Optimized joint analysis of thoracic aortic 3D morphology and blood flow hemodynamics from MRI images – MAGNOLIA

Cardiovascular diseases at the early stages typically include alterations of the aorta, which is the largest artery connected to the heart and carries oxygenated blood towards the whole body. Indeed, each heartbeat generates mechanical stress and elevated pressures on the aortic wall that can be amplified by highly prevalent conditions such as hypertension, diabetes or obesity, and modify aortic geometry but also related wall elasticity and inner blood flow hemodynamics. However, to date, diagnosis, severity assessment and surgery decision-making in patients with aortic disease such as aneurysm are based on a simple geometrical measurement of aortic diameter from 2D medical images only, while rupture can still occur below recommended diameter thresholds. Our hypothesis is that exhaustive overview of aortic changes: function, wall tissue and blood flow hemodynamics, beyond and complementarily to geometry, would provide personalized and more accurate risk prediction, and thus better suited, individual patient management. In that setting, MRI is a unique opportunity to manage patients who can inherit such disease from their youngest age and require frequent follow-up: MRI provides non-invasive, radiation- and contrast agent-free evaluation of anatomy, function, tissue composition and hemodynamics, simultaneously within a 3D volume. However, MRI is associated with lengthy exams, high costs and long waiting lists. On the other hand, reading of the large and complex MRI image datasets can be challenging and tedious. The present proposal builds on our aim to accelerate aortic MRI, in terms of both image acquisition and analysis, in order to improve patient comfort as well as optimize radiologist and technologist time. More specifically, the objectives of MAGNOLIA are: 1) to design a software providing automated, reproducible, comprehensive and fast measurement of quantitative biomarkers of thoracic aortic 3D morphology, wall stiffness, hemodynamics and tissue characterization from MRI images, 2) to test recently available novel MRI techniques providing superior image quality or new information compared to current imaging methods. Our methodology will be trained on images acquired within the ongoing SEQUOIA protocol in healthy volunteers, and we will setup a prospective study, again in healthy volunteers, in order to evaluate feasibility and performances of the aforementioned latest MRI techniques. In the long term, such research will define the best compromise between scan and analysis time, and MRI diagnostic and prognostic value in aortic disease. MAGNOLIA will be led in Paris (Laboratory of Biomedical Imaging, Sorbonne University and Pitié-Salpêtrière hospital) by a young researcher who will collaborate with her coworkers to bring together complementary skills in MRI, image processing, artificial intelligence, software development and clinical expertise in radiology and cardiology. MAGNOLIA open-source, free software as well as tested and optimized MRI techniques will be used in future local MRI research studies, but also at the European and international levels through sharing with scientific community in order to favor MRI multicentric efforts and standardization. Findings will be reported in specialized journals and conferences, but communication with society is also planned.