Pulmonary Embolism Detection

This study developed a computer-aided diagnosis (CAD) system to assist radiologists in detecting pulmonary embolism (PE) on computed tomography pulmonary angiographic (CTPA) images using a 3D voxel clustering method and expectation-maximization analysis. The system demonstrated high detection rates for proximal and subsegmental emboli in patients without lung diseases (92.0% and 77.8%, respectively). The preliminary results suggest that the CAD system is a promising tool for PE detection on CTPA, with ongoing efforts to improve accuracy, particularly for subtle emboli.