This SBIR Phase II research project will incorporate machine-learning techniques into agneto-cardiography (MCG) that measures minute magnetic fields emitted by the heart's electrophysiological activity, based on SQUID technology and operable in typical (magnetically unshielded) hospital rooms, for early non-invasive diagnosis of heart disease. The overall objective of this project is to identify and localize, using MCG, cardiac ischemia, the leading cause of death in the US. The focus will be on excellent predictability, ease of tuning, and user transparency of machine learning tools. Upon successful completion of this project MCG has the potential to become the new gold standard for the detection of cardiac ischemia in patients presenting with suspicion of acute coronary syndrome.
Worldwide, the lack of inexpensive and non-invasive cardiac diagnostic techniques causes unnecessary delays in the recognition of acute coronary heart disease and its treatment. The feasibility of MCG to diagnose heart disease has been demonstrated. Machine learning tools provide quantitative methods for the automated diagnosis of heart disease. After successful completion of this project, physicians and nurses in leading U.S. hospitals can be trained in automated MCG diagnosis. It will also usher the use of machine learning tools for medical diagnosis in general.
