Acute mediastinitis: spectrum of computed tomography findings.

The objective of this study was to describe CT findings and to determine the diagnostic value of CT in diagnosis of acute mediastinitis. CT findings were retrospectively studied in 40 patients with suspected acute mediastinitis, including 28 postoperative patients, five with acute descending necrotizing mediastinitis (ADNM), and seven with suspected post-traumatic perforation of the esophagus. Findings included increased attenuation of mediastinal fat (100%), localized mediastinal fluid collections (55%), free gas bubbles in the mediastinum (57.5%), mediastinal lymph nodes (35%), pericardial effusions (27.5%), pleural effusions (85%), lung infiltrates (35%), sternal dehiscence (40%), and pleuromediastinal fistula (2.5%). The sensitivity and specificity of CT in postoperative patients in the first 17 days was 100% and 33% respectively, and after day 17, 100% and 90%. In patients with ADNM sensitivity was 100% while in patients with suspected esophageal perforation sensitivity and specificity were 100%. CT is a highly sensitive technique for the detection of mediastinitis of various causes. For the postoperative patients there is clear time dependence for CT interpretation and accuracy. In patients with suspected ADNM, and traumatic esophageal perforation CT is highly specific early after clinical presentation.

Multislice CT imaging of pulmonary embolism.

In recent years CT has been established as the method of choice for the diagnosis of central pulmonary embolism (PE) to the level of the segmental arteries. The key advantage of CT over competing modalities is the reliable detection of relevant alternative or additional disease causing the patient's symptoms. Although the clinical relevance of isolated peripheral emboli remains unclear, the alleged poor sensitivity of CT for the detection of such small clots has to date prevented the acceptance of CT as the gold standard for diagnosing PE. With the advent of multislice CT we can now cover the entire chest of a patient with 1-mm slices within one breath-hold. In comparison with thicker sections, the detection rate of subsegmental emboli can be significantly increased with 1-mm slices. In addition, the interobserver correlation which can be achieved with 1-mm sections by far exceeds the reproducibility of competing modalities. Meanwhile use of multislice CT for a combined diagnosis of PE and deep venous thrombosis with the same modality appears to be clinically accepted. In the vast majority of patients who receive a combined thoracic and venous multislice CT examination the scan either confirms the suspected diagnosis or reveals relevant alternative or additional disease. The therapeutic regimen is usually chosen based on the functional effect of embolic vascular occlusion. With the advent of fast CT scanning techniques, also functional parameters of lung perfusion can be non-invasively assessed by CT imaging. These advantages let multislice CT appear as an attractive modality for a non-invasive, fast, accurate, and comprehensive diagnosis of PE, its causes, effects, and differential diagnoses.