Session Item

Lung cancer could benefit essentially from particle therapy by taking advantage of the exclusive Bragg peak (BP). Besides the challenges of respiration motion, lung tissue has a very heterogeneous texture, which poses difficulties for available dose calculation algorithms based on patients’ CT density information even with the highest resolution of 1 mm. The lung tissue inhomogeneities have a size of 0.5 mm or less causing a varying density along the particles path, which is spreading the BP. The properties of heterogeneous materials mimicking lung tissue were quantified and their effect on the proton BP spreading was experimentally investigated and compared to MC dose calculations.

Following materials suitable as lung substitute were chosen based on their structure, density and compositions, namely cork (2 with different densities), konjac sponge, floral foam, pumice, balsawood and lung tissue equivalent plates (LTEP) (CIRS, USA). Integral depth dose (IDD) of proton beams 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 field ionisation BP chamber (Type 34070, PTW, Germany) with the materials positioned in front of the entrance window. σ, defined as                                                   , with the extraction for σ_mat² = σ_measurement²-σ_water² was calculated as a quantity for the material specific BP spreading. Based on the correlation between different thickness of LTEPs, σ was normalised to the same thickness of materials (1 g/cm²) and compared to the range of HU obtained from a 2x2x2 cm³ volume inside the material structure in the CT image. Furthermore, measurements were compared to Monte Carlo (MC) (v4.1) dose calculation performed with the TPS RayStation 6.99 (RaySearch, Sweden).

With a σ_mat of 0.9 to 1.1 mm a material was considered as homogeneous (LTEPs, balsawood), while heterogenous materials (floral foam, pumice, konjac sponge, corks) had a σ_mat between 2.4 and 3.2 mm. A linear relation between the HU range and squared σ_mat was observed as depicted in figure 1. For the LTEPs and the cork materials the linear fit agreed within 10% with the measurement data and within 30% for konjac sponge. For balsawood, floral foam and pumice σ_mat could not be predicted by the HU range. The TPS MC based dose calculation underestimated the spreading of the BP for the FWHM for all materials by at least 0.2 mm (e.g. ΔFWHM_{LTEPexhale2cm}=0.7 mm) and up to 3.7 mm for heterogeneous materials (e.g. ΔFWHM_Cork5.2cm=2.3 mm).

The MC dose calculation in the TPS cannot predict the BP spreading correctly. Using the HU variation in a material to estimate the BP spreading showed insufficient results for most materials. The tests with balsawood revealed that not just the HU variation but also the gradient of variation to the neighbouring HU voxels does influence the BP spreading, which is focus of current investigations.