ECG data dependency for atrial fibrillation detection based on residual networks

Applying masked autoencoder-based self-supervised learning for high-capability vision transformers of electrocardiographies

The generalization of deep neural network algorithms to a broader population is an important challenge in the medical field. We aimed to apply self-supervised learning using masked autoencoders (MAEs) to improve the performance of the 12-lead electrocardiography (ECG) analysis model using limited ECG data. We pretrained Vision Transformer (ViT) models by reconstructing the masked ECG data with MAE. We fine-tuned this MAE-based ECG pretrained model on ECG-echocardiography data from The University of Tokyo Hospital (UTokyo) for the detection of left ventricular systolic dysfunction (LVSD), and then evaluated it using multi-center external validation data from seven institutions, employing the area under the receiver operating characteristic curve (AUROC) for assessment. We included 38,245 ECG-echocardiography pairs from UTokyo and 229,439 pairs from all institutions. The performances of MAE-based ECG models pretrained using ECG data from UTokyo were significantly higher than that of other Deep Neural Network models across all external validation cohorts (AUROC, 0.913-0.962 for LVSD, p < 0.001). Moreover, we also found improvements for the MAE-based ECG analysis model depending on the model capacity and the amount of training data. Additionally, the MAE-based ECG analysis model maintained high performance even on the ECG benchmark dataset (PTB-XL). Our proposed method developed high performance MAE-based ECG analysis models using limited ECG data.

Diagnostic and Prognostic Electrocardiogram-Based Models for Rapid Clinical Applications

Leveraging artificial intelligence (AI) for the analysis of electrocardiograms (ECGs) has the potential to transform diagnosis and estimate the prognosis of not only cardiac but, increasingly, noncardiac conditions. In this review, we summarize clinical studies and AI-enhanced ECG-based clinical applications in the early detection, diagnosis, and estimating prognosis of cardiovascular diseases in the past 5 years (2019-2023). With advancements in deep learning and the rapid increased use of ECG technologies, a large number of clinical studies have been published. However, most of these studies are single-centre, retrospective, proof-of-concept studies that lack external validation. Prospective studies that progress from development toward deployment in clinical settings account for < 15% of the studies. Successful implementations of ECG-based AI applications that have received approval from the Food and Drug Administration have been developed through commercial collaborations, with approximately half of them being for mobile or wearable devices. The field is in its early stages, and overcoming several obstacles is essential, such as prospective validation in multicentre large data sets, addressing technical issues, bias, privacy, data security, model generalizability, and global scalability. This review concludes with a discussion of these challenges and potential solutions. By providing a holistic view of the state of AI in ECG analysis, this review aims to set a foundation for future research directions, emphasizing the need for comprehensive, clinically integrated, and globally deployable AI solutions in cardiovascular disease management.

A microfluidic model to study the effects of arrhythmic flows on endothelial cells

Atrial fibrillation (AF) is the most common type of cardiac arrhythmia and an important contributor to morbidity and mortality. Endothelial dysfunction has been postulated to be an important contributing factor in cardiovascular events in patients with AF. However, how vascular endothelial cells respond to arrhythmic flow is not fully understood, mainly due to the limitation of current in vitro systems to mimic arrhythmic flow conditions. To address this limitation, we developed a microfluidic system to study the effect of arrhythmic flow on the mechanobiology of human aortic endothelial cells (HAECs). The system utilises a computer-controlled piezoelectric pump for generating arrhythmic flow with a unique ability to control the variability in both the frequency and amplitude of pulse waves. The flow rate is modulated to reflect physiological or pathophysiological shear stress levels on endothelial cells. This enabled us to systematically dissect the importance of variability in the frequency and amplitude of pulses and shear stress level on endothelial cell mechanobiology. Our results indicated that arrhythmic flow at physiological shear stress level promotes endothelial cell spreading and reduces the plasma membrane-to-cytoplasmic distribution of β-catenin. In contrast, arrhythmic flow at low and atherogenic shear stress levels does not promote endothelial cell spreading or redistribution of β-catenin. Interestingly, under both shear stress levels, arrhythmic flow induces inflammation by promoting monocyte adhesion via an increase in ICAM-1 expression. Collectively, our microfluidic system provides opportunities to study the effect of arrhythmic flows on vascular endothelial mechanobiology in a systematic and reproducible manner.

Trends in Drug Spending of Oral Anticoagulants for Atrial Fibrillation, 2014-2021

INTRODUCTION

This study documents cost trends in oral anticoagulants (OAC) in patients with newly diagnosed atrial fibrillation.

METHODS

Using MarketScan databases, the mean annual patients' out-of-pocket costs, insurance payments, and the proportion of patients initiating OAC within 90 days from atrial fibrillation diagnosis were calculated from July 2014 to June 2021. Costs of OACs (apixaban, dabigatran, edoxaban, rivaroxaban, and warfarin) and the payments by three insurance types (commercial payers, Medicare, and Medicaid) were calculated. Patients' out-of-pocket costs and insurance payments were adjusted to 2021 prices. Joinpoint regression models were used to test trends of outcomes and average annual percent changes (AAPC) were reported. Data analyses were performed in 2022-2023.

RESULTS

From July 2014 to June 2021, the mean annual out-of-pocket costs of any OAC increased for commercial insurance (AAPC 3.0%) and Medicare (AAPC 5.1%) but decreased for Medicaid (AAPC -3.3%). The mean annual insurance payments for any OAC significantly increased for all insurance groups (AAPC 13.1% [95% CI 11.3-15.0] for Medicare; AAPC 11.8% [95% CI 8.0-15.6] for commercial insurance; and AAPC 16.3% [95% CI 11.3-21.4] for Medicaid). The initiation of any OAC increased (AAPC 7.3% for commercial insurance; AAPC 10.2% for Medicare; AAPC 5.3% for Medicaid).

CONCLUSIONS

There was a substantial increase in the overall cost burden of OACs and OAC initiation rates in patients with newly diagnosed atrial fibrillation in 2014-2021; these findings provide insights into the current and anticipated impact of rising drug prices on patients' and payers' financial burden.

Machine Learning for Detecting Atrial Fibrillation from ECGs: Systematic Review and Meta-Analysis

Background

Atrial fibrillation (AF) is a common arrhythmia that can result in adverse cardiovascular outcomes but is often difficult to detect. The use of machine learning (ML) algorithms for detecting AF has become increasingly prevalent in recent years. This study aims to systematically evaluate and summarize the overall diagnostic accuracy of the ML algorithms in detecting AF in electrocardiogram (ECG) signals.

Methods

The searched databases included PubMed, Web of Science, Embase, and Google Scholar. The selected studies were subjected to a meta-analysis of diagnostic accuracy to synthesize the sensitivity and specificity.

Results

A total of 14 studies were included, and the forest plot of the meta-analysis showed that the pooled sensitivity and specificity were 97% (95% confidence interval [CI]: 0.94-0.99) and 97% (95% CI: 0.95-0.99), respectively. Compared to traditional machine learning (TML) algorithms (sensitivity: 91.5%), deep learning (DL) algorithms (sensitivity: 98.1%) showed superior performance. Using multiple datasets and public datasets alone or in combination demonstrated slightly better performance than using a single dataset and proprietary datasets.

Conclusions

ML algorithms are effective for detecting AF from ECGs. DL algorithms, particularly those based on convolutional neural networks (CNN), demonstrate superior performance in AF detection compared to TML algorithms. The integration of ML algorithms can help wearable devices diagnose AF earlier.

Automatic segmentation of atrial fibrillation and flutter in single-lead electrocardiograms by self-supervised learning and Transformer architecture

OBJECTIVES

Automatic detection of atrial fibrillation and flutter (AF/AFL) is a significant concern in preventing stroke and mitigating hemodynamic instability. Herein, we developed a Transformer-based deep learning model for AF/AFL segmentation in single-lead electrocardiograms (ECGs) by self-supervised learning with masked signal modeling (MSM).

MATERIALS AND METHODS

We retrieved data from 11 open-source databases on PhysioNet; 7 of these databases included labeled ECGs, while the other 4 were without labels. Each database contained ECG recordings with durations of ≥30 s. A total of 24 intradialytic ECGs with paroxysmal AF/AFL during 4 h of hemodialysis sessions at Seoul National University Hospital were used for external validation. The model was pretrained by predicting masked areas of ECG signals and fine-tuned by predicting AF/AFL areas. Cross-database validation was used for evaluation, and the intersection over union (IOU) was used as a main performance metric in external database validation.

RESULTS

In the 7 labeled databases, the areas marked as AF/AFL constituted 41.1% of the total ECG signals, ranging from 0.19% to 51.31%. In the evaluation per ECG segment, the model achieved IOU values of 0.9254 and 0.9477 for AF/AFL segmentation and other segmentation tasks, respectively. When applied to intradialytic ECGs with paroxysmal AF/AFL, the IOUs for the segmentation of AF/AFL and non-AF/AFL were 0.9896 and 0.9650, respectively. Model performance by different training procedure indicated that pretraining with MSM and the application of an appropriate masking ratio both contributed to the model performance. It also showed higher IOUs of AF/AFL labels than in previous studies when training and test databases were matched.

CONCLUSION

The present model with self-supervised learning by MSM performs robustly in segmenting AF/AFL.

The combination of CD138/MUM1 dual-staining and artificial intelligence for plasma cell counting in the diagnosis of chronic endometritis

OBJECTIVE

To investigate the utility of combination of CD138/MUM1 dual-staining (DS) and artificial intelligence (AI) for plasma cell (PC) counting in the diagnosis of chronic endometritis (CE).

METHODS

Two hundred ninety-eight infertile women underwent endometrial biopsy were included. In 100 women, three successive sections were cut from each paraffin-embedded tissue block for CD138 immunohistochemical (IHC) single-staining (SS), MUM1 SS and CD138/MUM1 DS. The prevalence of CE and the sensitivity/specificity in the diagnosis of CE with different methods was studied. These sections diagnosed as CE with DS were collected to train artificial intelligence (AI) diagnostic system. In other 198 women, their tissue sections stained with CD138/MUM1 DS were used to test the AI system in the diagnosis of CE.

RESULTS

CD138/MUM1 DS revealed that the cell membranes and nuclei of PCs were simultaneously labelled by CD138 and MUM1, respectively. The positive rate of ECs identified by CD138/MUM1 DS (38%, 38/100) was lower than CD138 SS (52%, 52/100) and MUM1 SS (62%, 62/100) (p < .05). The sensitivity, specificity and accuracy of CD138/MUM1 DS in the diagnosis of ECs reached 100%. The sensitivity, specificity and accuracy rates of AI diagnostic system of ECs were 100%, 83.3% and 91.4%, respectively. The 17 cases over-diagnosed as EC with the AI were corrected quickly by pathologists reviewing these false PC pictures listed by the AI.

CONCLUSION

The combination of CD138/MUM1 DS and AI is a promising method to improve the accuracy and efficiency of CE diagnosis.

Wearable Devices Combined with Artificial Intelligence-A Future Technology for Atrial Fibrillation Detection?

Atrial fibrillation (AF) is the most common cardiac arrhythmia in the world. The arrhythmia and methods developed to cure it have been studied for several decades. However, professionals worldwide are still working to improve treatment quality. One novel technology that can be useful is a wearable device. The two most used recordings from these devices are photoplethysmogram (PPG) and electrocardiogram (ECG) signals. As the price lowers, these devices will become significant technology to increase sensitivity, for monitoring and for treatment quality support. This is important as AF can be challenging to detect in advance, especially during home monitoring. Modern artificial intelligence (AI) has the potential to respond to this challenge. AI has already achieved state of the art results in many applications, including bioengineering. In this perspective, we discuss wearable devices combined with AI for AF detection, an approach that enables a new era of possibilities for the future.

State-of-the-art Deep Learning Methods on Electrocardiogram Data: A Systematic Review (Preprint)

Background

Electrocardiogram (ECG) is one of the most common noninvasive diagnostic tools that can provide useful information regarding a patient’s health status. Deep learning (DL) is an area of intense exploration that leads the way in most attempts to create powerful diagnostic models based on physiological signals.

Objective

This study aimed to provide a systematic review of DL methods applied to ECG data for various clinical applications.

Methods

The PubMed search engine was systematically searched by combining “deep learning” and keywords such as “ecg,” “ekg,” “electrocardiogram,” “electrocardiography,” and “electrocardiology.” Irrelevant articles were excluded from the study after screening titles and abstracts, and the remaining articles were further reviewed. The reasons for article exclusion were manuscripts written in any language other than English, absence of ECG data or DL methods involved in the study, and absence of a quantitative evaluation of the proposed approaches.

Results

We identified 230 relevant articles published between January 2020 and December 2021 and grouped them into 6 distinct medical applications, namely, blood pressure estimation, cardiovascular disease diagnosis, ECG analysis, biometric recognition, sleep analysis, and other clinical analyses. We provide a complete account of the state-of-the-art DL strategies per the field of application, as well as major ECG data sources. We also present open research problems, such as the lack of attempts to address the issue of blood pressure variability in training data sets, and point out potential gaps in the design and implementation of DL models.

Conclusions

We expect that this review will provide insights into state-of-the-art DL methods applied to ECG data and point to future directions for research on DL to create robust models that can assist medical experts in clinical decision-making.