Discordance between CT and Angiography in the PIOPED II Study

Abstract

To retrospectively evaluate the causes of discordant computed tomographic (CT)-angiographic readings from the Prospective Investigation of Pulmonary Embolism Diagnosis, or PIOPED, II study. Institutional review board approval was obtained for this HIPAA-compliant study. Of 1036 patients suspected of having pulmonary embolism who were examined with CT, 226 underwent angiography; 206 patients had concordant results and 20 had discordant results according to two independent readers. Of these 20 patients, 10 were men and 10 were women (mean age, 49 years). Among the 20 studies with discordant results, central readers identified seven cases as negative and 13 as positive for pulmonary embolism at CT; these findings were reversed at angiography. Side-by-side comparisons of discordant studies were performed in consensus. The time between CT and angiography and all locations of pulmonary embolism vascular territory were recorded. The McNemar binomial test was used. One patient had false-positive findings at angiography, 13 patients had false-negative findings at angiography, and two patients had false-negative findings at CT. Four patients had true-negative findings at CT; however, findings were positive for thrombus at angiography. The sensitivity for the detection of pulmonary embolism was 87% for CT and 32% for angiography (P=.007). The largest missed thrombus at angiography was subsegmental in eight patients, segmental in two patients, and lobar in three patients; at CT it was subsegmental in two patients. The mean time between CT and angiography was 40 hours+/-21 (standard deviation) (range, 10-97 hours). In the interval between CT and angiography, thrombi can remain the same, resolve, develop, or result from angiography.