From imaging to functional outcome in pulmonary embolism

The interaction between mechanical obstruction and outcome in pulmonary embolism (PE) is not well quantified. Therefore a simple prognostic tool that can be used quickly in the clinical setting remains elusive. Several scoring systems have been proposed to address this problem. However, they are unable to adequately capture the functional outcomes in PE so have not been adopted widely clinically. Here we present an image-based computational model that correlates very well with measures of RV dysfunction. The model extracts the geometric features of the lung, airways, blood vessels and emboli from CTPA (computed tomography pulmonary angiogram) imaging and simulates function (perfusion, ventilation and gas exchange) within these geometries. This results in subject-specific predictions of function in 9 patients with acute PE. There is a high correlation between model results and indicators of right heart dysfunction (p=0.001 in the case of the ratio between right and left ventricular volumes and p<0.03 in the case of systolic pulmonary artery pressure estimated from echocardiography). An existing scoring system that accounts only for the mechanical obstruction of capillary bed performs less well than the model (p=0.04 in the case of the ratio between right and left ventricular volumes and p=0.23 in the case of systolic pulmonary artery pressure estimated from echocardiography). This suggests that the functional impact of occlusion must be accounted to construct useful PE scoring systems.