POS0672 RHEUMATOLOGISTS OVERCALL SACROILIITIS ON X-RAY AND MRI IN AXIAL SPONDYLOARTHRITIS PATIENTS: DATA FROM THE BELGIAN INFLAMMATORY ARTHRITIS AND SPONDYLITIS COHORT (BE-GIANT)

Background Imaging of the sacroiliac joints, especially with magnetic resonance imaging (MRI), is an important tool for early diagnosis axial spondyloarthritis (axSpA). Interpretation joint can vary according to readers’ experience, but it currently unknown if and how assessment differs between academic hospitals community based rheumatological care. Objectives To investigate (1) agreement local central reading images (X-ray MRI) from axSpA patients, (2) explore potential differences patients diagnosed in hospital compared centres. Methods The BelGian Inflammatory Arthritis spoNdylitis cohort (Be-GIANT) includes newly biological-naïve that fulfil ASAS classification criteria, at outpatient clinic eight centres Flanders. X-ray MRI joints (SIJ) enrolled November 2010 August 2020 were assessed by rheumatologist (‘local reading’) two calibrated readers (‘central definite radiographic sacroiliitis modified New York criteria (X-SIJ) active ASAS/OMERACT definition a positive (MRI-SIJ). Central resolved discrepant cases consensus. Inter-reader reliability was Cohen’s Kappa, % overall, negative agreement. Results Among 271 included (n=205 hospital, n=66 hospital), 231 X-SIJ 208 MRI-SIJ available (Table 1). disagreed on 30/231 (13%) (κ=0.44, moderate); 4/231 (1.7%) reclassified as 26/231 (11.3%) not showing sacroiliitis. Overall higher rheumatologists (90.5% vs. 70.7%, p<0.001). 53/208 (25.4%) (κ=0.36, fair); majority (51/208, 24.5%). agreed proportion versus (77.2% 63.4%, p=0.07). Conclusion In prevalence low. Sacroiliitis overcalled both non-academic settings, underscoring need continuous educational trainings. Table 1. Agreement centers. All Academic Community Local (N=231 ) + 41 (18%) 19 (8%) 30 (16%) 14 (7%) 11 (27%) 5 (12%) - 190 (82%) 212 (92%) 160 (84%) 176 (93%) (73%) 36 (88%) Positive Negative Kappa (95% CI) 87.0% 50.0% 92.5% 0.44 (0.28 – 0.60) 90.5% 59.1% 94.6% 0.55 (0.37 0.72) 70.7% 25.0% 81.8% 0.10 (-0.20,0.40) (N=208 181 (87%) 132 (63%) 151 (90%) 115 (69%) 17 (41%) 27 76 (37%) 16 (10%) 52 (31%) 24 (59%) 74.5% 83.1% 48.5% 0.36 (0.25 0.48) 77.2% 85.7% 44.1% 0.35 (0.20 0.49) 63.4% 68.1% 57.1% 0.32 (0.10 0.55) REFERENCES: NIL. Acknowledgements: Disclosure Interests None Declared.