Computational methods of modeling vascular geometry and tracking pulmonary motion from medical images
Abstract/Contents
- Abstract
- Modern anatomical medical imaging technologies, such as computed tomography and magnetic resonance, capture structures of the human body in exquisite detail. Computational anatomy is a developing discipline to extract and characterize the anatomy from images. Unfortunately, anatomical images do not reveal the functional behavior. Computational physiology shows great potential to link the structure-function relationship by considering both the anatomical information and the physical governing laws. The simulated physiology can be used to assess physiological states, and more importantly predict the outcomes of interventions. On the other hand, advances in the functional imaging techniques provide measured physiology information and should be utilized together with computational physiology. In the theme of computational anatomy and physiology, this dissertation describes computational methods of modeling vascular geometry for image-based blood flow computation and tracking pulmonary motion for image-guided radiation therapy. Blood flow computation is a useful tool to quantify in vivo hemodynamics. The essential first step is to model vascular geometry from medical imaging data. I have developed a new workflow for this task. The geometric model construction is based on 3D image segmentation and geometric processing. To represent the topology of the constructed model, I have developed a novel centerline extraction method. To account for compliant vessels, methods to assign spatially-varying mechanical properties of the vessel wall are also developed. The workflow greatly increases the modeling efficiency. The combination of the patient-specific geometry and wall deformation can enhance the fidelity of blood flow simulation. Image-based blood flow computation also holds great promise for device design and surgical procedure evaluation. Next, I have developed novel virtual intervention methods to deploy stents or stent grafts to patient-specific pre-operative geometric models constructed from medical images. These methods enable prospective model construction and may be used to evaluate the outcomes of alternative treatment options. Respiratory motion is closely related to the physiology of the lung. Finally, I have developed a novel framework to track patient-specific pulmonary motion from 4D computed tomography images. A large set of vascular junction structures in the lung are identified as landmarks and tracked to obtain their motion trajectories. This framework can provide accurate motion information, which is important in radiation therapy to reduce healthy tissue irradiation while allowing target dose escalation. This work demonstrates the importance of the geometry and motion modeling tools in computational anatomy and physiology. Accurate physiological information, whether simulated or measured, will benefit the diagnosis and treatment of various diseases.
Description
| Type of resource | text |
|---|---|
| Form | electronic; electronic resource; remote |
| Extent | 1 online resource. |
| Publication date | 2011 |
| Issuance | monographic |
| Language | English |
Creators/Contributors
| Associated with | Xiong, Guanglei | |
|---|---|---|
| Associated with | Stanford University, Program in Biomedical Informatics. | |
| Primary advisor | Taylor, Charles A. (Charles Anthony) | |
| Primary advisor | Xing, Lei | |
| Thesis advisor | Taylor, Charles A. (Charles Anthony) | |
| Thesis advisor | Xing, Lei | |
| Thesis advisor | Paik, David | |
| Thesis advisor | Zarins, Christopher K | |
| Advisor | Paik, David | |
| Advisor | Zarins, Christopher K |
Subjects
| Genre | Theses |
|---|
Bibliographic information
| Statement of responsibility | Guanglei Xiong. |
|---|---|
| Note | Submitted to the Program in Biomedical Informatics. |
| Thesis | Thesis (Ph.D.)--Stanford University, 2011. |
| Location | electronic resource |
Access conditions
- Copyright
- © 2011 by Guanglei Xiong
- License
- This work is licensed under a Creative Commons Attribution Non Commercial 3.0 Unported license (CC BY-NC).
Also listed in
Loading usage metrics...