Objective: To evaluate the usefulness of magnetic resonance imaging (MRI) in predicting destructive rheumatoid arthritis (RA) in the knee joint, and to compare this method with clinical examination and conventional radiography (CR).
Methods: Clinical evaluations of the knee joint, followed by MRI and CR examinations were performed in 30 patients with early RA. The MRI examination included evaluation of inflammation using a synovitis score and evaluation of destruction with an erosion score. The first examinations were performed within 14 months from disease onset. Twenty-eight patients were re-examined after 1 year, and 23 patients after 3 years. 'Disease activity score' (DAS), using a 28 joints score (DAS28); health assessment questionnaire (HAQ); rheumatoid factor (RF); and C-reactive protein (CRP) were also analysed.
Results: At baseline, MRI found synovitis in 29 patients, of whom 18 also had clinical synovitis. At baseline five patients had 17 MRI erosions, whereas on CR two patients had one erosion each. After 1 year 17 of 35 and after 3 years 28 of 55 MRI erosions were detected also on CR. In only one case CR showed an erosion that was not visible on MRI. The MRI synovitis score (reflecting the extent of the synovitis) at baseline correlated significantly with the number of erosions on MRI both at year 1 and 3, and with the number of erosions on CR at 3 years. In logistic multiple regression analyses the MRI-synovitis score proved to be the best independent predictor of erosiveness.
Conclusion: MRI was superior to clinical examination and CR in detecting erosions. MRI synovitis score was the best independent predictor of erosiveness in the knee joint in patients with early RA.