Objective: To compare different magnetic resonance imaging (MRI) based algorithms for assessment of spinal inflammation in patients with ankylosing spondylitis (AS) being treated with disease modifying drugs.
Methods: Eleven patients (10 men, 1 woman) who fulfilled modified New York diagnostic criteria and had severe disease [Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) > 4] were given intravenous infusion of infliximab (Remicade, 5 mg/kg) for 96 weeks. Whole-spine MRI was done at 0, 24, and 54 weeks. Measurements of the Ankylosing Spondylitis Spinal MRI Activity Score (ASspiMRI), paravertebral inflammatory lesion count (pILC), contrast:noise ratio (CNR) measurements of defined inflammatory lesions, and other scores together with C-reactive protein concentration were made at each visit. Examinations were anonymized and randomly presented twice to 2 radiologists. The significance of any changes in scores, their correlation with the BASDAI, and interobserver and intraobserver correlations were calculated.
Results: The mean (+/- SD) BASDAI improved from 7.2 (+/- 1.5) to 1.3 (+/- 0.9) after 54 weeks (p < 0.001), and the ASspiMRI score improved from 12.0 (+/- 8.0) to 0.2 (+/- 0.5) (p < 0.001). Correlations between ASspiMRI score and BASDAI were 0.831, 0.746, and 0.369 (p < 0.001 each). The pILC improved significantly (p < 0.01). CNR showed no correlation with any clinical score.
Conclusion: The ASspiMRI score performed best for assessment and quantification of spinal inflammation and disease activity in patients with AS, but should also quantify paravertebral inflammatory lesions, since we could show that this will significantly improve its correlation to clinical scores and increase its sensitivity to mild inflammatory processes.