Office of Technology Transfer – University of Michigan

Novel method for assessment and prediction of cardiovascular status using signal processing and machine learning

Technology #6215

Questions about this technology? Ask a Technology Manager

Download Printable PDF

Categories
Researchers
Kevin Ward
Managed By
Drew Bennett
Associate Director - Software Licensing 734-615-4004
Patent Protection
US Patent Pending
Publications
Predicting defibrillation success with a multiple-domain model using machine learning
BMC Medical Informatics and Decision Making 2012 doi:10.1186/1472-6947-12-116,12:116, Oct 2012, 2012

Cardiac arrest accounts for over 15% of all deaths in Western countries. Cardiac arrest is caused by abrupt loss of heart function leading to cardiac dysrhythmia. Some cases of cardiac arrest such as such as ventricular fibrillation are treatable with CPR (cardiopulmonary resuscitation) and immediate defibrillation. However chances of successful defibrillation deteriorates over time and with repetitive unsuccessful shocks to the heart. Several approaches are being developed that use cardiac signal processing tools to assist the defibrillation process and determine the counter shock success. However, clinical transition of these methods have been proved to be impracticable or precluded due to low specificities.

Novel method for assessment and prediction of cardiovascular status using signal processing and machine learning tools

A method to guide resuscitation efforts and predict counter-shock success was developed by researcher at Virginia Commonwealth University. This unique approach employed real-time computational ventricular fibrillation waveform analysis of electrocadiograms, with and without addition of the end-tidal carbon dioxide (ETCO2) signal, and advanced machine learning algorithms to predict cardiac status of a victim. The invention also allows incorporation of other basic physiologic signals, which can be collected during cardiopulmonary resuscitation attempts and in the post-resuscitation period. The information obtained can then be used to help guide therapy and predict decompensation.

Applications:

  • Monitoring cardiovascular status
  • Assessing and analyzing cardiac health
  • Emergency patient care
  • Cardiac therapy
  • Use in ER or First responder setting
  • Use in Remote monitoring and telecardiology services
  • Application in cardiac research
  • Use in cardiac safety & efficacy trials

Advantages:

  • Improved cardiac arrest survival rate
  • Ability to incorporate diverse physiologic signals facilitating a holistic assessment of cardiovascular status
  • Guided resuscitation procedure enabling ease-of-use by untrained individuals
  • Immediate counter shock success prediction
  • Automated real-time analysis