Hranek, A. (2019). Experimental validation of the ripple filter effect of lung tissue equivalent materials in a proton beam [Diploma Thesis, Technische Universität Wien]. reposiTUm. https://doi.org/10.34726/hss.2019.69502
Ion beam therapy oers the advantage of an increased dose deposition at the end of the particle's range (Bragg peak (BP)), which results in a sharp fall o behind the tumour and potentially allows a better sparing of the surrounding healthy tissue. Lung cancer could pro t essentially from particle therapy by decreasing the dose to the ipsilateral lung and avoiding the contralateral lung completely. Besides the challenges of respiration motion, lung tissue has a very heterogeneous texture, which poses diculties for available dose calculation algorithms. This varying density aects the stopping power along the particles path and results in a spreading of the BP. Clinically used dose calculation algorithms implemented in treatment planning systems (TPS) are based on the stopping power information received from a patient's CT images. Their resolution is limited to about 1 mm while the size of inhomogeneities for lung tissue is in the sub-millimetre region. The aim of this thesis was to investigate the proton beam in heterogeneous materials mimicking lung tissue, quantify its properties, try to predict the eect on the beam and further compare them to Monte Carlo dose calculations of the TPS. Materials, which were suitable as lung substitute, were chosen based on their structure and density, namely cork (2 with dierent densities), konjac sponge, oral foam, pumice and balsawood, all available with dierent thickness. Additionally, lung tissue equivalent plates (CIRS, USA) were included representing the correct atomic composition but not representing the realistic texture of the heterogeneities. Integral depth dose (IDD) of a proton beam with 97.4 MeV and 148.2 MeV were measured in a water phantom (MP3-PL, PTW, Germany) with a reference ionisation chamber (Type 34080, PTW, Germany) and a eld ionisation BP chamber (Type 34070, PTW, Germany). The investigated materials were positioned in front of the entrance window of the water phantom and their IDDs compared to reference measurements without additional material in the beam path. From the dose curves dierent points of interest were derived, the so called RX, which is the range at X% of the height of the peak. Those were the necessary basic information for calculations of the water equivalent thickness (WET) and (derived from the FWHM of the BP). Based on the correlation between dierent thickness of lung equivalent plates results for WET and were normalised to the same width (in mm) and thickness (in g=cm2) of materials for comparison purposes. The heterogeneity of the materials was quanti ed by the modulation power. A correlation between the Houns eld units (HU) and heterogeneities was investigated by comparing to the HU spectrum of a material. Furthermore, measurements were compared to Monte Carlo (v4.1) dose calculation in RayStation 6.99 (RaySearch, Sweden). Lateral pro les were measured, where instead of the eld ionisation chamber a microDiamond chamber (Type 60019, PTW, Germany) was used, to investigate the eects of inhomogeneities in transversal beam direction. The depths at which the pro les were acquired were depending on the material's R80 for comparability between the materials. The investigations of the WET normalised to the same thickness (1 g=cm2) yielded an average WET of all materials excluding the pumice of 9:8 mm. The values for the lung tissue equivalent plates, the cork samples and the konjac sponge were within 5 %,about 6 % for the balsawood and the oral foam and about 10 % for the pumice. The findings of the WET were independent of the heterogeneity of the material in terms of range shifting but expcted by the thickness determination of the material. In case of pumice the deviation of the WET from the average value was caused by the atomic composition. The derived values of the modulation power as quantity of the material heterogeneity were between 0:09 and 0:13 mm for the lung tissue equivalent plates, between 0:61 and 0:66 mm for the cork with higher density, 0:80 mm for the cork with lower density, 0:12 mm for the balsawood, 0:68 mm for the oral foam, 1:06 mm for the konjac sponge and 1:15 mm for the pumice. For the heterogeneous materials with a modulation power larger than 0:13 mm the heterogeneous contribution to the spreading of the BP was two to three times larger than the homogeneous contribution. This heterogeneity was correlated to the variation of HU within the material via a square root function, under the precondition that the inhomogeneities were distinguishable in the CT. This was the case for the cork samples, the konjac sponge and by lacking of inhomogeneity for the lung tissue equivalent plates. Neither the WET nor were found to be energy dependent at clinically relevant energies. Further did the inhomogeneities not eect the lateral spreading of the beam within the measured accuracy. The comparison of measurements to Monte Carlo dose simulation in RayStation yielded that the treatment planning system was capable of predicting the range of homogeneous and inhomogeneous materials equally well. However, it was not capable to consider the Bragg peak spreading caused by the material heterogeneity and lead to an underestimation. Future research aims at verifying will be to verify the results of lung substitutes for real lung tissue and quantify its modulation power to allow corrections of the TPS dose calculation.
en
Additional information:
Abweichender Titel nach Übersetzung der Verfasserin/des Verfassers