Comparison of bone scan, computed tomography, and magnetic resonance imaging in the diagnosis of active sacroiliitis.

Quantitative bone scan (QBS), computed tomography (CT), and magnetic resonance imaging (MRI) have each been used to confirm the diagnosis of active sacroiliitis (SI) in patients with low back pain (LBP). The authors prospectively evaluated 19 patients referred for symptoms of possible inflammatory LBP (group I), 26 seronegative spondyloarthropathy (SNSP) patients with LBP (group II, inflammatory or mechanical), and 5 SNSP patients without LBP (group III) to determine which radiological scan alone or in combination with other serological tests (Westergren erythrocyte sedimentation rate, C-reactive protein, HLA-B27, immunoglobulin A) was most useful in confirming a clinical diagnosis of active inflammatory SI. All patients were followed up for a minimum of 1 year to confirm the clinical diagnosis and evaluate response to therapy. Eight of 19 group I patients had active SI clinically or on plain radiographs on follow-up evaluation. Of these patients, 5 had abnormal QBS (71%), 3 had abnormal CT scans (38%), and 8 had abnormal MRI scans (100%, type I lesions). These type I MRI lesions were indicative of active inflammation manifested as subcortical bone marrow edema. The remaining 11 group I patients had negative scans for SI. Ten of 26 group II patients with LBP had SI diagnosed clinically and confirmed with positive QBS (60%), CT (100%), and MRI (100%, type I lesions). The remaining 16 group II patients had mechanical LBP without active SI clinically and had negative QBS (88%), CT (19%), and MRI (100%, normal or type II lesions). These type II MRI lesions represented old postinflammatory lesions with either fibrosis or fat replacement. All 5 group III patients had negative scans for active SI. Three patients (2 group I and group II) with inflammatory SI treated with sulfasalazine showed marked improvement on serial MRI scans. Westergren erythrocyte sedimentation rate, C-reactive protein, immunoglobulin A, and CT scan alone or in combination with other tests were not reliable predictors of active SI. Positive QBS and HLA-B27 tests were the best combination of screening tests with 82% predictability of inflammatory SI in whites, and QBS alone had an 80% predictability in black patients. However, MRI, which had 100% predictability, was the best single test for confirming active inflammatory SI.