The thoracic aorta is an anatomical structure that is subject to constant motion. Stent-grafting changes the deformation patterns on the vessel surface, and the nature of these changes is suspected to correlate with treatment outcome. However, they are currently only poorly understood. We propose a method for quantifying the changes in vessel motion caused by stent-grafting in the aortic arch. For this, we automatically build a surface model of the vessel and its movement from gated computed tomography sequences that depict the deformation of the thoracic aorta during the cardiac cycle. We reconstruct trajectories corresponding to the displacement of a dense sampling of the vessel surface. From these, we construct a manifold of vessel motion patterns. Models acquired before and after stent-grafting are registered and we analyze the changes to the deformation in the embedding space. Results for four clinical cases indicate that the local density in the embedding space is a marker of local deformation complexity, and sites of potential complications.1