MRI-defined sacroiliitis and spinal inflammation in chronic back pain: real-world patterns in axial spondyloarthritis and axial psoriatic arthritis.

OBJECTIVES: To evaluate the diagnostic performance of MRI-defined sacroiliitis for axial spondyloarthritis (axSpA) in routine clinical practice and to characterize inflammatory spinal lesions (ISLs) and clinical factors associated with sacroiliitis across non-psoriatic axSpA (nPsA-axSpA), axial psoriatic arthritis (axPsA) and non-inflammatory chronic back pain (NI-CBP). METHODS: We conducted a retrospective study of adults with chronic back pain who underwent SI joint and whole-spine MRI for suspected axSpA. Following longitudinal follow-up, patients were classified by expert rheumatologists as nPsA-axSpA, axPsA or NI-CBP. Two blinded musculoskeletal radiologists independently assessed MRI scans for sacroiliitis and ISLs using Assessment of SpondyloArthritis international Society/OMERACT definitions. Diagnostic accuracy metrics for MRI-defined sacroiliitis were calculated using expert diagnosis as the reference standard and multivariate logistic regression was used to identify predictors of sacroiliitis. RESULTS: Ninety-five patients were included (29 nPsA-axSpA, 21 axPsA, 45 NI-CBP). MRI-defined sacroiliitis showed high specificity (91%) and moderate sensitivity (77-83%) for axSpA diagnosis. Independent predictors of sacroiliitis included age [odds ratio (OR) 1.05/year], regular physical activity (OR 3.50), expert-assessed inflammatory back pain (OR 35.11) and the number of SpA features (OR 1.72). ISLs were present in 40% of patients overall and were most prevalent in axPsA (76%). In axPsA, ISLs frequently occurred in the absence of sacroiliitis and displayed a broader and more heterogeneous distribution. CONCLUSION: MRI-defined sacroiliitis is highly specific but insufficient as a stand-alone diagnostic tool for axSpA in real-world settings and should be interpreted in context.