Analysis of interval measurements on CSE multilead reference ECGs

Area of Science:

• Biomedical Engineering
• Medical Imaging and Signal Processing

Background:

• The International Electrotechnical Commission (IEC) is developing standards for interpretive electrocardiographs.
• The Common Standards of Electrocardiography (CSE) multilead measurement database is utilized for this standardization effort.

Purpose of the Study:

• To evaluate if original electrocardiogram (ECG) data offers advantages over artificial (cycle-repeated) ECGs for standardization.
• To compare measurement accuracy and noise levels between clean and noisy/nonsinus rhythm ECGs.

Main Methods:

• Processed 125 original (MO1) and 125 artificial (MA1) ECG series from the CSE database.
• Analyzed 100 clean ECGs and 25 noisy/nonsinus rhythm ECGs for P duration, PR interval, QRS duration, and QT interval.
• Compared measured intervals against CSE reference values and assessed noise content in median cardiac complexes.

Main Results:

• Original ECG data yielded consistently smaller measurement differences across all intervals compared to artificial ECG data.
• Median cardiac complexes from original ECG data exhibited significantly lower noise levels than those from artificial ECG data.
• Clean ECGs demonstrated lower noise levels than noisy ECGs in both original and artificial datasets.

Conclusions:

• Original ECG data is superior to artificial ECG data for accurate interval measurements in interpretive electrocardiograph standardization.
• Utilizing clean ECG data is essential for reducing noise and improving the reliability of ECG measurements.
• These findings support the use of original, clean ECG data in the development of robust IEC standards for electrocardiographs.