Access Type

Open Access Dissertation

Date of Award

January 2017

Degree Type

Dissertation

Degree Name

Ph.D.

Department

Medical Physics

First Advisor

Charles D. Bloch

Second Advisor

Jacob Burmeister

Abstract

Purpose: Accurate dose calculation is one of the most necessary components of radiation therapy. While the commercially available photon dose calculation algorithm offerings have improved considerably in last decade, proton dose calculations are still performed using the analytical dose calculation algorithms. The goal of this work is to validate a newly available commercial Monte Carlo (MC) dose calculation algorithm using measurements and simulations in GATE software. A secondary goal is to compare and contrast the performance of analytical algorithm against MC algorithm. Finally, GATE simulations are used to evaluate a newly available ceramic marker for ocular melanoma proton therapy.

Methods: An analytical and MC beam models of a full 360 degree gantry at SCCA proton therapy center were commissioned in RayStation treatment planning system. Measurements were performed using variety of detectors such as parallel plate ion chambers, 2D ion chamber arrays, Bragg peaks chamber, 2D high resolution scintillation imager, and radiographic film. The analytical beam model was put to a series of tests that involved verification of point doses, PDDs, profiles, and doses in patient specific plans. The comparison of analytical and RayStation MC (RS-MC) algorithm was carried out by measurements in homogenous, heterogeneous, and anthropomorphic phantoms. For comparisons against simulations, a beam model was developed in GATE MC Toolkit using the measured beam data. For evaluation of ceramic marker, a custom phantom with styrofoam insert with embedded marker was created. Simulation and measurements were made for marker with clip in parallel, perpendicular and transverse orientation relative to beam.

Results: For the analytical algorithm, evaluation of point doses in water showed dose differences>3% for proton ranges>30 cm, field sizes>15 x15 cm2, and depths>25 cm. When a range shifter was employed, analytical algorithm showed dose up to 10% dose difference in the entrance region for air gaps>30 cm. In oblique beam conditions, analytical algorithm showed broadening in distal penumbra by up to 5 mm. RS-MC algorithm matched measurements and GATE simulations to within 3% at all points for SOBP depth doses, beam with a range shifter, and oblique incidence. RS-MC also predicted accurate doses in inhomogeneous phantom, where the dose profile created at the interface matched to measurements and simulations with 100% gamma index (GI) pass rate at 3% dose and 3 mm distance-to-agreement (DTA). In anthropomorphic phantom, 6/7 planes had GI> 90% using 3% dose and 3 mm DTA for RS-MC. Corresponding numbers for analytical algorithm showed only 3/7 planes with GI >90%. The ceramic marker showed considerable dose attenuation behind the marker that worsened when marker was placed close to the distal edge. The transverse marker orientation in the clinical SOBP beam showed dose reduction of up to 61% and range pull-back of 2.4 mm.

Conclusions: The RS-MC algorithm demonstrated improved dosimetric accuracy over analytical algorithm in presence of homogenous, heterogeneous and anthropomorphic phantoms. The computation performance of RS-MC was similar to RS-PBA algorithm. For complex disease sites like breast, head and neck, and lung cancer, the RS-MC algorithm will provide significantly more accurate treatment planning. The dose attenuation behind ceramic marker was found to be a function of marker orientation and location within the SOBP. The analysis showed that small volumes behind the marker can see severe under-dosing. This effect should be taken into account during the treatment planning.

Download

Share

COinS