Rapid diagnosis of myocardial infarction (MI) using electrocardiography (ECG) is the cornerstone of effective treatment and prevention of mortality; however, conventional interpretation methods has low reliability for detecting MI and is difficulty to apply to limb 6-lead ECG based life type or wearable devices. We developed and validated a deep learning-based artificial intelligence algorithm (DLA) for detecting MI using 6-lead ECG. A total of 412,461 ECGs were used to develop a variational autoencoder (VAE) that reconstructed precordial 6-lead ECG using limb 6-lead ECG. Data from 9536, 1301, and 1768 ECGs of adult patients who underwent coronary angiography within 24 h from each ECG were used for development, internal and external validation, respectively. During internal and external validation, the area under the receiver operating characteristic curves of the DLA with VAE using a 6-lead ECG were 0.880 and 0.854, respectively, and the performances were preserved by the territory of the coronary lesion. Our DLA successfully detected MI using a 12-lead ECG or a 6-lead ECG. The results indicate that MI could be detected not only with a conventional 12 lead ECG but also with a life type 6-lead ECG device that employs our DLA.

Heart disease (HD) is the leading cause of global death; there are several mortality prediction models of HD for identifying critically-ill patients and for guiding decision making. The existing models, however, cannot be used during initial treatment or screening. This study aimed to derive and validate an echocardiography-based mortality prediction model for HD using deep learning (DL). In this multicenter retrospective cohort study, the subjects were admitted adult (age ≥ 18 years) HD patients who underwent echocardiography. The outcome was in-hospital mortality. We extracted predictor variables from echocardiography reports using text mining. We developed deep learning-based prediction model using derivation data of a hospital A. And we conducted external validation using echocardiography report of hospital B. We conducted subgroup analysis of coronary heart disease (CHD) and heart failure (HF) patients of hospital B and compared DL with the currently used predictive models (eg, Global Registry of Acute Coronary Events (GRACE) score, Thrombolysis in Myocardial Infarction score (TIMI), Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) score, and Get With The Guidelines-Heart Failure (GWTG-HF) score). The study subjects comprised 25 776 patients with 1026 mortalities. The areas under the receiver operating characteristic curve (AUROC) of the DL model were 0.912, 0.898, 0.958, and 0.913 for internal validation, external validation, CHD, and HF, respectively, and these results significantly outperformed other comparison models. This echocardiography-based deep learning model predicted in-hospital mortality among HD patients more accurately than existing prediction models and other machine learning models.

Out-of-hospital cardiac arrest (OHCA) is a major healthcare burden, and prognosis is critical in decision-making for treatment and the withdrawal of life-sustaining therapy. This study aimed to develop and validate a deep-learning-based out-of-hospital cardiac arrest prognostic system (DCAPS) for predicting neurologic recovery and survival to discharge. The study subjects were patients from the Korea OHCA registry who experienced return of spontaneous circulation (ROSC) after OHCA. A total of 36,190 patients were exclusively divided into a set of 28,045 subjects for derivation data and 8,145 subjects for validation data. We used information available for the time of ROSC as predictor variables, and the endpoints were neurologic recovery (cerebral performance category 1 or 2) and survival to discharge. The DCAPS was developed using the derivation data and represented the favorability of prognosis with a score between 0 and 100. The area under the receiver operating characteristic curve (AUROC) of DCAPS for predicting neurologic recovery for the validation data was 0.953 [95% confidence interval 0.952-0.954]; these results significantly outperformed those of logistic regression (0.947 [0.943-0.948]), random forest (0.943 [0.942-0.945]), support vector machine (0.930 [0.929-0.932]), and conventional methods of a previous study (0.817 [0.815-0.820]). The AUROC of the DCAPS for survival to discharge was 0.901 [0.900-0.903], and this result significantly outperformed those of other models as well. The DCAPS predicted neurologic recovery and survival to discharge of OHCA patients accurately and outperformed the conventional method and other machine-learning methods.

Screening and early diagnosis for heart failure (HF) are critical. However, conventional screening diagnostic methods have limitations, and electrocardiography (ECG)-based HF identification may be helpful. This study aimed to develop and validate a deep-learning algorithm for ECG-based HF identification (DEHF). The study involved 2 hospitals and 55,163 ECGs of 22,765 patients who performed echocardiography within 4 weeks were study subjects. ECGs were divided into derivation and validation data. Demographic and ECG features were used as predictive variables. The primary endpoint was detection of HF with reduced ejection fraction (HFrEF; ejection fraction [EF]≤40%), and the secondary endpoint was HF with mid-range to reduced EF (≤50%). We developed the DEHF using derivation data and the algorithm representing the risk of HF between 0 and 1. We confirmed accuracy and compared logistic regression (LR) and random forest (RF) analyses using validation data. The area under the receiver operating characteristic curves (AUROCs) of DEHF for identification of HFrEF were 0.843 (95% confidence interval, 0.840–0.845) and 0.889 (0.887–0.891) for internal and external validation, respectively, and these results significantly outperformed those of LR (0.800 [0.797–0.803], 0.847 [0.844–0.850]) and RF (0.807 [0.804–0.810], 0.853 [0.850–0.855]) analyses. The AUROCs of deep learning for identification of the secondary endpoint was 0.821 (0.819–0.823) and 0.850 (0.848–0.852) for internal and external validation, respectively, and these results significantly outperformed those of LR and RF. The deep-learning algorithm accurately identified HF using ECG features and outperformed other machine-learning methods.

This study aimed to develop and validate deep-learning-based artificial intelligence algorithm for predicting mortality of AHF (DAHF). 12,654 dataset from 2165 patients with AHF in two hospitals were used as train data for DAHF development, and 4759 dataset from 4759 patients with AHF in 10 hospitals enrolled to the Korean AHF registry were used as performance test data. The endpoints were in-hospital, 12-month, and 36-month mortality. We compared the DAHF performance with the Get with the Guidelines–Heart Failure (GWTG-HF) score, Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) score, and other machine-learning models by using the test data. Area under the receiver operating characteristic curve of the DAHF were 0.880 (95% confidence interval, 0.876–0.884) for predicting in-hospital mortality; these results significantly outperformed those of the GWTG-HF (0.728 [0.720–0.737]) and other machine-learning models. For predicting 12- and 36-month endpoints, DAHF (0.782 and 0.813) significantly outperformed MAGGIC score (0.718 and 0.729). During the 36-month follow-up, the high-risk group, defined by the DAHF, had a significantly higher mortality rate than the low-risk group(p<0.001). DAHF predicted the in-hospital and long-term mortality of patients with AHF more accurately than the existing risk scores and other machine-learning models.

Conventional risk stratification models for mortality of acute myocardial infarction (AMI) have potential limitations. This study aimed to develop and validate deep-learning-based risk stratification for the mortality of patients with AMI (DAMI). The data of 22,875 AMI patients from the Korean working group of the myocardial infarction (KorMI) registry were exclusively divided into 12,152 derivation data of 36 hospitals and 10,723 validation data of 23 hospitals. The predictor variables were the initial demographic and laboratory data. The endpoints were in-hospital mortality and 12-months mortality. We compared the DAMI performance with the global registry of acute coronary event (GRACE) score, acute coronary treatment and intervention outcomes network (ACTION) score, and the thrombolysis in myocardial infarction (TIMI) score using the validation data. In-hospital mortality for the study subjects was 4.4% and 6-month mortality after survival upon discharge was 2.2%. The areas under the receiver operating characteristic curves (AUCs) of the DAMI were 0.905 [95% confidence interval 0.902–0.909] and 0.870 [0.865–0.876] for the ST elevation myocardial infarction (STEMI) and non ST elevation myocardial infarction (NSTEMI) patients, respectively; these results significantly outperformed those of the GRACE (0.851 [0.846–0.856], 0.810 [0.803–0.819]), ACTION (0.852 [0.847–0.857], 0.806 [0.799–0.814] and TIMI score (0.781 [0.775–0.787], 0.593[0.585–0.603]). DAMI predicted 30.9% of patients more accurately than the GRACE score. As secondary outcome, during the 6-month follow-up, the high risk group, defined by the DAMI, has a significantly higher mortality rate than the low risk group (17.1% vs. 0.5%, p < 0.001). The DAMI predicted in-hospital mortality and 12-month mortality of AMI patients more accurately than the existing risk scores and other machine-learning methods.

In‐hospital cardiac arrest is a major burden to public health, which affects patient safety. Although traditional track‐and‐trigger systems are used to predict cardiac arrest early, they have limitations, with low sensitivity and high false‐alarm rates. We propose a deep learning–based early warning system that shows higher performance than the existing track‐and‐trigger systems. This retrospective cohort study reviewed patients who were admitted to 2 hospitals from June 2010 to July 2017. A total of 52 131 patients were included. Specifically, a recurrent neural network was trained using data from June 2010 to January 2017. The result was tested using the data from February to July 2017. The primary outcome was cardiac arrest, and the secondary outcome was death without attempted resuscitation. As comparative measures, we used the area under the receiver operating characteristic curve (AUROC), the area under the precision–recall curve (AUPRC), and the net reclassification index. Furthermore, we evaluated sensitivity while varying the number of alarms. The deep learning–based early warning system (AUROC: 0.850; AUPRC: 0.044) significantly outperformed a modified early warning score (AUROC: 0.603; AUPRC: 0.003), a random forest algorithm (AUROC: 0.780; AUPRC: 0.014), and logistic regression (AUROC: 0.613; AUPRC: 0.007). Furthermore, the deep learning–based early warning system reduced the number of alarms by 82.2%, 13.5%, and 42.1% compared with the modified early warning system, random forest, and logistic regression, respectively, at the same sensitivity. An algorithm based on deep learning had high sensitivity and a low false‐alarm rate for detection of patients with cardiac arrest in the multicenter study.

Triage is important in identifying high-risk patients amongst many less urgent patients as emergency department (ED) overcrowding has become a national crisis recently. This study aims to validate that a Deep-learning-based Triage and Acuity Score (DTAS) identifies high-risk patients more accurately than existing triage and acuity scores using a large national dataset. We conducted a retrospective observational cohort study using data from the Korean National Emergency Department Information System (NEDIS), which collected data on visits in real time from 151 EDs. The NEDIS data was split into derivation data (January 2014-June 2016) and validation data (July-December 2016). We also used data from the Sejong General Hospital (SGH) for external validation (January-December 2017). We predicted in-hospital mortality, critical care, and hospitalization using initial information of ED patients (age, sex, chief complaint, time from symptom onset to ED visit, arrival mode, trauma, initial vital signs and mental status as predictor variables). A total of 11,656,559 patients were included in this study. The primary outcome was in-hospital mortality. The Area Under the Receiver Operating Characteristic curve (AUROC) and Area Under the Precision and Recall Curve (AUPRC) of DTAS were 0.935 and 0.264. It significantly outperformed Korean triage and acuity score (AUROC:0.785, AUPRC:0.192), modified early warning score (AUROC:0.810, AUPRC:0.116), logistic regression (AUROC:0.903, AUPRC:0.209), and random forest (AUROC:0.910, AUPRC:0.179). Deep-learning-based Triage and Acuity Score predicted in-hospital mortality, critical care, and hospitalization more accurately than existing triages and acuity, and it was validated using a large, multicenter dataset.