Description and evaluation of a method for computer analysis of the exercise electrocardiogram

Area of Science:

• Cardiology
• Medical Informatics
• Biomedical Engineering

Background:

• Exercise electrocardiography (ECG) is crucial for diagnosing cardiac ischemia.
• Computer-assisted analysis of ECG data can potentially improve diagnostic accuracy.
• Validation of automated ECG analysis is essential for clinical adoption.

Purpose of the Study:

• To develop and evaluate a new computer program for real-time, 12-lead exercise ECG analysis.
• To assess the program's ability to detect ischemic changes during exercise stress tests.
• To compare the diagnostic performance of the automated program with visual interpretation.

Main Methods:

• Development of an "on line" computer program for analyzing 12-lead exercise ECG.
• The program quantifies parameters like QRS duration, ST segment level/slope, and T wave amplitude.
• Quantitative data from 200 exercise stress tests were compared against visual analysis and clinical data.

Main Results:

• The program successfully detects normal QRS complexes and ectopic beats.
• Specific criteria for ST level and ST slope were identified as optimal for ischemia detection.
• Computer analysis achieved a sensitivity of 92% for ischemia, an improvement from 86.6% with visual reading, while maintaining 94% specificity.

Conclusions:

• The developed computer program offers an effective method for "on line" exercise ECG analysis.
• Automated analysis demonstrates improved sensitivity in diagnosing cardiac ischemia compared to visual interpretation.
• This technology holds promise for enhancing the diagnostic capabilities of exercise stress testing.