Date of Award

Fall 2023

Document Type


Degree Name

Master of Science (MS)


Biomedical Engineering

First Advisor

Taly Gilat-Schmidt


Computed Tomography imaging is an important diagnostic tool but carries some risk due to radiation dose used to form the image. Currently, CT scanners report a measure of radiation dose for each scan that reflects the radiation emitted by the scanner, not the radiation dose absorbed by the patient. The radiation dose absorbed by organs, known as organ dose, is a more relevant metric that is important for risk assessment and CT protocol optimization. Tools for rapid organ-dose estimation are available but are limited to using general patient models. These publicly available tools are unable to model patient-specific anatomy and positioning within the scanner. To address these limitations, the Personalized Rapid Estimator of Dose in Computed Tomography (PREDICT) dosimetry tool was recently developed. This study validated the organ doses estimated by ‘PREDICT’ with ground truth values. The patient-specific PREDICT performance was also compared to two publicly available phantom-based methods: VirtualDose and NCICT. The PREDICT tool demonstrated lower organ dose errors compared to the phantom-based methods, demonstrating the benefit of patient-specific modeling. This study also developed a method to extract the walls of cavity organs, such as the bladder and the intestines, and quantified the effect of organ wall extraction on organ dose. The study found that the exogenous material within the cavity organ can affect organ dose estimate, therefore demonstrating the importance of boundary wall extraction in dosimetry tools such as PREDICT.