What is it about?
In this work, we explored using a type of artificial intelligence called a 3D Convolutional Neural Network (3D-CNN) to automatically classify the heart's left ventricle function based on a standard diagnostic test called transthoracic echocardiography (TTE). The left ventricle is an essential part of the heart because it pumps blood to the rest of the body. The function of the left ventricle is measured by the ejection fraction (EF), the amount of blood it pumps out with each beat. It is important to accurately measure the EF because problems with the left ventricle can lead to serious health problems. Doctors have to manually mark the left ventricle on TTE images to determine the EF, which can be time-consuming and prone to errors. To train our 3D-CNN, we created a dataset of TTE exams from a medical centre. We selected specific images from each exam and processed them to remove unnecessary information. We then used this dataset to teach the 3D-CNN how to classify the EF. We tested the 3D-CNN on a separate dataset to see how well it could classify the EF without human input. We found that the 3D-CNN could accurately classify the EF into four categories: very low, low-normal, high-normal, and very high. These results suggest that 3D-CNNs have the potential to classify the EF from TTE exams automatically and could be used to help doctors diagnose and treat patients with heart problems. This work is an essential first step towards developing automated tools to assist doctors in evaluating heart function.
Featured Image
Photo by Ali Hajiluyi on Unsplash
Why is it important?
This work is important because it presents a new approach for automating the process of evaluating the function of the left ventricle of the heart. Currently, this process involves manual annotation of images, which can be time-consuming and prone to errors. By using artificial intelligence, this work has the potential to improve the accuracy and efficiency of diagnosing and treating heart problems. Additionally, this work is timely because it addresses a major global health issue - cardiovascular diseases are the leading cause of death worldwide. By improving the accuracy and efficiency of diagnosing and treating heart problems, this work has the potential to make a significant impact on public health.
Perspectives
Read the Original
This page is a summary of: Ejection Fraction Classification in Transthoracic Echocardiography Using a Deep Learning Approach, June 2018, Institute of Electrical & Electronics Engineers (IEEE),
DOI: 10.1109/cbms.2018.00029.
You can read the full text:
Contributors
The following have contributed to this page