Leveraging Machine Learning And Pre-Hospital 12-Lead ECG For Acute Coronary Syndrome Prediction
DOI:
https://doi.org/10.64252/g60ktm78Keywords:
12 Lead ECG Acute Coronary Syndrome Support Vector Machine Ejection FractionAbstract
Background: Heart Failure is present in 1 out of 10 patients presenting Acute Coronary Syndrome. ECG predictors of reduced LVEF provide essential non-invasive triage capabilities, especially when echocardiography is not available instantly within reach. 12 lead ECG is easily accessible during the preliminary checkup of patient No direct comparison of the current electrocardiogram (ECG) interpretation program exists. Additional ECG training can improve accuracy. A systematic review and meta-analysis observed that across various training like pretraining and post-training and the median accuracy of ECG interpretation was 54% and 67% respectively [1].
Method: Here we proposed machine learning based approach to identify the highly predictive ECG Marker of Suspected Acute Coronary Syndrome i.e. the Ejection Fraction found in checkups [2]. This research proposes a modified technique for identifying cardiac abnormalities and QRS complexes, Leveraging Machine Learning and SVM (Support Vector Machine) classifiers.
Results: The presented technique surpasses existing methods in both sensitivity and specificity achieving accuracy of 98.4 % for cardiac irregularities for the standard 12-Lead ECG Georgia database