Do musculoskeletal ultrasound and magnetic resonance imaging identify synovitis and tenosynovitis at the same joints and tendons? A comparative study in early inflammatory arthritis and clinically suspect arthralgia

Abstract Objective Ultrasound (US) and magnetic resonance imaging (MRI) are recommended in the diagnostic process of rheumatoid arthritis. Research on its comparability in early disease phases is scarce. Therefore, we compared synovitis and tenosynovitis detected by US and MRI on joint/tendon level. Methods Eight hundred forty joints and 700 tendons of 70 consecutive patients, presenting with inflammatory arthritis or clinically suspect arthralgia, underwent US and MRI of MCP (2–5), wrist and MTP (1–5) joints at the same day. Greyscale (GS) and power Doppler (PD) synovitis were scored according to the modified Szkudlarek method (combining synovial effusion and hypertrophy) and the recently published EULAR-OMERACT method (synovial hypertrophy regardless of the presence of effusion) on static images. US-detected tenosynovitis was scored according to the OMERACT. MRI scans were scored according to the RAMRIS. Test characteristics were calculated on joint/tendon level with MRI as reference. Cut-off for US scores were ≥ 1 and ≥ 2 and for MRI ≥ 1. Results Compared to MRI, GS synovitis according to EULAR-OMERACT (cut-off ≥ 1) had a sensitivity ranging from 29 to 75% for the different joint locations; specificity ranged from 80 to 98%. For the modified Szkudlarek method, the sensitivity was 68–91% and specificity 52–71%. PD synovitis had a sensitivity of 30–54% and specificity 97–99% compared to MRI. The sensitivity to detect GS tenosynovitis was 50–78% and the specificity 80–94%. For PD tenosynovitis, the sensitivity was 19–58% and specificity 98–100%. Conclusion Current data showed that US is less sensitive than MRI in the early detection of synovitis and tenosynovitis, but resulted in only few non-specific findings. The higher sensitivity of MRI is at the expense of less accessibility and higher costs.