An ECG Analysis with Artificial Intelligence to Predict Short- and Long-Term Prediction of Acute Heart Failure
Nov 14, 2024
A research team led by Professors Cho Young-jin, Yoon Min-jae, and Choi Dong-joo of the Department of Circulatory Medicine at Bundang Seoul National University Hospital, and Professor Kim Joong-hee of the Department of Emergency Medicine (joint research by professors Lee Chan-joo and Kang Seok-min of Severance Hospital) analyzed ECG with artificial intelligence to accurately predict the short- and long-term prognosis of patients with acute heart failure.
Heart failure, in which blood is not properly supplied to body tissues due to structural and functional abnormalities of the heart, causes extreme fatigue, decreased exercise capacity, and swelling, and even if treated, caution is required due to the high risk of readmission and death.
In order to diagnose heart failure and predict the prognosis, detailed examinations such as blood tests, electrocardiogram, chest X-rays, and echocardiography are performed, and these tests are often difficult to fully utilize in clinical settings due to realistic limitations such as time and cost.
Accordingly, the research team analyzed 47,000 ECG test results that record the electrical activity of the heart in the form of wavelengths using a deep learning algorithm and developed a model to predict the condition of acute heart failure patients based on ECG results. The electrocardiogram test can be performed relatively simply and the test results can be checked quickly at low cost, so it is particularly useful for people with heart disease.
The research team tried to accurately predict the prognosis of patients with acute heart failure through AI-based quantitative electrocardiogram (QCG), which numerically displays various urgent heart-related indicators such as heart shock, cardiac arrest, and reduced left ventricular ejection rate, and conducted a study that applied them to 1,254 patients with acute heart failure admitted to Seoul National University Bundang Hospital and Severance Hospital.
As a result, it was found that the prediction rate of AI-based quantitative electrocardiogram developed by the research team was significantly ahead of biomarkers such as NT-proBNP or left ventricular ejection rate in predicting cardiac causes during hospitalization, and that quantitative electrocardiogram was highly predictable in predicting long-term mortality.
This study is meaningful because it suggests that the results of electrocardiogram analysis using artificial intelligence in heart failure, which was difficult to predict the prognosis without a detailed examination, can be used simply and conveniently for predicting the prognosis.
Cho Young-jin, professor of circulatory medicine at Seoul National University Bundang Hospital, said "Through artificial intelligence, we have confirmed that even a simple electrocardiogram can help predict the prognosis of heart failure patients."We will continue our research to improve the use of artificial intelligence-based electrocardiogram so that the prognosis of heart disease patients can be predicted more accurately."
Meanwhile, the research results were published in the international journal 『Journal of Medical Internet Research』, and the research team's artificial intelligence-based electrocardiogram analysis solution was developed under the name 『ECG Buddy" and was licensed as a second-class medical device by the Ministry of Food and Drug Safety and selected as a new medical technology for evaluation deferral.
In addition to this study, ECG Buddy has been confirmed for clinical usefulness in various heart diseases, such as its excellent performance in predicting the risk of coronary artery disease in patients with stable angina.
Heart failure, in which blood is not properly supplied to body tissues due to structural and functional abnormalities of the heart, causes extreme fatigue, decreased exercise capacity, and swelling, and even if treated, caution is required due to the high risk of readmission and death.
In order to diagnose heart failure and predict the prognosis, detailed examinations such as blood tests, electrocardiogram, chest X-rays, and echocardiography are performed, and these tests are often difficult to fully utilize in clinical settings due to realistic limitations such as time and cost.
Accordingly, the research team analyzed 47,000 ECG test results that record the electrical activity of the heart in the form of wavelengths using a deep learning algorithm and developed a model to predict the condition of acute heart failure patients based on ECG results. The electrocardiogram test can be performed relatively simply and the test results can be checked quickly at low cost, so it is particularly useful for people with heart disease.
The research team tried to accurately predict the prognosis of patients with acute heart failure through AI-based quantitative electrocardiogram (QCG), which numerically displays various urgent heart-related indicators such as heart shock, cardiac arrest, and reduced left ventricular ejection rate, and conducted a study that applied them to 1,254 patients with acute heart failure admitted to Seoul National University Bundang Hospital and Severance Hospital.
As a result, it was found that the prediction rate of AI-based quantitative electrocardiogram developed by the research team was significantly ahead of biomarkers such as NT-proBNP or left ventricular ejection rate in predicting cardiac causes during hospitalization, and that quantitative electrocardiogram was highly predictable in predicting long-term mortality.
This study is meaningful because it suggests that the results of electrocardiogram analysis using artificial intelligence in heart failure, which was difficult to predict the prognosis without a detailed examination, can be used simply and conveniently for predicting the prognosis.
Cho Young-jin, professor of circulatory medicine at Seoul National University Bundang Hospital, said "Through artificial intelligence, we have confirmed that even a simple electrocardiogram can help predict the prognosis of heart failure patients."We will continue our research to improve the use of artificial intelligence-based electrocardiogram so that the prognosis of heart disease patients can be predicted more accurately."
Meanwhile, the research results were published in the international journal 『Journal of Medical Internet Research』, and the research team's artificial intelligence-based electrocardiogram analysis solution was developed under the name 『ECG Buddy" and was licensed as a second-class medical device by the Ministry of Food and Drug Safety and selected as a new medical technology for evaluation deferral.
In addition to this study, ECG Buddy has been confirmed for clinical usefulness in various heart diseases, such as its excellent performance in predicting the risk of coronary artery disease in patients with stable angina.
|
bellho@sportschosun.com
