Predicting thrombus formation the left atrium using mechanistic modelling and gaussian processes
Dr Adelaide De Vecchi ( Department of Digital Twins for Healthcare, King’s College, London )
- 14:00 14th March 2025 ( week 8, Hilary Term 2025 )(this is a virtual seminar)
Thrombus formation in the left atrium is a multiscale, multifactorial process based on the interplay between blood flow and structural dynamics in the myocardium. Here we present our work on modelling both the patient-specific blood flow, the myocardial wall dynamics, and the biochemical reactions that occur once the coagulation cascade is activated. We also discuss how probabilistic approaches such as Gaussian Process Emulators can be used to estimate likelihood of thrombus formation from clinical and imaging data. The goal of this work is to develop a personalised computational framework to support and complement the clinical criteria currently used to stroke risks by incorporating personalised predictions of function.
Speaker bio
Since completing my PhD in aeronautics at Imperial College London, I have been working on precision cardiovascular medicine, first at the University of Oxford, then at King’s College London. This defined my research interest in data-driven models for disease prediction. I am now Head of the Department of Digital Twins for Healthcare at King’s, which focuses on the development of Digital Twins for precision medicine combining machine learning with probabilistc and deterministic modelling appraoches. My specific research interests include modelling risks of thrombus formation, predicting vascular remodelling in pediatric congenital heart diseases and machine learning based assessment of pulmonary hypertension.Share this:
