Building a Hybrid Compton Camera System for Improving Medical Imaging Applications

Hart, Shanyn-Dee

Building a Hybrid Compton Camera System for Improving Medical Imaging Applications

Thesis / Dissertation

2025

Publisher

University of Cape Town

Department

Department of Physics

Faculty

Faculty of Science

Abstract

Compton cameras present a powerful advancement in gamma-ray imaging, offering en-hanced resolution and efficiency over conventional imaging methods for medical applications such as proton therapy and other targeted radiotherapies. Unlike mechanically collimated imaging systems, a two-stage Compton camera uses a combination of scatterer and absorber detectors to reconstruct gamma-ray paths through Compton scattering events, allowing for greater flexibility in capturing images from uncollimated gamma rays. This study investi-gated the development, optimisation, and performance of three novel two-stage Compton camera prototypes designed to advance gamma-ray detection specifically for medical imag- ing applications. The prototypes — referred to as CC1, CC2, and CC3 — each featured unique configurations of scatterer and absorber detectors, tailored to explore and maximise imaging capabilities. The CC1 design integrated a hybrid setup, combining a commercial Polaris-J CZT de-tector as the scatterer layer with a 2” × 2” LaBr3:Ce detector as the absorber. CC2 and CC3 incorporated novel MacroPixel LB-14x25c-SiPM-T scintillation detector assemblies using a va-riety of different physical configurations. CC2 was evaluated for three distinct geometries, utilising a 4 × 1 matrix of detector assemblies in both the scatterer and absorber layers. Mean- while, CC3 employed a 6 × 1 matrix in the scatterer layer with a 2” × 2” LaBr3:Ce detector as the absorber. To ensure optimal performance, each detector system — including the 2” × 2” LaBr3:Ce, LaBr3:Ce and SrI2:Eu MacroPixel LB-14x25c-SiPM-T, and Polaris-J CZT detector — was char-acterised for energy resolution, detection efficiency, and timing precision prior to their use as Compton camera imaging detectors. A digital data acquisition system was used to opti-mise the energy resolution through signal shaping, achieving measurements of 4.07(9)% for the 2” × 2” LaBr3:Ce detector, 3.41(20)% and 3.82(18)% for the LaBr3:Ce and SrI2:Eu detector assemblies, respectively, and 1.38(25)% for the CZT detector at 662 keV. The fast timing capabil-ities of the 2” × 2” LaBr3:Ce detector produced an excellent timing resolution of 218.08(11) ps. Characterisation indicated that the SrI2:Eu detector, while promising in terms of energy reso-lution, experienced significant pile-up effects at close source-to-detector distance, rendering it unsuitable for Compton camera applications. The detection performance of each prototype was evaluated through TOPAS Monte Carlo simulations, validated by experimental measurements with gamma-ray sources and proton beam tests conducted at iThemba LABS in Cape Town, South Africa. The simulations were well validated in terms of full-energy peak detection efficiency and energy resolution, al-though charge build-up effects led to minor discrepancies for the CZT and SrI2:Eu detectors. Key figures of merit were chosen to be image resolution, angular resolution, and double-scatter Compton event efficiency, used to assess each Compton camera design. Image resolu-tion determines the system's ability to produce clear, spatially accurate images, essential for pinpointing gamma-ray sources in medical imaging. Angular resolution, assessed by compar-ing the angle derived from the kinematics of the Compton scattering equation with the angle reconstructed from the geometrical positions of interactions, reflects the system's ability to distinguish between gamma-ray paths and accurately resolve closely spaced sources. Lower values of angular resolution translate to sharper images with less ambiguity, enhancing di- agnostic clarity. Efficiency, defined as the fraction of detected gamma rays that contribute to image formation, influences both imaging speed and quality. An iterative three-dimensional filtered back projection method was chosen as the optimal method of image reconstruction, a technique that calculates likely gamma-ray paths and reconstructs an image by refining these projections. The results showed that CC1 achieved reasonable performance for each figure of merit; however, timing synchronisation issues limited its clinical viability. The setup of the CC1 lay-ers included a Polaris-J CZT detector with an onboard data acquisition system, which was configured to send synchronisation pulses to the absorber's data acquisition system. Mi- crosecond timing delays were encountered and were corrected extensively using background subtraction, time-walk correction, and synchronisation techniques for the data acquisition sys- tems of CC1. Ultimately, it was not possible to track double scatter Compton camera events across CC1. Although CC2 and CC3 were assessed through simulation, their figures of merit demonstrated potential for superior imaging capabilities when compared to the other designs, primarily due to their innovative configurations and timing resolution. The 4 × 2 matrix of LaBr3:Ce MacroPixel LB-14x25c-SiPM-T detector assemblies in CC2 design exhibited the greatest promise for clinical application, balancing the benefits of low-voltage operation, modular scalability and Compton camera performance. In summary, while CC1 performed strongly in simulation, CC2's optimised geometry and modular design make it a suitable candidate for further development as a high-resolution gamma-ray imaging sys-tem, particularly in advanced medical diagnostics.