Our goal is to develop a computerized algorithm for the electrocardiogram (ECG) that will facilitate the reliable and early detection of acute myocardial infarction (AMI). The widely recognized value of thrombolytic therapy when delivered early in the management of patients with acute MI has made the development of such an algorithm especially important. Current algorithms for ECG interpretation do not provide the levels of sensitivity and specificity that are desirable. Our phase 1 findings demonstrate the feasibility of significant improvements in accuracy of identification of acute MI that may be achieved by taking into account age and gender and by factoring in such confounders as coexisting LVH, ST changes due to conditions other than acute myocardial infarction, and influences of prior infarcts. In phase 2, we plan to further improve corrections based on age and gender using a larger and more diverse database of ECGs. We will expand our database to include a much larger number of African American individuals and develop algorithms to account for differences between this population and Caucasians. We propose also to strengthen our algorithms that correct for several confounding factors. The databases we will use have associated non-ECG clinical information that provide confirmation of acute MI that will be used as the """"""""gold standard"""""""" for assessing accuracy. Validation will also include comparisons of accuracy achieved by our developed system compared to the most recent commercially available ECG interpretive systems. The demonstration of significant diagnostic improvement will lead to considerable commercial opportunities based upon the high prevalence and incidence of acute myocardial infarction.
