I recently read Charles S. White, MD’s guest editorial in this journal [Chest radiology: Not just an academic subspecialty, Nov. 2017]. In making his case for thoracic imaging fellowships, Charlie noted that many radiologists think that “anyone can read chest imaging.” To his credit, he thoroughly debunked that contention.
Nevertheless, it hit a nerve. Actually, several nerves.
Many radiologists, including many of those within my department and elsewhere, also believe that any radiologist can read body imaging—meaning sonography, CT, and MR of the chest, abdomen, and pelvis. Anytime, anywhere, no matter how seemingly simple or how apparently complicated.
I understand—particularly as a body imager with clinical and academic interests in emergency radiology and other areas beyond torso imaging—that the vast majority of community and academic radiology departments, including mine, don’t have the luxury of having every examination interpreted by a fellowship-trained attending radiologist with extensive experience, particularly after hours.
However—and no offense meant to those whose responsibility it is to staff departments—in my experience, body imaging becomes a dumping ground.
Whereas breast imaging or interventional radiology cannot and will not be staffed by “just anyone,” as they both require subspecialty training, experience in specific procedures, and certifications, in marked contrast it’s often a case of, “Well, let’s put in the musculoskeletal radiologist or the nuclear medicine physician” or, yes, even the interventional radiologist, when body imaging needs to be covered.
When someone applies for a job here at my institution, after years in neuroradiology or pediatric radiology, they’re told, “We expect you to do body imaging, too,” and that’s OK. So why did I do a fellowship in body imaging (albeit a relatively general one)? Why doesn’t 21 years of attending-level experience matter if “just anyone” can do it?
It doesn’t, until problems arise and mistakes are made. And then it matters. A lot.
Lately, it seems that more and more of the scans I interpret are NOT straightforward, prompting me to make comments to my residents like, “Aren’t you glad you weren’t alone on call for this one?” and “If I were just coming out of training, this patient’s images would really trouble me, and even at my level they still do.”
Challenging cross-sectional images at my community/university hybrid hospital today alone included—no exaggeration—a spontaneously ruptured spleen in a patient with lymphoma; an atypical presentation of aspergillosis in a patient with vasculitis and emphysema; appendiceal mucoceles without and with appendicitis; complications of pulmonary embolism; and Hermansky-Pudlak syndrome (Good thing I went to that RSNA film panel two years ago—I was all over that last scan).
I realize this is a very complicated and nuanced topic that gets to the heart of what we are as radiologists—diagnostic, interventional, or both—who should be doing what, when, and why.
In fact, body imaging—and even more so emergency radiology—is among the most general subspecialties in radiology. Does the body imager have to deal with, on a routine or semi-routine basis, breast, neuro-, musculoskeletal, and vascular imaging findings? Yes. Similarly, isn’t there a portion of the lungs on each and every thoracic spine CT and MR examination which neuroradiologists have to interpret? Yes.These divisions are somewhat arbitrary, and we all have to know something about everything—or at least know when to ask a colleague for help. Did I do a fellowship in skin imaging, and yet don’t I frequently see cutaneous and subcutaneous findings on imaging examinations? No and yes.
Which brings me to my final point, recently emphasized in a meeting with our incoming department chairman, Michael Recht, MD: increased radiologist subspecialization may help to hold back the threat of artificial intelligence (AI) as our imminent replacement. This brings to mind a rather challenging patient who had not just one but three factors within the abdomen contributing to her extensive deep venous thrombosis and pulmonary embolism. Could Watson and his AI colleagues handle this one? Probably not now, and hopefully not for at least a few more years.
When Watson can beat the best of us at film panels and “Diagnosis, Please,” then perhaps, as with Ken Jennings on “Jeopardy!,” it will be time for us to lay down our Dictaphones.
So, is it true that “anyone can cook,” as stated by Auguste Gusteau, the chef from “Ratatouille”? Well, maybe, but certainly not at the level of a Thomas Keller or an Eric Ripert. Is it true that “any radiologist can read body imaging”?
Well, maybe yes. But then again, probably not.Back To Top
Katz DS. Some body imagers are more equal than others. Appl Radiol. 2018;47(2):4-5.
Dr. Katz is Director of Body Imaging and Vice Chair for Clinical Research and Education, NYU Winthrop Hospital, Mineola, Long Island, New York, and Professor of Radiology at Stony Brook University School of Medicine, Stony Brook, NY.