Objective: To assess the intra- and inter-observer reliability of Ultrasound (US) in scoring
B-mode, Doppler synovitis and combined B-mode and Doppler synovitis scores in different peripheral
joints of Rheumatoid Arthritis (RA) patients.
Methods: Four rheumatologists with a formal training in Musculoskeletal US (MSKUS) particularly
focus on definitions and scoring synovitis on B-mode and Doppler mode participated in a patient-
based reliability exercise on 16 active RA patients. The four rheumatologists independently
and consecutively performed a B-mode and Power Doppler (PD) US assessment of 7 joints of each
patient in two rounds in a blinded fashion. Each joint was semi quantitatively scored from 0 to 3 for
B-mode Synovitis (BS), Doppler Synovitis (DS), and combined B-mode/Doppler synovitis (CS).
Intraobserver reliability was assessed by Cohen’s κ. Interobserver reliability was assessed by unweight
Results: The mean prevalence of synovitis on B-mode was 83% of joints; scores ranging from
grade 1 in 18% of joints, to grade 3 in 33%. In 55% of joints synovial PD signal was detected and
the distribution of scores range from 14% of joints for grade 3, to 26% for grade 2.
After a total of 448 joints scanned with 896 adquired images our intraobserver and interobserver
reliability was good to excellent for most of the joints.
Conclusion: Formal, structured and continuous training in musculoskeletal ultrasound would bring
a good to excellent reproducibility in rheumatological hands with a high reliability in real time acquisition
BS, DS and CS modalities for scoring synovitis in patients with active rheumatoid arthritis.