• Early Rheumatoid Arthritis
• Magnetic Resonance Imaging
• Outcomes research

Efficient methods for distinguishing rheumatoid arthritis (RA) at an earlier phase from other diseases are strongly desired since early therapeutic intervention improves clinical and radiographic outcomes of RA.1–4 The clinical 2010 RA classification criteria was established based upon the consensus that RA is an inflammatory disease that develops persistent and/or erosive arthritis.2 ,3 Our series of studies as well as the article describing European League Against Rheumatism recommendations for the use of imaging for the clinical management of RA mention that MRI can be used to improve the certainty of a diagnosis of RA above clinical criteria.5 ,6 The present study was undertaken to investigate whether MRI findings of wrist and finger joints improve the diagnostic performance of 2010 RA classification criteria. One hundred sixty-six patients with early arthritis, who do not fulfil the 1987 RA criteria or other international criteria for rheumatic disease at entry with disease duration less than 6 months (median disease duration at entry was 2 months), were consecutively enrolled from the Nagasaki Early Arthritis Cohort at our institution as previously described.6 A total 166 patients are enrolled including 13 patients without obvious swollen joints and 2 patients with typical plain radiographic erosion. Each patient provided a signed consent form to participate in the study, which was approved by the Institutional Review Board of Nagasaki University.

All of the subjects underwent physical examination, blood tests and gadolinium diethylenetriamine penta-acetic acid-enhanced MRI (1.5T system, Sigma, GE Medical Systems, Milwaukee, Wisconsin, USA) of wrist and finger joints on the same day as described previously.6–8 Reference standard RA of the present study was considered as reported previously9 ,10 by the following two definitions: patients disease-modifying antirheumatic drugs (DMARDs) introduced within the 1st year or those who fulfil the 1987 RA criteria at 1 year (table 1). Figure 1 showed the classification of patients by 2010 RA classification criteria at entry or 1987 RA criteria at 1 year. We investigated the diagnostic performance of 2010 RA criteria with or without the finding of MRI-detected pathology. 2010 RA classification criteria classified RA with a sensitivity of 61.9%, a specificity of 82.6%, a positive predictive value of 83.3%, a negative predictive value of 60.6% and an accuracy of 70.5% if reference standard RA is considered as patients DMARDs introduced within the 1st year (table 1). The results were similar if RA is considered as the patients who fulfil the 1987 RA criteria at 1 year (table 1). As compared with symmetrical synovitis and bone erosion, bone marrow oedema was the most useful MRI finding since the positive predictive value of bone marrow oedema (84.9%) was higher than symmetrical synovitis (72.0%) or bone erosion (81.0%) in the reference standard patients with RA (DMARDs introduced within the 1st year). The results were similar if RA is considered as the patients who fulfil the 1987 RA criteria at 1 year (data not shown). We used a decision-tree algorithm that involves initially applying 2010 RA classification criteria, and if the patient does not fulfil these criteria, the MRI-detected bone marrow oedema rule is introduced. The tree algorithm has been shown to differentiate patients more efficiently than the 2010 RA classification criteria alone, exhibiting better sensitivity, negative predictive value and accuracy for the classification of reference standard RA (table 1). The present findings are the first evidence that the diagnostic probability of early RA using the 2010 RA classification criteria is improved by combining these criteria with MRI-detected bone marrow oedema of the wrist and finger joints. Our study may strengthen the statements of the European League Against Rheumatism recommendations for the use of imaging.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Classification of the patients at entry by 2010 rheumatoid arthritis (RA) classification criteria or at 1 year by 1987 RA criteria. 2010 RA classification criteria or 1987 RA criteria was applied towards 166 patients. The former was applied at entry whereas the latter at 1 year, respectively.