Obtenção de medidas de valvas mitrais a partir de ecocardiograma utilizando processamento digital de imagens

Abstract

Valves separate atria and ventricles of the human heart. They are important for controlling blood flow. Therefore, the detection of valve diseases is important for human health. Currently, the doctor, from the video obtained in the echocardiogram exam, performs an image freeze, manually measures the area of the mitral valve, and then finalizes the diagnosis. This procedure takes time and may contain some values of areas that are not real, as the video may be frozen at an inappropriate time. Thus, this work aims to detect the maximum opening area of the mitral valves, in an automated way, using digital image processing, which consists, in short terms, of signal processing in matrix format. This information is very importance for the identification of stenosis (inability of the valve to open completely due to a process of histological alteration that causes loss of elasticity and alters its form). Therefore, from the videos of 3D transesophageal echocardiogram exams, it was possible to develop a graphical interface, in Python language, capable of loading the videos, performing digital processing, providing the value of the maximum valve opening area and the following diagnosis: “THERE IS STENOSIS” or “THERE IS NO STENOSIS”. The values selected manually, by cardiologists, and the values of the system created were compared, showing that the system fullfield its proposal. In addition, tests were made with 3D printing applications using the 3D transesophageal echocardiogram exams. It turns out that it is possible to obtain prototypes of mitral valves from this examination, maintaining their anatomical characteristics and printing in their real size.