BACKGROUND:Atrial fibrillation (AF) is the most common cardiac arrhythmia associated with an increased risk of stroke and heart failure. AF is often asymptomatic and paroxysmal, making diagnosis challenging. Artificial intelligence (AI) applied to electrocardiogram (ECG) interpretation is a promising approach for improved diagnosis. Although ECG-AI studies have shown promise, the common practice of evaluating based on data from single institutions may overestimate performance. External validation is essential to ensure that AI models generalize well to diverse settings and populations. OBJECTIVE:This study aimed to externally validate an ECG-AI model for predicting 1-year AF risk. METHODS:In this retrospective study, ECG data from 3 clinical sites were aggregated and patients' charts were manually abstracted to define inclusion and exclusion criteria (age 65+ years with no previous AF or history of pacer/defibrillator use) and endpoints (new AF diagnosis within 1 year or 1 year of AF-free follow-up). The sensitivity and specificity of a risk score from an ECG-AI model were evaluated against prespecified minimum values of 20% and 85%, respectively. RESULTS:The analysis included 4017 patients, with 240 (6.0%) developing AF within 1 year. The ECG-AI model returned an "increased risk" result for 391 patients (9.7%), including 74 who developed AF (sensitivity 31%; 95% confidence interval 25-37). A total of 3626 patients had a "no increased risk" result, with 3460 remaining free from AF (specificity 92%; 95% confidence interval 91-92). CONCLUSION:The results validate the performance of the Tempus ECG-AF model and support its clinical use for AF risk stratification.

To broaden our understanding of bradyarrhythmias and conduction disease, we performed common variant genome-wide association analyses in up to 1.3 million individuals and rare variant burden testing in 460,000 individuals for sinus node dysfunction (SND), distal conduction disease (DCD) and pacemaker (PM) implantation. We identified 13, 31 and 21 common variant loci for SND, DCD and PM, respectively. Four well-known loci (SCN5A/SCN10A, CCDC141, TBX20 and CAMK2D) were shared for SND and DCD, while others were more specific for SND or DCD. SND and DCD showed a moderate genetic correlation (rg = 0.63). Cardiomyocyte-expressed genes were enriched for contributions to DCD heritability. Rare-variant analyses implicated LMNA for all bradyarrhythmia phenotypes, SMAD6 and SCN5A for DCD and TTN, MYBPC3 and SCN5A for PM. These results show that variation in multiple genetic pathways (for example, ion channel function, cardiac developmental programs, sarcomeric structure and cellular homeostasis) appear critical to the development of bradyarrhythmias. Genome-wide analyses identify variants associated with sinus node dysfunction, distal conduction disease and pacemaker implantation, implicating ion channel function, cardiac developmental programs and sarcomeric structure in bradyarrhythmia susceptibility.

Atrial fibrillation (AF) is a prevalent and morbid abnormality of the heart rhythm with a strong genetic component. Here, we meta-analyzed genome and exome sequencing data from 36 studies that included 52,416 AF cases and 277,762 controls. In burden tests of rare coding variation, we identified novel associations between AF and the genes MYBPC3, LMNA, PKP2, FAM189A2 and KDM5B. We further identified associations between AF and rare structural variants owing to deletions in CTNNA3 and duplications of GATA4. We broadly replicated our findings in independent samples from MyCode, deCODE and UK Biobank. Finally, we found that CRISPR knockout of KDM5B in stem-cell-derived atrial cardiomyocytes led to a shortening of the action potential duration and widespread transcriptomic dysregulation of genes relevant to atrial homeostasis and conduction. Our results highlight the contribution of rare coding and structural variants to AF, including genetic links between AF and cardiomyopathies, and expand our understanding of the rare variant architecture for this common arrhythmia.

BACKGROUND: Population genomic screening for desmosome variants associated with arrhythmogenic right ventricular cardiomyopathy (ARVC) may facilitate early disease detection and protective intervention. The validated ARVC risk calculator offers a novel means to risk stratify individuals with diagnosed ARVC, but predicted risk in the context of genomic screening identification has not been explored. METHODS: Individuals harboring a pathogenic/likely pathogenic variant in a desmosome gene ( PKP2 , DSP , DSG2 , or DSC2 ) were identified through the Geisinger MyCode Genomic Screening and Counseling program. The ARVC risk calculator was applied to patients with a subsequent evaluation of right ventricular function. This predicted risk was compared with outcomes in the first 5 years (range, 0.3–5.0 years) after genetic result return. RESULTS: Of 254 individuals with a clinically confirmed pathogenic/likely pathogenic desmosome variant, 113 (median age, 56 [interquartile range, 42–66]; 71% female) had cardiac imaging in follow-up and no prior sustained ventricular arrhythmia (VA). Eighty-two (73%) had no ARVC task force criteria (TFC) besides the variant (possible diagnosis), 22 (19%) had a single additional minor criterion (borderline diagnosis), and 9 (8%) met criteria for definite diagnosis. The median 5-year predicted VA risk was 3.9% (2.3%–6.6%), notably lower than that of the calculator derivation cohort (20.6%). The risk of fast VA was 1.6% (1.0%–2.9%). The predicted VA risk was higher in individuals with any nongenetic ARVC task force criteria (6.3% [2.5–13.2%]) versus those without (3.7% [2.2–5.6%]; P =0.01), and in individuals with DSP variants (6.1% [3.9–7.8%] versus PKP2 3.4% [2.2–5.3%]; P =0.01). Over a median 3.0 years of follow-up (≤5 years only), no sustained VA events were observed in this cohort. CONCLUSIONS: The predicted 5-year risk of VA in individuals ascertained via population genomic screening for desmosome variants is low (3.9%; 1.6% for fast VA) but may vary by affected gene and ARVC task force criteria burden.

Heart failure is a complex trait, influenced by environmental and genetic factors, affecting over 30 million individuals worldwide. Here we report common-variant and rare-variant association studies of all-cause heart failure and examine how different classes of genetic variation impact its heritability. We identify 176 common-variant risk loci at genome-wide significance in 2,358,556 individuals and cluster these signals into five broad modules based on pleiotropic associations with anthropomorphic traits/obesity, blood pressure/renal function, atherosclerosis/lipids, immune activity and arrhythmias. In parallel, we uncover exome-wide significant associations for heart failure and rare predicted loss-of-function variants in TTN, MYBPC3, FLNC and BAG3 using exome sequencing of 376,334 individuals. We find that total burden heritability of rare coding variants is highly concentrated in a small set of Mendelian cardiomyopathy genes, while common-variant heritability is diffusely spread throughout the genome. Finally, we show that common-variant background modifies heart failure risk among carriers of rare pathogenic truncating variants in TTN. Together, these findings discern genetic links between dysregulated metabolism and heart failure and highlight a polygenic component to heart failure not captured by current clinical genetic testing.

Low ejection fraction (EF), an indicator of impaired heart function, often goes undiagnosed and can lead to avoidable heart failure and arrhythmias. We developed and externally validated a deep learning model for detecting low EF from 12-lead electrocardiograms. The model achieved 85.8% sensitivity and 83.0% specificity on an independent validation cohort, with consistent results across demographic subgroups. These findings supported FDA 510(k) clearance of the model. Clinical net benefit analysis further showed that the model provides greater clinical value than default screening approaches, confirming its meaningful potential impact for clinical practice. ### Competing Interest Statement The authors have declared no competing interest. ### Funding Statement This study did not receive any funding ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: This study was conducted on de-identified health information subject to an IRB exempt determination (Advarra Pro00072742) and did not involve human subjects research. I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes Deidentified data used in the research were collected in a real-world health care setting and are subject to controlled access for privacy and proprietary reasons. When possible, derived data supporting the findings of this study have been made available within the paper and its Supplementary Figures/Tables.

BACKGROUND:Many medications are associated with long QTc. Current long QTc predictors have limited generalizability and/or modest performance. OBJECTIVE:This study aimed to compare the performance of machine learning-enabled approaches to drug-induced long QTc prediction with current risk scores, including Tisdale and RISQ-PATH. METHODS:We identified patients who began QTc-prolonging medications (QTdrugs) and had follow-up 12-lead electrocardiograms (ECGs) within 1 year. Using 5-fold cross-validation, we trained XGBoost (XGB), an ECG-based deep neural network, and combined models using electronic health record data and ECG traces to predict QTc of ≥500 ms within 1 year. We assessed performance using the area under the receiver operating characteristic (AUROC) curves and compared them with corresponding Tisdale and RISQ-PATH scores. RESULTS:QTdrug records were identified for 345,371 patients, 5.7% of whom had long QTc within 1 year. In the subset with baseline ECGs available (N = 182,448; 7.7% events), both the XGB and deep neural network models demonstrated high performance (AUROC 0.869 and 0.864, respectively) but their combination yielded no significant improvement (AUROC 0.874). Therefore, focusing on the XGB model, we observed superior performance vs RISQ-PATH in the overall population (AUROC 0.859 vs 0.701), as well as Tisdale in predominantly inpatients (N = 110,558; 8.8% events; AUROC 0.855 vs 0.770). Positive predictive value was 61.5% vs 28.35% and 54.0% vs 28.3% at equivalent operating points for the XGB model vs RISQ-PATH and Tisdale, respectively. CONCLUSION:Development and retrospective validation of 3 machine learning-based models for predicting drug-induced long QTc at the time of new QTdrug starts demonstrated superior performance compared with current clinical risk calculators and may be useful tools to support medical decision making when initiating new therapies.

The majority of biomedical studies use limited datasets that may not generalize over large heterogeneous datasets that have been collected over several decades. The current paper develops and validates several multimodal models that can predict 1-year mortality based on a massive clinical dataset. Our focus on predicting 1-year mortality can provide a sense of urgency to the patients. Using the largest dataset of its kind, the paper considers the development and validation of multimodal models based on 25,137,015 videos associated with 699,822 echocardiography studies from 316,125 patients, and 2,922,990 8-lead electrocardiogram (ECG) traces from 631,353 patients. Our models allow us to assess the contribution of individual factors and modalities to the overall risk. Our approach allows us to develop extremely low-parameter models that use optimized feature selection based on feature importance. Based on available clinical information, we construct a family of models that are made available in the DISIML package. Overall, performance ranges from an AUC of 0.72 with just ten parameters to an AUC of 0.89 with under 105k for the full multimodal model. The proposed approach represents a modular neural network framework that can provide insights into global risk trends and guide therapies for reducing mortality risk.

Heart failure (HF) is a major contributor to global morbidity and mortality. While distinct clinical subtypes, defined by etiology and left ventricular ejection fraction, are well recognized, their genetic determinants remain inadequately understood. In this study, we report a genome-wide association study of HF and its subtypes in a sample of 1.9 million individuals. A total of 153,174 individuals had HF, of whom 44,012 had a nonischemic etiology (ni-HF). A subset of patients with ni-HF were stratified based on left ventricular systolic function, where data were available, identifying 5,406 individuals with reduced ejection fraction and 3,841 with preserved ejection fraction. We identify 66 genetic loci associated with HF and its subtypes, 37 of which have not previously been reported. Using functionally informed gene prioritization methods, we predict effector genes for each identified locus, and map these to etiologic disease clusters through phenome-wide association analysis, network analysis and colocalization. Through heritability enrichment analysis, we highlight the role of extracardiac tissues in disease etiology. We then examine the differential associations of upstream risk factors with HF subtypes using Mendelian randomization. These findings extend our understanding of the mechanisms underlying HF etiology and may inform future approaches to prevention and treatment.

BACKGROUND AND AIMS:Classification and risk stratification in aortic (AR), mitral (MR), and tricuspid regurgitation (TR) remains a significant clinical challenge. This study aimed to develop an artificial intelligence (AI) system to assess valvular regurgitation and stratify MR-progression risk. METHODS:Using transthoracic echocardiograms (TTEs) at two sites (internal development/test, external test), the DELINEATE-Regurgitation system was developed to classify AR, MR, and TR severity using colour Doppler videos. Methods of summating video-level classifications into study-level predictions were tested, comparing single-view with multiview approaches integrating predictions across multiple videos. Model agreement with cardiologists was assessed by weighted kappa. A separate AI system (DELINEATE-MR-Progression) analysing colour Doppler videos was developed to predict which patients with mild, mild-moderate, and moderate MR were most likely to progress to moderate-severe or severe MR with analysis by Kaplan-Meier and Cox proportional hazards models. RESULTS:A total of 71 660 TTEs with 1 203 980 colour Doppler videos were included. The weighted kappa in internal/external test sets for regurgitation classification was 0.81/0.76 for AR, 0.76/0.72 for MR, and 0.73/0.64 for TR using a multiview approach taking all colour Doppler videos in a study, demonstrating substantial agreement with cardiologist interpretation with superiority of multiview over single view approaches. In the progression analysis, the AI score stratified MR-progression risk even when controlled for clinical factors known to be associated with MR progression [internal test set hazard ratio 4.1 (95% confidence interval 2.5-6.6)]. CONCLUSIONS:An AI system can accurately classify AR, MR, and TR and predict MR progression beyond currently known risk factors.

Early detection of structural heart disease is critical to improving outcomes, but widespread screening remains limited by the cost and accessibility of imaging tools such as echocardiography1,2. Recent advances in machine learning applied to heart rhythm recordings have shown promise in identifying disease3,4, although previous work has been limited by development in narrow populations or targeting only select heart conditions5. Here we introduce a deep learning model, EchoNext, trained on more than 1 million heart rhythm and imaging records across a large and diverse health system to detect many forms of structural heart disease. The model demonstrated high diagnostic accuracy in internal and external validation, outperforming cardiologists in a controlled evaluation and showing consistent performance across different care settings and racial and/or ethnic groups. The models were prospectively evaluated in a clinical trial of patients without previous cardiac imaging, successfully identifying previously undiagnosed heart disease. These findings support the potential of artificial intelligence to expand access to heart disease screening at scale. To enable further development and transparency, we have publicly released model weights and a large, annotated dataset linking heart rhythm data to imaging-based diagnoses.

BACKGROUND:Substantial data support a heritable basis for supraventricular tachycardias, but the genetic determinants and molecular mechanisms of these arrhythmias are poorly understood. We sought to identify genetic loci associated with atrioventricular nodal reentrant tachycardia (AVNRT) and atrioventricular accessory pathways or atrioventricular reciprocating tachycardia (AVAPs/AVRT).METHODS:We performed multiancestry meta-analyses of genome-wide association studies to identify genetic loci for AVNRT (4 studies) and AVAP/AVRT (7 studies). We assessed evidence supporting the potential causal effects of candidate genes by analyzing relations between associated variants and cardiac gene expression, performing transcriptome-wide analyses, and examining prior genome-wide association studies.RESULTS:Analyses comprised 2384 AVNRT cases and 106 489 referents, and 2811 AVAP/AVRT cases and 1,483 093 referents. We identified 2 significant loci for AVNRT, which implicate NKX2-5 and TTN as disease susceptibility genes. A transcriptome-wide association analysis supported an association between reduced predicted cardiac expression of NKX2-5 and AVNRT. We identified 3 significant loci for AVAP/AVRT, which implicate SCN5A, SCN10A, and TTN/CCDC141. Variant associations at several loci have been previously reported for cardiac phenotypes, including atrial fibrillation, stroke, Brugada syndrome, and electrocardiographic intervals.CONCLUSIONS:Our findings highlight gene regions associated with ion channel function (AVAP/AVRT), as well as cardiac development and the sarcomere (AVAP/AVRT and AVNRT) as important potential effectors of supraventricular tachycardia susceptibility.

Dilated cardiomyopathy (DCM) is a leading cause of heart failure and cardiac transplantation. We report a genome-wide association study and multi-trait analysis of DCM (14,256 cases) and three left ventricular traits (36,203 UK Biobank participants). We identified 80 genomic risk loci and prioritized 62 putative effector genes, including several with rare variant DCM associations (MAP3K7, NEDD4L and SSPN). Using single-nucleus transcriptomics, we identify cellular states, biological pathways, and intracellular communications that drive pathogenesis. We demonstrate that polygenic scores predict DCM in the general population and modify penetrance in carriers of rare DCM variants. Our findings may inform the design of genetic testing strategies that incorporate polygenic background. They also provide insights into the molecular etiology of DCM that may facilitate the development of targeted therapeutics.