Agreement of ejection fraction measured by coronary computed tomography (CT) and cardiac ultrasound in evaluating...

Area of Science:

• Cardiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Cardiac ultrasound, vital for cardiovascular medicine, has high technical operator requirements limiting its use.
• Chronic heart failure diagnosis relies on accurate assessment of cardiac function.

Purpose of the Study:

• To investigate the agreement between ejection fraction measured by coronary computed tomography (CT) and cardiac ultrasound.
• To evaluate the diagnostic performance of CT-derived ejection fraction (EF-CT) in patients with chronic heart failure.

Main Methods:

• A retrospective study of 343 patients with suspected heart failure symptoms.
• Coronary CT images analyzed using Accurate Left Ventricle (AccuLV) AI software for EF-CT calculation.
• Comparison of EF-CT with ultrasound-derived ejection fraction (EF) using correlation and Bland-Altman analysis.

Main Results:

• A good correlation (R square = 0.704, r=0.837) was found between EF-CT and EF.
• High diagnostic performance for EF measured by coronary CT: 91% sensitivity and 96% positive diagnostic value.
• AI analysis demonstrated excellent agreement between EF-CT and EF.

Conclusions:

• Accurate Left Ventricle (AccuLV) AI software shows excellent agreement with cardiac ultrasound for ejection fraction assessment.
• AI-enhanced coronary CT can potentially serve as a clinical diagnostic reference for left ventricular function.