Abstract Book

S315

ESTRO 37

volume parameters used in classical NTCP models (VxGy and Dx%). Receiving Operating Characteristic (ROC) curves were also established. Results As compared to children not treated by radiation therapy, and when controlling for chemotherapy, children who received less than 1 Gy in average, to the cerebral arteries had a 1.8 (95%CI: 0.80-3.9) times higher risk of stroke, whereas those who received more than 20 Gyin aveage to cerebral arteries had a risk 26 (95%CI : 10-68) times higher. These risk factors were higher for ischemic strokes than for haemorrhagic ones. To take into account dose-volume parameters allowed to improve these risk predictions. High radiation doses strongly increased stroke risk. As an exemple,to have received during childhood 30 Gy or more to only 1 to 5% of the cerebral arteries volume increased the long term risk of stroke by a factor 2.8 (1.3-5.5). Conclusion This study confirms that average radiation dose to the cerebral arteries is not the best parameter for predicting long term risk of stroke. Classical dose-volume parameters may be used for such predictions, even if established on antropomorphic phantoms in retrospective studies OC-0601 Development and Monte Carlo simulations of a novel 3D range-modulator for proton therapy Y. Simeonov 1 , U. Weber 2 , P. Penchev 1 , T. Printz Ringbæk 1 , K. Zink 1,3,4 1 Institute of Medical Physics and Radiation Protection - IMPS, LSE, Gießen, Germany 2 GSI Helmholtzzentrum für Schwerionenforschung GmbH, Biophysics division, Darmstadt, Germany 3 Frankfurt Institute for Advanced Studies FIAS, Physics department, Frankfurt, Germany 4 University Hospital Gießen-Marburg, Department of Radiotherapy and Radiooncology, Marburg, Germany Purpose or Objective Pencil beam scanning, though state-of-the-art in particle therapy, hasn’t yet established in the treatment of moving targets, due to the large number of different iso- energy layers and the associated long irradiation time. The aim of this work was to design a 3D range-modulator, capable of creating a homogeneous and highly conformal dose distribution with only one fixed energy, thus reducing considerably the treatment time. Material and Methods In a first step, a novel 2D range-modulator was developed, capable of modulating a 151.77 MeV 1 H pristine Bragg Peak (BP) into a 5 cm homogeneous spread-out Bragg Peak (SOBP). Contrary to conventional ridge filters, the 2D modulator consists of many pyramid- shaped pins with ~2 mm 2 base area (Fig. 1a). This new design introduces an additional degree of freedom, as the height of each individual pin can now be independently varied in the transverse direction. The modulator was triangulated and manufactured in rapid prototyping technique. Using the FLUKA Monte Carlo (MC) package, a sophisticated model was developed in order to investigate the modulating properties of the range- modulator and calculate the resulting dose distribution. The model incorporates intensity modulated scanning and is able to handle the complex geometry contour of the modulator. In order to compare the MC simulations with measurements, the model was precisely tuned to Proffered Papers: PH 11: Emerging technologies

accurately reproduce the depth dose distribution and lateral beam profiles in air of the Marburg Ion-Beam Therapy Centre (MIT). The integral depth dose distribution after the modulator was subsequently measured at MIT using the PTW PEAKFINDER and used to validate the FLUKA model. In a second step, taking advantage of the additional degree of freedom, a new 3D range-modulator was developed (Fig.1b). The position-dependent height of each single pin was calculated in such a way so that the final form of the modulator corresponds to a spherical target of 5 cm diameter. When irradiated, the 3D range- modulator creates a quasi-static irradiation field, tightly shaped around the target. The dose distribution behind the modulator was simulated with FLUKA.

Fig. 1 A single pin (a) and a 3D range-modulator (b) for a spherical tumour with 5 cm diameter. Results Fig. 2a shows a very good agreement between the measured and simulated SOBP for the 2D modulator. The simulation results from the spherical 3D range-modulator (Fig. 2b) show a homogeneous 3D dose distribution conformed not only to the distal, but also to the proximal edge of the target.

Fig. 2 Measured vs. simulated SOBP (a); 1 H, E=151.77 MeV. Monte Carlo simulated X-Z midplane profile (b). Conclusion Utilizing state-of-the-art 3D printing technique to manufacture complex modulators is possible. Combining the advantages of very short treatment time, the 3D range-modulator could be an alternative to treat lung tumours with the same conformity as full raster- scanning treatment. Further measurements must be