Jianxin Zhou, Ph.D. Candidate
Dr. Angela Di Fulvio, Director of Research
August 12, 2024 | 9:00am - 11:00am CST
This final examination will be held in 101A Talbot Laboratory.
Zoom: Meeting ID: 984 956 7087 | Password: 982527
Machine-learning models and shielding materials for radiation protection of patients and personnel in ion therapy
ABSTRACT: Compared to traditional X-ray therapy, ion therapy improves dose conformality, increases biological effectiveness, and spares healthy tissues surrounding the tumor volume. However, widespread adoption of ion therapy is prevented by technological challenges pertaining to the radiation protection of patients and personnel, including 1) optimal tumor volume coverage and minimization of stray radiation delivered to healthy tissues, 2) implementation of adaptive treatment to account for anatomy changes, and 3) shielding of highly-penetrating secondary radiation. The work described in this dissertation focuses on proton therapy and aims to improve these three aspects enabling a safer and more effective treatment workflow.
Proton beams exhibit the highest energy deposition, i.e., the Bragg peak, at the end of their path in tissue, which is also referred to as its range. It is important to verify the location of the range during the treatment to ensure tumor coverage. This dissertation investigates the feasibility of using a novel organic scintillator, deuterated stilbene (stilbene-d12), capable of detecting and discriminating both gamma rays and neutrons to perform real-time range verification. The simulated results demonstrate that the fall-off of the remotely measured neutron flux can be used to locate the beam range with an uncertainty of 1~mm (1 standard deviation).
In proton therapy, the Bragg peak needs to achieve full coverage of the tumor volume to avoid overdosing organs at risk (OAR) surrounding the tumor while not treating the target volume effectively. Therefore, accurate organ contouring throughout the treatment is needed. Adaptive radiation treatment (ART) can account for patients' anatomy changes during the whole treatment by repeating the imaging and organ contouring for each treatment session. However, APT is impractical in most medical centers, due to the long manual contouring time. We developed a volumetric segmentation model based on V-Net to automatically contour organs. We improved its computation time performance by replacing the voxel-based data format with a point-cloud format and implementing a point-cloud-based segmentation model. Our model achieves an 89% average segmentation accuracy of the prostate and a 1.5-second segmentation speed, outperforming voxel-based state-of-the-art segmentation models.
Proton beams can produce highly penetrating secondary neutrons and gamma rays during the treatment by interacting with the patient and surrounding structure. Therefore, adequate and specific radiation shielding is needed to protect personnel. In this dissertation, we investigated the feasibility of using novel geopolymer (GP)-based materials for improved and custom radiation shielding. This work enhances the radiation shielding ability of GP with the addition of suitable dispersants. We demonstrated that tungsten, polyethylene, and boron-loaded GP composites outperform high-density concrete for gamma ray, fast neutron, and thermal neutron shielding, respectively. Therefore, using the developed GP composites as shielding materials can significantly reduce the construction footprint of ion therapy units.