Scintigraphic probability and angiographic diagnostic certainty in acute pulmonary embolism

The objectives of this study were to determine if diagnostic certainty on angiography correlates with scintigraphic probability for the diagnosis of pulmonary embolism. From a total of 160 consecutive patients who underwent both nuclear imaging and invasive selective pulmonary angiography, we reviewed the xenon-133 ventilation images in 2 posterior oblique views and the Tc-99m macroaggregated serum albumin perfusion scans and angiograms of 40 patients (15 men, 25 women; average age 57 years) who were discharged from the hospital on anticoagulants with a diagnosis of pulmonary embolism. The angiograms were reviewed and the diagnosis of embolism was considered certain in the presence of an intraluminal filling defect, a trailing embolus, or a branch occlusion equal to or larger than a segmental branch (n=29; 73%), and uncertain when the studies were reinterpreted as either equivocal or negative or in the presence of a single, small subsegmental filling defect of questionable clinical significance. The ventilation-perfusion scans were read as high (n=18; 45%), intermediate (n=10; 25%), or low (n=12; 30%) probability. The proportion of patients with diagnostic certainty on angiography in the high-, intermediate-, and low-probability scintigraphic subgroups was, respectively, 100% (18 of 18), 70% (7 of 10), and 33% (4 of 12) (P=0.004). In patients diagnosed with pulmonary embolism based on selective angiography, a lower probability of pulmonary embolism on ventilation-perfusion scintigraphy correlates with a lesser degree of diagnostic certainty on angiography and a higher incidence of single subsegmental emboli.