Prediction of pulmonary embolism outcome and severity by computed tomography

Pulmonary embolism (PE) is the third most common cause of cardiovascular death. When PE is fatal, right ventricular failure usually occurs within the first few hours, so right ventricular dysfunction (RVD) should be diagnosed rapidly to identify patients who could benefit from fibrinolytic therapy.

To determine whether quantification of PE with computed tomography (CT) pulmonary angiography and ventricular measurements is of value in the diagnosis of PE severity and prediction of patient outcome.

Multidetector-row CT studies of 48 hospitalized patients with proven pulmonary embolism were reviewed. Pulmonary artery (PA) obstructive index was derived for each patient on the basis of location and degree of obstruction. Ventricular measurements, including right ventricular (RV) short axis, left ventricular (LV) short axis, and RV/LV ratio, were evaluated and compared among survivors and nonsurvivors. Also, the ventricular measurements were correlated with the PA obstructive index. RV/LV ratio and related PA obstructive index were correlated to the patient outcome and hospital stay.

Comparison of the PA obstructive index and ventricular measurements between survivors and nonsurvivors showed significant difference in PA obstructive index (P<0.001), RV short axis (P<0.001), and RV/LV ratio (P=0.03), and no significant difference was noted in LV short axis (P=0.95). Good correlation was found between PA obstructive index and LV short axis (-0.326), RV short axis (0.539), and RV/LV ratio (0.696). A significant difference was found between the PA obstructive indexes of the different categories of RV/LV ratio (P<0.001).

PA obstructive index of more than 50% and RV/LV ratio >1.5 are useful diagnostic criteria for severe PE and poor patient outcome.