Clinical significance, challenges and limitations in using artificial intelligence for electrocardiography-based diagnosis

Advancements in Artificial Intelligence for Precision Diagnosis and Treatment of Myocardial Infarction: A Comprehensive Review of Clinical Trials and Randomized Controlled Trials

Coronary artery disease (CAD) is still a serious global health issue that has a substantial impact on death and illness rates. The goal of primary prevention strategies is to lower the risk of developing CAD. Nevertheless, current methods usually rely on simple risk assessment instruments that might overlook significant individual risk factors. This limitation highlights the need for innovative methods that can accurately assess cardiovascular risk and offer personalized preventive care. Recent advances in machine learning and artificial intelligence (AI) have opened up interesting new avenues for optimizing primary preventive efforts for CAD and improving risk prediction models. By leveraging large-scale databases and advanced computational techniques, AI has the potential to fundamentally alter how cardiovascular risk is evaluated and managed. This review looks at current randomized controlled studies and clinical trials that explore the application of AI and machine learning to improve primary preventive measures for CAD. The emphasis is on their ability to recognize and include a range of risk elements in sophisticated risk assessment models.

Biometric contrastive learning for data-efficient deep learning from electrocardiographic images

OBJECTIVE

Artificial intelligence (AI) detects heart disease from images of electrocardiograms (ECGs). However, traditional supervised learning is limited by the need for large amounts of labeled data. We report the development of Biometric Contrastive Learning (BCL), a self-supervised pretraining approach for label-efficient deep learning on ECG images.

MATERIALS AND METHODS

Using pairs of ECGs from 78 288 individuals from Yale (2000-2015), we trained a convolutional neural network to identify temporally separated ECG pairs that varied in layouts from the same patient. We fine-tuned BCL-pretrained models to detect atrial fibrillation (AF), gender, and LVEF < 40%, using ECGs from 2015 to 2021. We externally tested the models in cohorts from Germany and the United States. We compared BCL with ImageNet initialization and general-purpose self-supervised contrastive learning for images (simCLR).

RESULTS

While with 100% labeled training data, BCL performed similarly to other approaches for detecting AF/Gender/LVEF < 40% with an AUROC of 0.98/0.90/0.90 in the held-out test sets, it consistently outperformed other methods with smaller proportions of labeled data, reaching equivalent performance at 50% of data. With 0.1% data, BCL achieved AUROC of 0.88/0.79/0.75, compared with 0.51/0.52/0.60 (ImageNet) and 0.61/0.53/0.49 (simCLR). In external validation, BCL outperformed other methods even at 100% labeled training data, with an AUROC of 0.88/0.88 for Gender and LVEF < 40% compared with 0.83/0.83 (ImageNet) and 0.84/0.83 (simCLR).

DISCUSSION AND CONCLUSION

A pretraining strategy that leverages biometric signatures of different ECGs from the same patient enhances the efficiency of developing AI models for ECG images. This represents a major advance in detecting disorders from ECG images with limited labeled data.

Artificial intelligence-enhanced electrocardiography for accurate diagnosis and management of cardiovascular diseases

Electrocardiography (ECG), improved by artificial intelligence (AI), has become a potential technique for the precise diagnosis and treatment of cardiovascular disorders. The conventional ECG is a frequently used, inexpensive, and easily accessible test that offers important information about the physiological and anatomical state of the heart. However, the ECG can be interpreted differently by humans depending on the interpreter's level of training and experience, which could make diagnosis more difficult. Using AI, especially deep learning convolutional neural networks (CNNs), to look at single, continuous, and intermittent ECG leads that has led to fully automated AI models that can interpret the ECG like a human, possibly more accurately and consistently. These AI algorithms are effective non-invasive biomarkers for cardiovascular illnesses because they can identify subtle patterns and signals in the ECG that may not be readily apparent to human interpreters. The use of AI in ECG analysis has several benefits, including the quick and precise detection of problems like arrhythmias, silent cardiac illnesses, and left ventricular failure. It has the potential to help doctors with interpretation, diagnosis, risk assessment, and illness management. Aside from that, AI-enhanced ECGs have been demonstrated to boost the identification of heart failure and other cardiovascular disorders, particularly in emergency department settings, allowing for quicker and more precise treatment options. The use of AI in cardiology, however, has several limitations and obstacles, despite its potential. The effective implementation of AI-powered ECG analysis is limited by issues such as systematic bias. Biases based on age, gender, and race result from unbalanced datasets. A model's performance is impacted when diverse demographics are inadequately represented. Potentially disregarded age-related ECG variations may result from skewed age data in training sets. ECG patterns are affected by physiological differences between the sexes; a dataset that is inclined toward one sex may compromise the accuracy of the others. Genetic variations influence ECG readings, so racial diversity in datasets is significant. Furthermore, issues such as inadequate generalization, regulatory barriers, and interpretability concerns contribute to deployment difficulties. The lack of robustness in models when applied to disparate populations frequently hinders their practical applicability. The exhaustive validation required by regulatory requirements causes a delay in deployment. Difficult models that are not interpretable erode the confidence of clinicians. Diverse dataset curation, bias mitigation strategies, continuous validation across populations, and collaborative efforts for regulatory approval are essential for the successful deployment of AI ECG in clinical settings and must be undertaken to address these issues. To guarantee a safe and successful deployment in clinical practice, the use of AI in cardiology must be done with a thorough understanding of the algorithms and their limits. In summary, AI-enhanced electrocardiography has enormous potential to improve the management of cardiovascular illness by delivering precise and timely diagnostic insights, aiding clinicians, and enhancing patient outcomes. Further study and development are required to fully realize AI's promise for improving cardiology practices and patient care as technology continues to advance.

Advancements in Pancreatic Cancer Detection: Integrating Biomarkers, Imaging Technologies, and Machine Learning for Early Diagnosis

Artificial intelligence (AI) has come to play a pivotal role in revolutionizing medical practices, particularly in the field of pancreatic cancer detection and management. As a leading cause of cancer-related deaths, pancreatic cancer warrants innovative approaches due to its typically advanced stage at diagnosis and dismal survival rates. Present detection methods, constrained by limitations in accuracy and efficiency, underscore the necessity for novel solutions. AI-driven methodologies present promising avenues for enhancing early detection and prognosis forecasting. Through the analysis of imaging data, biomarker profiles, and clinical information, AI algorithms excel in discerning subtle abnormalities indicative of pancreatic cancer with remarkable precision. Moreover, machine learning (ML) algorithms facilitate the amalgamation of diverse data sources to optimize patient care. However, despite its huge potential, the implementation of AI in pancreatic cancer detection faces various challenges. Issues such as the scarcity of comprehensive datasets, biases in algorithm development, and concerns regarding data privacy and security necessitate thorough scrutiny. While AI offers immense promise in transforming pancreatic cancer detection and management, ongoing research and collaborative efforts are indispensable in overcoming technical hurdles and ethical dilemmas. This review delves into the evolution of AI, its application in pancreatic cancer detection, and the challenges and ethical considerations inherent in its integration.

Artificial intelligence-based identification of left ventricular systolic dysfunction from 12-lead electrocardiograms: external validation and advanced application of an existing model

Aims

The diagnostic application of artificial intelligence (AI)-based models to detect cardiovascular diseases from electrocardiograms (ECGs) evolves, and promising results were reported. However, external validation is not available for all published algorithms. The aim of this study was to validate an existing algorithm for the detection of left ventricular systolic dysfunction (LVSD) from 12-lead ECGs.

Methods and results

Patients with digitalized data pairs of 12-lead ECGs and echocardiography (at intervals of ≤7 days) were retrospectively selected from the Heart Center Leipzig ECG and electronic medical records databases. A previously developed AI-based model was applied to ECGs and calculated probabilities for LVSD. The area under the receiver operating characteristic curve (AUROC) was computed overall and in cohorts stratified for baseline and ECG characteristics. Repeated echocardiography studies recorded ≥3 months after index diagnostics were used for follow-up (FU) analysis. At baseline, 42 291 ECG-echocardiography pairs were analysed, and AUROC for LVSD detection was 0.88. Sensitivity and specificity were 82% and 77% for the optimal LVSD probability cut-off based on Youden's J. AUROCs were lower in ECG subgroups with tachycardia, atrial fibrillation, and wide QRS complexes. In patients without LVSD at baseline and available FU, model-generated high probability for LVSD was associated with a four-fold increased risk of developing LVSD during FU.

Conclusion

We provide the external validation of an existing AI-based ECG-analysing model for the detection of LVSD with robust performance metrics. The association of false positive LVSD screenings at baseline with a deterioration of ventricular function during FU deserves a further evaluation in prospective trials.

Multichannel high noise level ECG denoising based on adversarial deep learning

This paper proposes a denoising method based on an adversarial deep learning approach for the post-processing of multi-channel fetal electrocardiogram (ECG) signals. As it's well known, noise leads to misinterpretations of fetal ECG signals and thus limits the use of fetal electrocardiography for healthcare applications. Therefore, denoising algorithms are essential for the exploitation of non-invasive fetal ECG. The proposed method is based on the combination of three end-to-end trained sub-networks to convert noisy fetal ECG signals into clean signals. The first two sub-networks are linked by skip connections and form a deep convolutional network that downsamples the noisy signals into a latent representation and subsequently upsamples this latent representation to recover clean signals. The third sub-network aims to boost the decoder sub-network to generate realistic clean signals. Experiments carried out on synthetic and real data showed that the proposed method improved by the signal-to-noise (SNR) of fetal ECG signals with input SNR ranging from [Formula: see text] to 0 dB by an average of 20 dB, and improve fetal signal quality by significantly increasing the number of true detected QRS complexes and halving QRS complex detection errors.

Discrete Wavelet Transform based ECG classification using gcForest: A deep ensemble method

BACKGROUND

Cardiovascular diseases (CVDs) are the leading global cause of mortality, necessitating advanced diagnostic tools for early detection. The electrocardiogram (ECG) is pivotal in diagnosing cardiac abnormalities due to its non-invasive nature.

OBJECTIVE

This study aims to propose a novel approach for ECG signal classification, addressing the challenges posed by the complexity of ECG signals associated with various diseases.

METHODS

Our method integrates Discrete Wavelet Transform (DWT) for feature extraction, capturing salient features of cardiovascular diseases. Subsequently, the gcForest model is employed for efficient classification. The approach is tested on the MIT-BIH Arrhythmia Database.

RESULTS

The proposed method demonstrates promising results on the MIT-BIH Arrhythmia Database, achieving a test accuracy of 98.55%, recall of 98.48%, precision of 98.44%, and an F1 score of 98.46%. Additionally, the model exhibits robustness and low sensitivity to hyper-parameters.

CONCLUSION

The combined use of DWT and the gcForest model proves effective in ECG signal classification, showcasing high accuracy and reliability. This approach holds potential for improving early detection of cardiovascular diseases, contributing to enhanced cardiac healthcare.

Healthcare Big Data in Hong Kong: Development and Implementation of Artificial Intelligence-Enhanced Predictive Models for Risk Stratification

Routinely collected electronic health records (EHRs) data contain a vast amount of valuable information for conducting epidemiological studies. With the right tools, we can gain insights into disease processes and development, identify the best treatment and develop accurate models for predicting outcomes. Our recent systematic review has found that the number of big data studies from Hong Kong has rapidly increased since 2015, with an increasingly common application of artificial intelligence (AI). The advantages of big data are that i) the models developed are highly generalisable to the population, ii) multiple outcomes can be determined simultaneously, iii) ease of cross-validation by for model training, development and calibration, iv) huge numbers of useful variables can be analyzed, v) static and dynamic variables can be analyzed, vi) non-linear and latent interactions between variables can be captured, vii) artificial intelligence approaches can enhance the performance of prediction models. In this paper, we will provide several examples (cardiovascular disease, diabetes mellitus, Brugada syndrome, long QT syndrome) to illustrate efforts from a multi-disciplinary team to identify data from different modalities to develop models using territory-wide datasets, with the possibility of real-time risk updates by using new data captured from patients. The benefit is that only routinely collected data are required for developing highly accurate and high-performance models. AI-driven models outperform traditional models in terms of sensitivity, specificity, accuracy, area under the receiver operating characteristic and precision-recall curve, and F1 score. Web and/or mobile versions of the risk models allow clinicians to risk stratify patients quickly in clinical settings, thereby enabling clinical decision-making. Efforts are required to identify the best ways of implementing AI algorithms on the web and mobile apps.

An Open-Source Graphical User Interface-Embedded Automated Electrocardiogram Quality Assessment: A Balanced Class Representation Approach

The rise in cardiovascular diseases necessitates accurate electrocardiogram (ECG) diagnostics, making high-quality ECG recordings essential. Our CNN-LSTM model, embedded in an open-access GUI and trained on balanced datasets collected in clinical settings, excels in automating ECG quality assessment. When tested across three datasets featuring varying ratios of acceptable to unacceptable ECG signals, it achieved an F1 score ranging from 95.87% to 98.40%. Training the model on real noise sources significantly enhances its applicability in real-life scenarios, compared to simulations. Integrated into a user-friendly toolbox, the model offers practical utility in clinical environments. Furthermore, our study underscores the importance of balanced class representation during training and testing phases. We observed a notable F1 score change from 98.09% to 95.87% when the class ratio shifted from 85:15 to 50:50 in the same testing dataset with equal representation. This finding is crucial for future ECG quality assessment research, highlighting the impact of class distribution on the reliability of model training outcomes.

Predictive factors for type A aortic dissection mortality based on electrocardiogram parameters and clinical presentations

BACKGROUND

Aortic dissection is a rare but potentially lethal disorder and may be associated with electrocardiogram (ECG) changes. In this study, we aim to investigate ECG-related parameters alongside clinical presentations of type A aortic dissection to come up with the predictive factors for the severity of the disease and its mortality rate.

METHODS

In this retrospective study, 201 patients with type A aortic dissection were studied between March 2015 and March 2020. Two expert cardiologists blinded to the diagnosis studied former and new patients' ECGs and recorded changes.

RESULTS

Two-hundred and one patients, including 143 (71.1%) men and 58 (28.9%) women, presented with acute dissection of the aorta, were studied. Forty-four (21.8%) and 84 (41.7%) patients had ST-segment elevation and depression in ECG, respectively. Bivariate analysis revealed that higher heart rate (p = 0.006), longer QTc (p = 0.044), and ST-segment elevation in aVR lead (p = 0.044) were associated with mortality in the patients. Multivariate regression showed higher heart rate (OR = 1.022, CI = 1.003-1.041, p = 0.012) and ST-segment elevation in aVR (OR = 4.854, CI = 2.255-10.477, p < 0.001) were independently associated with increased odds of mortality in aortic dissection patients. ROC curve analysis showed heart rate equal to or >60 per minute (AUC = 0.625, sensitivity = 86%, specificity = 10%, p = 0.019) and ST-segment elevation in aVR >0.5 mm (AUC = 0.854, sensitivity = 75%, specificity = 92%, p < 0.001) were associated with a higher mortality rate.

CONCLUSION

Heart rate equal or >60 and ST-segment elevation >0.5 mm in aVR lead can be used as predictive factors for mortality of patients with type A aortic dissection.

Application of artificial intelligence in the diagnosis of sleep apnea

STUDY OBJECTIVES

Machine learning (ML) models have been employed in the setting of sleep disorders. This review aims to summarize the existing data about the role of ML techniques in the diagnosis, classification, and treatment of sleep-related breathing disorders.

METHODS

A systematic search in Medline, EMBASE, and Cochrane databases through January 2022 was performed.

RESULTS

Our search strategy revealed 132 studies that were included in the systematic review. Existing data show that ML models have been successfully used for diagnostic purposes. Specifically, ML models showed good performance in diagnosing sleep apnea using easily obtained features from the electrocardiogram, pulse oximetry, and sound signals. Similarly, ML showed good performance for the classification of sleep apnea into obstructive and central categories, as well as predicting apnea severity. Existing data show promising results for the ML-based guided treatment of sleep apnea. Specifically, the prediction of outcomes following surgical treatment and optimization of continuous positive airway pressure therapy can be guided by ML models.

CONCLUSIONS

The adoption and implementation of ML in the field of sleep-related breathing disorders is promising. Advancements in wearable sensor technology and ML models can help clinicians predict, diagnose, and classify sleep apnea more accurately and efficiently.

CITATION

Bazoukis G, Bollepalli SC, Chung CT, et al. Application of artificial intelligence in the diagnosis of sleep apnea. J Clin Sleep Med. 2023;19(7):1337-1363.

Animal Disease Models and Patient-iPS-Cell-Derived In Vitro Disease Models for Cardiovascular Biology-How Close to Disease?

Currently, zebrafish, rodents, canines, and pigs are the primary disease models used in cardiovascular research. In general, larger animals have more physiological similarities to humans, making better disease models. However, they can have restricted or limited use because they are difficult to handle and maintain. Moreover, animal welfare laws regulate the use of experimental animals. Different species have different mechanisms of disease onset. Organs in each animal species have different characteristics depending on their evolutionary history and living environment. For example, mice have higher heart rates than humans. Nonetheless, preclinical studies have used animals to evaluate the safety and efficacy of human drugs because no other complementary method exists. Hence, we need to evaluate the similarities and differences in disease mechanisms between humans and experimental animals. The translation of animal data to humans contributes to eliminating the gap between these two. In vitro disease models have been used as another alternative for human disease models since the discovery of induced pluripotent stem cells (iPSCs). Human cardiomyocytes have been generated from patient-derived iPSCs, which are genetically identical to the derived patients. Researchers have attempted to develop in vivo mimicking 3D culture systems. In this review, we explore the possible uses of animal disease models, iPSC-derived in vitro disease models, humanized animals, and the recent challenges of machine learning. The combination of these methods will make disease models more similar to human disease.