Cervical myelopathy: a comparison of magnetic resonance and myelography

Fifty-seven patients with a strong clinical suspicion of cervical myelopathy were studied with body coil magnetic resonance (MR) and conventional myelography or CT myelography. Eight patients were believed to have normal studies with both modalities. There were six patients with syringomyelia; four with an intramedullary tumor; one with an arteriovenous malformation; 19 with cervical spondylosis at multiple levels; eight with cervical spondylosis at a single level; four with extensive rheumatoid arthritis; four with extradural neoplasm; two with trauma; and one patient with an epidural abscess. In this study, body coil MR was the superior examination for the evaluation of an intramedullary process. It was as diagnostic as myelography in one case of an extramedullary intradural lesion. In patients with extradural disease, body coil MR was the superior study in 45%, equivalent to myelography in 37%, and, although still diagnostic, inferior to myelography in 17%. In 8% of the cases, body coil MR was at best equivocal, whereas myelography was diagnostic. It appears that in technically adequate studies, MR is at least equivalent to myelography in its ability to delineate disease. A superior MR study provides a better appraisal of the size and character of the spinal cord as well as the degree of both anterior and posterior defects on the subarachnoid space and neural structures. In addition, MR is as good as conventional myelography for the identification of extrinsic cervical cord lesions producing cervical myelopathy. Finally, an additional small group of 30 patients were studied with a prototype surface coil to determine its advantages relative to body coil imaging. Each patient had correlative myelography. As with body coil MR, imaging with the surface coil was believed to be more informative than conventional myelography in four patients with intramedullary lesions. The remaining 26 patients suffered from cervical spondylosis. Surface coil MR was believed to be more informative than myelography in six cases (23%), equivalent to myelography in 19 (73%), and less diagnostic than myelography in one (4%). The improved spatial resolution with the use of the surface coil was believed to increase the accuracy of MR.