Novel geometric deep learning methods for patient-specific risk assessment and progression of cardiovascular diseases based on multimodal imaging techniques
Organization:
Funded by: | VASCUL-AID (Horizon Europe) |
PhD: |  |
Supervisors: | chair MIA:  Daily supervisor:  |
Collaboration: |
Description:
The aim of VASCUL-AID is to predict the risk of cardiovascular events and progression of the vascular diseases abdominal aortic aneurysm (AAA) and peripheral arterial disease (PAD). This information will be used improve the patient’s quality of life and care and assisting clinicians to make better-informed decisions involving the patient.
Automatic analysis and outcome prediction of cardiovascular diseases based on medical images is a crucial step in personalized treatment. In recent years, geometric deep learning (GDL) has been shown to be a very efficient approach to modelling patient-specific hemodynamic and biomechanical processes. These techniques can work directly on the 3D data represented in the form of meshes, point clouds or 3D spatial graphs as well as exploiting the symmetries thereof for better efficiency.
In this project, we aim to develop novel geometric deep learning methods integrating various imaging techniques for risk and progression assessment of cardiovascular diseases, in particular AAA and PAD. We will look into deep learning models that utilize geometric and physics-based features to accurately model hemodynamics and biomechanics in (automatically extracted) 3D shapes of vasculatures. Moreover, we aim to study the temporal aspect of the disease progression by taking into account short-term (heart cycle) and long-term (longitudinal data) changes in vasculature shape and its biomechanics. Together with 15 international partner organizations, we aim to leverage geometric deep learning to improve personalized treatment of CVD patients.
Output:
Pictures:
