AI-based Prediction of Cardiac Function Using Echocardiography and Body Composition Data (ECHO-FIT Study)

Background and Rationale:

Heart failure represents a significant global health burden, characterized by high morbidity and mortality rates. While echocardiography remains the gold standard for heart failure diagnosis and monitoring, providing crucial measurements like left ventricular ejection fraction (LVEF) and diastolic function assessment, its widespread implementation is limited by resource constraints and operator dependency. Bioelectrical impedance analysis (BIA) offers a promising complementary approach, providing rapid and non-invasive assessment of body composition parameters that have shown correlations with cardiovascular outcomes. This study seeks to leverage the potential synergy between echocardiographic findings and body composition data to develop a more accessible screening tool for cardiac dysfunction.

Study Objectives:

• Primary: To develop and validate a predictive model for left ventricular function by integrating body composition data from the QCCUNIQ BC 720 device with standard echocardiographic parameters.
• Secondary:

• To investigate correlations between body composition indices and echocardiographic measurements
• To evaluate the utility of body composition analysis in identifying high-risk cardiovascular patients
• To assess the model's potential as a screening tool in resource-limited settings

Methodology:

This single-center, prospective observational study will enroll 2,000 adults (≥20 years) undergoing routine echocardiography, equally divided between those with normal cardiac function (LVEF ≥50%) and heart failure (LVEF <50%). Participants will undergo body composition analysis using the QCCUNIQ BC 720 device within one week of their echocardiogram.

Data Collection and Analysis:

Comprehensive data collection will include standard echocardiographic parameters (LVEF, diastolic function, structural measurements) and detailed body composition analysis (fat mass, skeletal muscle mass, total body water). Statistical analysis will employ both traditional regression methods and advanced machine learning algorithms to develop the predictive model. Model validation will utilize k-fold cross-validation, with performance assessed through standard metrics including sensitivity, specificity, and area under the curve (AUC).

Inclusion criteria

• Aged 20 years or older.
• Undergoing a standard echocardiographic examination.
• Providing consent to undergo body composition analysis.
• Signing the informed consent form to voluntarily participate in the study.

Exclusion criteria

• Having a physical or mental condition that makes it impossible to conduct an echocardiogram or perform body composition analysis.
• Deemed inappropriate for study participation by the researcher (e.g., unable to cooperate).