The Importance of Dialogue in Radiology

Overreliance on written communication is hazardous

Various medical imaging techniques, such as CT scans, MRIs, and ultrasounds, are overseen by a physician specifically trained for the job: the radiologist. The radiologist is responsible for ensuring that the imaging procedure is conducted correctly and safely, and for interpreting the images obtained. He or she generates a written report detailing their observations (the “findings”) and the findings’ clinical significance (the “impression”).

In many cases, clinicians do not comprehensively review the images themselves, and seldom speak with the radiologist. As a consequence, the clinician's understanding of the written radiology report is a crucial cognitive task. Important decisions often center around this understanding. Misunderstanding or ignoring parts of the report can be a significant source of clinical error.

I believe that more direct interaction between clinicians and radiologists can reduce these errors. Instead of relying on a single, static written report, I suggest a dialogue (verbal or written), at least in cases where important decisions hang in the balance.

A few years ago, a 24 year old man presented to the emergency department with intractable abdominal pain. CT showed enlarged lymph nodes ("adenopathy") in the abdomen. The radiologist interpreted this as “probable lymphoma.” The patient was also discovered to have acute hepatitis C. He continually insisted on more pain medication, but because he had a history of opioid abuse, this was suspected to be drug-seeking behavior. The senior liver consultant attributed the adenopathy to the hepatitis, and suggested the patient return in 2 months. So the patient was discharged. A few weeks later he returned with uncontrolled pain. The adenopathy had increased and he now had extensive small bowel wall edema. Surgery revealed venous infarct, meaning his small bowel had died because its veins had been compressed by the markedly enlarged lymph nodes. His pain had probably been from reduced blood flow to his tissues (“ischemia”) all along. Most of the small bowel had to be surgically removed. Biopsy of the adenopathy revealed anaplastic large cell lymphoma. The patient died a year later after enduring short bowel syndrome and many other complications.

There is no record to suggest that the clinicians ever reviewed the original CT images with a radiologist. It is unknown whether they looked at the images themselves. They failed to understand that, while adenopathy can indeed be seen with hepatitis, nodes in that disease are typically scattered near the entry of vessels and bile ducts into the liver (the “porta hepatis”). Bulky adenopathy in the mid-abdomen membrane that supplies blood to the small bowel (the “mesentery”) is not a feature of hepatitis.

Details matter, patterns matter. The term adenopathy by itself tells us little. Where precisely is it? How large are the nodes? What shape are they? Are these possibly prominent but still normal? Is this pattern compatible with the proposed cause(s)?

It remains a mystery why the original radiologist's suggestion of lymphoma was pushed aside. Perhaps it is relevant to note that the liver specialist was a very prominent professor; could his attribution of "adenopathy" to hepatitis been judged more credible than that of the radiologist? And how much did the history of drug abuse diminish the patient's credibility? There is no way to be sure now.

I feel certain that if the specialist had consulted with a radiologist, the lymphoma could have been addressed weeks earlier. How much that would have made a difference to the patient is less clear. What is clear is that this was a tragedy.

When a major decision is at hand, or when a patient's situation is puzzling, I strongly encourage clinicians to look at the images side by side with a radiologist.  I have seen countless instances where this kind of consultation resulted in a revision of the surgeon's or other clinician's plans. If a side by side meeting is impractical, even a written exchange of questions and answers can be helpful, or a phone call. The conclusions of any dialogue should be noted in writing, just as the rest of the clinician's reasoning is documented (or should be) in their clinical notes.

In my experience working in a busy medical center, I sometimes had the opportunity to hold side by side sessions with inpatient teams. Some non-radiologist physicians were good at understanding the images we reviewed together, while others seemed to be staring blankly at them. But because we could have a dialogue, we could work to arrive at a mutual understanding. This often helped the team decide on their next actions. 

This admonition for dialogue goes against the tide of current practice. Physicians are already busy enough. The pressures of so-called "efficiency" are against spending "non-productive" time; but these terms make sense only when talking about money, not about the patient-doctor relationship (or doctor-doctor relationship for that matter).

Instead, most communication between caregivers takes place via the written record. Text reports simply can't capture all of the nuance on the images. They cannot dynamically address the questions of the clinician, cannot persuade or argue. There is no dialogue.

As a patient, you should be aware of this dynamic. In all likelihood, what your physician is telling you about your imaging test results, he or she has gleaned from reading a written report. When a big decision is at stake, I recommend asking your physician to review the images side by side with a radiologist.

If you have access to the written report, read it. Ask your physician about anything you don't understand. You will be surprised for example at how often the stated reason for the study (the "indication") is wrong or incomplete. If this is the case, ask whether having the correct indication might have influenced the radiologist's interpretation. Non-radiologists do not seem to understand how important the given history is to the radiologist, as evidenced by the fact that they provide so little to them. If the radiologist mentions a possible diagnosis, ask your physician about where that fits into his or her thinking.

Ideally, patients should have the option of reviewing images directly with the interpreting radiologist but in the current system this almost never happens. Radiologists already have too many interruptions; but is this kind of consultation really an interruption? Or does it instead address the core meaning of being a radiologist, a physician? Personally, my direct interaction with other doctors and patients was the best part of my job.

Many factors determine whether a radiology test will contribute meaningfully and accurately to the care of the patient. I discuss each of the following questions in more detail in my paper “Towards better metainterpretation: improving the clinician's interpretation of the radiology report” and suggest how to proceed when the answers are unclear:

– Was the appropriate test ordered?

– Were proper clinical history and indications provided to the radiologist?

– Was the exam technically adequate?

– Has comparison with old studies been done?

– Might there be radiologist perceptual error at play?

– Does the radiologist’s interpretation make sense?

– What exactly is the nature of any “incidental” findings? (“Incidental” can have two different meanings: trivial or unexpected; obviously these two meanings have different implications for action or follow-up.)

–  What actions should be considered based on this report?