While highly specific for detecting thromboembolism in proximal pulmonary arteries, spiral computed tomography (spiral CT) cannot reliably exclude pulmonary embolism. Therefore "negative" spiral CT in patients with high clinical probability of acute pulmonary embolism should be considered as non-conclusive. The goal of our study was to check whether echo/Doppler could stratify patients with suspected pulmonary embolism according to the chance of obtaining a conclusive spiral CT result. Echo/Doppler recordings of 51 patients (27 F, mean aged 50 +/- 19) admitted to ICU with high probability of acute pulmonary embolism were compared with the results of spiral CT of pulmonary artery. Pulmonary embolism was revealed by spiral CT in 36 pts. who at echocardiography were found to have shorter acceleration time (AcT 86 +/- 27 vs 117 +/- 7 ms, p < 0.00001) and right ventricle enlargement (PK 30 +/- 5 vs 25 +/- 2 mm p < 0.0001). The velocity of tricuspid regurgitation (TVPG) could not be measured in 19 patients. Among patients with long AcT (> or = 120 ms) or without right ventricle enlargement 8 of 14 and 14 of 28 patients, respectively had no intrapulmonary clots at spiral CT. Echo/Doppler helps in preselection of patients in whom confirmation of PE with spiral CT is highly probable. In patients without indirect echocardiographic sings of pulmonary embolism, despite high clinical probability the chance of a conclusive spiral CT is 40-50%. This should be kept in mind when selecting the most effective diagnostic strategy in individual cases.