An Artificial Intelligence-Based Automated Echocardiographic Analysis: Enhancing Efficiency and Prognostic Evaluation...

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Imaging

Background:

• Manual echocardiography measurements are time-consuming and variable.
• Assessing cardiac parameters is crucial for patient care.
• ST-segment elevation myocardial infarction (STEMI) requires timely and accurate assessment.

Purpose of the Study:

• Evaluate an AI-based automated system for echocardiographic analysis in STEMI patients.
• Assess the feasibility, accuracy, and predictive value of AI measurements.
• Compare AI-derived cardiac parameters with conventional manual methods.

Main Methods:

• Developed an AI system using a nationwide echocardiographic dataset.
• AI automatically identified views, segmented, and tracked left ventricle (LV) and left atrium (LA).
• Compared AI measurements (LV ejection fraction, global longitudinal strain, LA volume index, reservoir strain) with manual methods in 632 STEMI patients.

Main Results:

• AI system achieved 98.5% accuracy in view identification.
• Strong correlations (Pearson 0.81-0.92, ICC 0.74-0.90) between AI and manual measurements.
• AI-derived measurements independently predicted clinical outcomes, similar to manual methods.

Conclusions:

• A fully automated AI system for LV and LA analysis is feasible in STEMI patients.
• AI provides accurate echocardiographic measurements comparable to conventional methods.
• AI offers a promising solution for comprehensive analysis, reduced workload, and improved patient care.