Automated J wave detection from digital 12-lead electrocardiogram

Area of Science:

• Cardiology
• Biomedical Engineering
• Signal Processing

Background:

• J waves are a subtle finding on electrocardiograms (ECGs), often associated with specific cardiac conditions.
• Accurate J wave detection is crucial for diagnosis but can be challenging due to inter-observer variability.
• Automated methods offer a potential solution for consistent and efficient J wave identification.

Purpose of the Study:

• To develop and validate an automated method for J wave detection in ECG tracings.
• To assess the accuracy of the automated J wave detection method using signal processing and functional data analysis.
• To explore the potential utility of this automated approach for other complex ECG waveform analysis.

Main Methods:

• Developed an automated J wave detection algorithm using signal processing and functional data analysis techniques.
• Utilized two sets of ECG tracings (training n=100, test n=116) from a core laboratory.
• ECGs were recorded on GE MAC 1200 at 500Hz sampling rate and processed using GE Marquette 12-SL software.

Main Results:

• The automated method achieved 100% sensitivity and 94% specificity in the training set.
• In the test set, the method demonstrated 89% sensitivity and 86% specificity for J wave detection.
• The automated J wave detection showed good accuracy, indicating its potential clinical utility.

Conclusions:

• The developed automated method for J wave detection exhibits good accuracy.
• This approach may be valuable for the consistent identification of J waves in clinical practice.
• The signal processing techniques used show promise for detecting other complex ECG waveforms.