The Spine Blog

Friday, August 22, 2014

Do MRI findings predict outcomes in cervical myelopathy?

Treatment decision-making can be challenging in some cases of cervical spondylotic myelopathy (CSM). The spine community agrees that surgery is indicated in the case of a healthy, middle-aged patient with clinically significant myelopathy. However, there is scant data to guide treatment in patients with mild signs and symptoms, and the natural history of mild myelopathy is unknown. Choosing treatment for the frail, elderly patient with severe disease is also challenging as the benefits may not justify the substantial risks of extensive surgery. Given these difficulties in deciding whether to perform surgery for myelopathy, efforts have been made to determine if demographic, clinical, or radiographic characteristics can predict outcomes after surgery. In the August 15 issue, Dr. Fehlings and his colleagues from Toronto analyzed the associations between MRI findings and baseline and post-operative disease severity in 134 CSM patients. Transverse area (TA) of the spinal cord at the level of maximum compression as well as the presence of intramedullary T2 hyperintensity and/or T1 hypointensity were determined. These were correlated with physical exam findings, mJOA scores, Nurick grade, SF-36 scores, and timed walking test. The investigators found that there were moderate correlations between TA and physical exam findings, Nurick grade, and mJOA scores, both pre- and post-operatively. Correlations between intramedullary cord signal changes and disease characteristics were much weaker and generally not statistically significant. Interestingly, neither TA or cord signal changes were associated with change scores, indicating that patients with severe cord compression and signal changes could still have significant improvement with surgery.


This study is helpful in that indicates that the degree of cord compression is fairly well-correlated with the clinical severity of myelopathy. It should serve as a reminder to the spine physician to be suspicious of the diagnosis of CSM in patients with pronounced findings and mild cord compression on MRI. It does not shed much light on the prognostic role of cord signal change as the study found minimal correlation between cord signal change and disease severity or degree of improvement. Such a result could be due to a true lack of correlation in the study population or represent a Type II error due to lack of power. Some prior studies have suggested that cord signal changes were associated with less post-operative improvement, though that finding has not been consistent. While studies such as this one provide some prognostic information to spine surgeons and patients about what to expect post-operatively, the lack of a non-operative cohort limits how much it can assist with surgical decision-making. For patients struggling with the decision about whether or not to undergo surgery, prognostic models need to be able predict both surgical and non-operative outcomes in order to determine the likely benefit of surgery. Studies that could provide these data are difficult to perform as most spine surgeons feel patients with a diagnosis of CSM should be treated surgically, so there is a lack of equipoise to perform a randomized clinical trial. The best option going forward is probably an observational study in which patients who are felt to be surgical candidates yet decline surgery are followed to get a better sense of the natural history of CSM using modern outcome measures. Until such a study is performed, patients and spine surgeons are left with predictive data from only one side of the treatment decision-making equation.


Please read Dr. Fehlings’s article on this topic in the August 15 issue. Does this change how you see the prognostic role of MRI findings in CSM? Let us know by leaving a comment on The Spine Blog.

Adam Pearson, MD,MS

Associate Web Editor