Advancing Endovascular Management of Thoracic Aortic Disease

Abstract: Despite technological advances and new endograft designs, endovascular treatment of the thoracic aorta still has important limitations. The aims of this thesis were to obtain further understanding on specific limitations in treating complex thoracic aortic pathologies with current technologies, to gain more insights into the hemodynamic consequences of thoracic endografting, and to explores the potential of deep learning algorithm-based automatic assessment of follow-up CTA imaging.Limitations imposed by the aortic arch branches on the proximal landing zone (PLZ) remain the biggest challenge for thoracic endovascular aortic repair (TEVAR). When left subclavian artery (LSA) preservation is required to obtain an adequate PLZ, single-branched endografts, such as the thoracic branch endoprosthesis (TBE), offer a complete endovascular solution, but still are limited by the lack of understanding of the hemodynamic effects and long-term data on clinical performance. Paper I explored the impact of TBE implantation on the LSA hemodynamics using computational fluid dynamic (CFD) analysis. It was shown TBE implantation produces modest hemodynamic disturbances which are unlikely to result in clinically relevant changes.Paper II evaluated the anatomic feasibility of TBE in blunt traumatic aortic injury patients who would require LSA revascularization. Only 32% of these patients had met all the anatomic requirements, justifying the need for additional designs. It was also shown that significant morphologic differences in arch anatomy exist between thoracic aortic pathologies.A major challenge in treating aortic dissection is unsatisfactory compliance of the distal portion of stent-grafts in a dissected aorta. Paper III evaluated the hemodynamic effect of a novel dissection-specific stent-graft (DSSG) with the aim to prevent distal stent graft-induced new entry (dSINE). The CFD analysis showed changes in shear-stress distribution different from that with standard thoracic stent-graft, transitioning high wall shear stress gradient zones into the stent-covered aorta. This transition may help prevent intimal injury and consequent dSINE development.Despite close surveillance imaging, early signs of endograft component separation resulting in type IIIa endoleak and sac repressurization are easily missed. Paper IV developed an AI-assisted fully automated CT image assessment method for early detection of endograft component separation and explored its potential.In depth understanding of the limitations of TEVAR helps continue progress in providing optimal patient- and pathology-specific endovascular solutions. CFD-based hemodynamic analysis and AI-driven automated image assessment have the potential to aid in TEVAR optimization.