Master of Science - Research
Centre for Medical Radiation Physics - Faculty of Engineering
Cullen, Ashley James, Strip detector for high spatial resolution dosimetry in radiation therapy, Master of Science - Research thesis, Centre for Medical Radiation Physics - Faculty of Engineering, University of Wollongong, 2009. http://ro.uow.edu.au/theses/848
Radiation therapy is an established modality in the treatment of tumours. With treatments ever evolving and increasing in terms of their complexities, the need arises to ensure the best quality treatment is delivered; the survival of the patient relies upon it. A modern treatment such as Intensity Modulated Radiation Therapy employs steep dose gradients varying dynamically to deliver complex dose profiles, whilst the experimental Microbeam Radiation Therapy (MRT) involves the delivery of an array of intense beams tens-of-microns wide separated by several hundred microns. In both cases, conventional dosimetry is inadequate in providing both spatial and temporal information about complex dose profiles.
The silicon strip detector was created to fill this void in current dosimetry techniques. Designed to withstand the intense beam of a synchrotron wriggler x-rays whilst not significantly perturbing the beam, the detector provides linear sensitive volume elements two hundred microns in pitch. This enables the ability to perform high spatial resolution dosimetry in real time.
This thesis investigates the use of the silicon strip detector as an on-line dosimetry system for MRT with applications to clinical radiotherapy. Of particular interest is the distribution and magnitude of energy deposition within the detector and the perturbative effect the strip detector has on a synchrotron wriggler x-ray beam.
Monte Carlo simulations are performed to investigate the properties of radiation incident upon the detector. These seek to understand how a beam traversing the detector interacts with it, as well as the effects the detector has on the transmitted beam and its properties.
The energy deposition spectrum within the detector was found to be predominant at low energies of below 100 keV. The deposition of dose through the detector was found to be largely constant with depth through the central axis of a beam, dropping to ~10-3 of the central value at 50 µm off the central beam axis for an infinitesimally thin pencil beam and ~10-4 at 100 µm off-axis. Energy deposition laterally through water was determined as dominated by secondary electrons from the beam-edge to 150 µm, and Compton photons thereafter.
The depth dose of a MRT pencil beam was found to have an average decrease in dose of (1.44 ± 0.15) % (95% C.I.) when the strip detector was introduced into the beam. The probability of interaction of incident photons with the detector for the MRT spectrum was determined with GEANT4 and theoretically with a comparison made. The overall interaction probability of an MRT photon is (1.97 ± 4.43×10-4) % (95% C.I.).
A simulation to determine the PVDR (peak-to-valley dose ratio) in a MRT field was created, however only the peak dose could be determined due to an inadequate primary photon count. The peak dose was found to decrease by 1.41 ± 0.03 % (95% C.I.). Qualitative film measurements (deliberate overexposure of peak regions) displayed an increase of valley dose to film at 10 mm in water with the strip detector in the beam, but no such phenomenon at a depth of 1 mm. A simplified GEANT4 simulation was created to replicate such results with only five beamlets. Peak-to-valley dose ratio calculations from the simulation show no discernible effect at 1 mm depth, but a discernible increase in the PVDR at 10 mm depth; replicating experimental results.
Since charge collection across a semiconductor device is often complex and dynamically varies with the bias conditions. Ion beam induced charge collection (IBICC) studies seek to investigate the charge collection properties of the detector in various voltage-biasing conditions. It was found that the application of reverse bias to strips adjacent to that being read out reduced the effective sensitive volume of the read-out strip. This provides evidence for a proposal to incorporate biased guard-ring structures to prevent charge sharing, and improve the confinement of the sensitive volume.
Finally, clinical irradiations are performed with a highly collimated orthovoltage x-ray beam down to beam sizes of hundreds of microns. The spatial resolution of the detector in this configuration was found to be 500 µm. The efficacy of the detector in contemporary radiotherapy is also investigated through the use of a 6 MV linear accelerator’s photon beam. Excellent agreement was found between strip detector read-outs and reference data for the linear accelerator. A quadratic relation was found between dose rate and charge per strip.
02Whole.pdf (7155 kB)