Background/Purpose: Early detection of erosive progression is an important application of MRI in RA. For research purposes the OMERACT Rheumatoid Arthritis MRI Scoring system (RAMRIS) is commonly applied, with semi-quantitative assessment of erosive volume. However, over time new erosive lesions can occur without sufficient increase in volume to result in a higher score. This could allow for subtle erosive progression to go undetected. We aimed to examine if quantifying the number of erosions in addition to erosive volume could provide improved sensitivity to change.

Methods: In a cohort consisting of 84 patients, all fulfilling the 1987 ACR criteria for RA, with disease duration of < 1 year, MRI of the dominant wrist was acquired at 0, 3, 6, and 12 months. A trained reader, blinded for patient data, scored the MRI images in known chronological order according to the RAMRIS method. In addition, we registered the number of erosive lesions at each site. The 78 patients who had completed at least the baseline and 6- or 12-month exams were by each method identified as either progressors or non-progressors, using an increase of 1 unit as cut-off value. The results were compared using cross tables and Spearman correlation.

Results: The median baseline age (25th, 75th percentile) was 58.1 (47.3, 66.4) years, disease duration was 107 (70, 186) days, 77% were female, and 55% were anti-CCP positive. Median baseline RAMRIS erosion score was 8 (5, 11) and the median number of MRI erosions was 10 (6, 14). Median 1 year change in RAMRIS erosion score was 1 (0, 2), and median change in number of erosions was 1 (0, 2). Both methods identified an equal number of patients as progressors, and agreement was observed in 82% of cases. Seven patients were progressors by the RAMRIS method only and 7 by the counting erosive lesions method only (table). The Spearman correlation coefficient of the two scores was 0.75, p<0.001 (figure).  

Conclusion: The RAMRIS erosion score was generally low, both with respect to baseline level as well as progression rate. By supplementing the assessment with quantification of the number of erosive lesions, we could in this material identify a limited number of additional patients as progressors. Further longitudinal studies could show to what extent patients with this type of early minimal erosive lesions on MRI progress on conventional radiographs, and if this is clinically important.