Year

2015

Degree Name

Doctor of Philosophy

Department

School of Physics

Abstract

Intraocular melanoma is a rare form of cancer with a high mortality rate and limited effective treatment options. It causes death in 45% of patients within 15 years, typically within 6 - 9 months of the first signs of metastatic spread. Plaque brachytherapy is one of the most common treatments, and has proven to be quite successful at local tumour control. It involves temporarily stitching a small, radionuclide-bearing device, called an eye plaque, on the surface of the eye above the tumour.

The accuracy of treatment planning and dose verification procedures remains suboptimal, due to the very high gradient of the radiation field immediately adjacent to the sources. Ideally, the dose distribution should be directly measured using the actual plaque and brachytherapy sources via a high-resolution spatial dosimetry sensor. The reverse-biased silicon diode has proven to be particularly well suited to real-time, accurate dosimetry due to the high tissue-equivalence of silicon.

The current recommended treatment planning process is based on TG-43U1 formalism, which assumes infinite water surroundings and does not account for material heterogeneities. Plaque brachytherapy treatment, however, includes regions of high Zeff materials, including bone and the plaque, which have been shown to significantly affect dose. An accurate dose verification system for plaque brachytherapy would have great potential for improved treatment efficacy and reduced damage to healthy tissue through dose optimisation and encouraging patient-specific plaque customisation.

The project aimed to develop a high-precision, volumetric dosimetry system, the PANOPTES, based on silicon detector technology. Two variants of the concept were developed and validated experimentally; the first with a single-diode detector, and the second with a pixelated detector array.

When tested with a brachytherapy plaque fitted with ten 125I seeds, the dose measured with the single-diode detector was within 4% of the TG-43U1 formalism at distances greater than 3.0 mm from the plaque.

Medipix detectors demonstrated the potential of using pixelated detectors and spectroscopic dosimetry in 125I plaque brachytherapy by showing a correlation within 1% between event counting and charge modes. This illustrated that event counting would be a feasible dosimetry technique in the subsequent design of the PANOPTES.

Depth-dose measurements using the PANOPTES showed agreement within ~ 4% with Monte Carlo simulations, Plaque Simulator software and the TG-43U1 formalism along the central plaque axis. A volumetric dosimetry tool, such as the PANOPTES, could offer much to improve current dose verification and treatment planning practises in eye plaque brachytherapy.

Share

COinS