Scoring ultrasound synovitis in rheumatoid arthritis: a EULAR-OMERACT ultrasound taskforce<b>—</b>Part 1: definition and development of a standardised, consensus-based scoring system

TLDR

Further evaluation is needed to assess the performance of the new ultrasound scoring system before it can be applied in clinical trials. The study aimed to develop a consensus‑based ultrasound definition and quantification system for rheumatoid arthritis synovitis. A multistep, iterative process was employed, involving baseline agreement assessment of grey‑scale and power Doppler scoring on static images and patient scans, followed by a Delphi exercise to refine definitions and scoring components. Baseline reliability varied, being higher for static images; power Doppler scoring showed excellent intra‑ and inter‑reader agreement, whereas grey‑scale scoring was more variable; after refinement, the consensus definition incorporated hypoechoic synovial hypertrophy and power Doppler signal with a semi‑quantitative severity score, yielding a standardized system that combines individual and combined components.

Abstract

Objectives To develop a consensus-based ultrasound (US) definition and quantification system for synovitis in rheumatoid arthritis (RA). Methods A multistep, iterative approach was used to: (1) evaluate the baseline agreement on defining and scoring synovitis according to the usual practice of different sonographers, using both grey-scale (GS) (synovial hypertrophy (SH) and effusion) and power Doppler (PD), by reading static images and scanning patients with RA and (2) evaluate the influence of both the definition and acquisition technique on reliability followed by a Delphi exercise to obtain consensus definitions for synovitis, elementary components and scoring system. Results Baseline reliability was highly variable but better for static than dynamic images that were directly acquired and immediately scored. Using static images, intrareader and inter-reader reliability for scoring PD were excellent for both binary and semiquantitative (SQ) grading but GS showed greater variability for both scoring systems (κ ranges: −0.05 to 1 and 0.59 to 0.92, respectively). In patient-based exercise, both intraobserver and interobserver reliability were variable and the mean κ coefficients did not reach 0.50 for any of the components. The second step resulted in refinement of the preliminary Outcome Measures in Rheumatology synovitis definition by including the presence of both hypoechoic SH and PD signal and the development of a SQ severity score, depending on both the amount of PD and the volume and appearance of SH. Conclusion A multistep consensus-based process has produced a standardised US definition and quantification system for RA synovitis including combined and individual SH and PD components. Further evaluation is required to understand its performance before application in clinical trials.