Lin, Jingying2021-09-242021-09-242020-06https://hdl.handle.net/11299/224604University of Minnesota Ph.D. dissertation. June 2020. Major: Biophysical Sciences and Medical Physics. Advisors: E. Ritenour, Jafar Golzarian. 1 computer file (PDF); viii, 139 pages.Embolotherapy is an effective interventional therapy for patients with intermediate-stage hepatocellular carcinoma. Taking advantage of hypervascular characteristics from hepatocellular carcinoma, the embolic agents injected into hepatic arteries via a catheter block the tumor-feeding arteries. The degree of targeted embolization depends on how successfully the embolic agents are transported to the tumor-feeding arteries and lodged near the tumors. In addition, being able to minimize and detect non-targeted embolization is also a major concern for radiologists. Due to limitations on the ability to monitor blood dynamics, the influence of local hemodynamics near a tumor during embolization is not well understood. Since blood flow is the only propulsive force on the embolic agents after they are released into the hepatic arteries, hemodynamic changes caused by the interaction between embolic agents and blood flow is crucial to the success of embolotherapy. Understanding the hemodynamic effects of various factors and how they interact is the key to linking this complex situation to treatment outcomes and could assist radiologists in improving their practice. So far, the factors that influence the hemodynamics have been investigated through pre-clinical experiments and clinical case studies, such as catheter placement, and microsphere characteristics (size, quantity, and type), catheter type, microsphere injection techniques (injection timing, velocity, and interval). Utilizing computational analysis and simulation on particle-hemodynamics inside hepatic vascular models has been an effective preclinical approach to investigate the interaction among hemodynamic factors. One of the key components for reliable hemodynamic simulation is a geometric domain in which physiological boundary conditions can be applied to match the simulation-specified flow conditions. A geometric domain - also called geometric configuration, computational domain, or mathematical model - is defined as a three-dimensional (3D) vascular geometry. Currently, hepatic vascular geometries used for studying computational hemodynamic simulation are rare, have limited or truncated vasculature, and often lack the vasculature that microspheres are targeting. Existing imaging technology cannot provide a process for transmitting emboli. However, visualization of emboli transmission can be achieved through simulation models. In this dissertation, a set of procedures was developed to construct a visualized embolization simulation model based on detailed and coherent anatomic vascular structures. Aiming at the limitations of the existing vascular models, we built a vascular network model based on a real vascular structure with details and continuity to form the geometric domain necessary for the simulation model. First, tissue samples filled with contrast were established by selecting a commonly used preclinical animal model to ensure faithful vascular structure, and the samples were processed to meet the requirements of a detailed and coherent vascular network. Due to limitations of the imaging system, multi-resolution micro-CT images were used to model the microvasculature by modeling the vasculature in each individual liver lobe and registering them back to the entire liver to achieve fidelity and detailed spatial vasculature under physiological conditions. Finally, the vascular modeling method, including vascular segmentation and vascular centerline extraction, was used to automatically generate the vascular network. The ability of an imaging system to faithfully reproduce an image of an object is described by the modulation transfer function (MTF). The MTF of an imaging system is a measure of how well sine waves of various frequencies that describe the spatial distribution transmission of X- rays through an object are represented faithfully in the image. An ideal MTF is a horizontal straight line, which means that no matter how an object’s size varies, the imaging system would modulate and deliver a faithful profile of the object onto the image. However, the MTF of the micro-CT scanner reflects a Gaussian blurring function. This blurring function convolves with the object’s true cross-sectional profile, resulting in the blurred output: as the object becomes smaller, the intensity peak of the object is reduced significantly, with a broadened full width at half maximum (FWHM). In our reconstructed vascular model, the over-estimation of radii on smaller vessels was corrected by the Gaussian correction function. To evaluate the accuracy of the correction, registration of partial microvasculature from higher resolution micro-CT to the individual lobes were studied to quantify the comparison of the vessel segment radius between the corrected size and the measurements from high-resolution images. Because liver tumors have hypervascularity features, modeling the cancerous vasculature is particularly important. The purpose of this study is to investigate the feasibility of modeling the vascular network for a tumor vascular model that can be used for blood flow simulation to aid embolization. To visualize the real-time trajectories of embolic agents as they are transported inside arteries, a computer model based upon micro-CT images of hepatic microvasculature in a rabbit was constructed. Under the effect of simulated blood flow, a particle-based simulation model that provides embolic agents’ transportation within the constructed model was then developed. In the final model, the simulated transmission of emboli is visualized by a web-based application. This application can be used not only to directly visualize the transmission of emboli but also to analyze hemodynamics in different situations, providing a reference for actual clinical operations. The work presented here is intended to provide a basis for building visual simulation models for individual human patient’s unique vascular structures at some point in the future.en3D hepatic vascular modelHemodynamic flow simulationHepatocellular carcinomaMicro-CT imagingTransarterial embolizationVessel segmentationThesis or Dissertation