A Neural Network Approach to Assess Myocardial Infarction
Ezquerra, Norberto F.
MetadataShow full item record
The assessment of myocardial infarction is a complex information intensive process that involves the analysis and interpretation of cardiovascular nuclear medicine images. For a number of years, a knowledge-based approach has been under development jointly between Georgia Tech and Emory University to assist in making this clinical assessment, using images obtained from Thallium-201 single-photon emission computed tomography (SPECT) images. This paper discusses recent attempts to extend this knowledge-based system to incorporate the concept of myocardial thickening as a possible measure of myocardial viability, using Tc-99m and connectionist methods. The implementation of neural networks, its linkage to the knowledge-based system, and the use of Sestamibi Tc-99 (instead of T1-201 imagery), introduce novel informatics methods to diagnostic cardiology.