Biomechanical Modelling

Computational modelling and simulations are paramount in a vast range of medical applications.

Recent advances in personalised healthcare require computational tools combined with the individual anatomy and physio-pathological data to support diagnosis, therapeutic inference, surgical planning and device design.

Biomedical fluid mechanics is also rapidly advancing, driven by unprecedented volumes of data from field measurements and experiments at multiple spatiotemporal scales.

Machine learning and AI offer a set of tools to extract information from large-scale data to be translated into knowledge about the underlying vascular fluid mechanics and about the associated predictors and biomarkers of clinical relevance.

The aim of this project is to combine traditional methodologies employed in computational modelling and biomechanical simulation with novel machine learning approaches for interventional neuro- and cardio-vascular applications, towards risk prediction, long-life monitoring and personalised healthcare. In brief, we are developing a state-of-the art deep learning algorithm to extract the vascular tree from the MRA/CTA volumetric image. This tree will later be represented in terms of vertex and edges. Then, we apply a mathematical model (continumm fluid mechanics) to simulate cerebral blood flow along this graph representation.


Elastic registration of geodesic vascular graphs

Moriconi, S., Zuluaga, M.A., Jäger, H.R., Nachev, P., Ourselin, S. and Cardoso, M.J., 2018, September. In International Conference on Medical Image Computing and Computer-Assisted Intervention (pp. 810-818). Springer, Cham.