AI Program PanEcho Revolutionizes Echocardiogram Interpretation, Promising Faster Heart Care

AI Program PanEcho Demonstrates High Accuracy in Echocardiogram Interpretation

Researchers at Yale School of Medicine have developed an artificial intelligence program called PanEcho that could revolutionize the interpretation of echocardiograms, potentially leading to faster and more efficient heart care. The findings were presented at the American Heart Association's Scientific Sessions 2024 in Chicago 1

1

2

PanEcho's Capabilities and Performance

PanEcho is designed to independently interpret echocardiography videos, building upon previous AI applications in cardiology that were limited to single views of the heart and disease-specific criteria. The program's comprehensive reporting capabilities cover all major findings from any set of echocardiography videos 1

1

The AI system's diagnostic performance was evaluated using the area under the receiver operating characteristic curve (AUC), where 1.0 represents 100% accuracy. PanEcho demonstrated impressive results across various diagnostic tasks:

• Average score of 0.91 AUC across 18 different diagnostic classification tasks
• 0.95 AUC for detecting increased left ventricle size
• 0.98 AUC for identifying left ventricle systolic dysfunction
• 0.91 AUC for detecting left ventricle hypertrophy
• 0.93 AUC for identifying right ventricle systolic dysfunction
• 0.99 AUC for identifying severe aortic stenosis
• 0.96 AUC for identifying mitral stenosis 2

  2

Accuracy in Quantitative Measurements

PanEcho also demonstrated high accuracy in estimating continuous echocardiographic parameters:

• 4.4% mean absolute error when estimating left ventricle ejection fraction
• 1.3 mm mean absolute error when estimating left ventricle intraventricular septum thickness
• 1.2 mm mean absolute error when estimating left ventricle posterior wall thickness 1

  1

  2

  2

These measurements are crucial for accurately assessing left ventricular structure and function, which is a major aspect of heart health.

Related Stories

Yale Researchers Develop AI Tool for Rapid and Accurate Echocardiogram Interpretation

AI Revolutionizes Heart Disease Detection: From ECGs to Hidden Conditions

AI Model Detects Multiple Heart Failure Types from Routine ECGs, Including Hard-to-Identify HFpEF

AI Model Revolutionizes Early Detection of Cardiac Amyloidosis

Potential Impact on Patient Care

Gregory Holste, M.S.E., a researcher with the Cardiovascular Data Science (CarDS) Lab at Yale School of Medicine, highlighted the potential of PanEcho in clinical settings: "PanEcho has the potential to be used in simplified, AI-assisted screening echocardiograms. In settings where expert readers may not be readily accessible, PanEcho could rapidly rule out abnormalities that would otherwise require urgent referral" 1

1

Study Background and Future Directions

The PanEcho AI model was developed using 1.23 million echocardiogram videos from nearly 34,000 transthoracic echocardiography tests conducted at Yale-New Haven Health System hospitals and outpatient clinics between 2016 and 2022. The study included data from 26,067 unique individuals with an average age of 67 2

2

While the current validation is retrospective, the next step is to prospectively validate PanEcho's application in real-world patient care environments. Researchers also aim to evaluate its use with portable echocardiogram machines in emergency rooms and smaller medical clinics, where the potential for positive impact with AI tools is greatest 1

1

2

The research team hopes that the public release of the PanEcho AI model will encourage the research community to move towards flexible, multi-task, multi-view approaches for echocardiogram interpretation, potentially leading to more widespread adoption of AI-assisted cardiac imaging analysis 1

1

2