ECG-Based Deep Learning and Clinical Risk Factors to Predict Atrial Fibrillation.

BACKGROUND: Artificial intelligence (AI)-enabled analysis of 12-lead ECGs may facilitate efficient estimation of incident atrial fibrillation (AF) risk. However, it remains unclear whether AI provides meaningful and generalizable improvement in predictive accuracy beyond clinical risk factors for AF. METHODS: We trained a convolutional neural network (ECG-AI) to infer 5-year incident AF risk using 12-lead ECGs in patients receiving longitudinal primary care at Massachusetts General Hospital (MGH). We then fit 3 Cox proportional hazards models, composed of ECG-AI 5-year AF probability, CHARGE-AF clinical risk score (Cohorts for Heart and Aging in Genomic Epidemiology-Atrial Fibrillation), and terms for both ECG-AI and CHARGE-AF (CH-AI), respectively. We assessed model performance by calculating discrimination (area under the receiver operating characteristic curve) and calibration in an internal test set and 2 external test sets (Brigham and Women's Hospital [BWH] and UK Biobank). Models were recalibrated to estimate 2-year AF risk in the UK Biobank given limited available follow-up. We used saliency mapping to identify ECG features most influential on ECG-AI risk predictions and assessed correlation between ECG-AI and CHARGE-AF linear predictors. RESULTS: The training set comprised 45 770 individuals (age 55±17 years, 53% women, 2171 AF events) and the test sets comprised 83 162 individuals (age 59±13 years, 56% women, 2424 AF events). Area under the receiver operating characteristic curve was comparable using CHARGE-AF (MGH, 0.802 [95% CI, 0.767-0.836]; BWH, 0.752 [95% CI, 0.741-0.763]; UK Biobank, 0.732 [95% CI, 0.704-0.759]) and ECG-AI (MGH, 0.823 [95% CI, 0.790-0.856]; BWH, 0.747 [95% CI, 0.736-0.759]; UK Biobank, 0.705 [95% CI, 0.673-0.737]). Area under the receiver operating characteristic curve was highest using CH-AI (MGH, 0.838 [95% CI, 0.807 to 0.869]; BWH, 0.777 [95% CI, 0.766 to 0.788]; UK Biobank, 0.746 [95% CI, 0.716 to 0.776]). Calibration error was low using ECG-AI (MGH, 0.0212; BWH, 0.0129; UK Biobank, 0.0035) and CH-AI (MGH, 0.012; BWH, 0.0108; UK Biobank, 0.0001). In saliency analyses, the ECG P-wave had the greatest influence on AI model predictions. ECG-AI and CHARGE-AF linear predictors were correlated (Pearson r: MGH, 0.61; BWH, 0.66; UK Biobank, 0.41). CONCLUSIONS: AI-based analysis of 12-lead ECGs has similar predictive usefulness to a clinical risk factor model for incident AF and the approaches are complementary. ECG-AI may enable efficient quantification of future AF risk.

Deep learning on resting electrocardiogram to identify impaired heart rate recovery.

Background and Objective: Postexercise heart rate recovery (HRR) is an important indicator of cardiac autonomic function and abnormal HRR is associated with adverse outcomes. We hypothesized that deep learning on resting electrocardiogram (ECG) tracings may identify individuals with impaired HRR. Methods: We trained a deep learning model (convolutional neural network) to infer HRR based on resting ECG waveforms (HRRpred) among UK Biobank participants who had undergone exercise testing. We examined the association of HRRpred with incident cardiovascular disease using Cox models, and investigated the genetic architecture of HRRpred in genome-wide association analysis. Results: Among 56,793 individuals (mean age 57 years, 51% women), the HRRpred model was moderately correlated with actual HRR (r = 0.48, 95% confidence interval [CI] 0.47-0.48). Over a median follow-up of 10 years, we observed 2060 incident diabetes mellitus (DM) events, 862 heart failure events, and 2065 deaths. Higher HRRpred was associated with lower risk of DM (hazard ratio [HR] 0.79 per 1 standard deviation change, 95% CI 0.76-0.83), heart failure (HR 0.89, 95% CI 0.83-0.95), and death (HR 0.83, 95% CI 0.79-0.86). After accounting for resting heart rate, the association of HRRpred with incident DM and all-cause mortality were similar. Genetic determinants of HRRpred included known heart rate, cardiac conduction system, cardiomyopathy, and metabolic trait loci. Conclusion: Deep learning-derived estimates of HRR using resting ECG independently associated with future clinical outcomes, including new-onset DM and all-cause mortality. Inferring postexercise heart rate response from a resting ECG may have potential clinical implications and impact on preventive strategies warrants future study.

Cohort design and natural language processing to reduce bias in electronic health records research.

Electronic health record (EHR) datasets are statistically powerful but are subject to ascertainment bias and missingness. Using the Mass General Brigham multi-institutional EHR, we approximated a community-based cohort by sampling patients receiving longitudinal primary care between 2001-2018 (Community Care Cohort Project [C3PO], n = 520,868). We utilized natural language processing (NLP) to recover vital signs from unstructured notes. We assessed the validity of C3PO by deploying established risk models for myocardial infarction/stroke and atrial fibrillation. We then compared C3PO to Convenience Samples including all individuals from the same EHR with complete data, but without a longitudinal primary care requirement. NLP reduced the missingness of vital signs by 31%. NLP-recovered vital signs were highly correlated with values derived from structured fields (Pearson r range 0.95-0.99). Atrial fibrillation and myocardial infarction/stroke incidence were lower and risk models were better calibrated in C3PO as opposed to the Convenience Samples (calibration error range for myocardial infarction/stroke: 0.012-0.030 in C3PO vs. 0.028-0.046 in Convenience Samples; calibration error for atrial fibrillation 0.028 in C3PO vs. 0.036 in Convenience Samples). Sampling patients receiving regular primary care and using NLP to recover missing data may reduce bias and maximize generalizability of EHR research.

Physiology as a Lingua Franca for Clinical Machine Learning.

The intersection of medicine and machine learning (ML) has the potential to transform healthcare. We describe how physiology, a foundational discipline of medical training and practice with a rich quantitative history, could serve as a starting point for the development of a common language between clinicians and ML experts, thereby accelerating real-world impact.

Re-examining physician-scientist training through the prism of the discovery-invention cycle.

The training of physician-scientists lies at the heart of future medical research. In this commentary, we apply Narayanamurti and Odumosu's framework of the "discovery-invention cycle" to analyze the structure and outcomes of the integrated MD/PhD program. We argue that the linear model of "bench-to-bedside" research, which is also reflected in the present training of MD/PhDs, merits continual re-evaluation to capitalize on the richness of opportunities arising in clinical medicine. In addition to measuring objective career outcomes, as existing research has done, we suggest that detailed characterization of researchers' efforts using both qualitative and quantitative techniques is necessary to understand if dual-degree training is being utilized. As an example, we propose that the application of machine learning and data science to corpora of biomedical literature and anonymized clinical data might allow us to see if there are objective "signatures" of research uniquely enabled by MD/PhD training. We close by proposing several hypotheses for shaping physician-scientist training, the relative merits of which could be assessed using the techniques proposed above. Our overarching message is the importance of deeply understanding individual career trajectories as well as characterizing organizational details and cultural nuances to drive new policy which shapes the future of the physician-scientist workforce.

OpenWorm: overview and recent advances in integrative biological simulation of Caenorhabditis elegans.

The adoption of powerful software tools and computational methods from the software industry by the scientific research community has resulted in a renewed interest in integrative, large-scale biological simulations. These typically involve the development of computational platforms to combine diverse, process-specific models into a coherent whole. The OpenWorm Foundation is an independent research organization working towards an integrative simulation of the nematode Caenorhabditis elegans, with the aim of providing a powerful new tool to understand how the organism's behaviour arises from its fundamental biology. In this perspective, we give an overview of the history and philosophy of OpenWorm, descriptions of the constituent sub-projects and corresponding open-science management practices, and discuss current achievements of the project and future directions.This article is part of a discussion meeting issue 'Connectome to behaviour: modelling C. elegans at cellular resolution'.

Integrative biological simulation praxis: Considerations from physics, philosophy, and data/model curation practices.

Integrative biological simulations have a varied and controversial history in the biological sciences. From computational models of organelles, cells, and simple organisms, to physiological models of tissues, organ systems, and ecosystems, a diverse array of biological systems have been the target of large-scale computational modeling efforts. Nonetheless, these research agendas have yet to prove decisively their value among the broader community of theoretical and experimental biologists. In this commentary, we examine a range of philosophical and practical issues relevant to understanding the potential of integrative simulations. We discuss the role of theory and modeling in different areas of physics and suggest that certain sub-disciplines of physics provide useful cultural analogies for imagining the future role of simulations in biological research. We examine philosophical issues related to modeling which consistently arise in discussions about integrative simulations and suggest a pragmatic viewpoint that balances a belief in philosophy with the recognition of the relative infancy of our state of philosophical understanding. Finally, we discuss community workflow and publication practices to allow research to be readily discoverable and amenable to incorporation into simulations. We argue that there are aligned incentives in widespread adoption of practices which will both advance the needs of integrative simulation efforts as well as other contemporary trends in the biological sciences, ranging from open science and data sharing to improving reproducibility.

Unit testing, model validation, and biological simulation.

The growth of the software industry has gone hand in hand with the development of tools and cultural practices for ensuring the reliability of complex pieces of software. These tools and practices are now acknowledged to be essential to the management of modern software. As computational models and methods have become increasingly common in the biological sciences, it is important to examine how these practices can accelerate biological software development and improve research quality. In this article, we give a focused case study of our experience with the practices of unit testing and test-driven development in OpenWorm, an open-science project aimed at modeling Caenorhabditis elegans. We identify and discuss the challenges of incorporating test-driven development into a heterogeneous, data-driven project, as well as the role of model validation tests, a category of tests unique to software which expresses scientific models.

Specification of photonic circuits using quantum hardware description language.

Following the simple observation that the interconnection of a set of quantum optical input-output devices can be specified using structural mode VHSIC hardware description language, we demonstrate a computer-aided schematic capture workflow for modelling and simulating multi-component photonic circuits. We describe an algorithm for parsing circuit descriptions to derive quantum equations of motion, illustrate our approach using simple examples based on linear and cavity-nonlinear optical components, and demonstrate a computational approach to hierarchical model reduction.