Degree

Doctor of Philosophy (PhD)

Department

Physics and Astronomy

Document Type

Dissertation

Abstract

Healthy vasculature is critical to sustaining the function of normal tissues in the human body. Radiation therapy for cancer causes injury of the vasculature of non-cancerous tissues. These changes have been associated with potentially deadly conditions such as necrosis of the brain tissue. There are currently no computational methods to study the effects of radiation vascular injury in whole-organ vasculatures because of the large number of vessels involved. The goal of this work was to test the feasibility of simulating radiation damage to whole-brain vascular networks and calculate the resulting change in blood flow. To accomplish this, we developed algorithms to create a fractal-like geometry with 17 billion vessels, simulate the radiation dose to those vessels, and calculate the resulting change in blood flow. Computational performance metrics were measured for each algorithm individually and as a complete pipeline to determine the computational feasibility. Using a modular system containing these algorithms, we demonstrated that it is computationally feasible to predict the effects of radiation on blood flow in whole-organ vasculature. The system required 90 hours to perform the simulation for 2 million protons incident on an 8.5 billion vessel network using 128 compute nodes. Furthermore, the dose calculations were determined to be the most time consuming part of this system. The vessel-geometry algorithm and blood-flow algorithm both demonstrated the ability to reach 17 billion vessels. With future improvements, whole-organ simulations of vascular injury have the potential to elucidate the importance of vasculature to the development of radiation late effects.

Date

5-28-2019

Committee Chair

Newhauser, Wayne D.

DOI

10.31390/gradschool_dissertations.4941

Download

Share

COinS