Pitfalls of computed tomography in diagnosis of discitis.

Sixteen patients with discitis from January 1980 through December 1983 underwent 18 CT scans for initial evaluation. In six scans the study produced a false negative result (sensitivity 63%, 11/16). In three of these six the scan was performed at the wrong disc level, and in three the error was interpretive. During the same time period 6 patients had a CT diagnosis of discitis which proved incorrect (positive predictive value 63%), three of which had fractures, two had normal post discectomy changes, and one had a neuropathic arthropathy. These studies were reviewed in a blinded fashion along with 30 CT scans of post operative patients without clinical or laboratory evidence of discitis. The CT findings in the discitis patients were: (a) anterior paravertebral soft tissue swelling with obliteration of paravertebral fat planes, (b) fragmentation or erosions of vertebral end plates, and (c) paravertebral fluid collection (abscess). Both (a) and (b) were seen in 13/15 patients, (a) alone in 1/15, (b) alone in 1/15, and all three (a, b, c) in 2/15. The CT scan is diagnostic of discitis in those with all three findings. In those patients with only (a) or both (a) and (b), the CT can be suggestive of discitis in the proper clinical setting when correlated with plain film findings: however, these CT findings are also observed in other conditions. Involvement of the spinal canal by inflammatory mass was seen in 6/16 patients with discitis. Low attenuation (hypodensity) of the affected disc was not observed.