Digital twins for stroke prediction and management

Stroke remains a leading global cause of mortality and disability, with optimal outcomes dependent on rapid risk identification, precise acute management, tailored interventions, and sustained rehabilitation. Digital twins, which are dynamic computational replicas of individual patients integrating multimodal data and mechanistic models, represent an innovative approach to simulating disease trajectories and treatment responses before real world implementation.

This narrative review synthesises current evidence on digital twin applications across the stroke care continuum, examining their role in risk prediction, acute decision support, clinical trial optimisation, intensive care monitoring, and personalised neurorehabilitation.

We explored publications from PubMed, Nature Portfolio journals, IEEE and ACM databases spanning 2021 to 2025, identifying relevant publications across technical, clinical, and implementation dimensions. Search terms combined “digital twin” with stroke-specific and enabling-technology terminology. The review integrates findings across technical, clinical, and implementation dimensions to provide a holistic perspective on this rapidly evolving field.

Emerging evidence demonstrates feasibility and early clinical utility of digital twins across multiple stroke phases. Applications include atrial fibrillation related embolic risk stratification through patient specific left atrial modelling, imaging informed cerebrovascular hemodynamics for acute triage, disease digital twin generators for trial enrichment, and robotic rehabilitation systems with adaptive learning. However, most studies remain at preclinical or early pilot stages, with few meeting strict real-time digital twin criteria requiring bidirectional data flow and continuous updating.

Digital twins offer transformative potential for stroke care through personalised risk assessment, hyperacute decision support, and adaptive rehabilitation. Translation to routine practice requires rigorous prospective validation, standardised data models, robust governance frameworks, and demonstrated clinical utility across diverse populations. A clear translational roadmap, from current component technologies to fully closed-loop systems, is essential to guide investment and prioritise research.