Electrocardiograms with pacemakers: accuracy of computer reading

Area of Science:

• Cardiology
• Biomedical Engineering
• Artificial Intelligence in Medicine

Background:

• Electronic pacemakers (PMs) are increasingly implanted.
• Electrocardiograms (ECGs) are crucial for assessing cardiac function, including pacemaker function.
• Automated ECG interpretation algorithms are being developed to aid clinicians.

Purpose of the Study:

• To evaluate the accuracy of a computer algorithm in reading ECGs with electronic pacemakers.
• To identify specific errors made by the algorithm in interpreting pacemaker activity.

Main Methods:

• Screening ECGs for electronic pacing spikes.
• Comparing computer-derived interpretations with expert cardiologists' readings.
• Analyzing discrepancies and types of errors in automated interpretations.

Main Results:

• Computer interpretations required cardiologist revision in 61.3% of cases.
• The algorithm failed to recognize pacemakers in 18.4% of ECGs.
• Common errors included misidentifying paced beats, failing to detect underlying rhythms, and misclassifying pacemaker types.

Conclusions:

• Current ECG reading algorithms need significant refinement for accurate pacemaker interpretation.
• Improvements are essential in recognizing underlying rhythms, pacing spikes, and pacing modes.
• Enhanced algorithms are necessary due to the growing number of pacemaker patients and ECGs.