Investigation and development towards a clinically practical proton computed tomography system for treating cancer

Cotterill, John Vincent (2021). Investigation and development towards a clinically practical proton computed tomography system for treating cancer. University of Birmingham. Ph.D.

Preview

Cotterill2021PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (26MB) | Preview

Abstract

Imaging of a patient for the planning of a proton therapy treatment for cancer is currently conducted using x-ray Computed Tomography (CT). Uncertainties arise from the conversion of the measured parameters of the CT to proton stopping powers required for treatment planning. This conversion results in a 3.5% uncertainty on the range of the proton and is equivalent to a 1.2cm uncertainty on a 230MeV proton beam in water. This uncertainty on the distal edge of the beam can result in excess dose to healthy tissue or insufficient dose to the tumour.

The uncertainties associated with this image conversion may be removed if the proton stopping powers are measured directly using proton Computed Tomography (pCT). No such system currently exists in operation clinically. The high spatial density of incident protons, high current, and large energy range, make designing a system difficult, especially whilst remaining practical for mounting on a rotating gantry.

This work describes studies into proposed solutions and specifications of a device suitable for clinical operation. Individual proton tracking is achieved using silicon strip trackers; two placed before the patient, and two after. Each tracker consists of three silicon strip sensors, in an x-u-v orientation, with each layer rotated at 120 degrees with respect to each other. Simulations were performed using Geant4 to optimise the strip pitch and silicon thickness through examination of the achievable position and tracking resolutions. These studies found that a strip pitch of 200 micrometres and a silicon thickness of 150 micrometres was optimal for the tracking system when considering additional clinical constraints.

One proposed solution for a more compact system was the Hybrid Approach, in which the energy is measured using a single layer of silicon pixels. The average signal for each pixel was assigned to every proton traversing it, from which the Water Equivalent Path Length (WEPL) could be inferred. This system configuration was examined using simulation and experiment, with promising results for the approach, but distortion to the internal structures of an imaged object in a 2D proton radiograph was observed. This reduction in image quality was confirmed to originate from the proton scatter.

A deeper examination of the proton scatter led to the development of an alternative Scattering Approach to pCT. The device using this approach is more clinically practical as only proton tracking devices are required. Results from both simulation and experiment showed significant improvements to the position of internal structures in 2D radiographs in comparison to the Hybrid approach. A 3D reconstruction algorithm was also developed to produce a pCT, and compared with an x-ray CT, both using experimental data. The reconstructed image showed internal structure on a 200 micrometre scale, an improvement of five times that which is typically found clinically with x-rays.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):