Entropy Limited's Entropy Minimax multivariate statistical analysis technology has been applied to heart disease in several ways. Examples include identifying patterns in heartbeat abnormality, detecting possible problems in heart valves and determining heart patient survival classification.
Patterns of Survival in Patients with Defined Coronary Artery Disease
In cooperation with Duke University Medical Center, Entropy Limited developed computer algorithms detailing a set of patterns in non-mathematical medical language for predicting related survival classification based on catheterization and clinical data from over 2000 medically and surgically treated coronary artery disease patients. The patterns were tested on a new set of about 1000 independent cases, and found to be accurate both with respect to short term (two-year) and longer term (20% of normal remaining lifetime) survival.
Spectral Analysis of Heart Valve Sounds
EL and the Department of Metallurgy and Materials Science at Carnegie-Mellon University worked together to conduct a spectral analysis of heart valve sounds in response to audio pulses as a means of assessing valve reliability by detecting cracks and incipient cracks in the valves.
ECG Waveform Pattern Analysis
The Biomedical Technology Program at Carnegie-Mellon University used Entropy Limited's minimax technologies to identify patterns in ECG waveforms discriminating normal and abnormal heartbeats.
Principal Components Analysis of Heartbeat Waveforms
In cooperation with the Biomedical Technology Program at Carnegie-Mellon University, EL developed an automatic phase alignment method for principal components (Karhunen-Loeve) expansion analysis applicable to heartbeat waveforms.