Cardiovascular digital health journal最新文献

Background

Emerging evidence indicates that a high atrial fibrillation (AF) burden is associated with adverse outcome. However, AF burden is not routinely measured in clinical practice. An artificial intelligence (AI)-based tool could facilitate the assessment of AF burden.

Objective

We aimed to compare the assessment of AF burden performed manually by physicians with that measured by an AI-based tool.

Methods

We analyzed 7-day Holter electrocardiogram (ECG) recordings of AF patients included in the prospective, multicenter Swiss-AF Burden cohort study. AF burden was defined as percentage of time in AF, and was assessed manually by physicians and by an AI-based tool (Cardiomatics, Cracow, Poland). We evaluated the agreement between both techniques by means of Pearson correlation coefficient, linear regression model, and Bland-Altman plot.

Results

We assessed the AF burden in 100 Holter ECG recordings of 82 patients. We identified 53 Holter ECGs with 0% or 100% AF burden, where we found a 100% correlation. For the remaining 47 Holter ECGs with an AF burden between 0.01% and 81.53%, Pearson correlation coefficient was 0.998. The calibration intercept was -0.001 (95% CI -0.008; 0.006), and the calibration slope was 0.975 (95% CI 0.954; 0.995; multiple R² 0.995, residual standard error 0.017). Bland-Altman analysis resulted in a bias of -0.006 (95% limits of agreement -0.042 to 0.030).

Conclusion

The assessment of AF burden with an AI-based tool provided very similar results compared to manual assessment. An AI-based tool may therefore be an accurate and efficient option for the assessment of AF burden.

Background

Differentiating among cardiac diseases associated with left ventricular hypertrophy (LVH) informs diagnosis and clinical care.

Objective

To evaluate if artificial intelligence–enabled analysis of the 12-lead electrocardiogram (ECG) facilitates automated detection and classification of LVH.

Methods

We used a pretrained convolutional neural network to derive numerical representations of 12-lead ECG waveforms from patients in a multi-institutional healthcare system who had cardiac diseases associated with LVH (n = 50,709), including cardiac amyloidosis (n = 304), hypertrophic cardiomyopathy (n = 1056), hypertension (n = 20,802), aortic stenosis (n = 446), and other causes (n = 4766). We then regressed LVH etiologies relative to no LVH on age, sex, and the numerical 12-lead representations using logistic regression (“LVH-Net”). To assess deep learning model performance on single-lead data analogous to mobile ECGs, we also developed 2 single-lead deep learning models by training models on lead I (“LVH-Net Lead I”) or lead II (“LVH-Net Lead II”) from the 12-lead ECG. We compared the performance of the LVH-Net models to alternative models fit on (1) age, sex, and standard ECG measures, and (2) clinical ECG-based rules for diagnosing LVH.

Results

The areas under the receiver operator characteristic curve of LVH-Net by specific LVH etiology were cardiac amyloidosis 0.95 [95% CI, 0.93–0.97], hypertrophic cardiomyopathy 0.92 [95% CI, 0.90–0.94], aortic stenosis LVH 0.90 [95% CI, 0.88-0.92], hypertensive LVH 0.76 [95% CI, 0.76-0.77], and other LVH 0.69 [95% CI 0.68-0.71]. The single-lead models also discriminated LVH etiologies well.

Conclusion

An artificial intelligence–enabled ECG model is favorable for detection and classification of LVH and outperforms clinical ECG-based rules.

Background

Accurately determining arrhythmia mechanism from a 12-lead electrocardiogram (ECG) of supraventricular tachycardia can be challenging. We hypothesized a convolutional neural network (CNN) can be trained to classify atrioventricular re-entrant tachycardia (AVRT) vs atrioventricular nodal re-entrant tachycardia (AVNRT) from the 12-lead ECG, when using findings from the invasive electrophysiology (EP) study as the gold standard.

Methods

We trained a CNN on data from 124 patients undergoing EP studies with a final diagnosis of AVRT or AVNRT. A total of 4962 5-second 12-lead ECG segments were used for training. Each case was labeled AVRT or AVNRT based on the findings of the EP study. The model performance was evaluated against a hold-out test set of 31 patients and compared to an existing manual algorithm.

Results

The model had an accuracy of 77.4% in distinguishing between AVRT and AVNRT. The area under the receiver operating characteristic curve was 0.80. In comparison, the existing manual algorithm achieved an accuracy of 67.7% on the same test set. Saliency mapping demonstrated the network used the expected sections of the ECGs for diagnoses; these were the QRS complexes that may contain retrograde P waves.

Conclusion

We describe the first neural network trained to differentiate AVRT from AVNRT. Accurate diagnosis of arrhythmia mechanism from a 12-lead ECG could aid preprocedural counseling, consent, and procedure planning. The current accuracy from our neural network is modest but may be improved with a larger training dataset.

Background

Paroxysmal atrial fibrillation (AF) often eludes early diagnosis, resulting in significant morbidity and mortality. Artificial intelligence (AI) has been used to predict AF from sinus rhythm electrocardiograms (ECGs), but AF prediction using sinus rhythm mobile electrocardiograms (mECG) remains unexplored.

Objective

The purpose of this study was to investigate the utility of AI to predict AF events prospectively and retrospectively using sinus rhythm mECG data.

Methods

We trained a neural network to predict AF events from sinus rhythm mECGs obtained from users of the Alivecor KardiaMobile 6L device. We tested our model on sinus rhythm mECGs within ±0–2 days, ±3–7 days, and ±8–30 days from AF events to determine the optimal screening window. Finally, we tested our model on mECGs from before an AF event to determine whether AF can be predicted prospectively.

Results

We included 73,861 users with 267,614 mECGs (mean age 58.14 years; 35% women). Users with paroxysmal AF contributed 60.15% of mECGs. Model performance on the test set comprising control and study samples across all windows of interest showed an area under the curve (AUC) score of 0.760 (95% confidence interval [CI] 0.759–0.760), sensitivity of 0.703 (95% CI 0.700–0.705), specificity of 0.684 (95% CI 0.678–0.685), and accuracy of 69.4% (95% CI 0.692–0.700). Model performance was better on ±0–2 day samples (sensitivity 0.711; 95% CI 0.709–0.713) and worse on the ±8–30 day window (sensitivity 0.688; 95% CI 0.685–0.690), with performance on the ±3–7 day window falling in between (sensitivity 0.708; 95% CI 0.704–0.710).

Conclusion

Neural networks can predict AF using a widely scalable and cost-effective mobile technology prospectively and retrospectively.

Background

Cardiovascular implantable electronic devices (CIEDs) capture an abundance of data for clinicians to review and integrate into the clinical decision-making process. The multitude of data from different device types and vendors presents challenges for viewing and using the data in clinical practice. Efforts are needed to improve CIED reports by focusing on key data elements used by clinicians.

Objective

The purpose of this study was to uncover the extent to which clinicians use the specific types of data elements from CIED reports in clinical practice and explore clinicians’ perceptions of CIED reports.

Methods

A brief, web-based, cross-sectional survey study was deployed using snowball sampling from March 2020 through September 2020 to clinicians who are involved in the care of patients with CIEDs.

Results

Among 317 clinicians, the majority specialized in electrophysiology (EP) (80.1%), were from North America (88.6%), and were white (82.2%). Over half (55.3%) were physicians. Arrhythmia episodes and ventricular therapies rated the highest among 15 categories of data presented, and nocturnal or resting heart rate and heart rate variability were rated the lowest. As anticipated, clinicians specializing in EP reported using the data significantly more than other specialties across nearly all categories. A subset of respondents offered general comments describing preferences and challenges related to reviewing reports.

Conclusion

CIED reports contain an abundance of information that is important to clinicians; however, some data are used more frequently than others, and reports could be streamlined for users to improve access to key information and facilitate more efficient clinical decision making.

Cuff-based home blood pressure (BP) devices, which have been the standard for BP monitoring for decades, are limited by physical discomfort, convenience, and their ability to capture BP variability and patterns between intermittent readings. In recent years, cuffless BP devices, which do not require cuff inflation around a limb, have entered the market, offering the promise of continuous beat-to-beat measurement of BP. These devices take advantage of a variety of principles to determine BP, including (1) pulse arrival time, (2) pulse transit time, (3) pulse wave analysis, (4) volume clamping, and (5) applanation tonometry. Because BP is calculated indirectly, these devices require calibration with cuff-based devices at regular intervals. Unfortunately, the pace of regulation of these devices has failed to match the speed of innovation and direct availability to patient consumers. There is an urgent need to develop a consensus on standards by which cuffless BP devices can be tested for accuracy. In this narrative review, we describe the landscape of cuffless BP devices, summarize the current status of validation protocols, and provide recommendations for an ideal validation process for these devices.

Background

The QT interval in the electrocardiogram (ECG) is a fundamental risk measure for arrhythmic adverse cardiac events. However, the QT interval depends on the heart rate and must be corrected accordingly. The present QT correction (QTc) methods are either simple models leading to under- or overcorrection, or impractical in requiring long-term empirical data. In general, there is no consensus on the best QTc method.

Objective

We introduce a model-free QTc method—AccuQT—that computes QTc by minimizing the information transfer from R-R to QT intervals. The objective is to establish and validate a QTc method that provides superior stability and reliability without models or empirical data.

Methods

We tested AccuQT against the most commonly used QT correction methods by using long-term ECG recordings of more than 200 healthy subjects from PhysioNet and THEW databases.

Results

AccuQT overperforms the previously reported correction methods: the proportion of false-positives is reduced from 16% (Bazett) to 3% (AccuQT) for the PhysioNet data. In particular, the QTc variance is significantly reduced and thus the RR-QT stability is increased.

Conclusion

AccuQT has significant potential to become the QTc method of choice in clinical studies and drug development. The method can be implemented in any device recording R-R and QT intervals.

Background

Telerobotic surgery could improve access to specialty procedures such as cardiac catheter ablation in rural and underserved regions in the United States and worldwide. Advancements in telecommunications, internet infrastructure, and surgical robotics are lowering the technical hurdles for this future healthcare delivery paradigm. Nonetheless, important questions remain regarding the safe implementation of telerobotic surgery in rural community hospital settings.

Objective

The purpose of this study was to pilot test a system and methods to explore telerobotic cardiac catheter ablation in a rural community hospital setting.

Methods

We assembled a portable preclinical telerobotic catheter ablation system from commercial-grade components using third-party vendors. We then carried out 4 telerobotic surgery simulations with an urban surgeon and a rural community hospital operating room (OR) team spanning a distance of more than 2000 miles. Two challenge scenarios were incorporated into the simulations, including loss of network connection and cardiac perforation with subsequent life-threatening tamponade physiology. An ethnographic analysis was then performed.

Results

Interviews and observations suggested that rural OR teams readily adapt to the telesurgery context. However, participant perceptions of team trust, communication, and emergency management were significantly altered by the remote location of the surgeon. In addition, most participants believed the OR team would have been better equipped for the challenges had they received formal training or had prior experience with the procedure being simulated.

Conclusion

We demonstrate the utility and feasibility of a system and methods for studying specialty telerobotic surgery in a rural hospital OR setting.