Surgical Simulation for Vascular Interventional Radiology Procedures

Vascular diseases have been the leading cause of death worldwide. Interventional procedures are an increasingly promising therapy for treating vascular diseases, which are usually done by a guidewire-catheter combination under the fluoroscopic guidance. However, due to the complicated vascular network, bending of surgical instruments and the risk of vessel injury, these techniques need to be performed by highly trained and experienced specialists. Virtual reality based training of these procedures offers high flexibility and cost effective alternative. Furthermore, it allows training evaluation and accelerates learning process without risk to patients, therefore has distinct advantages than traditional training methods on animals or cadavers.;In order to build a high fidelity interventional simulator for physician training and surgery planning, accurate reconstruction of three dimensional vascular network, real-time simulation of angiographic medium propagation and physics-based simulation of interaction between surgical instruments and vessel wall are absolutely indispensable. Thus, first, a methodology for geometric vascular modeling is proposed. As the reconstructed models are essential for many subsequent applications such as deformable modeling and visualization, a series of methods are proposed based on the parallel transport frames in order to maintain high mesh quality of these models. An improved bifurcation modeling method and two novel trifurcation modeling methods are developed based on 3D Bezier curve segments in order to ensure the continuous surface transition at furcations. To solve the twisting problem caused by frame mismatch of two successive furcations, a frame blending scheme is implemented. A curvature based adaptive sampling scheme combined with a mesh quality guided frame tilting algorithm is developed to construct an evenly distributed, non-concave and self-intersection free surface mesh. In terms of surface mesh quality criteria, our methodology can generate vascular models with better mesh quality than previous methods.;Second, we extend our geometric modeling method for illustrative visualization of vasculature, which is an indispensable component in medical education and training. Illustration of vasculature accentuates depth perception and provides a specific manner to identify the branching pattern and topology of vascular structure, which is crucial for therapy planning and real surgery in order to give an effective treatment. With advanced GPU acceleration techniques including render to texture (RTT), framebuffer object (FBO) and fast image convolution, a real-time visualization can be achieved.;Third, an interactive simulation of angioplasty procedure is reported. To achieve an efficient modeling of soft tissue deformation and virtual device mechanics, mass spring models are adopted to construct the deformable models of vessel wall and stent. By designing a quasi-equilateral triangular mesh model of blood vessel and stent, a linear spring coefficients setting method is adopted to achieve the same accuracy compared with finite element method. With the employment of Physics Processing Unit (PPU), a real-time simulation of the interaction between blood vessel wall and surgical device is developed for vascular interventional radiology simulation.;Fourth, in order to clearly visualize vascular networks and the placement of instruments while treating the lesion, a physics-based simulation for angiography procedure is presented based on navier-stokes equation and semi-lagrangian method. The multi-scale vessel grid is reconstructed for flow distribution, and point sprites based rendering is adopted to preserve real-time visualization of the procedure. The experiments demonstrate that our results are more realistic compared to previous methods and are closer to the real angiography procedure.;Finally, the system of vascular interventional radiology simulator is discussed by integrating all presented techniques and designing a trackball mouse based hardware sensors. Training experiments demonstrate that the presented techniques benefit rapid development of realistic and interactive vascular interventional radiology simulators.