Project description: In radiation therapy of cancer, cone beam CT images are used to position the patient immediately before each treatment session. Cone beam CT images are acquired by directing a cone of kilovoltage x rays at the patient from different angles during a full 360-degree rotation. At each angle, the x-rays transmitted through the patient are detected using a flat panel detector past the patient. The data from all angles are fed into image reconstruction software to reconstruct the 3D volume image of the patient. The "cone" nature in cone beam CT imaging makes the images suffer from a large scatter component and a higher patient imaging dose. This project aims to  develop a prototype that explores the relatively novel concept of "volume of interest" cone beam CT imaging where the x-ray beam is dynamically collimated to the patient areas of interest, and the image reconstruction is modified to handle missing information due to the collimation during acquisition. The project has a strong experimental component to develop the prototype and to interface it with the state-of-the-art radiation treatment facilities at The Ottawa Hospital Cancer, as well as programming component using Python.


Supervisor: Ali Elsayed, elaliattoh [dot] ca