New VCG and ECG Indexes for Early Identification of Acute Myocardial Infarction Patients

Abstract

A novel electrocardiographic (ECG) and vectorcardiographic (VCG) signal processing technique is proposed for identification of patients with Acute Myocardial Infarction (AMI). To this end, we propose a method based on two QRSloops features: a) Volume, and b) Planar Area; along with five orthogonal ECG indexes: c) ST-T vector difference, d) T vector difference, and e-g) the area under the T-wave in X, Y and Z leads. Our method was tested using ECG recordings of 97 AMIs and 52 healthy subject from the PhysioNet PTB database. The results indicate that these indexes show significant statistical differences (p-values <0.05) between the two populations (Healthy versus AMI population, with the exception of c). To validate the suggested index a 70/30 cross validation technique was used. Moreover, combining all parameters (with the exception of the T-wave area in Y and Z leads), it was possible to classify the AMI and healthy subjects with a sensitivity = 93.90%, a specificity = 93.44% and an accuracy = 93.73%, applying a linear discriminant classifier method. We conclude that the proposed technique could be used as an alternative diagnostic technique in the emergency room.