Diagnostic accuracy of computer aided electrocardiogram analysis in dogs

Area of Science:

• Veterinary Cardiology
• Biomedical Engineering
• Diagnostic Imaging

Background:

• Electrocardiograms (ECGs) are crucial for diagnosing cardiac conditions in animals.
• Accurate interpretation of ECGs requires specialized expertise, potentially leading to delays in diagnosis and treatment.
• Automated ECG analysis tools can potentially improve efficiency and accuracy in veterinary cardiology.

Purpose of the Study:

• To evaluate the diagnostic performance of a computerized ECG algorithm in dogs.
• To compare the algorithm's interpretation against a consensus panel of board-certified veterinary cardiologists.
• To determine the algorithm's utility in identifying arrhythmias and other ECG anomalies.

Main Methods:

• A retrospective cohort study involving 399 canine ECG recordings.
• ECGs were analyzed by a panel of seven veterinary cardiologists for various parameters and overall diagnosis.
• The computerized algorithm's interpretations were compared against the cardiologists' consensus findings.

Main Results:

• The algorithm demonstrated high sensitivity (99.7%) and specificity (99.5%) for detecting arrhythmias.
• For identifying any ECG anomaly, the algorithm's sensitivity was 71.3% and specificity was 35.1%.
• The algorithm excelled in arrhythmia detection but was less comprehensive for all ECG abnormalities.

Conclusions:

• The computerized ECG algorithm is a highly sensitive tool for detecting arrhythmias in dogs.
• The algorithm is not as effective in identifying all types of ECG anomalies.
• This tool can aid in the triage of ECGs within a veterinary clinical workflow, supporting cardiologists' interpretations.