CERVICAL EPIDURAL SPINAL INFECTION - MR-IMAGING CHARACTERISTICS

OBJECTIVE. Cervical epidural spinal infection, which includes diskitis , osteomyelitis, and/or epidural space infection, is an uncommon disea se, and the MR imaging features have not been fully described. Accordi ngly, the objective of this study was to determine the MR imaging find ings in a large series of patients with this condition, with emphasis on patients with pyogenic infection. MATERIALS AND METHODS. During a 5 -1/2-year interval, 12,695 MR examinations of the spine were performed in our department. A computer search retrospectively identified 60 pa tients with proved epidural spinal infection; 19 (32%) had involvement of the cervical spine. Pyogenic infection was present in approximatel y 90% of cases. One patient had an epidural abscess without an underly ing osseous abnormality. The remaining 18 patients had cervical diskit is/osteomyelitis, as well as inflammation in the epidural space. All M R examinations were performed with a 1.5-T superconducting magnet by u sing a combination of spin-echo, fast spin-echo, and/or gradient-echo pulse sequences; 12 patients received IV gadopentetate dimeglumine. Th e MR examinations were retrospectively reviewed by a neuroradiologist for multiple imaging characteristics, including extent and location of osseous abnormalities, extent and location of inflammation in the epi dural space, likelihood of abscess formation, and frequency of spinal cord compromise. RESULTS. On MR images, an average of three vertebral bodies showed abnormal signal of the bone marrow. Inflammation in the epidural space extended an average of four levels and was most frequen tly anterior in location; MR images in six (50%) of 12 patients who re ceived contrast material showed a large, peripherally enhancing epidur al abscess. Thirteen (72%) of the 18 infections with osseous involveme nt occurred at or below the level of C4; the C5 and C6 levels were aff ected most frequently. Cord compression was identified in 14 (74%) of 19 patients; cord hyperintensity was seen in 12 (63%) of 19 patients a nd was well defined on T2-weighted FSE images. CONCLUSION. The results of our study show that cervical epidural spinal infection is a more a ggressive disease than previously recognized. MR findings are characte rized by multilevel involvement, an anterior location, and frequent ab scess formation. MR evidence of isolated involvement of the epidural s pace, without underlying osseous abnormality, is rare. There is a pred ilection for infection to affect the lower cervical segments. We obser ved a strong association between cervical epidural spinal infection an d spinal cord compromise, which makes prompt diagnosis and treatment m andatory.