OBJECTIVE: To evaluate whether different types of sacroiliac (SI) joint lesions identified by magnetic resonance imaging (MRI) could differentiate axial spondyloarthritis (SpA) from conditions with buttock or pelvic pain attributable to other reasons, including postpartum women and healthy subjects.

METHODS: The study was designed as a prospective, cross-sectional study involving 204 participants, comprising patients with axial SpA (n = 41) and control groups of subjects with or without SI joint pain, including patients with lumbar disc herniation (n = 25), women with (n = 46) or without (n = 14) postpartum buttock/pelvic pain (having given birth within the preceding 4-16 months), hospital cleaning staff (n = 26), long-distance runners (n = 23), and healthy men (n = 29). Participants underwent clinical examination and MRI, and MRIs were evaluated in a blinded manner by 2 readers according to the Spondyloarthritis Research Consortium of Canada (SPARCC) SI joint inflammation and structural lesion scores. SPARCC score cutoff levels were defined as scores above a certain threshold. Primary analyses were based on reader agreement with regard to the presence of SI joint pathologic features on MRI ("concordant reads"). Sensitivity, specificity, and positive and negative predictive values were calculated.

RESULTS: SI joint ankylosis and backfill were detected by MRI only in patients with axial SpA (32% and 37%, respectively), while bone marrow edema (BME) and fat lesions were seen in all non-axial SpA control groups (3-39% with BME and 4-14% with fat lesions). SI joint erosion was present only in patients with axial SpA and in women with postpartum buttock/pelvic pain (at erosion score cutoffs of >1 and >4, 61% and 34%, respectively, in patients with axial SpA, and 9% and 2%, respectively, in women with postpartum buttock/pelvic pain). A SPARCC BME score of ≥5 was present only in patients with axial SpA (56%) and in women with postpartum buttock/pelvic pain (24%), while fat lesions were present, albeit rarely, at high SPARCC cutoff scores in nearly all groups. Of the 38 women from the non-postpartum control groups who had given birth (mean time since birth 9.7 years), 2 (5%) had BME, whereas none had SI joint erosion or fat lesions, and none had a BME score of ≥4.

CONCLUSION: BME and fat lesions were most pronounced in patients with axial SpA, but also occurred in other groups, particularly women with postpartum buttock/pelvic pain. Erosion above a certain SPARCC score threshold as well as backfill and ankylosis were highly specific for axial SpA.