Background Many studies have described the use of MRI in established rheumatoid arthritis (RA) and other rheumatic diseases. However there is a lack of studies determining its value in patients with early arthritis. Although some studies have described the correlation of presence or absence of MRI features with clinical diagnosis, few have taken into account (semi-) quantitative measures of MRI findings.
Objectives To evaluate the differential diagnostic value in early arthritis, with an emphasis on the early diagnosis of RA, and to describe the semiquantitative MRI findings observed in early arthritis patients with different initial diagnoses.
Methods Patients with recent onset arthritis from the Leiden early arthritis clinic received an MRI of the wrist, metacarpophalyngeal and metatarsophalyngeal joints on the most painful side. Imaging was performed on a ONI MSK Extreme 1.5T extremity MRI (GE, Wisconsin, USA). Acquired sequences were Coronal T1, coronal T2 fatsat and coronal and axial T1 fatsat after IV gadolinium contrast administration. MRI parameters evaluated were synovitis, tenosynovitis, bone marrow edema and erosions. They were scored according to OMERACT RAMRIS score by a trained reader blinded to clinical data (intrareader correlation coefficient total RAMRIS 0,98). Diagnoses were established by the treating rheumatologists, who were blinded to MRI findings, after two weeks with all clinical and radiological data available. Analyses were performed by Receiver Operator Characteristic (ROC) curves and logistic regression analysis.
Results 58 consecutively included early arthritis patients were studied between August 2010 and February 2011. 43% were male, the mean age was 53 years. Initial diagnosis was rheumatoid arthritis according to ACR87 criteria in 12 patients, undifferentiated arthritis in 27, psoriatric arthritis in 8, osteoarthritis in 5 patients and other diagnoses were made in 6 patients. MRI parameters did not differ between RA and any of the other diagnostic groups, except for tenosynovitis score which was lower in osteoarthritis (1,00 vs 5,83, p<0.01). No significant positive determinants were found for RA. ROC curve analysis showed an area under the ROC for the diagnosis of RA of 0,66 for synovitis, 0,69 for tenosynovitis, 0,44 for bone marrow edema, 0,49 for erosions and 0,61 for the total RAMRIS score.
Conclusions None of the individual parameters evaluated nor the total score adequately discriminated between RA and other diagnostic groups in early arthritis. Further longitudinal studies are required that combine MRI data with clinical data and determine the value of MRI specifically in the group of undifferentiated arthritis patients, in whom the need for supplemental diagnostic tools is the highest.
Disclosure of Interest W. Stomp Grant/Research support from: This research was supported by the Center for Translational Molecular Medicine., A. Krabben: None Declared, D. van der Heijde: None Declared, J. Bloem: None Declared, T. Huizinga: None Declared, M. Reijnierse: None Declared, A. van der Helm-van Mil: None Declared