CT pulmonary angiography in the detection of pulmonary emboli: a meta-analysis of sensitivities and specificities

OBJECTIVE: To determine the overall sensitivity and specificity for CT pulmonary angiography (CTPA) in the diagnosis of pulmonary emboli (PE) using a meta-analysis of the published literature. MATERIALS AND METHODS: A Medline search was constructed to include all English language publications indexed in the Index Medicus from 1990 to 2000, which included the terms CT, PE and pulmonary angiography. Studies selected were designed principally to compare CTPA in the overall detection of PE as confirmed by an abnormal fluoroscopic pulmonary angiogram or a high probability V:Q scan. Results were corrected for the patient sample size in the respective studies prior to pooling the data. In the absence of an accepted technique for calculating a ROC curve in the meta-analysis of imaging studies, a previously untested theoretical technique was used to obtain a composite ROC curve. RESULTS: Twelve studies of CTPA comprising a total of 1250 patients were analyzed. The overall sensitivity and specificity for CTPA after correction for study size was 74.1% and 89.5% with a range of 57-100% and 68-100%, respectively, for the detection of PE. No trend was detected with respect to the year of publication or sample size. CONCLUSIONS: CTPA has acceptable sensitivity and specificity with a strong ROC curve making it a good first line investigation for PE.