Multilead ECG delineation using spatially projected leads from wavelet transform loops.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Cardiology

Background:

• Accurate electrocardiogram (ECG) interpretation relies on precise delineation of waveform boundaries, such as the QRS complex and T-wave.
• Existing single-lead (SL) delineation methods often lack robustness and accuracy due to limited spatial information.

Purpose of the Study:

• To propose and evaluate a novel multilead (ML) based automatic strategy for enhanced ECG boundary delineation.
• To improve the accuracy, robustness, and stability of QRS and T-wave boundary localization.

Main Methods:

• A novel ML strategy was developed, leveraging spatial information from three orthogonal leads.
• A derived lead, optimized for delineation, was created by combining information from the orthogonal leads.
• A single-lead (SL) wavelet-transform-based delineation system was applied to this optimal derived lead.

Main Results:

• The ML strategy demonstrated reduced error dispersion compared to SL results.
• The proposed method provided more robust, accurate, and stable boundary locations than individual ECG leads.
• The ML approach outperformed strategies relying on lead selection rules after SL delineation.

Conclusions:

• The novel ML-based strategy offers a significant improvement for automatic ECG boundary delineation.
• This approach enhances the reliability of QRS and T-wave boundary identification, crucial for clinical diagnostics.
• The method's ability to integrate spatial information leads to superior performance over traditional SL techniques.