SUBJECT: Ph.D. Dissertation Defense
BY: Tiffany Tsui
TIME: Friday, March 17, 2023, 7:00 a.m.
PLACE: Virtual, Virtual
TITLE: An Investigative Study of a Novel Technique of Material Decomposition for Dual Energy Cone Beam Computed Tomography - Technique and Clinical Applications
COMMITTEE: Dr. Chris Wang, Chair (NRE/MP)
Dr. Tianye Niu (NRE/MP)
Dr. Eric Elder (Emory Radiation Oncology)
Dr. Mohammad Khan (Emory Radiation Oncology)
Dr. Sean Cavanaugh (GensisCare Radiation Oncology)


As radiation therapy (RT) techniques advance and the modalities become more complex (i.e., high dose per fraction, higher dose rate, and more conformity with steeper isodose lines), accurate image-guided RT methods become highly desirable. Kilovoltage (kV) cone-beam computed tomography (CBCT) is a common volumetric guidance tool before and during RT. It offers great soft-tissue contrast and submillimeter spatial resolution. Despite the widespread usage of kV CBCT, there are still areas to improve (e.g., scatter, soft-tissue delineation, and low contrast detectability). Dual-energy CBCT (DECBCT), using the multi-material decomposition (MMD) technique, has the potential to improve image quality by determining electron densities more accurately compared to single-energy CT images. The potential of DEDECT in RT was investigated in four clinical applications, including: (1) enhancing pre-treatment IGRT image quality, (2) quantitative evaluation of invasive breast cancer biomarkers, (3) hydrogel placement for better general pelvis anatomy definition, and (4) adaptive radiotherapy. In this study, an image-domain dual-energy MMD technique was used, with the condition of material sparsity assumption. The material sparsity constraint (MSC), which is represented by the L0 norm, operates under the condition that each voxel of the CBCT image contains no more than two different basis materials. The MSC method, which is validated on phantom data, shows a higher volume fraction accuracy (VFA) from 92.0% to 98.5% as compared to a traditional two-material assumption (TMA) method. The high decomposition image quality obtained from the new techniques, therefore, successfully demonstrated a promising outcome for the four clinical applications listed above.