We clinicians do our best to practice evidence-based medicine. We seek objective evidence to diagnose disease and monitor treatment. Symptoms and complaints are subjective. They may suggest a disease is present or a pathologic process is responsible, but we need objective data to confirm our subjective impression. The facts, ma’am, nothing but the facts.

But understanding and interpreting objective data is not always straightforward. A case in point is spine MRIs. What can be more objective than an MRI picture showing a herniated disk? A patient comes in with low back pain and leg pain. They do not get better with usual conservative treatment. An MRI shows a bulging disk. Clearly this finding is abnormal and is responsible for the patient’s symptoms, right? The etiology has been determined objectively and treatment for the pathology is indicated, since it will resolve the problem. Right? Not always.

Spine MRIs are notorious for generating objective data that is meaningless and irrelevant to the problem being evaluated. If MRIs are done on patients with no back symptoms of any kind, only 30% of those MRIs are normal. Half of the MRIs show herniated disks. Yet the patients have no back symptoms of any kind.

It is also true that when patients have significant back symptoms, the results of MRI have little bearing on their symptoms. A study out of Europe published in the NEJM in March demonstrated this clearly. The authors looked at 283 patients with significant back pain and radicular symptoms which corresponded to pathology found on MRI. The patients were randomized to surgical vs. conservative treatment. One year later, symptoms were reassessed and the MRI was repeated.

Overall 84% of patients reported a favorable outcome at one year. Of these, 35% still had herniated disks by MRI. Among those patients who reported an unfavorable outcome, 33% had herniated disks by MRI. Looking at those patients with a herniated disk on the follow-up MRI, 85% reported a favorable outcome. For those with no disk herniation, 83% reported a favorable outcome.

The authors did not report on whether patients did better with surgery or with conservative treatment, since their intent was to determine the significance of MRI findings. On this issue, the study showed that MRI findings are not helpful in monitoring treatment outcomes.

MRIs are truly miracles of modern technology. With no radiation exposure, they can look inside the human body and yield amazing anatomic images from beneath the skin surface. But these images are not always significant. MRI findings can be red herrings or incidentalomas. We clinicians always need to be cautious in interpreting MRI findings. While they may show objective pathology, that pathology is not necessarily problematic.

Pictures are not always worth a thousand words.

Richard Fleming, MD