Digital twin of aortic dissection – DigiTAD

The ambition of the DigiTAD project is to develop a digital twin of residual aortic dissection. To achieve this goal, original quantitative medical imaging processing combined with advanced patient-specific numerical modelling will be associated with innovative deep learning approaches. A database of patients with residual aortic dissections, which favorable or unfavorable evolution will be known a posteriori, will be built up through quantitative medical imaging acquisitions carried out by the 2 university hospital centers participating in the project. It will include demographic, clinical, morphological, hemodynamical and structural data with a three-year follow-up. Data from medical imaging – CT angiography, 2D+time flow MRI, 4D Flow MRI and cineMRI - will be used to define the morphology and fluid and structural boundary conditions of each patient. Combined with in situ X-ray microtomography mechanical characterization on thrombus, they will also be used to build a representative thrombus model if necessary. Advanced patient-specific 3D numerical modelling will then be implemented and validated using 4D MRI data. Deep learning approaches, using the numerical simulation results, will allow to obtain a digital twin of patients with residual aortic dissection fully biomimetic of their pathology. These original approaches will overcome the limits of long computing times involved in numerical modelling, which are a hindrance to their use in clinical practice and will enable rapid analysis of new patients. The DigiTAD project will be a proof of concept for the future development of a non-invasive in vivo diagnostic tool predicting the evolution of a patient with residual aortic dissection