LAUSR

Background A fatty corner lesion (FCL), as defined by fat infiltration at the vertebral corner in T1 Magnetic Resonance Imaging (MRI) sequence, has been reported to be useful in Spondyloarthritis (SpA) diagnosis. Objectives Our objective is to systematically evaluate the diagnostic utility of FCLs in a large, multicenter back pain-MRI cohort in Hong Kong. Methods Three hundred and one patients with back pain were recruited from 3 rheumatology centres. Clinical, biochemical and radiological parameters were collected and all patients underwent a whole spine and sacroiliac (SI) joints MRI. FCLs from C4 to L5 levels were scored. Results were compared to expert-diagnosed axial SpA to determine the usefulness of FCLs in disease diagnosis. Results Anterior FCLs of whole or thoracic spine were found to be useful in axial SpA diagnosis. (AUC 0.622; p=0.003, AUC 0.640; p=0.001 respectively) Applying FCLs to the Modified New York (MNY) criteria and Assessment of SpondyloArthritis international Society (ASAS) axial SpA criteria, a minimum of 5 FCLs at whole spine level had a sensitivity of 66.4% and 91.6%, specificity of 98.4% and 91.9% respectively. Applying a minimum of 3 FCLs at thoracic spine level to the MNY criteria and ASAS axial SpA criteria yielded a sensitivity of 68.5% and 92.0%, specificity of 95.2% and 93.5% respectively. Three FCLs improves both classification criteria. Conclusions FCLs are useful in axial SpA diagnosis. A minimum of 3FCLs at the thoracic level is useful for the disease diagnosis. Disclosure of Interest None declared