This is a multicenter collaborative study to determine the sensitivity, specificity and predictive values of contrast enhanced spiral computed tomography (spiral CT) for the diagnosis of acute pulmonary embolism (PE). Pulmonary embolism is common, yet frequently undiagnosed and fatal. Ventilation/ perfusion lung scans are the most commonly used test. However, they are nondiagnostic in 72 percent of patients with suspected PE and in 57 percent of patients with subsequently proven PE. Pulmonary angiography is the definitive diagnostic test in such patients, but there is hesitancy to use it because of morbidity, discomfort, cost, and lack of availability is community hospitals. Noninvasive leg tests, particularly venous ultrasound, permit a strategy of management by treatment on the basis of venous thromboembolic disease (PE or DVT). Spiral CT is a nearly noninvasive test that offers the possibility of a definitive diagnosis of PE by showing the outline of the thrombus in a pulmonary artery. However, its utility in the diagnosis of PE is not established. Even though spiral CT has not been adequately validated, many hospitals now use it as a first diagnostic test for PE, or as a definitive test in patients in whom the ventilation/perfusion lung scan is nondiagnostic. This could lead to over- or undertreatment, which have serious potential complications. In view of the potential applicability of spiral CT, this investigation is proposed to test its sensitivity, specificity, positive and negative predictive values, using contemporary diagnostic reference standard methods. The strengths of this proposal are: 1) the importance of the problem of PE diagnosis, 2) the potential of a widely available technology for the diagnosis, 3) a protocol that parallels standard diagnostic strategy without subjecting patients to risk only for data acquisition, 4) a strong team of investigators, which has collaborated previously in successful research in PE.
