Extraction and Quantification Clusters of Three-Dimensional Lorenz Plots.

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• The Lorenz plot (LP) is a visualization tool for analyzing electrocardiogram (ECG) signals and heart rate variability.
• Analyzing long-time ECG signals, particularly RR-interval time series, is crucial for diagnosing cardiac arrhythmias.
• Premature complexes in ECG data present a challenge for traditional analysis methods.

Purpose of the Study:

• To develop and evaluate a three-dimensional Lorenz plot (3DLP) method for analyzing RR-interval time series.
• To accurately differentiate between normal and abnormal cardiac rhythm patterns, specifically eccentric clusters (ECs) indicative of arrhythmias.
• To assess the diagnostic capability of 3DLP for identifying ventricular extrasystoles.

Main Methods:

• Constructed 3D LPs using three successive RR intervals (X, Y, Z axes) from 50 Holter records.
• Applied stereographic projection along the space diagonal to the 3D LPs.
• Utilized dot radii and azimuth frequency distribution to distinguish and identify eccentric clusters (ECs).
• Analyzed polar coordinates (radius and angle) of transformed 3DLP data.

Main Results:

• The 3DLP method distinguished eccentric clusters (ECs) from centric clusters with 94 ± 6.0% accuracy.
• Eccentric scatter-dots were identified by azimuth frequency distribution with 93 ± 13.3% accuracy.
• A specific angular threshold (APF < -2.8°) in CPN ECs supported ventricular extrasystoles diagnosis with high sensitivity (1.0) and specificity (0.92).

Conclusions:

• Transformed polar coordinates of 3DLP effectively extract and quantify clusters for arrhythmia diagnosis.
• The 3DLP method demonstrates high accuracy in identifying cardiac arrhythmias and distinguishing premature complexes.
• This approach may offer additional prognostic information beyond arrhythmia diagnosis.