Daily Cardiology Research Analysis

BACKGROUND AND AIMS: The 12-lead electrocardiogram (ECG) remains a cornerstone of cardiac diagnostics, yet existing artificial intelligence (AI) solutions for automated interpretation often lack generalizability, remain closed source, and are primarily trained using supervised learning (SL), which requires extensive labelled datasets and may limit adaptability across diverse clinical settings. Self-supervised learning (SSL) can potentially overcome these limitations by learning robust representations from unlabelled data. To address these challenges, this study developed and compared two open-source foundational ECG models: DeepECG-SL, a supervised multilabel ECG model, and DeepECG-SSL, a self-supervised model. METHODS: Both models were trained on over 1 million ECGs using a standardized preprocessing pipeline and automated free-text extraction from ECG reports to predict 77 cardiac conditions. DeepECG-SSL leveraged unlabelled data through self-supervised contrastive learning and masked lead modelling before fine-tuning for downstream tasks, while DeepECG-SL was trained directly on labelled diagnostic data in an end-to-end fashion. Performance was evaluated across seven private, multilingual healthcare systems and four public ECG repositories, with assessment of fairness by age and sex, and investigation of privacy vulnerabilities as well as memory and compute requirements. RESULTS: DeepECG-SSL achieved micro-averaged area under the receiver operating characteristic curves (AUROCs) across all 77 cardiac conditions for ECG interpretation of 0.990 [95% confidence interval (CI): 0.990, 0.990] on the internal dataset (MHI-ds), 0.981 (95% CI: 0.981, 0.981) on external public datasets (UKB, CLSA, MIMIC-IV and PTB), and 0.983 (95% CI: 0.983, 0.983) on external private datasets (UW, UCSF, JGH, NYP, MGH, CSH and CHUM), while DeepECG-SL demonstrated AUROCs of 0.992 (95% CI: 0.992, 0.992), 0.980 (95% CI: 0.980, 0.980), and 0.983 (95% CI: 0.983, 0.984), respectively. Fairness analyses revealed minimal disparities (true-positive rate and false-positive rate difference <0.1) across age and sex groups for both models. DeepECG-SSL demonstrated superior performance on limited-data digital biomarker tasks, with the largest improvements in long QT syndrome (LQTS) genotype classification (AUROC 0.931 vs 0.850, P = .026, n = 127 ECGs) and 5 year atrial fibrillation risk prediction (AUROC 0.742 vs 0.734, P < 0.001, n = 132 050 ECGs), while achieving superior performance in left ventricular ejection fraction ≤40% classification (AUROC 0.926 vs 0.917, P < 0.001, n = 25 252 ECGs) and comparable performance in LQTS detection (AUROC 0.767 vs 0.735, P = 0.117, n = 934 ECGs). CONCLUSIONS: This study establishes SSL as a promising paradigm for ECG analysis, particularly in settings with limited annotated data, enhancing accessibility, generalizability, and fairness in AI-driven cardiac diagnostics. By releasing model weights, preprocessing tools, and validation code, this work aims to support robust, data-efficient AI diagnostics across diverse clinical environments and questions.

BACKGROUND: Atrial fibrillation (AF), the most common cardiac arrhythmia, is a major cause of stroke and often remains undiagnosed due to its paroxysmal and frequently asymptomatic nature. Wearables provide a scalable, noninvasive screening tool. OBJECTIVES: This trial evaluated new onset AF detection in patients at elevated stroke risk using remote smartwatch-based screening. METHODS: This prospective multicenter randomized controlled trial included patients ≥65 years with elevated stroke risk (CHA RESULTS: Between November 2022 and December 2023, 437 patients were randomized (219 intervention, 218 control); the median age was 75 years, 46.7% were female and the median CHA CONCLUSIONS: This randomized controlled trial provides evidence that 6-month smartwatch-based AF screening enhances the detection rate of new onset AF compared with standard care in patients at elevated stroke risk. (Detection and Quantification of Atrial Fibrillation in High-risk Patients Using a Smartwatch Wearable [Apple Watch] [EQUAL]; NCT05686330).

BACKGROUND AND AIMS: LDL-C and non-HDL-C do not fully capture coronary heart disease (CHD) risk attributed to all apoB-containing lipoproteins. Use of apolipoprotein B (apoB) as a marker of total atherogenic particle number improves risk prediction, but risk may still be underestimated when triglyceride-rich lipoproteins (TRL/remnants) and lipoprotein(a) [Lp(a)] are elevated. The aim was to formulate a new metric-risk-weighted apoB (RW-apoB)-designed to capture risk from LDL, TRL/remnants, and Lp(a) in a single number. METHODS: Based on previously published estimates of the relative atherogenicity of LDL, TRL/remnant, and Lp(a) particles, RW-apoB was developed (using UK Biobank data) as an atherogenicity-weighted apoB-sum calculated as: RW-apoB = 11.65×TG(mmol/L) + 0.215×lipoprotein(a)(nmol/L) + 0.736×apoB(mg/dL). RESULTS: Assigning RW-apoB to individuals substantially reclassified their risk status. Compared with ranking by measured apoB, 52% of individuals were up- or down-ranked by ≥10 percentiles. About one-third of those in the top RW-apoB quintile-with elevated TRL and Lp(a) and a CHD event rate of 5.4%-were misclassified as lower risk by apoB. Conversely, individuals in the top measured apoB quintile but with low TRL and Lp(a) had a lower event rate (3.9%) and were correctly down-ranked. RW-apoB improved risk prediction, significantly increasing Harrell's C-index relative to apoB (P < .0001). In statin-treated subjects, RW-apoB was potentially a better index of residual risk. RW-apoB consistently outperformed apoB as a risk predictor in Cox models across the UK Biobank and three other large population cohorts. CONCLUSIONS: RW-apoB represents not only particle number but also accounts for the higher atherogenicity of TRL and Lp(a). It offers clinically meaningful improvements in CHD risk stratification.