Research: AI Can Make Echocardiography More Efficient for Diagnostics.

Research: AI Can Make Echocardiography More Efficient for Diagnostics.

Share this
A medical technician performs a diagnostic imaging test on a patient.

Novel research from MedStar Health helps pave the road for the future of echocardiography, through the development and implementation of machine learning tools that improve the accuracy of this cardiac imaging modality and the efficiency in the echocardiography lab workflow.


Could AI help produce higher-quality, more reliable images of the heart? Our research, published in the Journal of the American Society of Echocardiography (JASE), JAMA Cardiology, and Circulation Imaging indicates AI processes can help improve image capture, analysis, and could assist doctors in processing the resulting data.

An echocardiogram is an ultrasound of the heart that produces volumes of data that help physicians make cardiac diagnoses. During an echocardiogram, a sonographer uses a wand-like transducer to direct sound waves that bounce off the heart to create pictures, much like an ultrasound shows images of a fetus in pregnancy. 

An echocardiogram can examine the anatomy and the blood flow through the heart to look for heart diseases and other conditions. More than 7 million echocardiograms are performed in the U.S. each year—our investigations suggest that AI could help create more precise images and improve data analysis by making it more accurate and efficient through novel algorithms and models. 

AI and manual measurements are accurate.

Our studies compared traditional, manual measurement against automated AI calculations to determine left ventricular ejection fraction (LVEF) and global longitudinal strain (GLS). 

These critical measurements can help doctors identify normal hearts and diagnose heart conditions such as cardiac amyloidosis, a type of cardiomyopathy also known as “stiff heart syndrome,” or COVID-related heart disease.  

In a study recently published in Echocardiography, we have reviewed the images of 51 patients with confirmed cardiac amyloidosis and found both manual and AI-automated methods could accurately assess LVEF and GLS with comparable sensitivity and specificity.

Our research indicates that if AI assistance is widely adopted in echocardiogram labs, such conditions can be identified more efficiently without sacrificing accuracy and reproducibility.

Advancing data equity and safety with AI.

Algorithms can help eliminate subjective judgments that unconscious biases could influence. For example, AI could undertake several processes simultaneously, acquiring images, taking measurements, generating a report, and cross-referencing these measurements against other databases to help interpret the results. In addition to increasing efficiency, this method could reduce opportunities for subjective human biases to impact analysis. 

Echocardiograms produce a significant volume of information on the structure and function of the heart. Algorithms could analyze this data much faster than humans and plotting that information against outcomes data could help inform diagnosis at high speed, without losing accuracy. 

With the continued development of AI, there may come a time when our field of echocardiography is different from what it is today. AI will not replace the expert judgment of physicians but will help us provide better care and it may free our time for other important work, like spending more time communicating with patients and their families. 

We’re working on several different studies, learning all we can about how AI can help us become better healthcare providers, diagnose earlier, and personalize treatment while we work to advance equity in a responsible manner through data safety and appropriate regulations.

Want more information about the MedStar Health Research Institute?

Discover how we’re innovating for tomorrow.

Explore With Us

Stay up to date and subscribe to our blog

Latest blogs