AB1175 Development and preliminary validation of an omeract mri enthesitis scoring system for the ankle in spondyloarthritis

Background Enthesitis is regarded the primary lesion in spondyloarthritis (SpA) and MRI allows sensitive visualisation of entheseal inflammation/damage, but no validated, internationally accepted scoring system exists. Objectives To develop and perform preliminary validation of a novel OMERACT MRI scoring system for assessing ankle enthesitis in SpA patients, and to improve this through iterative multi-reader scoring exercises and calibration sessions. Methods A systematic literature review of MRI studies on enthesitis in SpA identified key inflammatory and structural pathologies. Definitions were agreed by consensus within the OMERACT MRI in arthritis working group. Then, in a first internet-based multireader exercise the Achilles tendon and plantar fascia entheses in 10 ankle MRIs (sagittal T1W, sagittal and Axial T2W fat suppressed) were scored by 15 readers (3 radiologists and 12 rheumatologists), with varying expertise in ankle MRI, for tendon/fascia thickness/signal change, tendon/peritendon signal alteration, retrocalcaneal bursitis, bone spur, erosion and bone marrow oedema, using semi-quantitative scores (0–3: no/mild/moderate/severe pathology). After a subsequent calibration session leading to minor modifications of assessed parameters, 16 ankle MRIs (specifications as above), were scored by 15 readers in exercise 2 with a modified score sheet. Rules were agreed for scoring pathologies. In both exercises, scores for each reader for individual variables were compared and discussed, and mean scores for each variable were determined using descriptive statistics, as were sum scores. Inter-reader agreement was calculated using two-way consistency single measures intra-class correlation coefficient (ICC 3,1) for inflammatory and structural lesions. Results Exercise 1: Mean pairwise inter-reader ICC for combined score of inflammatory and structural variables was 0.65 (range 0.10–0.94), with 75% of values being good/very good (≥0.50). Discussion of results led to minor modifications of parameters to be assessed. Exercise 2: Inter-reader agreement (ICC) for all inflammatory variables combined ranged from 0.26–0.93 among reader pairs (mean 0.64; median 0.66; IQR 0.46–0.79). For structural variables combined ICC ranged from 0.05–0.91 among reader pairs (mean 0.45; median 0.45; IQR 0.2–0.6). Conclusions Initial steps in developing an OMERACT MRI heel enthesitis scoring system have demonstrated overall moderate reliability of the proposed variables. Further modification, refinement, calibration and validation (ongoing) are needed before this system is ready for use in SpA clinical trials. Disclosure of Interest None declared