Artificial Intelligence-Powered Left Ventricular Ejection Fraction Analysis Using the LVivoEF Tool for COVID-19...

Area of Science:

• Cardiology
• Artificial Intelligence in Medicine
• Point-of-Care Ultrasound

Background:

• COVID-19 can lead to cardiac complications, including impaired left ventricular ejection fraction (LVEF).
• Accurate and timely LVEF assessment is crucial for managing COVID-19 patients.
• Traditional echocardiography requires specialized expertise and equipment, limiting point-of-care application.

Purpose of the Study:

• To prospectively evaluate the accuracy of an AI-based tool for LVEF assessment using a hand-held ultrasound device (HUD) in COVID-19 patients.
• To determine if reduced LVEF, as assessed by AI-HUD, predicts adverse composite endpoints in COVID-19 patients.

Main Methods:

• Prospective investigation involving 42 COVID-19 patients.
• Real-time LVEF assessment using an AI-equipped HUD compared to blinded echocardiographer assessment.
• Analysis of correlation and agreement between AI-HUD and echocardiographer LVEF measurements.

Main Results:

• Excellent correlation (r=0.774, p<0.001) and substantial agreement (kappa=0.797, p<0.001) between AI-HUD and echocardiographer LVEF assessments.
• AI-HUD demonstrated high sensitivity (72.7%) and specificity (100%) for detecting LVEF < 50%.
• AI-based reduced LVEF (<50%) was significantly associated with worse composite endpoints (adjusted OR=6.40, p=0.041).

Conclusions:

• AI-integrated HUD provides reliable, real-time LVEF assessment in COVID-19 patients.
• This AI-HUD tool can serve as a valuable decision support system.
• AI-based LVEF assessment may identify COVID-19 patients at risk for adverse outcomes.