SAT0671 Initial development of a whole-body magnetic resonance imaging inflammation index for active disease of peripheral joints and entheses in patients with inflammatory arthritis

Background Magnetic resonance imaging (MRI) allows objective assessment of inflammation in peripheral joints and entheses. MRI scoring systems have until now focused on assessing specific parts of the musculoskeletal system in detail, e.g. the Rheumatoid Arthritis MRI Scoring System (RAMRIS), which is applied to wrist and metacarpophalangeal joints and adjacent tendon sheaths. The interest in a whole-body MRI approach is growing as modern MRI scanners now permit whole-body scanning within an acceptable time frame, and future improvements in MRI hardware and pulse sequences are expected to improve scan time and image resolution further. Objectives To develop a whole-body MRI scoring system for inflammation of peripheral joints and entheses and to investigate its feasibility and reliability. Methods Definitions of the key pathologies and locations for assessment have been agreed upon in the OMERACT MRI Working Group1. In a first round in June 2017, 9 readers (AJM/DG/FG/IE/MO/PB/SJP/SK/WPM) scored MR images of 2 patients with spondyloarthritis using a draft web-based scoring system. Results were discussed and the scoring system was slightly modified. Hereafter, in a second round in October 2017, 14 MRI readers (3 musculoskeletal radiologist (IE/JLJ/RGL) and 11 rheumatologists with varying exposure to MRI (AJM/DG/FG/MS/MO/PB/RP/SJP/SK/VF/WPM), scored 5 similar patients by the modified scoring system. Using a semiquantitative scale 0–3 (none/mild/moderate/severe), synovitis and osteitis were scored separately for 83 joints, and soft tissue inflammation and osteitis were scored separately for 33 entheses. Discrepancies between readers were discussed during an online meeting to obtain consensus, to train inexperienced readers, and to identify potential pitfalls when applying the scoring system. Results Inter-reader reliability was overall moderate for joint scores and poor for enthesis scores; however, among the 3 musculoskeletal radiologists, enthesis scores were as reliable as joint scores (Table). Reliability did not improve between the first and second round, possibly because patients with several very conspicuous inflammatory lesions were selected as cases in the first round. Conclusions It is feasible to perform online multi-reader scoring exercises of whole-body MRI using a web-based scoring interface. MRI readers need to be further trained and calibrated in the semiquantitative scoring approach used to increase inter-reader reliability. Reference [1] Ostergaard M, et al. J Rheumatol2017;44:1699–1705. Disclosure of Interest None declared