Arthritis & Rheumatism, Volume 60,
October 2009 Abstract Supplement
The 2009 ACR/ARHP Annual Scientific Meeting
Philadelphia October 16-21, 2009.
Does Fat Infiltration in the Sacroiliac Joint Contribute to the Diagnostic Utility of MRI in Ankylosing Spondylitis?
Weber1, Ulrich, Pedersen2, Susanne J., Hodler1, Juerg, Ostergaard2, Mikkel, Lambert3, Robert GW, Maksymowych3, Walter P.
Fat infiltration in bone marrow of the sacroiliac joint (SIJ) and spine is frequently detected on T1-weighted (T1W) MRI in patients with AS but also in healthy individuals and those with other causes of back pain. There have been no systematic studies to address its diagnostic utility in AS. We aimed to evaluate the diagnostic utility of fat infiltration in the SIJ per se and to determine whether any morphological features characterize pathological infiltration.
Six experienced readers (4 rheumatologists, 2 radiologists) from 3 international centres, blinded to patient and diagnosis, independently assessed MRI scans (T1W and STIR) from the following subjects all aged < 45 years: 30 patients with AS and symptom duration 5 years or less; 34 controls (26 with mechanical low back pain (mLBP)). MRI scans of the SIJ were read systematically as described in a standardized online training module using an online data entry system. In addition to deciding on the presence/absence of AS readers were asked to record the following: the primary MRI feature on which the diagnosis of AS was based, the presence/absence of fat infiltration, the degree to which fat infiltration was considered due to AS (010 scale), its location (sacral, iliac, both), and associated morphological features (bone edema, erosion, sclerosis, ankylosis, distinct border, infiltration adjacent to iliac/sacral joint surface, homogeneity of infiltration). We calculated sensitivity, specificity, and likelihood ratios.
Fat infiltration was considered present in 28/30 (93.3%) of AS patients by a majority (>=4) of readers and in 29/30 (96.7%) by at least 2 readers but stated to be the principal diagnostic feature by only 2 readers in only 3 patients. Only one AS patient was considered to have no infiltration by any reader. It was also present in 6/34 (17.6%) of controls by a majority (>=4) of readers and in 13/34 (38.2%) by at least 2 readers (sensitivity 96.7%, specificity 61.8% for AS as recorded by at least 2 readers). Only 20/34 (58.8%) of controls were considered to have no fat infiltration by any reader. Applying a pre-specified definition of any 2 readers recording fat infiltration and each reader scoring >=7 for fat infiltration considered due to AS gave a sensitivity of 80% and specificity of 97.1%. At least 1 associated feature was recorded in all patients by all readers. Table values represent concordant data from at least 2 readers.
|Both sacral and iliac||70%||94.1%||11.9||0.32|
|Adjacent to sacral/iliac endplate||90%||88.2%||7.6||0.11|
Fat infiltration in the SIJ per se has high sensitivity but low specificity for AS. Its diagnostic utility primarily reflects the presence of associated abnormalities, especially erosions, characteristic of AS.
To cite this abstract, please use the following information:
Weber, Ulrich, Pedersen, Susanne J., Hodler, Juerg, Ostergaard, Mikkel, Lambert, Robert GW, Maksymowych, Walter P.; Does Fat Infiltration in the Sacroiliac Joint Contribute to the Diagnostic Utility of MRI in Ankylosing Spondylitis? [abstract]. Arthritis Rheum 2009;60 Suppl 10 :541