The effect of confounding data features on a deep learning algorithm to predict complete coronary occlusion in a retrospective observational setting

Machine learning based hybrid anomaly detection technique for automatic diagnosis of cardiovascular diseases using cardiac sympathetic nerve activity and electrocardiogram

OBJECTIVES

Coronary artery diseases (CADs) are the leading cause of death worldwide and early diagnosis is crucial for timely treatment. To address this, our study presents a novel automated Artificial Intelligence (AI)-based Hybrid Anomaly Detection (AIHAD) technique that combines various signal processing, feature extraction, supervised, and unsupervised machine learning methods. By jointly and simultaneously analyzing 12-lead cardiac sympathetic nerve activity (CSNA) and electrocardiogram (ECG) data, the automated AIHAD technique performs fast, early, and accurate diagnosis of CADs.

METHODS

In order to develop and evaluate the proposed automated AIHAD technique, we utilized the fully labeled STAFF III and PTBD databases, which contain the 12-lead wideband raw recordings non-invasively acquired from 260 subjects. Using these wideband raw recordings, we developed a signal processing technique that simultaneously detects the 12-lead CSNA and ECG signals of all subjects. Using the pre-processed 12-lead CSNA and ECG signals, we developed a time-domain feature extraction technique that extracts the statistical CSNA and ECG features critical for the reliable diagnosis of CADs. Using the extracted discriminative features, we developed a supervised classification technique based on Artificial Neural Networks (ANNs) that simultaneously detects anomalies in the 12-lead CSNA and ECG data. Furthermore, we developed an unsupervised clustering technique based on Gaussian mixture models (GMMs) and Neyman-Pearson criterion, which robustly detects outliers corresponding to CADs.

RESULTS

Using the automated AIHAD technique, we have, for the first time, demonstrated a significant association between the increase in CSNA signals and anomalies in ECG signals during CADs. The AIHAD technique achieved highly reliable detection of CADs with a sensitivity of 98.48 %, specificity of 97.73 %, accuracy of 98.11 %, positive predictive value of 97.74 %, negative predictive value of 98.47 %, and F1-score of 98.11 %. Hence, the automated AIHAD technique demonstrates superior performance compared to the gold standard diagnostic test ECG in the diagnosis of CADs. Additionally, it outperforms other techniques developed in this study that separately utilize either only CSNA data or only ECG data. Therefore, it significantly increases the detection performance of CADs by taking advantage of the diversity in different data types and leveraging their strengths. Furthermore, its performance is comparatively better than that of most previously proposed machine and deep learning methods that exclusively used ECG data to diagnose or classify CADs. Additionally, it has a very low implementation time, which is highly desirable for real-time detection of CADs.

CONCLUSIONS

The proposed automated AIHAD technique may serve as an efficient decision-support system to increase physicians' success in fast, early, and accurate diagnosis of CADs. It may be highly beneficial and valuable, particularly for asymptomatic patients, for whom the diagnostic information provided by ECG alone is not sufficient to reliably diagnose the disease. Hence, it may significantly improve patient outcomes by enabling timely treatments and considerably reducing the mortality of cardiovascular diseases (CVDs).

Physicians and Machine-Learning Algorithm Performance in Predicting Left-Ventricular Systolic Dysfunction from a Standard 12-Lead-Electrocardiogram

Early detection of left ventricular systolic dysfunction (LVSD) may prompt early care and improve outcomes for asymptomatic patients. Standard 12-lead ECG may be used to predict LVSD. We aimed to compare the performance of Machine Learning Algorithms (MLA) and physicians in predicting LVSD from a standard 12-lead ECG. By utilizing a dataset of 13,820 pairs of ECGs and echocardiography, a deep residual convolutional neural network was trained for predicting LVSD (ejection fraction (EF) < 50%) from ECG. The ECGs of the test set (n = 850) were assessed for LVSD by the MLA and six physicians. The performance was compared using sensitivity, specificity, and C-statistics. The interobserver agreement between the physicians for the prediction of LVSD was moderate (κ = 0.50), with average sensitivity and specificity of 70%. The C-statistic of the MLA was 0.85. Repeating this analysis with LVSD defined as EF < 35% resulted in an improvement in physicians’ average sensitivity to 84% but their specificity decreased to 57%. The MLA C-statistic was 0.88 with this threshold. We conclude that although MLA outperformed physicians in predicting LVSD from standard ECG, prior to robust implementation of MLA in ECG machines, physicians should be encouraged to use this approach as a simple and readily available aid for LVSD screening.

Clinical validation of a 13-lead electrocardiogram derived from a self-applicable 3-lead recording for diagnosis of myocardial supply ischaemia and common non-ischaemic electrocardiogram abnormalities at rest

Aims

In this study, we compare the diagnostic accuracy of a standard 12-lead electrocardiogram (ECG) with a novel 13-lead ECG derived from a self-applicable 3-lead ECG recorded with the right exploratory left foot (RELF) device. The 13th lead is a novel age and sex orthonormalized computed ST (ASO-ST) lead to increase the sensitivity for detecting ischaemia during acute coronary artery occlusion.

Methods and results

A database of simultaneously recorded 12-lead ECGs and RELF recordings from 110 patients undergoing coronary angioplasty and 30 healthy subjects was used. Five cardiologists scored the learning data set and five other cardiologists scored the validation data set. In addition, the presence of non-ischaemic ECG abnormalities was compared. The accuracy for detection of myocardial supply ischaemia with the derived 12 leads was comparable with that of the standard 12-lead ECG (P = 0.126). By adding the ASO-ST lead, the accuracy increased to 77.4% [95% confidence interval (CI): 72.4-82.3; P < 0.001], which was attributed to a higher sensitivity of 81.9% (95% CI: 74.8-89.1) for the RELF 13-lead ECG compared with a sensitivity of 76.8% (95% CI: 71.9-81.7; P < 0.001) for the 12-lead ECG. There was no significant difference in the diagnosis of non-ischaemic ECG abnormalities, except for Q-waves that were more frequently detected on the standard ECG compared with the derived ECG (25.9 vs. 13.8%; P < 0.001).

Conclusion

A self-applicable and easy-to-use 3-lead RELF device can compute a 12-lead ECG plus an ischaemia-specific 13th lead that is, compared with the standard 12-lead ECG, more accurate for the visual diagnosis of myocardial supply ischaemia by cardiologists.

Heart age estimated using explainable advanced electrocardiography

Electrocardiographic (ECG) Heart Age conveying cardiovascular risk has been estimated by both Bayesian and artificial intelligence approaches. We hypothesised that explainable measures from the 10-s 12-lead ECG could successfully predict Bayesian 5-min ECG Heart Age. Advanced analysis was performed on ECGs from healthy subjects and patients with cardiovascular risk or proven heart disease. Regression models were used to predict patients’ Bayesian 5-min ECG Heart Ages from their standard, resting 10-s 12-lead ECGs. The difference between 5-min and 10-s ECG Heart Ages were analyzed, as were the differences between 10-s ECG Heart Age and the chronological age (the Heart Age Gap). In total, 2,771 subjects were included (n = 1682 healthy volunteers, n = 305 with cardiovascular risk factors, n = 784 with cardiovascular disease). Overall, 10-s Heart Age showed strong agreement with the 5-min Heart Age (R² = 0.94, p < 0.001, mean ± SD bias 0.0 ± 5.1 years). The Heart Age Gap was 0.0 ± 5.7 years in healthy individuals, 7.4 ± 7.3 years in subjects with cardiovascular risk factors (p < 0.001), and 14.3 ± 9.2 years in patients with cardiovascular disease (p < 0.001). Heart Age can be accurately estimated from a 10-s 12-lead ECG in a transparent and explainable fashion based on known ECG measures, without deep neural network-type artificial intelligence techniques. The Heart Age Gap increases markedly with cardiovascular risk and disease.

Prospects and Pitfalls of Machine Learning in Nutritional Epidemiology

Nutritional epidemiology employs observational data to discover associations between diet and disease risk. However, existing analytic methods of dietary data are often sub-optimal, with limited incorporation and analysis of the correlations between the studied variables and nonlinear behaviours in the data. Machine learning (ML) is an area of artificial intelligence that has the potential to improve modelling of nonlinear associations and confounding which are found in nutritional data. These opportunities notwithstanding, the applications of ML in nutritional epidemiology must be approached cautiously to safeguard the scientific quality of the results and provide accurate interpretations. Given the complex scenario around ML, judicious application of such tools is necessary to offer nutritional epidemiology a novel analytical resource for dietary measurement and assessment and a tool to model the complexity of dietary intake and its relation to health. This work describes the applications of ML in nutritional epidemiology and provides guidelines to avoid common pitfalls encountered in applying predictive statistical models to nutritional data. Furthermore, it helps unfamiliar readers better assess the significance of their results and provides new possible future directions in the field of ML in nutritional epidemiology.

Machine Learning for Clinical Decision-Making: Challenges and Opportunities in Cardiovascular Imaging

The use of machine learning (ML) approaches to target clinical problems is called to revolutionize clinical decision-making in cardiology. The success of these tools is dependent on the understanding of the intrinsic processes being used during the conventional pathway by which clinicians make decisions. In a parallelism with this pathway, ML can have an impact at four levels: for data acquisition, predominantly by extracting standardized, high-quality information with the smallest possible learning curve; for feature extraction, by discharging healthcare practitioners from performing tedious measurements on raw data; for interpretation, by digesting complex, heterogeneous data in order to augment the understanding of the patient status; and for decision support, by leveraging the previous steps to predict clinical outcomes, response to treatment or to recommend a specific intervention. This paper discusses the state-of-the-art, as well as the current clinical status and challenges associated with the two later tasks of interpretation and decision support, together with the challenges related to the learning process, the auditability/traceability, the system infrastructure and the integration within clinical processes in cardiovascular imaging.

Machine learning with electrocardiograms: A call for guidelines and best practices for 'stress testing' algorithms

This paper provides a brief description of how computer programs are used to automatically interpret electrocardiograms (ECGs), and also provides a discussion regarding new opportunities. The algorithms that are typically used today in hospitals are knowledge engineered where a computer programmer manually writes computer code and logical statements which are then used to deduce a possible diagnosis. The computer programmer's code represents the criteria and knowledge that is used by clinicians when reading ECGs. This is in contrast to supervised machine learning (ML) approaches which use large, labelled ECG datasets to induct their own 'rules' to automatically classify ECGs. Although there are many ML techniques, deep neural networks are being increasingly explored as ECG classification algorithms when trained on large ECG datasets. Whilst this paper presents some of the pros and cons of each of these approaches, perhaps there are opportunities to develop hybridised algorithms that combine both knowledge and data driven techniques. In this paper, it is pointed out that open ECG data can dramatically influence what international ECG ML researchers focus on and that, ideally, open datasets could align with real world clinical challenges. In addition, some of the pitfalls and opportunities for ML with ECGs are outlined. A potential opportunity for the ECG community is to provide guidelines to researchers to help guide ECG ML practices. For example, whilst general ML guidelines exist, there is perhaps a need to recommend approaches for 'stress testing' and evaluating ML algorithms for ECG analysis, e.g. testing the algorithm with noisy ECGs and ECGs acquired using common lead and electrode misplacements. This paper provides a primer on ECG ML and discusses some of the key challenges and opportunities.