Pulmonary embolism: comparison of MR images with radionuclide and angiographic studies.

PURPOSE: To assess the accuracy and potential of magnetic resonance (MR) imaging in evaluation of patients in whom pulmonary embolism is suspected. MATERIALS AND METHODS: Blinded, prospective interpretations of multiphasic, cardiac-gated spin-echo MR images were compared with retrospective chart review in 86 patients in whom pulmonary embolism was suspected. In 64 patients, the presence or absence of pulmonary emboli was established with x-ray angiography (n = 34) or ventilation-perfusion (V-P) scans and concurrent clinical impression (n = 30). RESULTS: In the subgroup with angiographic proof, MR imaging had a sensitivity of 90%, specificity of 77%, positive predictive value of 86%, and negative predictive value of 83%. In 21 patients with intermediate probability of pulmonary embolism on V-P scans and angiograms, MR images enabled diagnosis of pulmonary embolism in 12 of 12 patients (sensitivity, 100%) and absence of pulmonary embolism in seven of nine patients (specificity, 78%). CONCLUSION: MR imaging reliably depicts large and medium-size pulmonary emboli, regardless of infiltrates or effusion; hence, it may clarify findings on V-P scans that show intermediate probability of pulmonary embolism or are at variance with the clinical impression.

Low resolution spin echo: a simple timesaving technique for MRI liver exams.

MR evaluation of the liver at mid-field strength requires relatively lengthy T2-weighted sequences (T2W) for differentiation of benign cavernous hemangiomas from malignant lesions. A short duration T2W, which maintains standard signal-to-noise ratio and also contrast relationships, can be easily implemented by increasing the pixel size in the phase-encoding direction (thus reducing spatial resolution) and proportionally decreasing the number of phase-encoding steps in the matrix (thus reducing acquisition time). Blinded interpretations of a quick (4 min), low resolution (3.4 mm x 1.7 mm pixel) T2W sequence (matrix 64 x 256, FOV 21.7 cm phase x 43.5 cm frequency) were compared to the 17 min standard resolution (1.7 mm x 1.7 mm) T2W sequence (256 x 256 matrix, FOV 43 x 43 cm) in 25 patients suspected of having liver metastasis. Lesions felt to be cavernous hemangiomas showed a 100% (24/24) agreement for interpreter "A" and 96% (22/23) agreement for interpreter "B" when 4 min low resolution T2W was compared to the standard 17 min sequence. Sensitivity (for all types of lesions) of the low resolution T2W sequence ranged from 100% (31/31) for interpreter "A" to 80% (28/35) for interpreter "B." Missed lesions (interpreter "B") were either partially obscured by excessive fat (wrap around) (N = 4), less than 1 cm in size (N = 2), or degraded by motion artifact (N = 1). Thus in many situations low resolution T2 may provide a substantial timesaving alternative to standard T2W particularly where T2W is used primarily for lesion classification in normal sized patients.

Osteomyelitis: characteristics and pitfalls of diagnosis with MR imaging.

Prospective and retrospective magnetic resonance (MR) imaging (0.35-T) interpretations were compared with final diagnoses in 110 patients suspected to have osteomyelitis. Diagnostic criteria of dark marrow on T1-weighted images and bright marrow on short-tau inversion-recovery images yielded a prospective sensitivity of 98% and a prospective specificity of 75%. Sixty percent of uncomplicated septic joint effusions demonstrated abnormal marrow signal intensity that was mistaken for osteomyelitis. Retrospective review revealed that overall specificity could be improved to 82% without loss of sensitivity if increased marrow signal intensity on T2-weighted images were included as an additional criterion. Specificity may be further increased by use of knowledge of morphologic patterns that distinguish various forms of osteomyelitis. Ten patients (9%) had potential pitfall diagnoses (eg, fracture, infarction, healed infection) that mimic osteomyelitis. MR imaging can be sensitive and specific for osteomyelitis if characteristic appearances and pitfall diagnoses are incorporated into the diagnostic criteria.

Differential diagnosis of bacterial myositis in AIDS: evaluation with MR imaging.

Magnetic resonance (MR) imaging was used to assess for the presence of bacterial myositis, rare outside the tropics, in 13 patients with either the acquired immunodeficiency syndrome (AIDS) (n = 11) or positive results of serologic tests for the human immunodeficiency virus but without other evidence of AIDS (n = 2). Bacterial myositis was diagnosed in six patients: in five it was caused by pyogenic bacteria, and in the other, by Mycobacterium tuberculosis; in each patient, little or no subcutaneous tissue alteration occurred. On T1-weighted images in three patients, muscle abscesses showed a rim of increased signal intensity corresponding to margins between drainable pus and edematous muscle. Subcutaneous tissues appeared normal in patients with bacterial myositis but was not in the others, in whom muscle abnormalities tended to be less prominent. The latter group included patients with lymphoma (n = 1), Kaposi sarcoma (n = 2), and carbunculosis (n = 1), and three patients in whom no diagnosis was made; lymphedema was presumed to account for imaging abnormalities in four of the latter group.

Deep venous thrombosis of extremities: role of MR imaging in the diagnosis.

Current noninvasive imaging techniques for diagnosis of deep venous thrombosis (DVT) of extremities are limited in their ability to demonstrate central vein involvement and to distinguish acute from chronic changes. The utility of spin-echo magnetic resonance (MR) imaging for DVT was evaluated in 100 patients suspected of having either upper- (n = 25) or lower-extremity (n = 75) DVT. Ninety-seven patients were imaged successfully. In a subset of 36 patients, prospective comparison of MR imaging with contrast venography revealed a sensitivity of 90%, specificity of 100%, and Kappa level of agreement of .752 (P less than .0001). MR imaging showed more central extent of thrombus than did venography in all five patients with upper-extremity DVT and in 13 of 25 patients (52%) with lower-extremity DVT. Although all patients in the study were evaluated for acute symptoms, 13 of 59 (22%) MR imaging studies positive for DVT demonstrated chronic disease. MR images demonstrated ancillary abnormalities in 18 of 41 (44%) patients who did not have DVT. Thus, MR imaging has a role as the definitive examination when the results of initial screening studies are unsatisfactory, or as a first-line examination if (a) there is suspicion of upper-extremity or pelvic vein thrombosis, (b) there is a history of prior DVT that necessitates distinction of acute from chronic changes, or (c) other tests are unavailable.