Background MSUS helps the management of rheumatic diseases being able to show the level of tissue inflammation as Doppler signal. This signal might be scored in semi-quantitative and quantitative way. Several semi-quantitative scoring systems (SQS) have been developed but no one is subjective free. Quantitative Doppler score (QS) seems to be more objective (Terslev et al) being based on the computerized evaluation of the ratio between colored and non-colored pixels (CR). The present QS is based on the assessment of a single still image – usual the one with the highest Doppler signal; this approach does not exclude the subjective vulnerabilities (e.g. frame selection). We hypothesized that a Continuous quantitative assessment - CQS - (i.e. a quantitative assessment of all frames included in a video-loop) might be superior to single frame QS.
Objectives To test the ability of MS sonographers to identify the highest Doppler signal (highest CR) images from a group of frames extracted from the same loop. To test the reliability of a CQS method in rheumatoid arthritis
Methods 14 sonographers with high (4), medium (4) and basic (6) expertise in MSUS have participated in 3 phases of this project. In the 1st exercise we asked each participant to identify the frame with highest CR from a group of 4 frames (extracted from the same loop); the 4 frames had a CR decreasing in steps of 10%. The exercise included ten such groups. The 2nd exercise used a CR decreasing in steps of 2%. In the 3rd exercise the sonographers scored independently 10 MSUS video-clips and 10 static images of RA patients; the 0-3 SQS have been used. A blinded researcher analyzed each video and still image in CQS system twice; the average and the peak of CR have been noted for videos. CQS and SQS scores have been compared.
Results In 1st exercise the highest CR frame was identified in 7.5 (1.2) cases; in the 2nd exercise the performance was lower: 2.4 (2,1) cases out of 10. In both cases the agreement was poor. In the 3rd exercise the test-retest coefficients for CQS: 0.985 and 0.949 for average and, respectively, peak CRs. The average SQS scores correlated well with CQS CR (r=0.74 for average and r=0.76 for peak CR). The correlation was better for video clips (r=0.815) than for static images (r=0.546) - for average CR.
Conclusions Continuous quantitative assessment might be superior to single frame quantitative assessment in regular sonographers as the testers shown a moderate to poor performance in identification of the most significant picture. The test-retest performance and the higher correlation of CQS of video clips with SQS scores might recommend it as a reliable technique to assess Doppler loading in MSUS images.
Acknowledgements Sf. Maria residents.
Disclosure of Interest None declared