The RetiAF score demonstrated a robust performance across multi-ethnic, multi-center datasets, and the potentials of non-invasive retinal imaging as a scalable tool for AF and cardiovascular risk assessment, offering a promising alternative for large-scale screenings and personalized interventions.
Abstract
Atrial fibrillation (AF), a common cardiac arrhythmia, presents significant challenges for early detection and management due to its asymptomatic and paroxysmal characteristics. In this study, we introduce the RetiAF score, a multimodal foundation-model-based biomarker derived from retinal fundus images for early detection of AF. We identified that the RetiAF score demonstrated a robust performance across multi-ethnic, multi-center datasets, achieving AUROCs of 0.8610 and 0.8019 on the UK Biobank (UKBB) development and internal testing datasets, and an AUROC of 0.7803 on an external dataset acquired in Shanghai, China. In addition, we identified that the RetiAF score consistently outperformed the traditional risk scores such as CHARGE-AF (AUROC: 0.7553) and C2HEST (AUROC: 0.7246) for the UKBB internal testing dataset. Multivariable logistic regression and propensity score analyses further demonstrated that the RetiAF score was independently associated with AF risk (p < 0.001). When stratified by higher C2HEST scores (≥3), the RetiAF score achieved an AUROC of 0.9619, highlighting its potential for identifying high-risk patients before the clinical onset of AF. The multimodal hybrid version of RetiAF (Hybrid_RetiAF) score, which incorporated clinical features (e.g., Age, BMI, etc) into the deep learning model, further enhanced predictive performance and achieved AUROCs of 0.8924 and 0.8381 on the UKBB Cohorts. As a secondary exploratory analysis, we evaluated whether RetiAF-derived scores were associated with chronic ischemic heart disease in UKBB, suggesting shared cardio-retinal risk information. These findings underscore the potentials of non-invasive retinal imaging as a scalable tool for AF and cardiovascular risk assessment, offering a promising alternative for large-scale screenings and personalized interventions.
The EHR-based machine learning model, FIND-AF 2.0, identifies a high-risk subpopulation for AF diagnosis among patients at elevated risk of stroke and could enable scalable, EHR-driven, risk-guided AF screening.
R. Nadarajah, Jianhua Wu, A. Wahab et al.· Circulation· 0 citations
External validation of the GLORIA-AF Stroke Weighted Risk Score supports its transportability and potential adjunctive role in guideline-directed thromboembolic risk assessment and generally greater net benefit than CHA2DS2-VA.
Background Asymptomatic cerebral infarction (ACI) is a frequent yet under-recognized complication following radiofrequency catheter ablation (RFCA) in patients with atrial fibrillation (AF). Early identification of patients at risk remains challenging, particularly in the absence of clinically overt neurological symptoms. This study aimed to develop a multi-parameter predictive model integrating clinical, procedural, and biomarker-based factors, with a particular focus on vascular endothelial growth factor (VEGF). Methods This retrospective cohort study included 300 consecutive AF patients undergoing first-time RFCA. Brain magnetic resonance imaging was performed within 24–72 h post-procedure to detect ACI. Clinical, procedural, and laboratory data were systematically collected. Serum VEGF levels were measured using an enzyme-linked immunosorbent assay. Univariable and multivariable logistic regression analyses were conducted to identify independent predictors of ACI. Model performance was evaluated using receiver operating characteristic (ROC) analysis. Results ACI was detected in 48 patients (16.0%). Patients with ACI were significantly older and had higher body mass index compared to those without ACI (p < 0.001). Serum VEGF levels were markedly elevated in the ACI group (350 ± 45 vs. 200 ± 30 pg/mL, p < 0.001). Multivariable analysis identified age (OR: 1.08, p = 0.01), BMI (OR: 1.15, p = 0.003), and VEGF (OR: 1.02, p < 0.001) as independent predictors. The predictive model demonstrated good discriminative ability with an AUC of 0.82. Additionally, RFCA was associated with significant improvements in cardiac function and autonomic regulation (p < 0.001). Conclusion ACI remains a clinically relevant complication following RFCA. A predictive model incorporating VEGF alongside clinical factors provides improved risk stratification. These findings highlight the importance of endothelial dysfunction in ACI pathogenesis and support the integration of biomarkers into clinical decision-making.
Siliang Han, Chunhong Chen, Zhe Wang et al.· Frontiers in Cardiovascular...· 0 citations
Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia worldwide and is associated with substantial morbidity, including ischemic stroke, heart failure, and cognitive decline. Despite established diagnostic and therapeutic strategies, AF remains frequently underdiagnosed and suboptimally managed, particularly in asymptomatic or paroxysmal cases, in which the episodic nature of the arrhythmia may make it difficult to detect using standard electrocardiography. Artificial intelligence (AI), including machine learning and deep learning, has emerged as a transformative technology across multiple aspects of AF care. AI-based electrocardiographic analysis and wearable technologies have demonstrated promising performance in detecting subclinical AF and facilitating scalable population screening. Multimodal models integrating clinical, imaging, and electrophysiological data have demonstrated improved accuracy in stroke prediction, recurrence risk estimation, and therapeutic planning. AI-assisted imaging has advanced atrial segmentation, fibrosis characterization, and ablation planning, whereas AI-driven mapping systems may enhance procedural efficiency and improve patient selection. Nevertheless, significant challenges remain, including limited external validation, data heterogeneity, concerns regarding interpretability, integration into clinical workflows, and ethical and regulatory considerations. Future efforts should prioritize explainable, clinically validated, and human-centered AI systems that are integrated into real-world clinical workflows. This review provides a comprehensive overview of current AI applications in AF management, including early detection, risk stratification, cardiovascular imaging, clinical decision support, and interventional electrophysiology.
V. Cicek, M. Hayıroğlu, Vanshali Sharma et al.· Balkan Medical Journal· 0 citations
In conclusion, AI-derived risk estimates improved physician risk discrimination in a structured simulated survey, particularly in non-specialist settings, supporting their potential role as a digital decision-support tool.
Yeji Kim, Bogeun Kim, J. Yoon et al.· npj Digital Medicine· 0 citations
