Background/Purpose:

To evaluate the agreement between ultrasound (US), whole body magnetic resonance imaging (WBMRI) and clinical assessment of joint inflammation and enthesitis in rheumatoid arthritis (RA) patients, by comparison on joint/enthesis level and by evaluating the correlation between composite scores at patient level.

Methods:

US, WBMRI and clinical assessment for tender joints (TJ) and swollen joints (SwJ) were performed in 19 RA patients (90% Women, median (range) age 55 (26-73), diseases duration 5.5 (1-42), SwJ(28) 5 (1-13), TJ(28) 7 (2-24) and DAS28-CRP 4.66 (3.48 -6.66))) fulfilling ACR 1987 criteria for RA. The 28 conventional joints, bilateral ankles and MTP 1-5, and the entheses of supraspinatus, gluteus muscles, quadriceps and Achilles tendon were assessed by WBMRI and US.

Joint inflammation by US was graded 0-3 on B-mode and colour Doppler (CD), respectively, and subsequently converted to +/- for both components by defining US synovitis as B mode ≥2 or CD ≥1. US finding of enthesitis was defined by presence of CD activity <2mm from the cortical insertion, with or without erosions, or enthesophytes/calcifications. For WBMRI, joint inflammation was defined as presence of synovitis and/or osteitis, and enthesitis as presence of soft tissue inflammation and/or osteitis. For both modalities, the max score was 2 for a joint and 1 for an enthesis.

To assess the total inflammatory burden, a composite score was established as sum scores for the 28 conventional joints for US (US28) and for 26 joints (WBMRI26) for WMBRI – same 28 joints except elbows (due to poor image quality). The agreement between the clinical joint assessment, US and WBMRI for joint inflammation and enthesitis was calculated with Cohen’s kappa (κ). The correlations between US28, WBMRI26 and DAS28-CRP were calculated by Spearman correlation coefficient (rho).

Results:

US28 and WBMRI26 sum scores showed good correlation rho= 0.72 (p=0.003) (Fig. 1), whereas US28 and WBMRI26 did not correlate with DAS28 CRP (rho=-0.26, p=0.28; rho=0.20, p=0.47 respectively). Moderate-good agreement was found between US and WBMRI in wrists and MCP 1, 2 and 5(κ= 0.42–0.62) but poor in other joints (κ ≤ 0.37). Agreement between US and clinical joint tenderness was poor (all κ<0.31; except κ=0.42 for wrist), but better between US and clinical swelling for shoulders, elbows, MCP1 and PIP5 (κ=0.42–0.66) while ≤0.36 for other joints. Agreement between WBMRI and clinical joint swelling or tenderness showed κ<0.35 in all joints.

Enthesitis was rare in RA patients, when defined by positive Doppler signal on US (only 3 cases were detected) and hence poor agreement (κ<0.1 at all entheses) was found between US and WBMRI.

Conclusion:

WBMRI and US sum scores of joint inflammation showed good correlation in RA patients but the agreement at joint level was variable. The agreement on enthesitis between MRI and US was low, but findings were minimal. No correlation with DAS28 was found for either modality.

« Back to 2018 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/agreement-between-ultrasound-and-whole-body-magnetic-resonance-imaging-assessment-of-joint-inflammation-and-enthesitis-in-rheumatoid-arthritis-patients/