Detection of Left Ventricular Hypertrophy with Electrocardiographic Signals

Left ventricular hypertrophy (LVH) is associated with an increased risk of cardiovascular disease and is a sign of subclinical organ damage. Electrocardiograms (ECGs) are cheap, non-invasive, and easy to repeat, according to medical experts. They are often the initial line of defense in the fight against heart disease. Nowadays, there are a number of criteria used to evaluate LVH by ECG. The RS peak voltage combination of the multi-lead ECG usually needs to be higher than one or more diagnostic thresholds in order to meet these requirements. Segmenting the ECG beats of each case (LVH or non-LVH) was done using 24 features that were automatically retrieved from the 12-lead ECG's R-peak and S-valley amplitudes. After that, a dataset with these attributes was used to train a backpropagation neural network (BPN). In this way, we were able to create an algorithm that could identify LVH from ECG data. Until recently, echocardiogram (ECHO) was the go-to test for detecting LVH. In a group of people from Taiwan, there were 173 cases of LVH. We were able to detect 1466 ECG cycles of LVH patients after beat segmentation. This is because, due to changes in heart rate and other factors, each ECG sequence typically had 8 to 13 cycles (heartbeats). The findings showed that our BPN model for LVH detection has an accuracy of 0.961, a precision of 0.958, a sensitivity of 0.966, and a specificity of 0.956. Our BPN model outperforms seven other methods that use ECG criteria and other AI models that are based on ECG that have been reported for LVH identification in the past.