensive surgical procedure associated with a more advanced cancer. He adds that whenever the cost of follow-up for incidental findings comes into question, it behooves practitioners to look at the benefit of choosing such a path—and balance the harm and good that might result.
Adhering to nationally recommended guidelines will almost certainly increase costs, at least for certain types of incidental findings. Upon investigation, Rosman states, MGH’s recommendation rate for adrenal lesions was discovered to be less than half of what it should have been. Its clinical-recommendation rate (additional imaging procedures or referral to an endocrinologist) was almost 90% too low.
“From this, we know that if we follow guidelines, costs will rise, in the short term,” Rosman says. “We do not know what it will mean in the long term, but consistency, at a higher cost, is the right answer. It may be that consistency teaches us that our nationally recommended guidelines are wrong; we may recommend follow-up for a lesion type and discover that there is never a positive finding at six months. Perhaps, then, we could change the guideline’s recommendation to one year, thus lowering the cost without increasing morbidity.”
It’s not surprising that the prevalence and widespread implications of ancillary findings are spurring louder cries for solutions designed to address the incidentaloma problem, as some radiologists have come to call it. Existing solutions, the consensus holds, are far from adequate; radiologists need decision-support tools that will render the approach to these findings more consistent.
Berland points to a project at MGH, where a point-of-care decision-support tool has been developed to assist radiologists in determining a course of action after identifying pulmonary nodules on abdominal CT exams. The decision-support tool takes advantage of departmental guidelines based on the Fleischner Society’s criteria. Radiologists can access the application (which embeds text for the recommended imaging study directly into radiology reports) through an icon on the dashboard of every speech-recognition workstation in the department.
Michael Lu, MD, a fellow in thoracic and cardiac imaging at MGH, copresented “Follow-up of Pulmonary Nodule Detected on Abdominal CT: Cost Implications for Adhering to Nationally Recommended Best-practice Guidelines,” describing the MGH work, on December 3, 2013, at the annual meeting of the RSNA in Chicago, Illinois. During fall 2012, researchers at MGH conducted a single-center cohort study to determine whether it would improve concordance with the radiology department’s pulmonary-nodules guidelines. The study focused exclusively on adult patients in whom incidental solid, noncalcified lung nodules had been discovered by an abdominal radiologist on an abdominal CT exam.
It included 141 patients with lung nodules detected after deployment of the decision-support system and 268 controls imaged prior to deployment. The researchers reviewed each subject’s medical record to determine the size of the individual’s largest nodule, any malignancy within the past five years, whether the patient smoked, whether clinical decision support had previously been used, and what follow-up recommendations had been issued.
The research revealed that adherence to the department’s guidelines rose by 96% in the 40% of cases for which the decision-support system was used. “We found that a point-of-care decision-support tool improved adherence to recommendations for follow-up of incidental lung nodules,” Lu says. “This has profound implications for other evidence-based guidelines in radiology.” The decision-support tool has since been implemented in other radiology sections, including chest radiology.
Software solutions such as the one at MGH are definitely a step in the right direction. The radiology specialty as a whole, however, has “a lot of work to do,” Berland says, in the area of creating decision-support tools that will promote consistency in addressing incidental findings while simultaneously yielding the desired patient-care enhancement and cost control.
Integrating clinical understanding and natural language processing components into these tools will help to bridge the gap between having consensus recommendations in hand and being able to put them into practice in a consistent manner, Berland observes. This is because the natural language processing technology will recognize when it is advisable to follow a particular set of recommendations pertaining to incidental findings and will generate an educated suggestion of the specific recommendation for a given situation.
“Of course,” Berland concludes, “the radiologist would make the final decision. Having the assistance of that extra intelligence, though, can provide the necessary decision support, save time, improve accuracy, and make it more likely that the set of recommendations can be used when appropriate.”
- Orme NM, Fletcher JG, Siddiki HA, et al. Incidental findings in imaging research: evaluating incidence, benefit, and burden. Arch Intern Med. 2010;170(17):1525-1532.
- Berland LL, Silverman SG, Gore RM, et al. Managing incidental findings on abdominal CT: white paper of the ACR Incidental Findings Committee. J Am Coll Radiol. 2010:7(10):754-773.
- Berland LL. Overview of white papers of the ACR Incidental Findings Committee II on adnexal, vascular, splenic, nodal, gallbladder, and biliary findings. J Am Coll Radiol. 2013;10(9):672-674.