Significance of CSF area measurements in cervical spondylitic myelopathy

Mild clinical myelopathy can occur without cord compression, and asymptomatic cord compression seen on MRI is common. The aim of this study was to ascertain the MRI features which best correlate with early clinical myelopathy. The study was conducted on three groups: group A, 20 patients with clinical myelopathy and MRI evidence of cervical spondylosis; group B, 20 patients without myelopathy, but with other clinical and MRI evidence of cervical spondylosis; and group C, 10 normal volunteers with no MRI evidence of spondylosis. The cross-sectional area (CSA) of the spinal cord (SP-CSA), spinal canal (SC-CSA) and CSF space (CSF-CSA) were measured on T1-weighted axial images at the level of the most severe spinal canal stenosis. The severity of myelopathy was assessed using a simple scoring system giving a score from 0 (normal) to 11 (severe). Subjective demonstration of cord compression on sagittal images was an insensitive indicator of clinical myelopathy. All three measures of cross-sectional area were significantly smaller in Group A than in B (p<0.01). The reduction in SP-CSA was the only independent prognosticator for severity of myelopathy (p<0.005) accounting for 63% of the variation in myelopathy score. All three variables showed a significant correlation with the presence of myelopathy (p<0.01); however, logistic regression analysis showed a decrease in CSF-CSA to be the only independent significant prognosticator of the presence of clinical myelopathy (p<0.02). Reduction of the CSF space to less than 0.7 cm2 was associated with a 90% chance of clinical myelopathy (specificity 83%).