MRI of musculoskeletal extraspinal tuberculosis.

PURPOSE: The aim of this study was to describe the MR findings in extraspinal musculoskeletal tuberculosis (EMT).

METHOD: A retrospective review was conducted of the MR findings of 18 patients with microbiologically and/or pathologically proven EMT. All MR studies were performed using T1-and T2-weighted spin echo sequences. T1-weighted spin echo sequences after Gd-DTPA injection were obtained for 12 patients. The MR images were evaluated for abnormalities in joints, bones, and soft tissues, and the results were grouped by anatomic localization, frequency distribution of structures affected, and morphologic patterns of involvement.

RESULTS: Isolated soft tissue tuberculosis was found in 10 (55.5%) patients and involvement of more than one structure in 8 (44.4%). Pyomyositis (n = 6) and arthritis with involvement of adjacent soft tissues (n = 7) were the most common forms of presentation. One patient presented with isolated fascial superficial tissue involvement in one leg. Isolated pyomyositis involving one (n = 3) or two (n = 3) muscles was homogeneous in six cases and showed intermediate (n = 6), low (n = 2), or high (n = 1) signal intensity on T1-weighted images and a high and very hyperintense signal on T2-weighted images. The tenosynovitis synovial fluid was homogeneous (n = 1) or heterogeneous with multiple tiny hypointense nodules (n = 1) on T2-weighted images. The subdeltoid bursitis fluid was characterized by homogeneous low signal intensity with a hyperintense rim (n = 2) on T1-weighted images and homogeneous (n = 1) or heterogeneous hyperintense signals with areas of low signal intensity (n = 1) on T2-weighted images. In tuberculous arthritis, the synovial joint fluid (n = 7) showed heterogeneous (n = 4) or homogeneous (n = 3) low signal intensity on T1-weighted images and high or very high signal intensity on T2-weighted images. Where involved, the adjacent muscle(s) (n = 8) were usually hypointense on T1-weighted images and very hyperintense on T2-weighted images. Associated cellulitis was found in arthritis with involvement of neighboring soft tissues (n = 5), pyomyositis (n = 2), and tenosynovitis (n = 1). The images obtained after Gd-DTPA showed peripheral (n = 10) or heterogeneous (n = 1) enhancement or no enhancement (n = 1).

CONCLUSION: The MR findings for EMT are variable. Although diagnosis is dependent largely on prior presumption and clinical context, MRI provides valuable guidelines in defining the extent of the lesions to select the appropriate treatment and for follow-up of abnormalities.