Clive Kearon

No single noninvasive test for pulmonary embolism is both sensitive and specific. Some tests are good for "ruling in" pulmonary embolism (e.g., helical CT) and some tests are good for "ruling out" pulmonary embolism (e.g., D-dimer); others are able to do both but are often nondiagnostic (e.g., ventilation-perfusion lung scanning). For optimal efficiency, choice of the initial diagnostic test should be guided by clinical assessment of the probability of pulmonary embolism and by patient characteristics that may influence test accuracy. This selective approach to testing enables pulmonary embolism to be diagnosed or excluded in a minimum number of steps. However, even with the appropriate use of combinations of noninvasive tests, it is often not possible to definitively diagnose or exclude pulmonary embolism at initial presentation. Most of these patients can be managed safely without treatment or pulmonary angiography by repeating ultrasound testing of the proximal veins after one and 2 weeks to detect evolving deep vein thrombosis. Helical CT and MRI are rapidly improving as diagnostic tests for pulmonary embolism and are expected to become central to its evaluation.