Improving endovascular intraoperative navigation with real-time skeleton-based deformation of virtual vascular structures

Endovascular surgery requires acquisition of intraoperative X-ray images, exposing patient and surgical staff to a considerable dose of radiations. This disadvantage, and the difficulty in using bidimensional projective images, motivates the integration of endovascular navigators in the clinical practice. This paper presents a real-time vascular deformation system to enhance the standard static virtual environment usually used in endovascular navigation. Our approach is based on a skeleton representation of the virtual vessel, linked to the 3D vascular structure via vertex color masks applied on the target vascular branches. This method allows the usage of multiple partial skeletons, each with its own deformation function and linking strategy; so, we can model different kind of deformations due to several factors (e.g., heartbeat, breathing etc.). The system has been tested modeling the deformation of renal arteries due to patient breathing: showing a good visual realism, and ensuring the necessary updating frequency for real-time simulation.