To many people die because of undiagnosed pulmonary embolism. Common pulmon
ary embolism is the most unexpected mortal event in necropsy, antemortem co
rrectly diagnosed in 18-39%. The diagnostic value of chest ultrasound (CUS)
has been investigated. Methods: 117 (68 women, 49 men) patients with clini
cal suspicion of pulmonary embolism underwent chest sonography and spiral c
omputed tomography (CT). Final diagnosis has been made by CT respective wit
h echocardiography, venous duplex sonography and fibrin dimer tests. Result
s: Finally, 70 patients suffered from pulmonary embolism. The chest sonogra
ms showed averaged 1.5 x 2.8 cm (0.5-8.5) large triangular or rounded hypoe
choic lesions, mean 2.6 pro patient, similar in form and size as in CT. Fre
sh reperfusionable infarcts were homogenous and hypoechoic. Older infarcts
were well demarcated, mainly wedge shaped. A hyperechoic reflex in the cent
er corresponds to the bronchiole: a sign of seg mental involvement. The sen
sitivity of chest ultrasound was 94%, the specificity 87%, positive predict
ive value 92%, negative predictive value 91%, accuracy 91%. Overall 61 pati
ents had PE in CT, in 47 (67%) cases a direct emboli detection was possible
. 14 patients had peripheral lung consolidations without detectable emboli,
but fibrin-dimer tests were positive in all cases, there was deep vein thr
ombosis diagnosed and they showed signs of PE in echocardiography. Spiral C
T showed a sensitivity of 85%, a specificity and a positive predictive valu
e of 100%, a negative predictive value of 83% and an accuracy of 92%. Concl
usion: CUS can improve diagnosis of pulmonary embolism. Sonography also rev
eals small infarcts which remain undetected with other imaging procedure su
ch as helical CT.