My Legs Only Hurt When I Stand Up!

Christopher G. Ullrich

In this issue of the AJNR, Hiwatashi et al provide an important reminder of several fundamental medical concepts that have sometimes been forgotten by many radiologists. This study looks at a group of 200 symptomatic spinal stenosis patients examined with MR imaging by using routine supine imaging techniques. By using a device that applied an axial load equivalent to 50% of patient body weight, additional supine axial loaded images were obtained. Upright MR imaging techniques take a similar approach and offer the additional possibility of dynamic imaging. A subset of 20 patients whose images showed substantial change after axial loading was then analyzed by three neurosurgeons with regard to treatment recommendations. On the basis of axial loaded images, changes in surgical therapy would have been made for as many as 10 of these patients. No attempt was made to determine actual improvements in outcomes for these patients.

Most of our statistics for effective imaging tests have tended to look at single variables, such as herniated disk versus no disk herniation, whereas lumbar spinal stenosis is a multivariable disease that can challenge correct interpretation of images. The lumbar spine is a dynamic structure that permits flexibility within well-defined physiologic limits, and, under normal circumstances, no neural compression occurs. As spinal integrity deteriorates, abnormal motion or structural shifting occurs, which may only be evident in a certain position or mechanical loading situation. These more subtle abnormalities can cause patient symptoms. Failure to recognize these more dynamic structural abnormalities can lead to suboptimal surgical therapy in some patients. It is interesting that about 5% of the patients in this study had such a change in recommended surgical therapy. Although this number is not large, this distinction could be very important for that patient group.

Physicians understand that appropriate patient treatment is determined by a combination of patient history, physical examination, and diagnostic testing. As radiologists, we naturally focus on diagnostic testing, and we have great confidence in the quality of our work. Many of us consider routine MR as the only diagnostic test that is needed for evaluating the lumbar spine. As Hiwatashi et al show, for many but not all patients, this assumption is correct. Well-trained spinal surgeons are aware of the dynamic element of lumbar spinal disease, and many of their procedures are predicated on stabilizing symptomatic “instability.” For them, diagnostic imaging is often done to confirm their clinical findings before surgery. If the imaging does not confirm their clinical opinion, additional testing may be needed. Some spine surgeons continue to use lumbar myelography as a problem-solving examination, particularly when they are concerned about dynamic changes in the spine or there is a significant discordance between the history, physical examination, and the routine MR imaging findings. Anyone who performs myelography has seen important structural findings in some patients that were “missed” on routine MR or CT images. Unfortunately many radiologists do not fully appreciate this point and treat myelography as a relatively unimportant and perhaps obsolete examination. The data in the article suggest that a further reduction in the need for myelography is possible with improved MR imaging strategies.

Because many spine surgeons perform imaging as a confirmatory test before surgery that they believe is indicated on the basis of history and physical examination, it can be rationally argued that for a small group of spinal stenosis patients it is more effective to perform CT myelography. This remains the best dynamic evaluation presently available at most medical centers, and the CT assessment of the lumbar spine is actually very effective, especially with multidetector CT and two-dimensional multiplanar reconstructions. If this is done as a presurgical study, additional imaging is almost never needed. In the previously instrumented patient, CT myelography is often the best study available. The same cannot be said for MR imaging.

With these concepts in mind, and with Hiwatashi’s information, it can be suggested that current routine lumbar MR imaging techniques are less than fully adequate for a small group of patients. With further refinements and validation of MR axial loaded and dynamic lumbar studies, the need for lumbar myelography will be further diminished. In the meantime, lumbar myelography is still valuable in deciding on clinical management and surgical approaches for spinal stenosis.