MRI of skeletal muscles in patients with idiopathic inflammatory myopathies: characteristic findings and diagnostic performance in dermatomyositis

Objective To define the characteristic findings on MRI of skeletal muscles in patients with dermatomyositis (DM) relative to those in patients with other idiopathic inflammatory myopathies (IIMs) and to assess their diagnostic performance in DM. Methods Thirty-six patients with DM, 17 patients with amyopathic DM, 19 patients with polymyositis and 16 patients with non-IIM classified by the 2017 European League Against Rheumatism/American College of Rheumatology criteria were included in this study. The following MRI findings (short-tau inversion recovery [STIR] and gadolinium-enhanced fat-suppressed T1-weighted imaging [Gd-T1WI]) for proximal limb muscles were compared between the disease groups and between myositis-specific autoantibodies/myositis-associated autoantibodies (MSAs/MAAs)-positive and MSAs/MAAs-negative groups: structures with high signal intensity (HSI) (subcutaneous, fascia, muscle); distributions of HSI areas in muscle (diffuse, patchy, peripheral) and patterns of HSI in muscle (honeycomb, foggy, strong HSI). Univariate, multivariate and receiver-operating characteristic [ROC] analyses were performed to assess the diagnostic performance of MRI in DM. Results The characteristic MRI findings in patients with DM were subcutaneous HSI, fascial HSI, peripheral distribution and honeycomb pattern. The MRI findings in the MSAs/MAAs-positive group included more frequent fascial HSI but less frequent foggy pattern compared with the MSAs/MAAs-negative group. Likelihood of DM score ≥ 3 (obtained by counting the number of characteristic MRI findings in patients with DM) showed good diagnostic performance in DM (STIR: sensitivity 72.2%, specificity 88.5%, area under ROC curve [AUC] 84.9%; Gd-T1WI: sensitivity 81.2%, specificity 91.5%, AUC 89.9%). Conclusion The characteristic MRI findings of skeletal muscles can predict patients with DM as well as patients with MSAs/MAAs.