AI-powered Digital Twin for Personalised Vascular Intervention Planning

Current clinical decision-making in vascular interventions relies heavily on physician experience and static imaging, making it difficult to apply clinical guidelines in a patient-specific and consistent way. This can lead to variability in treatment choices, suboptimal device selection and increased risk of complications. There is a clear need for tools that transform medical imaging and clinical data into actionable insights for personalised decision-making.

We aim to develop and clinically validate an AI-powered Digital Twin platform for personalised vascular treatment planning. Starting from CT imaging, the system reconstructs a patient-specific 3D model and simulates the interaction with different endovascular medical devices. The project will extend current capabilities by integrating new device classes and enabling advanced haemodynamic and biomechanical simulations to assess blood flow, wall stress and potential risks. By combining AI, physics-based modelling and clinical data, the platform provides quantitative indicators to support clinicians in selecting the most appropriate device and intervention strategy.

The technology is currently at TRL 6, with a prototype validated in research settings. The project aims to reach TRL 8 through multicentric clinical validation, integration into hospital workflows and deployment in real-world clinical environments.

The project will generate real-world data and evidence, improving decision-making, reducing variability and enhancing patient outcomes, while contributing to cost reduction and healthcare efficiency.

The solution aligns with PRECISEU priorities in health data and digital health, leveraging interoperable data infrastructures and AI for personalised medicine, promoting equity of access by standardising advanced decision-support tools across healthcare systems. It will accelerate the adoption of digital twin technologies in clinical practice, bridging the gap between research and real-world implementation.