CADENCE aims to develop advanced tools and AI models to better exploit quantitative Rubidium Positron Emission Tomography (PET) for cardiac imaging. Left ventricle Myocardial Perfusion Imaging is a powerful non-invasive tool to detect and localise heart diseases, widely used in clinics but generally simplified to a flattened and distorted 2D representation. This project aims to develop a fully automatic pipeline for 3D ventricle analysis from segmentation to multi-depth perfusion analysis. The resulting meshes are then used with graph neural networks to fully exploit the heart morphology for a variety of tasks.
The main goals of the project are as follows:
– Automatic Rubidium PET processing pipeline
– Major Adverse Cardiac Events prediction
– Rubidium only Viability prediction
– Epi-Endocardial perfusion analysis for diverse clinical tasks
Project Members:
-HES-SO Valais
-University of Lausanne (Pr. John Prior)
-CHUV (Pr. John Prior, MD Mario Jreige, MD Christel Kamani)
Team Members:
Adrien Depeursinge
Arthur Chevalley
Maria Martin Asiain
Acknowledgements:
This project has received funding from the Swiss National Science Foundation (SNSF) no. 219430