Validation of advanced ECG diagnostic software for the detection of prior myocardial infarction by using nuclear cardiac imaging.

Abstract The investigators report the diagnostic performance of the latest version (Version 2.5) of the recently developed Cardiovise algorithm for detecting prior myocardial infarction (MI). The Cardiovise 2.5 prior MI algorithm, a component of Cardiovise Cardiac Diagnostic System (Inovise Medical, Inc, Newberg, OR), uses scalar QRS, scalar T wave, and vectorcardiographic QRS criteria for detecting, sizing, and localizing prior MI. In this study only the detection part of the algorithm's performance was evaluated, using 105 patients with and 98 patients without prior MI as indicated by the results of cardiac imaging with Sestamibi. The specificity, and sensitivity of Cardiovise 2.5 for detecting prior MI in this population of patients are 97% and 79%, respectively. The sensitivity and overall diagnostic performance of Cardiovise 2.5 was significantly better than those of a total of 6 human readers (3 cardiologists and 3 primary care physicians) and to 2 commercially available ECG diagnostic algorithms.