A COMPREHENSIVE IMAGE-DRIVEN FRAMEWORK FOR PATIENT-SPECIFIC ASSESSMENT OF THE THORACIC AORTA

Cardiovascular diseases remain the leading cause of death worldwide, and thoracic aortic aneurysm (TAA) is among the most life-threatening conditions. Current clinical practice largely relies on maximum aortic diameter, which alone is often insufficient to predict rupture or dissection risk. This thesis proposes a patient-specific, image-driven framework that integrates morphometry, hemodynamics, and biomechanics to improve assessment and management of TAA. The work is organized into five chapters. First, it introduces the clinical limitations of diameter-based criteria and the need for advanced computational approaches. Then, it presents an automated framework for dynamic morphometric analysis based on ECG-gated CT images, using a 3D U-Net for segmentation of the thoracic aorta and left ventricle. A moving boundary CFD approach is subsequently developed to provide more accurate hemodynamic simulations. A dual method for in vivo biomechanical characterization of the aorta is also proposed, combining flow-based and image-based techniques. Overall, this research advances personalized cardiovascular modeling, offering promising tools for improved diagnosis, risk stratification, and treatment planning in TAA.