Clinical and MRI predictors of outcome after surgical intervention for cervical spondylotic myelopathy.

BACKGROUND: Cervical spondylotic myelopathy (CSM) is the most common cause of spinal cord dysfunction in older individuals. Controversy remains in terms of the optimal timing and indications for surgical intervention. In this context, it would be of benefit to define clinical and magnetic resonance imaging (MRI) predictors of outcome after intervention for CSM.
OBJECTIVE: We studied subjects with clinically documented cervical myelopathy to evaluate the relationship among preoperative MRI signal change, clinical findings, and outcome after surgical intervention.
METHODS: We performed a retrospective case study of 76 CSM patients who underwent cervical decompressive surgery and who had pre- and postoperative MRI studies available for review. Preoperative clinical findings and MRI abnormalities on T1- (T1WI) and T2-weighted (T2WI) images were correlated with outcomes (Nurick scores; Odom's criteria) following surgical intervention. Postoperative MRIs were performed 2-4 months postsurgery to assess for adequacy of decompression and resolution of preoperative signal changes. The pattern of spinal cord signal intensity was classified as: Group A (MRI N/N), no intramedullary signal intensity abnormality on T1WI or T2WI; Group B (MRI N/Hi), no intramedullary signal intensity abnormality on T1WI and high intramedullary signal intensity on T2WI; Group C (MRI Lo/Hi), low intensity intramedullary signal abnormality on T1WI and high intensity intramedullary signal abnormality on T2WI. Statistical analyses were performed using SAS (version 8.2).
RESULTS: We evaluated 76 patients (57% males, mean age 62 years, range 30-89) who experienced preoperative symptoms for an average of 6.5 months (range 1 month to 9 years). Preoperative MRI studies demonstrated the following: Group A (MRI N/N) = 45; Group B (MRI N/Hi) = 23; and Group C (MRI Lo/Hi) = 8. The mean postoperative follow-up period was 2.5 years (range 2 months to 8.5 years). A positive Babinski sign and the presence of intrinsic hand muscle atrophy showed the greatest association with abnormal preoperative MRI signal change. High preoperative Nurick score, clonus, and leg spasticity were associated with a less favorable postoperative outcome. In Group B (MRI N/Hi), 11/23 (52.17%) patients had recovery to MRI N/N (P < .0001) at their follow-up scan.
CONCLUSIONS: Patients with high intramedullary signal change on T2WI who do not have clonus or spasticity may experience a good surgical outcome and may have reversal of the MRI abnormality. A less favorable surgical outcome is predicted by the presence of low intramedullary signal on T1WI, clonus, or spasticity. These data suggest that there may be a window of opportunity to obtain optimal surgical outcomes in patients with CSM.