Feasibility of remote monitoring for fatal coronary heart disease using Apple Watch ECGs.

Background: Fatal coronary heart disease (FCHD) is often described as sudden cardiac death (affects >4 million people/year), where coronary artery disease is the only identified condition. Electrocardiographic artificial intelligence (ECG-AI) models for FCHD risk prediction using ECG data from wearable devices could enable wider screening/monitoring efforts. Objectives: To develop a single-lead ECG-based deep learning model for FCHD risk prediction and assess concordance between clinical and Apple Watch ECGs. Methods: An FCHD single-lead ("lead I" from 12-lead ECGs) ECG-AI model was developed using 167,662 ECGs (50,132 patients) from the University of Tennessee Health Sciences Center. Eighty percent of the data (5-fold cross-validation) was used for training and 20% as a holdout. Cox proportional hazards (CPH) models incorporating ECG-AI predictions with age, sex, and race were also developed. The models were tested on paired clinical single-lead and Apple Watch ECGs from 243 St. Jude Lifetime Cohort Study participants. The correlation and concordance of the predictions were assessed using Pearson correlation (R), Spearman correlation (ρ), and Cohen's kappa. Results: The ECG-AI and CPH models resulted in AUC = 0.76 and 0.79, respectively, on the 20% holdout and AUC = 0.85 and 0.87 on the Atrium Health Wake Forest Baptist external validation data. There was moderate-strong positive correlation between predictions (R = 0.74, ρ = 0.67, and κ = 0.58) when tested on the 243 paired ECGs. The clinical (lead I) and Apple Watch predictions led to the same low/high-risk FCHD classification for 99% of the participants. CPH prediction correlation resulted in an R = 0.81, ρ = 0.76, and κ = 0.78. Conclusion: Risk of FCHD can be predicted from single-lead ECGs obtained from wearable devices and are statistically concordant with lead I of a 12-lead ECG.

Association between deep neural network-derived electrocardiographic-age and incident stroke.

Background: Stroke continues to be a leading cause of death and disability worldwide despite improvements in prevention and treatment. Traditional stroke risk calculators are biased and imprecise. Novel stroke predictors need to be identified. Recently, deep neural networks (DNNs) have been used to determine age from ECGs, otherwise known as the electrocardiographic-age (ECG-age), which predicts clinical outcomes. However, the relationship between ECG-age and stroke has not been well studied. We hypothesized that ECG-age is associated with incident stroke. Methods: In this study, UK Biobank participants with available ECGs (from 2014 or later). ECG-age was estimated using a deep neural network (DNN) applied to raw ECG waveforms. We calculated the Δage (ECG-age minus chronological age) and classified individuals as having normal, accelerated, or decelerated aging if Δage was within, higher, or lower than the mean absolute error of the model, respectively. Multivariable Cox proportional hazards regression models adjusted for age, sex, and clinical factors were used to assess the association between Δage and incident stroke. Results: The study population included 67,757 UK Biobank participants (mean age 65 ± 8 years; 48.3% male). Every 10-year increase in Δage was associated with a 22% increase in incident stroke [HR, 1.22 (95% CI, 1.00-1.49)] in the multivariable-adjusted model. Accelerated aging was associated with a 42% increase in incident stroke [HR, 1.42 (95% CI, 1.12-1.80)] compared to normal aging. In addition, Δage was associated with prevalent stroke [OR, 1.28 (95% CI, 1.11-1.49)]. Conclusions: DNN-estimated ECG-age was associated with incident and prevalent stroke in the UK Biobank. Further investigation is required to determine if ECG-age can be used as a reliable biomarker of stroke risk.

Low-Density Lipoproteins, High-Density Lipoproteins (HDL), and HDL-Associated Proteins Differentially Modulate Chronic Myelogenous Leukemia Cell Viability.

Cellular lipid metabolism, lipoprotein interactions, and liver X receptor (LXR) activation have been implicated in the pathophysiology and treatment of cancer, although findings vary across cancer models and by lipoprotein profiles. In this study, we investigated the effects of human-derived low-density lipoproteins (LDL), high-density lipoproteins (HDL), and HDL-associated proteins apolipoprotein A1 (apoA1) and serum amyloid A (SAA) on markers of viability, cholesterol flux, and differentiation in K562 cells-a bone marrow-derived, stem-like erythroleukemia cell model of chronic myelogenous leukemia (CML). We further evaluated whether lipoprotein-mediated effects were altered by concomitant LXR activation. We observed that LDL promoted higher K562 cell viability in a dose- and time-dependent manner and increased cellular cholesterol concentrations, while LXR activation by the agonist TO901317 ablated these effects. LXR activation in the presence of HDL, apoA1 and SAA-rich HDL suppressed K562 cell viability, while robustly inducing mRNA expression of ATP-binding cassette transporter A1 (ABCA1). HDL and its associated proteins additionally suppressed mRNA expression of anti-apoptotic B-cell lymphoma-extra large (BCL-xL), and the erythroid lineage marker 5'-aminolevulinate synthase 2 (ALAS2), while SAA-rich HDL induced mRNA expression of the megakaryocytic lineage marker integrin subunit alpha 2b (ITGA2B). Together, these findings suggest that lipoproteins and LXR may impact the viability and characteristics of CML cells.

Enabling responsible public genomics.

As scientific understandings of genetics advance, researchers require increasingly rich datasets that combine genomic data from large numbers of individuals with medical and other personal information. Linking individuals' genetic data and personal information precludes anonymity and produces medically significant information--a result not contemplated by the established legal and ethical conventions governing human genomic research. To pursue the next generation of human genomic research and commerce in a responsible fashion, scientists, lawyers, and regulators must address substantial new issues, including researchers' duties with respect to clinically significant data, the challenges to privacy presented by genomic data, the boundary between genomic research and commerce, and the practice of medicine. This Article presents a new model for understanding and addressing these new challenges--a "public genomics" premised on the idea that ethically, legally, and socially responsible genomics research requires openness, not privacy, as its organizing principle. Responsible public genomics combines the data contributed by informed and fully consenting information altruists and the research potential of rich datasets in a genomic commons that is freely and globally available. This Article examines the risks and benefits of this public genomics model in the context of an ambitious genetic research project currently under way--the Personal Genome Project. This Article also (i) demonstrates that large-scale genomic projects are desirable, (ii) evaluates the risks and challenges presented by public genomics research, and (iii) determines that the current legal and regulatory regimes restrict beneficial and responsible scientific inquiry while failing to adequately protect participants. The Article concludes by proposing a modified normative and legal framework that embraces and enables a future of responsible public genomics.