Quickly finding a needle in a haystack: a new automated cardiac arrhythmia detection software for preclinical studies

Area of Science:

• Cardiovascular pharmacology
• Preclinical safety assessment
• Bioinformatics

Background:

• Drug-induced arrhythmias are a major safety concern in preclinical studies.
• Accurate assessment of arrhythmia incidence is crucial for risk evaluation.
• Manual ECG analysis for arrhythmias is time-consuming and lacks automation.

Purpose of the Study:

• To characterize the performance of the automated arrhythmia detection software, ARR30a.
• To evaluate ARR30a's accuracy in detecting major arrhythmia types in large and small animal species.

Main Methods:

• ARR30a was developed for automated detection of five major arrhythmia types.
• Sensitivity and predictivity were evaluated using 84 ECG recordings from various species and protocols.
• Automated detection was compared against manual analysis.

Main Results:

• ARR30a demonstrated high sensitivity and predictivity in large animals (dogs, non-human primates, pigs) for ventricular beats, AV-blocks II, and junctional beats.
• Lower sensitivity was noted in rats for junctional beats due to P-wave detection challenges.
• The software showed robustness to moderate noise levels and processed 24-hour recordings efficiently.

Conclusions:

• The validated ARR30a software enables comprehensive evaluation of ECG abnormalities and improves arrhythmia incidence quantification in preclinical studies.
• ARR30a enhances safety pharmacology and toxicology by providing automated, efficient arrhythmia detection.
• The accompanying RME10a tool allows manual validation and refinement of automated results.