Atrial fibrillation risk prediction from the 12-lead electrocardiogram using digital biomarkers and deep representation learning

Abstract

Aims This study aims to assess whether information derived from the raw 12-lead electrocardiogram (ECG) combined with clinical information is predictive of atrial fibrillation (AF) development. Methods We use a subset of the Telehealth Network of Minas Gerais (TNMG) database consisting of patients that had repeated 12-lead ECG measurements between 2010-2017 that is 1,130,404 recordings from 415,389 unique patients. Median and interquartile of age for the recordings were 58 (46-69) and 38% of the patients were males. Recordings were assigned to train-validation and test sets in an 80:20% split which was stratified by class, age and gender. A random forest classifier was trained to predict, for a given recording, the risk of AF development within 5-years. We use features obtained from different modalities, namely demographics, clinical information, engineered features, and features from deep representation learning. Results The best model performance on the test set was obtained for the model combining features from all modalities with an AUROC=0.909 against the best single modality model which had an AUROC=0.839. Conclusion Our study has important clinical implications for AF management. It is the first study integrating feature engineering, deep learning and EMR metadata to create a risk prediction tool for the management of patients at risk of AF. The best model that includes features from all modalities demonstrates that human knowledge in electrophysiology combined with deep learning outperforms any single modality approach. The high performance obtained suggest that structural changes in the 12-lead ECG are associated with existing or impending AF.

Artificial intelligence opportunities in cardio-oncology: Overview with spotlight on electrocardiography

“…One group assessed > 1 million raw 12-lead ECGs from > 415,000 unique patients, paired with their clinical data, to predict the development of atrial fibrillation [ 107 ] ( Table 4 ). Recordings were assigned to training, validation, and test sets, stratified by class, age, and gender.…”

Machine Learning to Support Triage of Children at Risk for Epileptic Seizures in the Pediatric Intensive Care Unit

“…Three different categories of features were considered: the clinical information of the patient, also called "metadata" (META), the Heart Rate Variability (HRV) features [13] and the ECG morphological features (MOR) [15]. The 16 META features include clinical data about the patient such as standard demographics (i.e.…”

“…The detection of the fiducial points on the ECG waveform was done using the popular open source wavedet algorithm [16]. A total of 74 features [15] were extracted from these fiducial points: 38 features extracted from interval duration (Table 6) and 36 from waves characteristics (Table 7).…”

“…The third one had 101 features and included the age of the patient and all the features extracted from the ECG recording i.e. HRV features [13] and MOR features [15]. The last model included all the 116 available features (META, HRV and MOR).…”

Using Minimum Redundancy Maximum Relevance Algorithm to Select Minimal Sets of Heart Rate Variability Parameters for Atrial Fibrillation Detection

“…The accuracy in 10-fold CV (by a patient) was 97.7%. Biton et al [ 29 ] extracted the following features from a 7–10 s 12-lead ECG: deep neural network features, morphology, HRV, and electronic medical record system (EMR) metadata. A subset of features was selected using MRMR to predict AF occurring within 5 years (59.6% sensitivity, 96.3% specificity in the test set).…”