Diagnostic utility of perilesional muscle edema in myositis ossificans.

Abstract

To investigate the value of extensive perilesional muscle edema for the differentiation between myositis ossificans (MO) and malignant intramuscular soft tissue tumors on MRI. Two blinded readers analyzed MR examinations of 90 consecutive patients with intramuscular soft tissue masses (group 1: MO, n = 20; group 2: malignant tumors, n = 70). Extent of edema around lesions was graded (0, none; 1, minimal edema; 2, moderate edema; 3, extensive edema). Edema-lesion ratio (ELR = ratio of the maximal diameter of the edema and the maximal diameter of the central lesion) was calculated. ROC analysis, Mann-Whitney U test, and Kappa test were used. A total of 70% and 60% of patients with MO had edema grade 3 (reader 1/reader 2), 30%/40% edema grade 2. For the patients with malignant tumors, it was 2.9%/1.4% (edema grade 3) and 16%/23% (edema grade 2). Interrater reliability was substantial (kappa = 0.66). Extent of edema was significantly higher for patients of group 1 (p < 0.0001, both readers). Mean ELR was 3.60 (group 1) and 1.35 (group 2), with statistically significant differences (p < 0.0001). Grade 3 edema showed a sensitivity/specificity of 70%/97.1% (reader 1) and 60%/99% (reader 2) for diagnosing MO. For ELR > 2.0, sensitivity was 90% and specificity 91% for diagnosing MO. Extensive perilesional muscle edema on MRI of more than double the size of the central lesion is highly specific, but not pathognomonic for myositis ossificans in the early/intermediate stage in the differentiation to malignant intramuscular soft tissue lesions.