Automated Detection of Acute Myocardial Infarction Using Asynchronous Electrocardiogram Signals-Preview of Implementing Artificial Intelligence With Multichannel Electrocardiographs Obtained From Smartwatches: Retrospective Study.

BACKGROUND: When using a smartwatch to obtain electrocardiogram (ECG) signals from multiple leads, the device has to be placed on different parts of the body sequentially. The ECG signals measured from different leads are asynchronous. Artificial intelligence (AI) models for asynchronous ECG signals have barely been explored. OBJECTIVE: We aimed to develop an AI model for detecting acute myocardial infarction using asynchronous ECGs and compare its performance with that of the automatic ECG interpretations provided by a commercial ECG analysis software. We sought to evaluate the feasibility of implementing multiple lead-based AI-enabled ECG algorithms on smartwatches. Moreover, we aimed to determine the optimal number of leads for sufficient diagnostic power. METHODS: We extracted ECGs recorded within 24 hours from each visit to the emergency room of Ajou University Medical Center between June 1994 and January 2018 from patients aged 20 years or older. The ECGs were labeled on the basis of whether a diagnostic code corresponding to acute myocardial infarction was entered. We derived asynchronous ECG lead sets from standard 12-lead ECG reports and simulated a situation similar to the sequential recording of ECG leads via smartwatches. We constructed an AI model based on residual networks and self-attention mechanisms by randomly masking each lead channel during the training phase and then testing the model using various targeting lead sets with the remaining lead channels masked. RESULTS: The performance of lead sets with 3 or more leads compared favorably with that of the automatic ECG interpretations provided by a commercial ECG analysis software, with 8.1%-13.9% gain in sensitivity when the specificity was matched. Our results indicate that multiple lead-based AI-enabled ECG algorithms can be implemented on smartwatches. Model performance generally increased as the number of leads increased (12-lead sets: area under the receiver operating characteristic curve [AUROC] 0.880; 4-lead sets: AUROC 0.858, SD 0.008; 3-lead sets: AUROC 0.845, SD 0.011; 2-lead sets: AUROC 0.813, SD 0.018; single-lead sets: AUROC 0.768, SD 0.001). Considering the short amount of time needed to measure additional leads, measuring at least 3 leads-ideally more than 4 leads-is necessary for minimizing the risk of failing to detect acute myocardial infarction occurring in a certain spatial location or direction. CONCLUSIONS: By developing an AI model for detecting acute myocardial infarction with asynchronous ECG lead sets, we demonstrated the feasibility of multiple lead-based AI-enabled ECG algorithms on smartwatches for automated diagnosis of cardiac disorders. We also demonstrated the necessity of measuring at least 3 leads for accurate detection. Our results can be used as reference for the development of other AI models using sequentially measured asynchronous ECG leads via smartwatches for detecting various cardiac disorders.

Machine Learning Approach for Active Vaccine Safety Monitoring.

BACKGROUND: Vaccine safety surveillance is important because it is related to vaccine hesitancy, which affects vaccination rate. To increase confidence in vaccination, the active monitoring of vaccine adverse events is important. For effective active surveillance, we developed and verified a machine learning-based active surveillance system using national claim data. METHODS: We used two databases, one from the Korea Disease Control and Prevention Agency, which contains flu vaccination records for the elderly, and another from the National Health Insurance Service, which contains the claim data of vaccinated people. We developed a case-crossover design based machine learning model to predict the health outcome of interest events (anaphylaxis and agranulocytosis) using a random forest. Feature importance values were evaluated to determine candidate associations with each outcome. We investigated the relationship of the features to each event via a literature review, comparison with the Side Effect Resource, and using the Local Interpretable Model-agnostic Explanation method. RESULTS: The trained model predicted each health outcome of interest with a high accuracy (approximately 70%). We found literature supporting our results, and most of the important drug-related features were listed in the Side Effect Resource database as inducing the health outcome of interest. For anaphylaxis, flu vaccination ranked high in our feature importance analysis and had a positive association in Local Interpretable Model-Agnostic Explanation analysis. Although the feature importance of vaccination was lower for agranulocytosis, it also had a positive relationship in the Local Interpretable Model-Agnostic Explanation analysis. CONCLUSION: We developed a machine learning-based active surveillance system for detecting possible factors that can induce adverse events using health claim and vaccination databases. The results of the study demonstrated a potentially useful application of two linked national health record databases. Our model can contribute to the establishment of a system for conducting active surveillance on vaccination.

Does Exposure to Computed Tomography Contrast Media Increase Risk of End-Stage Renal Disease?

BACKGROUND There are many studies on acute kidney injury (AKI) after exposure to contrast media in patients with chronic kidney disease (CKD). However, whether the risk of end-stage renal disease (ESRD) increases after exposure to contrast media in the long term, regardless of development of AKI after such exposure, has not been studied. MATERIAL AND METHODS The electronic health records of patients diagnosed with CKD and followed up from 2014 to 2018 at a tertiary university hospital were retrospectively collected. Patients were divided into patients who progressed to ESRD (ESRD group) and those who did not (non-ESRD group). Patients in the non-ESRD group were matched 1: 1 to those in the ESRD group by using disease risk score generation and matching. Multivariate logistic regression analysis was performed to assess the effect of contrast media exposure on progression to ESRD. RESULTS In total, 179 patients were enrolled per group; 178 (99.4%) were in CKD stage 3 or above in both groups. Average serum creatinine was 4.31±3.02 mg/dl and 3.64±2.55 mg/dl in the ESRD and non-ESRD groups, respectively (p=0.242). Other baseline characteristics were not statistically significant, except for the number of times contrast-enhanced computed tomography (CECT) was performed (0.00 [Interquartile range (IQR) 0.00-2.00] in the ESRD group and 0.00 [IQR 0.00-1.00] in the non-ESRD group [p=0.006]); in multivariate logistic regression, this number (OR=1.24, 95% CI=1.08-1.47, p=0.006) was significantly related to progression to ESRD. CONCLUSIONS The use of CECT increased the risk of ESRD 1.2-fold in advanced and stable CKD outpatients after 5-year follow-up.

AI-enabled ECG index for predicting left ventricular dysfunction in patients with ST-segment elevation myocardial infarction.

Electrocardiogram (ECG) changes after primary percutaneous coronary intervention (PCI) in ST-segment elevation myocardial infarction (STEMI) patients are associated with prognosis. This study investigated the feasibility of predicting left ventricular (LV) dysfunction in STEMI patients using an artificial intelligence (AI)-enabled ECG algorithm developed to diagnose STEMI. Serial ECGs from 637 STEMI patients were analyzed with the AI algorithm, which quantified the probability of STEMI at various time points. The time points included pre-PCI, immediately post-PCI, 6 h post-PCI, 24 h post-PCI, at discharge, and one-month post-PCI. The prevalence of LV dysfunction was significantly associated with the AI-derived probability index. A high probability index was an independent predictor of LV dysfunction, with higher cardiac death and heart failure hospitalization rates observed in patients with higher indices. The study demonstrates that the AI-enabled ECG index effectively quantifies ECG changes post-PCI and serves as a digital biomarker capable of predicting post-STEMI LV dysfunction, heart failure, and mortality. These findings suggest that AI-enabled ECG analysis can be a valuable tool in the early identification of high-risk patients, enabling timely and targeted interventions to improve clinical outcomes in STEMI patients.

Identifying Atrial Fibrillation With Sinus Rhythm Electrocardiogram in Embolic Stroke of Undetermined Source: A Validation Study With Insertable Cardiac Monitors.

BACKGROUND AND OBJECTIVES: Paroxysmal atrial fibrillation (AF) is a major potential cause of embolic stroke of undetermined source (ESUS). However, identifying AF remains challenging because it occurs sporadically. Deep learning could be used to identify hidden AF based on the sinus rhythm (SR) electrocardiogram (ECG). We combined known AF risk factors and developed a deep learning algorithm (DLA) for predicting AF to optimize diagnostic performance in ESUS patients. METHODS: A DLA was developed to identify AF using SR 12-lead ECG with the database consisting of AF patients and non-AF patients. The accuracy of the DLA was validated in 221 ESUS patients who underwent insertable cardiac monitor (ICM) insertion to identify AF. RESULTS: A total of 44,085 ECGs from 12,666 patient were used for developing the DLA. The internal validation of the DLA revealed 0.862 (95% confidence interval, 0.850-0.873) area under the curve (AUC) in the receiver operating curve analysis. In external validation data from 221 ESUS patients, the diagnostic accuracy of DLA and AUC were 0.811 and 0.827, respectively, and DLA outperformed conventional predictive models, including CHARGE-AF, C2HEST, and HATCH. The combined model, comprising atrial ectopic burden, left atrial diameter and the DLA, showed excellent performance in AF prediction with AUC of 0.906. CONCLUSIONS: The DLA accurately identified paroxysmal AF using 12-lead SR ECG in patients with ESUS and outperformed the conventional models. The DLA model along with the traditional AF risk factors could be a useful tool to identify paroxysmal AF in ESUS patients.

Predicting intraoperative hypotension using deep learning with waveforms of arterial blood pressure, electroencephalogram, and electrocardiogram: Retrospective study.

To develop deep learning models for predicting Interoperative hypotension (IOH) using waveforms from arterial blood pressure (ABP), electrocardiogram (ECG), and electroencephalogram (EEG), and to determine whether combination ABP with EEG or CG improves model performance. Data were retrieved from VitalDB, a public data repository of vital signs taken during surgeries in 10 operating rooms at Seoul National University Hospital from January 6, 2005, to March 1, 2014. Retrospective data from 14,140 adult patients undergoing non-cardiac surgery with general anaesthesia were used. The predictive performances of models trained with different combinations of waveforms were evaluated and compared at time points at 3, 5, 10, 15 minutes before the event. The performance was calculated by area under the receiver operating characteristic (AUROC), area under the precision-recall curve (AUPRC), sensitivity and specificity. The model performance was better in the model using both ABP and EEG waveforms than in all other models at all time points (3, 5, 10, and 15 minutes before an event) Using high-fidelity ABP and EEG waveforms, the model predicted IOH with a AUROC and AUPRC of 0.935 [0.932 to 0.938] and 0.882 [0.876 to 0.887] at 5 minutes before an IOH event. The output of both ABP and EEG was more calibrated than that using other combinations or ABP alone. The results demonstrate that a predictive deep neural network can be trained using ABP, ECG, and EEG waveforms, and the combination of ABP and EEG improves model performance and calibration.

Modeling Brain Volume Using Deep Learning-Based Physical Activity Features in Patients With Dementia.

There is a proven correlation between the severity of dementia and reduced brain volumes. Several studies have attempted to use activity data to estimate brain volume as a means of detecting reduction early; however, raw activity data are not directly interpretable and are unstructured, making them challenging to utilize. Furthermore, in the previous research, brain volume estimates were limited to total brain volume and the investigators were unable to detect reductions in specific regions of the brain that are typically used to characterize disease progression. We aimed to evaluate volume prediction of 116 brain regions through activity data obtained combining time-frequency domain- and unsupervised deep learning-based feature extraction methods. We developed a feature extraction model based on unsupervised deep learning using activity data from the National Health and Nutrition Examination Survey (NHANES) dataset (n = 14,482). Then, we applied the model and the time-frequency domain feature extraction method to the activity data of the Biobank Innovations for chronic Cerebrovascular disease With ALZheimer's disease Study (BICWALZS) datasets (n = 177) to extract activity features. Brain volumes were calculated from the brain magnetic resonance imaging of the BICWALZS dataset and anatomically subdivided into 116 regions. Finally, we fitted linear regression models to estimate each regional volume of the 116 brain areas based on the extracted activity features. Regression models were statistically significant for each region, with an average correlation coefficient of 0.990 ± 0.006. In all brain regions, the correlation was > 0.964. Particularly, regions of the temporal lobe that exhibit characteristic atrophy in the early stages of Alzheimer's disease showed the highest correlation (0.995). Through a combined deep learning-time-frequency domain feature extraction method, we could extract activity features based solely on the activity dataset, without including clinical variables. The findings of this study indicate the possibility of using activity data for the detection of neurological disorders such as Alzheimer's disease.

Electrocardiographic biomarker based on machine learning for detecting overt hyperthyroidism.

Aims: Although overt hyperthyroidism adversely affects a patient's prognosis, thyroid function tests (TFTs) are not routinely conducted. Furthermore, vague symptoms of hyperthyroidism often lead to hyperthyroidism being overlooked. An electrocardiogram (ECG) is a commonly used screening test, and the association between thyroid function and ECG is well known. However, it is difficult for clinicians to detect hyperthyroidism through subtle ECG changes. For early detection of hyperthyroidism, we aimed to develop and validate an electrocardiographic biomarker based on a deep learning model (DLM) for detecting hyperthyroidism. Methods and results: This multicentre retrospective cohort study included patients who underwent ECG and TFTs within 24 h. For model development and internal validation, we obtained 174 331 ECGs from 113 194 patients. We extracted 48 648 ECGs from 33 478 patients from another hospital for external validation. Using 500 Hz raw ECG, we developed a DLM with 12-lead, 6-lead (limb leads, precordial leads), and single-lead (lead I) ECGs to detect overt hyperthyroidism. We calculated the model's performance on the internal and external validation sets using the area under the receiver operating characteristic curve (AUC). The AUC of the DLM using a 12-lead ECG was 0.926 (0.913-0.94) for internal validation and 0.883(0.855-0.911) for external validation. The AUC of DLMs using six and a single-lead were in the range of 0.889-0.906 for internal validation and 0.847-0.882 for external validation. Conclusion: We developed a DLM using ECG for non-invasive screening of overt hyperthyroidism. We expect this model to contribute to the early diagnosis of diseases and improve patient prognosis.

Effectiveness of Transfer Learning for Deep Learning-Based Electrocardiogram Analysis.

OBJECTIVES: Many deep learning-based predictive models evaluate the waveforms of electrocardiograms (ECGs). Because deep learning-based models are data-driven, large and labeled biosignal datasets are required. Most individual researchers find it difficult to collect adequate training data. We suggest that transfer learning can be used to solve this problem and increase the effectiveness of biosignal analysis. METHODS: We applied the weights of a pretrained model to another model that performed a different task (i.e., transfer learning). We used 2,648,100 unlabeled 8.2-second-long samples of ECG II data to pretrain a convolutional autoencoder (CAE) and employed the CAE to classify 12 ECG rhythms within a dataset, which had 10,646 10-second-long 12-lead ECGs with 11 rhythm labels. We split the datasets into training and test datasets in an 8:2 ratio. To confirm that transfer learning was effective, we evaluated the performance of the classifier after the proposed transfer learning, random initialization, and two-dimensional transfer learning as the size of the training dataset was reduced. All experiments were repeated 10 times using a bootstrapping method. The CAE performance was evaluated by calculating the mean squared errors (MSEs) and that of the ECG rhythm classifier by deriving F1-scores. RESULTS: The MSE of the CAE was 626.583. The mean F1-scores of the classifiers after bootstrapping of 100%, 50%, and 25% of the training dataset were 0.857, 0.843, and 0.835, respectively, when the proposed transfer learning was applied and 0.843, 0.831, and 0.543, respectively, after random initialization was applied. CONCLUSIONS: Transfer learning effectively overcomes the data shortages that can compromise ECG domain analysis by deep learning.

Unsupervised feature learning for electrocardiogram data using the convolutional variational autoencoder.

Most existing electrocardiogram (ECG) feature extraction methods rely on rule-based approaches. It is difficult to manually define all ECG features. We propose an unsupervised feature learning method using a convolutional variational autoencoder (CVAE) that can extract ECG features with unlabeled data. We used 596,000 ECG samples from 1,278 patients archived in biosignal databases from intensive care units to train the CVAE. Three external datasets were used for feature validation using two approaches. First, we explored the features without an additional training process. Clustering, latent space exploration, and anomaly detection were conducted. We confirmed that CVAE features reflected the various types of ECG rhythms. Second, we applied CVAE features to new tasks as input data and CVAE weights to weight initialization for different models for transfer learning for the classification of 12 types of arrhythmias. The f1-score for arrhythmia classification with extreme gradient boosting was 0.86 using CVAE features only. The f1-score of the model in which weights were initialized with the CVAE encoder was 5% better than that obtained with random initialization. Unsupervised feature learning with CVAE can extract the characteristics of various types of ECGs and can be an alternative to the feature extraction method for ECGs.

Discovering hidden information in biosignals from patients using artificial intelligence.

Biosignals such as electrocardiogram or photoplethysmogram are widely used for determining and monitoring the medical condition of patients. It was recently discovered that more information could be gathered from biosignals by applying artificial intelligence (AI). At present, one of the most impactful advancements in AI is deep learning. Deep learning-based models can extract important features from raw data without feature engineering by humans, provided the amount of data is sufficient. This AI-enabled feature presents opportunities to obtain latent information that may be used as a digital biomarker for detecting or predicting a clinical outcome or event without further invasive evaluation. However, the black box model of deep learning is difficult to understand for clinicians familiar with a conventional method of analysis of biosignals. A basic knowledge of AI and machine learning is required for the clinicians to properly interpret the extracted information and to adopt it in clinical practice. This review covers the basics of AI and machine learning, and the feasibility of their application to real-life situations by clinicians in the near future.

Deep Learning Approach for Imputation of Missing Values in Actigraphy Data: Algorithm Development Study.

BACKGROUND: Data collected by an actigraphy device worn on the wrist or waist can provide objective measurements for studies related to physical activity; however, some data may contain intervals where values are missing. In previous studies, statistical methods have been applied to impute missing values on the basis of statistical assumptions. Deep learning algorithms, however, can learn features from the data without any such assumptions and may outperform previous approaches in imputation tasks. OBJECTIVE: The aim of this study was to impute missing values in data using a deep learning approach. METHODS: To develop an imputation model for missing values in accelerometer-based actigraphy data, a denoising convolutional autoencoder was adopted. We trained and tested our deep learning-based imputation model with the National Health and Nutrition Examination Survey data set and validated it with the external Korea National Health and Nutrition Examination Survey and the Korean Chronic Cerebrovascular Disease Oriented Biobank data sets which consist of daily records measuring activity counts. The partial root mean square error and partial mean absolute error of the imputed intervals (partial RMSE and partial MAE, respectively) were calculated using our deep learning-based imputation model (zero-inflated denoising convolutional autoencoder) as well as using other approaches (mean imputation, zero-inflated Poisson regression, and Bayesian regression). RESULTS: The zero-inflated denoising convolutional autoencoder exhibited a partial RMSE of 839.3 counts and partial MAE of 431.1 counts, whereas mean imputation achieved a partial RMSE of 1053.2 counts and partial MAE of 545.4 counts, the zero-inflated Poisson regression model achieved a partial RMSE of 1255.6 counts and partial MAE of 508.6 counts, and Bayesian regression achieved a partial RMSE of 924.5 counts and partial MAE of 605.8 counts. CONCLUSIONS: Our deep learning-based imputation model performed better than the other methods when imputing missing values in actigraphy data.

Construction of an Electrocardiogram Database Including 12 Lead Waveforms.

OBJECTIVES: Electrocardiogram (ECG) data are important for the study of cardiovascular disease and adverse drug reactions. Although the development of analytical techniques such as machine learning has improved our ability to extract useful information from ECGs, there is a lack of easily available ECG data for research purposes. We previously published an article on a database of ECG parameters and related clinical data (ECG-ViEW), which we have now updated with additional 12-lead waveform information. METHODS: All ECGs stored in portable document format (PDF) were collected from a tertiary teaching hospital in Korea over a 23-year study period. We developed software which can extract all ECG parameters and waveform information from the ECG reports in PDF format and stored it in a database (meta data) and a text file (raw waveform). RESULTS: Our database includes all parameters (ventricular rate, PR interval, QRS duration, QT/QTc interval, P-R-T axes, and interpretations) and 12-lead waveforms (for leads I, II, III, aVR, aVL, aVF, V1, V2, V3, V4, V5, and V6) from 1,039,550 ECGs (from 447,445 patients). Demographics, drug exposure data, diagnosis history, and laboratory test results (serum calcium, magnesium, and potassium levels) were also extracted from electronic medical records and linked to the ECG information. CONCLUSIONS: Electrocardiogram information that includes 12 lead waveforms was extracted and transformed into a form that can be analyzed. The description and programming codes in this case report could be a reference for other researchers to build ECG databases using their own local ECG repository.

Quantitative Evaluation of the Relationship between T-Wave-Based Features and Serum Potassium Level in Real-World Clinical Practice.

BACKGROUND: Proper management of hyperkalemia that leads to fatal cardiac arrhythmia has become more important because of the increased prevalence of hyperkalemia-prone diseases. Although T-wave changes in hyperkalemia are well known, their usefulness is debatable. We evaluated how well T-wave-based features of electrocardiograms (ECGs) are correlated with estimated serum potassium levels using ECG data from real-world clinical practice. METHODS: We collected ECGs from a local ECG repository (MUSE™) from 1994 to 2017 and extracted the ECG waveforms. Of about 1 million reports, 124,238 were conducted within 5 minutes before or after blood collection for serum potassium estimation. We randomly selected 500 ECGs and two evaluators measured the amplitude (T-amp) and right slope of the T-wave (T-right slope) on five lead waveforms (V3, V4, V5, V6, and II). Linear correlations of T-amp, T-right slope, and their normalized feature (T-norm) with serum potassium levels were evaluated using Pearson correlation coefficient analysis. RESULTS: Pearson correlation coefficients for T-wave-based features with serum potassium between the two evaluators were 0.99 for T-amp and 0.97 for T-right slope. The coefficient for the association between T-amp, T-right slope, and T-norm, and serum potassium ranged from -0.22 to 0.02. In the normal ECG subgroup (normal ECG or otherwise normal ECG), there was no correlation between T-wave-based features and serum potassium level. CONCLUSIONS: T-wave-based features were not correlated with serum potassium level, and their use in real clinical practice is currently limited.