Pulmonary embolism : Validation of diagnostic imaging methods in the clinical setting

Abstract: Pulmonary embolism (PE) is an elusive diagnosis and none of the existing imaging modalities have a 100% diagnostic specificity or sensitivity. Pulmonary arteriography (PA) is the most specific test although the improvement of computed tomography technique has made this a commonly used method. Lung scintigraphy often gives ambiguous results. Fibrin split products (D-dimer) are released into the blood in PE were elevated levels can be measured. However, D-dimer levels are elevated in venous thromboembolic (VTE) disease as well as in a number of other conditions. The aims of this thesis were to evaluate different radiological methods including pulmonary arteriography, lung scintigraphy and spiral computed tomography for the diagnosis of acute pulmonary embolism and to study if a clinical probability protocol or a simple blood test such as D- dimer could improve the diagnostic accuracy. Study I investigated the complication rate of PA in 707 patients. The overall complication rate was 1.6%, which is lower than previously reported. Study II assessed the interobserver variations in PA in 170 patients and compared the consensus results to a final outcome diagnosis. The mean interobserver agreement was 89%, higher for central vessel emboli, lower for peripheral locations. Study III investigated if the use of a combination of a clinical and scintigraphic protocol in relation to the final outcome could improve the diagnosis in patients with clinical suspicion of acute PE. A low combined probability yielded a negative predictive value (NPV) of 98%. The positive predictive value (PPV) was 100% if the combined probability was high. Study IV compared the diagnostic accuracy of contrast medium enhanced spiral computed tomography of the pulmonary arteries (s-CTPA) and a latex agglutination D-dimer assay in patients with suspected acute PE by using PA and clinical follow up as reference method. sCTPA had 95% NPV and 94% PPV. If a cut off level of 0.25 mg/L was used the corresponding figures for D-dimer were 92% and 63%. Study V investigated if 441 patients with a negative s-CTPA and without DVT symptoms, venous studies or anticoagulant treatment had a new episode of PE during three months follow up. Only 0.9% of the patients had proven PE during the follow up period. To conclude, the results of our studies show that PA is a safe method with good interobserver agreement and low complication rate. By applying a model of combined clinical and scintigraphic probabilities for PE, the diagnosis is ruled in when the combined probability is high, and ruled out when the combined probability is low. However, nearly half of the patients will still have an uncertain diagnosis if lung scintigraphy is used as diagnostic method. A low cut-off level of D-dimer can be used as a screening test to rule out PE, but can not confirm the diagnosis. s-CTPA has a high diagnostic accuracy when compared to PA. The overall results indicate that a negative s-CTPA result safely can rule out the existence of clinically significant, acute PE.