MS Patient, Ph.D.: Neuroradiology Reader Roulette
Neuroradiology readings: How often do they miss the gorilla in the film?
The lung CT scan is infamous: Right there in plain sight stands a gorilla, mid-dance amid the pulmonary tissues. Yet in spite of its cartoonish, mocking presence, one study found that a whopping 83% of radiologists missed seeing the thing in their review of the scan.
Shocking, yes?
To me, not so much.
The researcher who did the study, Trafton Drew of Harvard Medical School, has said that the radiologists missed the gorilla because they weren’t looking for a gorilla; they were looking for nodes. But what are radiologists looking for when they miss what they’re supposed to be finding in MRI imaging of the brain and spinal cord? Are they looking for gorillas and missing lesions?
I ask because the question pops up fairly frequently on multiple sclerosis patient message boards: A radiologist misses some important finding on the imaging, one that the patient learns out about later from another clinician who reads the imaging or after a second scan highlights it and warrants a comparison. In one case, a frustrated board commenter posted her imaging that a radiologist had concluded was fine, with no notable findings. Yet, as I pointed out to her, it looked like she had a Chiari malformation, something I knew about only because my youngest son has one. She took her films back to her neurologist, who confirmed it, and she went on to have treatment for the symptoms she was experiencing because of it.
In other cases, though, it’s just been an interesting pastime—rather like my experience with neurologists—of comparing radiology readings to economist predictions: No two conclusions from the same data set are in agreement. What one reader sees, another will overlook entirely or not view as important enough to mention.
As an example, let’s look at my spine. C-spine is great. No compressions, no visualizable lesions (although I do have L’hermittes). But my T-spine is a mess. Two herniations, one at 6-7 and one at 8-9. They don’t warrant surgery and aren’t causing my problems, say two out of three neurosurgeons. (Neurosurgeon 3 wanted to filet me.) I have an arachnoid cyst, too. And, at T2-3, I have a plaque “consistent with demyelinating disease.” And I have scoliosis, which I evidently developed in my late 30s. All of these findings have been present for years now. But of four spinal imaging studies I’ve had in that time, no single study has described each of these findings collectively. That tends to preclude getting a visual of an overall diagnostic picture.
Most readers catch the herniations, although one didn’t—or at least, the radiologist didn’t comment on them. Most catch the arachnoid cyst, although one didn’t. One reader who missed the cyst on 1.5-T imaging caught the T2-3 lesion but simply said, “There is a myelopathy of the upper t-spine.” Another reader of 3-T MRI images of my spine didn’t mention the myelopathy, while a second set of readers looking at the same films did and described its location (dorsal, T2-3) and its specific consistency with demyelinating disease.
Brain readings are similar. I have T2 hyperintensities in every image, but most readers comment only on the asymmetry of a couple around the ventricles and ignore the others, which number above a dozen, including juxtacortical hyperintensities, which have been described as being “specific for MS.”
Then there are the bilateral pleural effusions. They’ve been there for years. Been there since my first MRI in 2008, in fact. Most readers catch them—one was convinced I had lung cancer and said to my doctor at the time, “But she’s so young”—and some don’t. CT scans show nothing. Yet even though herniations or an arachnoid cyst or a plaque can go uncommented, the pleural effusions usually get a mention.
And they’re not even the gorilla the readers are trying to see.
Read other MS Patient, Ph.D. blog posts.