Interobserver reliability in ultrasound assessment of rheumatoid wrist joints.

OBJECTIVE

To evaluate interobserver reliability in the ultrasound assessment of synovitis in the radiocarpal (RC), midcarpal (MC) and ulnocarpal (UC) joints in RA.

METHODS

Ultrasound examinations of 295 rheumatoid wrist joints were performed over a three month period. The RC, MC and UC joints were examined using dorsal longitudinal ultrasound scans. Synovial thickening was assessed by quantitative measurement and a previously established semi-quantitative scoring system (Grades 0 to 3). Interobserver reliability was determined by the comparing the findings of two radiologists who were unaware of each other findings.

RESULTS

The intraclass correlation coefficient (ICC) between examiners for the quantitative measurement of synovitis in the RC, MC and UC recesses were 0.508, 0.346 and 0.240 (p<0.001), respectively. Weighted kappa values using the semi-quantitative scoring system were 0.308, 0.312 and 0.153 for the RC, MC and UC joints, respectively.

CONCLUSION

Interobserver reliability of the ultrasound assessment in rheumatoid wrists proved good for the quantitative measurement of synovitis in the RC joint, but poor agreement was found for the MC and UC joints. Using the semi-quantitative scoring system, interobserver agreement was poor for all three joints (RC, MC and UC).