A novel method for assessing cardiac function in patients with coronary heart disease based on wrist pulse analysis

Area of Science:

• Cardiology
• Biomedical Engineering
• Data Science

Background:

• Timely assessment of B-type natriuretic peptide (BNP) is crucial for managing chronic heart failure (CHF) risk in patients with coronary heart disease (CHD).
• Elevated BNP levels indicate increased risk and guide therapeutic interventions to reduce mortality in CHD patients.

Purpose of the Study:

• To evaluate the efficacy of wrist pulse signals for noninvasive cardiac monitoring in patients diagnosed with coronary heart disease (CHD).
• To explore the potential of advanced signal processing and machine learning techniques for assessing cardiac function using pulse wave analysis.

Main Methods:

• 419 patients with CHD were categorized into three groups based on B-type natriuretic peptide (BNP) levels.
• Noninvasive wrist pulse signals were acquired and analyzed using time-domain and multiscale entropy (MSE) methods to extract key features.
• Decision tree (DT) and random forest (RF) algorithms were utilized to build classification models for assessing cardiac status.

Main Results:

• Significant differences in pulse features were observed across the three BNP-defined groups, reflecting distinct cardiovascular pathological states.
• Random forest (RF) models demonstrated superior performance compared to decision tree (DT) models in classifying cardiac function.
• An optimal RF model, integrating both time-domain and MSE features, achieved high accuracy (90.90%) and F1-score (90.90%) for cardiac assessment.

Conclusions:

• Wrist pulse detection technology offers a viable and noninvasive method for assessing cardiac function in patients with coronary heart disease (CHD).
• This approach holds potential for improved cardiac monitoring, enabling earlier risk stratification and personalized management strategies for CHD patients.