Artificial Intelligence Decision Support System for Myocardial Infarction

AI Center for Medical Diagnosis

The First Hospital in Taiwan apply this system：

Acute myocardial infarction has always been a invisible killer of nationals’ health. According to the top 10 causes of death in 2019 announced by the Ministry of Health and Welfare, heart disease ranked the second, and has been for many years. Among them, coronary heart disease is even more common. Some patients often delay treatment due to atypical symptoms or delayed interpretation of ECG, which highlights the importance of early diagnosis and immediate getting through the blocked blood vessels.

In view of this, our hospital utilized the 12-lead ECG database as the training data set, and adopted the multi-layer neural network atrial fibrillation system, and the establishment of deep learning of the AI algorithm to analyze the changes in the ECG signal. The AI Decision Support System includes: AMI warning system based on the EKG AF detection, equipped with a mobile phone instant messaging mechanism, and supplemented by emergency ASAP scores to determine the risk level of patients with atypical symptoms.

Publications：

• Usefulness of Machine Learning-Based Detection and Classification of Cardiac Arrhythmias With 12-Lead Electrocardiograms. 2020. K.C. Chang et el. 10.1016/j.cjca.2020.02.096. Canadian Journal of Cardiology.

Verification results and effectiveness:
(as of 2020/8/14 statistical results)

Accuracy Rate：86%

Positive Predict Value：100%

Reduce D2B Time：From 61 mins even to 32 mins!

 Booth No. M804 

 Look for  

1. Stemi emergency warnings are currently presented in the form of mobile phone text messages. We are considering to cooperate with manufacturers of wearable devices to discuss advanced and more real-time warning procedures and expand the scope of use for those patients at home or rural hospitals.

2. Through this medical exhibition, we hope to find co-developed manufacturers or hospitals for technical communication or cooperation, in order to optimizing system or increase the types of diseases of this supporting system.