Background MRI is widely used to evaluate muscle inflammation in myositis. Muscle edema on STIR MRI sequences is thought to reflect active inflammation. However, it is unclear whether MRI has an added value over the cheaper measurement of serum creatine kinase (sCK) levels.
Objectives To assess the concordance between sCK and MRI edema in a cohort of patients with myositis at their first presentation to our centers.
Methods We enrolled in 2 Rheumatology centers 73 patients, 34 with dermatomyositis (DM) and 39 with polymyositis (PM) diagnosed according to Bohan and Peter criteria. 33% of patients were untreated. In all patients, sCK were measured and MRI sequences were acquired at the same time. MRI edema (1= present, 0= absent) was assessed bilaterally in 17 thigh and pelvic floor muscles. An MRI composite edema score (0-17) was calculated by adding the separate scores bilaterally and dividing them by two as described elsewhere (1). sCK was considered positive if values were above the upper limit of normal (190 U/l), while MRI was considered positive if the edema score was at least 1. The (single measures) intraclass correlation coefficient (ICC) between the Radiologists involved was 0.78. Muscle strength was measured by MMT (manual muscle testing) and graded according to the Medical Research Council extended scale (0-5). The ICC between the 2 physicians performing the MMT was 0.8.
Results sCK and MRI were discordant in 47% of patients with myositis, more frequently in DM (59%) than in PM (39%). 56% of patients with DM, but only 15% of those with PM had a positive MRI with normal sCK.
There was a significant correlation between MRI score and muscle strength of the hip flexors (Spearman's rho 0.26, p 0.028) but not between MRI score and overall muscle strength. MRI scores did not correlate with sCK. No significant differences were found between treated and untreated patients for any study variable (Mann-Whitney test, p>0.05).
Conclusions MRI is a useful tool to assess disease activity in myositis, especially in DM, where it can be identify a sizeable number of patients who have normal sCK. Further studies of larger cohorts are warranted to confirm our findings.
Clin Exp Rheumatol 2012; 30:570-3
Disclosure of Interest None declared