Dynamic simulation and digital twins in atrial fibrillation: Complementary roles and current limitations.

Abstract

Atrial fibrillation (AF) management relies on established risk scores, with CHA₂DS₂-VASc supporting thromboembolic risk assessment and HAS-BLED identifying modifiable bleeding risks rather than determining anticoagulation eligibility. Although widely validated and guideline-endorsed, these population-based models provide limited personalization and do not capture temporal dynamics, patient heterogeneity, or evolving disease trajectories. Advances in artificial intelligence, machine learning, dynamic simulation, and digital twin methodologies may enable more individualized and temporally adaptive decision-making. This narrative review evaluates the emerging role of dynamic simulation and digital twins in AF care, focusing on their potential to complement established frameworks, the evidence supporting their use, and barriers to clinical implementation. Current applications remain largely limited to feasibility studies, mechanistic modeling, and procedural use cases such as patient-specific ablation planning. Machine learning models may modestly improve prediction over traditional risk scores but are generally static or only episodically updated. Fully adaptive digital twins capable of recurrent updating and individualized clinical guidance have not been validated in routine practice, and reported gains primarily reflect predictive accuracy or simulation fidelity rather than improved patient outcomes. Major barriers include fragmented data ecosystems, limited longitudinal multimodal datasets, lack of prospective validation, unresolved regulatory and ethical issues, limited interpretability, clinician trust, and workflow integration. Digital twins and dynamic simulation should therefore be viewed as complementary rather than replacement tools: machine learning may refine risk estimation, mechanistic simulation may support procedural planning and investigation, and adaptive digital twins remain investigational without an established clinical role in AF.