Extraction of Quantitative Anatomical Information from Coronary Angiographies

Francisco      J. Novoa; Norberto      Ezquerra; Lola      Traba; Martin      Villar; Javier      Pereira; Jose      Manuel Vazquez-Rodriguez; Nicolas      Vazquez; Marcos      Martinez-Romero; Virgina      Mato

Abstract

Cardiovascular diseases and in particular severe coronary stenosis are the main cause of death in the western hemisphere. The diagnostic method considered as gold-standard in the quantification and location of the coronary lesions is the coronary angiography. This procedure is required in patients with a high probability of coronary heart disease. To treat patients, interventional cardiologists analyse the angiographic images, establish a disease diagnosis and may even provide a prognosis, depending on the location, severity, and extent of the coronary disease. From the late 1970s, a large number of studies have been carried out, aimed at building information systems that assess general practitioners in the diagnostic tasks. These systems are based on the quantification of the anatomical information in an objective way. The relevant information is extracted from the coronary angiographies using automatic or semiautomatic image segmentation techniques. The precision of the segmentation is a key element in the subsequent measurement of the stenosis and flow capacity of the arteries, which also enables the establishment of an index or score related to the prognosis of the disease. Multiscalar methods, matching filters, and morphologic mathematical methods present the best balance between precision and processing speed. The best results are frequently obtained through the common use of multiple techniques. The current paper presents a review of the best techniques used for the extraction of anatomical information from coronary angiographies over the past thirty years.

Keywords: Angiography, biomedical informatics, cardiovascular disease, medical imaging, segmentation, stenosis, line segments, Magnetic Resonancy Imaging, Structuring Element, Digital Substracted Angiography, Expectation-Maximization