Symmetrical compression distance for arrhythmia discrimination in cloud-based big-data services.

Area of Science:

• Cardiology
• Computer Science
• Data Science

Background:

• Cloud computing is transforming big data analysis, particularly in healthcare.
• The SCOOP platform offers a national big data service for implantable cardioverter defibrillators.
• Accurate classification of intracardiac electrograms (EGMs) is crucial for diagnosing cardiac arrhythmias.

Purpose of the Study:

• To propose a novel methodology for automatic EGM classification within a cloud computing system.
• To develop a computationally efficient, compression-based similarity measure (CSM) for EGM analysis.
• To evaluate the performance of the proposed methodology on a large EGM dataset.

Main Methods:

• Developed a new weighted fast compression distance (WFCD) as a compression-based similarity measure.
• Utilized simple machine learning techniques for classification.
• Applied the methodology to 6848 intracardiac electrograms (EGMs) from the SCOOP platform.
• Classified EGMs into seven arrhythmia types and a noise category.

Main Results:

• The WFCD demonstrated superior performance compared to existing CSMs.
• Achieved nearly 90% accuracy in arrhythmia classification when prior patient data was available.
• Reached 63% accuracy without prior patient data, outperforming majority class classification.
• The system effectively classified EGMs into distinct cardiac arrhythmia and noise classes.

Conclusions:

• The proposed cloud-based methodology offers a high-quality service for automatic EGM classification.
• The WFCD provides an efficient and effective approach for analyzing large EGM datasets.
• This technology can support physicians in improving patient diagnosis and knowledge of cardiac arrhythmias.