To evaluate the inter- and intra-observer agreement of ultrasonographic metacarpophalangeal joint static images in patients with rheumatoid arthritis by two investigators with different ultrasonographic experience.

Ultrasonography was performed by the senior on 386 metacarpophalangeal joints respectively in B-mode and 408 in power Doppler of 17 patients with active rheumatoid arthritis. A first interpretation was done and images were stored at examination time. Static images were then read twice by two independent investigators of different experiment (4-year and 1-year experience in musculoskeletal ultrasonography respectively for the senior and the junior).

For the intra-investigator reproducibility kappa or weighted kappa coefficient ranged from 0.74 to 0.99 for the junior and the senior for B and power Doppler mode. For inter-investigator reproducibility kappa or weighted kappa coefficient ranged from 0.61 and 0.98 for qualitative B and power Doppler modes and semi-quantitative power Doppler. The inter-investigator weighted kappa was 0.49 for semi-quantitative B-mode. The agreement between the dynamic exam and the static exam; results were very good for both readers in power Doppler Mode (0.89). In B-Mode, they didn't reach statistical significance. The reliability for measuring synovial thickness was excellent – rho=0.7609 (p=0.00001).

An experienced rheumatologist and a junior achieved high inter- and intra-observer agreement rates for the identification of synovitis and power Doppler activity for static images of ultrasonography. Trainees could use that method as a first step in US learning.