Much has been written about how radiologists are increasingly perceiving findings such as masses or lesions that are "incidental" to the reason the studies were ordered in the first place. Such findings are becoming considerably more frequent as an increasing amount of data is portrayed on advanced diagnostic imaging studies. And with more such findings flagged, in conjunction with of the use of artificial intelligence (AI) algorithms, radiologists' reports may convey a growing number of recommendations for follow up diagnostic testing or other actions such as specialty referrals, invasive procedures or other actions resulting from these incidental findings.
Largely, the academic literature has been positive about the benefits of these findings. But there is a contrary viewpoint that is emerging in the specialty.
In a clinical perspective article just accepted for publication in the American Journal of Roentgenology, Matthew S Davenport, M.D., Professor of Radiology and Urology, and Service Chief and Vice Chair in the Department of Radiology at Michigan Medicine, has fired a shot across the bow of the positive view of radiologists reporting many incidental findings to patients and their treating physicians. Dr. Davenport points to studies that have shown that intervention based upon incidental findings, including those that detect early cancers, can result in low-value care and actually cause harm. He cites several examples of detection of a cancer that—had it not been identified—would otherwise not have affected a patient’s life, or detection of a cancer for which intervention does not change disease trajectory. But, he notes, those findings can trigger "confirmatory testing or follow-up, cost, complications of diagnosis and therapy, and acute and chronic anxiety."
Dr. Davenport is aware how complicated it is for the radiologist to "unsee" an incidental finding. How does one determine which incidental findings create the risk of harm through over-diagnosis and over-treatment resulting from their detection of incidental findings? He advocates that radiologists develop incidental findings guidelines along the lines that exist in other medical and surgical specialty areas. He hopes for information technology solutions to help radiologists balance diagnostic sensitivity with other competing risks and determine which incidental findings are important and how best to manage (or ignore) those findings. Complicated indeed!
The essay does note that the professional liability environment complicates the moderation of reporting, especially when some incidental findings are cancer. He aspires to a more sophisticated understanding of the biases which predict low-value care—that early detection of some cancers can produce a paradoxically worse outcome than had those cancers never been detected at all. He rhetorically asks shouldn't this be the reasonable thing to expect of patients and the legal system to understand today? A valid question to ask.
But I believe Dr. Davenport's call for restraint in reporting will prove difficult for many radiologists who do not want to end up as defendants in medical malpractice actions. Most radiologists today, I think, may choose to report all findings.
This is a vitally important debate within the specialty that will be fascinating to watch.