The relationship between sacroiliac joint abnormalities on magnetic resonance imaging and the development of radiographic sacroliitis in patients with early axial spondyloarthritis

Objective: The diagnosis of ankylosing spondylitis is usually delayed since the development of radiographic sacroiliitis may take about a decade. Magnetic resonance imaging (MRI) is increasingly used to visualize active inflammation in sacroiliac joints (SIJ) and to facilitate earlier diagnosis. However the prognostic significance of MRI abnormalities is unclear. Thus we aimed to evaluate the relationship between MRI changes and the development of radiographic sacroiliitis. We also investigate the predictive factors for the radiographic sacroiliitis.Method: All MRI examinations ordered by the Rheumatology clinic between 26.06.2002 and 29.12.2006 were reviewed and patients who had both baseline and follow up (at least one year apart) plain radiographs were included in the study. ASAS/OMERACT suggestions were used for the definition of MRI findings relevant for sacroiliitis. MRIs were scored for the bone marrow edema using the Leeds scoring system and SIJ radiographs according to modified New York criteria. All MRIs and radiographs were evaluated by one of the investigator and intra-reader reliability was assessed by using linear kappa values. Demographic information, clinical and laboratory features were obtained from the patients records.Results: In the predefined period there were 472 MRI examinations, 50 of them were not MRI of SIJs. Of the remaining 422 patients (322 female [77%]) 167 (39.6%) were eligible for the analysis however one MRI could not be evaluated due top image quality. The demographic and disease related variables were similar for the patients included in the study and who were excluded from the analysis. However our study group was older than exluded patients. The results of the radiologic reports were available for 346 patients and interobserver agreement was moderate for inflammatory changes on MRI (? = 0,550, p<0.001) Intra-observer reliability was good (? = 0,750; p<0.001) for MRI inflammation and perfect for presence of radiographic sacroiliitis according to modified New York criteria. In the re-evaluation of baseline radiographs three patients has unequivocal sacroiliitis therefore they were excluded from the subsequent analysis. Baseline MRI examination showed bone marrow edema (BME) in 53 (32%), and in 24 of them there were ?grade 2 BME in at least one quadrant of SIJs. Ten patients developed radiographic sacroiliitis in the average four years of follow up. Both the presence of BME itself and high grade BME on MRI were correlated with the development of radiographic sacroiliitis. Radiographic sacroiliitis in the follow up was also correlated with the structural changes detected on MRI, male sex, HLA-B27 positivity, and uveitis. In the regression analysis moderate to severe BME and male sex were found to be independent predictors of the development of radiographic sacroilitis.Discussion: The occurrence of SIJ activity by MRI at baseline was related to future development of AS.