Precordial T wave inversion differentiation in pulmonary embolism vs acute coronary syndrome using machine learning

Abstract

T wave inversion is an electrocardiographic finding that is commonly encountered both in asymptomatic people and in patients with various clinical conditions, including acute coronary syndrome and pulmonary embolism, and differentiating these two clinical conditions from each other for doctors are very vital. Artificial intelligence is one of the new techniques for diagnosis that can be used in this field as well. Therefore, the aim of this study is to differentiate the inverted T wave morphology in precordial leads between acute pulmonary embolism and acute coronary syndrome caused by severe coronary artery stenosis via machine learning. Methods: In this study, one hundred patients with definite diagnosis of pulmonary embolism (PTE) and one hundred patients with diagnosis of acute coronary syndrome without ST segment elevation with coronary artery stenosis (ACS) diagnosed in coronary angiography and having inverted T wave in Precordial leads were selected. A questionnaire has been designed in which demographic information including age, sex, co-morbidities and laboratory and echocardiography findings are written, and then the ECG of these people is scanned and the ECG indicators are input to the algorithm and the type of disease is output to the machine. is defined Algorithms were trained using patient data and their performance was evaluated based on criteria such as accuracy, sensitivity and specificity. Results: Our results showed that for the differentiation of inverted T wave morphology in precordial leads between PTE and ACS, the best results are related to Ensemble sequence with accuracy rate of 84.44%, sensitivity 83.96%, specificity 84.87% and F1 score. 83.56%, and Neural Network with 84% accuracy, 85.14% sensitivity, 83.06% specificity, and 82.69% F1 score.