Objectives To compare a 0.2T extremity MRI to a 1.5T conventional MRI for detecting patients with synovitis, bone erosions and bone marrow oedema (BME) among patients presenting with inflammatory hand pain.
Methods Consecutively referred primary care patients participating in REACH1 with unexplained inflammatory hand pain <12 months had assessment of the wrist and MCP joints of the most painful hand on a 0.2T extremity MRI and a 1.5T conventional MRI. Both imaging procedures were performed using contrast and followed the recommendations of the OMERACT Rheumatoid Arthritis MRI Score (RAMRIS). The examinations were scored independently and in random time order by two musculoskeletal radiologists for synovitis, erosions and BME using the RAMRIS. A patient with synovitis was defined as ≥1 joint with a synovitis score ≥2, a patient with erosions was defined as ≥1 bone with an erosion score ≥2, and a patient with BME was defined as ≥1 bone with a BME score ≥1. We used cut-off points of 2 for synovitis and erosions, because grade 1 synovitis and erosions are often present in healthy controls.2-3 Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated for the 0.2T MRI with the 1.5T scanner as reference standard. Kappa values were calculated to determine inter-reader agreement in detecting patients with pathology.
Results 24 out of 40 patients presented with clinically swollen joints, while the remaining 16 patients had only painful joints. The 1.5T MRI detected 20 patients with synovitis, 19 patients with erosions, and 17 patients with BME. The sensitivity, specificity PPV and NPV of the 0.2T MRI are shown in the table. Inter-reader agreement for the proposed case definitions for the 0.2T and the 1.5T MRI respectively, were: 0.50 and 0.75 for scoring synovitis, 0.23 and 0.53 for scoring erosions, and 0.15 and 0.17 for scoring BME.
Conclusions 0.2T MRI is as sensitive as 1.5T MRI in detecting patients with synovitis and erosions, but had only moderate specificity, and lacks sensitivity for BME. Inter-reader reliability of 0.2T was moderate for synovitis, fair for erosions and low for BME.
Alves et al. Ann Rheum Dis 2011;70(9):1645.
Ejbjerg et al. Arthritis Rheum 2004;50(4):1097.
Palosaari et al. Scand J Rheumatol 2009;38(6):450.
Disclosure of Interest None Declared