A multilead scheme based on periodic component analysis for T-wave alternans analysis in the ECG

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• T-wave alternans (TWA) is an electrocardiogram (ECG) phenomenon associated with sudden cardiac death (SCD).
• Accurate detection of TWA is crucial for risk stratification in patients with cardiac conditions.

Purpose of the Study:

• To propose and evaluate a novel multilead TWA analysis scheme.
• To enhance the sensitivity and prognostic value of TWA detection compared to existing methods.

Main Methods:

• The proposed scheme integrates the Laplacian likelihood ratio (LLR) method with periodic component analysis (piCA), an eigenvalue decomposition technique.
• The method was tested using synthetic ECG signals and ECGs from stress tests.
• Performance was benchmarked against established LLR-based TWA detection schemes.

Main Results:

• The piCA-based multilead TWA analysis scheme demonstrated superior TWA detection capabilities.
• The new scheme outperformed previously reported LLR-based methods in identifying TWA.
• The findings suggest enhanced accuracy in identifying cardiac risk factors.

Conclusions:

• The developed piCA-based TWA analysis scheme offers improved detection performance.
• This method holds potential for enhancing the prognostic value of TWA testing.
• Further validation may lead to improved prediction of sudden cardiac death risk.