Intermachine agreement of twodimensional grey-scale and power Doppler ultrasound in the follow-up of arthritis patients

Objectives: Power Doppler ultrasound (PDUS) is used for the monitoring of treatment in patients with rheumatic diseases. Different ultrasound (US) equipment is used in specialised and outpatient centres, which raises the problem of the comparability of US examinations. The intermachine reliability of synovitis and Doppler findings was tested for two US machines.

Methods: 13 joints of patients with active arthritis were examined from one investigator during 4 follow up visits under treatment with the TNF-alpha inhibitor adalimumab. US was performed with a HDI 5000 (ATL/Philips, Bothell, WA/USA) using a 5–12MHz linear array transducer. Subsequently, a second US examination was carried out with an Image Point (Hewlett Packard, Bad Homburg/Germany) using a 7,5MHz linear transducer. 2D images were obtained by standardized scans. Synovitis was scored (0–3) and measured (mm), PDUS images were scored (0–3). Intrareader agreement was obtained by rescoring the images.

Results: Intermachine agreement was moderate for the grading of synovitis (k=0.58, Cohen's kappa value) and Doppler grading (k=0.57). Synovitis measurement showed a good correlation (rp=0.77, Pearsons correlation's coefficient, N=52). The intrareader agreement was 0.86 for synovitis and k=0.85 for 2D grading (N=104).

Conclusions: The intermachine agreement for the grading of synovitis and Doppler findings showed only moderate correlations for two different US machines and transducers. This should be considered while using semiquantitative grading systems for multicentre studies.