ESTRO 36 Abstract Book

S524 ESTRO 36 _______________________________________________________________________________________________

Material and Methods The kinetic M 1

cylinder applicator (Figure 1b) with 5 catheters. Inter- dwell distances of 2 and 5 mm were employed and the experiments performed for source activities between 5 and 10 Ci. The EPID response is proportional to the source activity so it is possible to obtain the activity by sending the source to pre-defined dwell position.

model, is based on the spherical harmonic expansion of the distribution function, solution of the linear Boltzmann equation. The first two angular moments equations, combined with the Continuous Slowing Down Approximation, are closed using the Boltzmann's principle of entropy maximization. The algorithm computes at the same time all primary and secondary particles created by the interactions of the beam with the medium. Thanks to the implementation of the interaction cross sections for electrons and photons in the energy range from 1keV up to 100 MeV, the algorithm can simulate different treatment techniques such as the external radiotherapy, brachytherapy or intra-operative radiation therapy. As a first validation step, a large number of heterogeneity shapes has been defined for various complex numerical phantoms both for electron and photon monoenergetic sources. Dose profiles at different positions have been measured in water phantoms including inhomogeneity of bone ( ρ = 1.85 g/cm 3 ), lung ( ρ = 0.3 g/cm 3 ) and air ( ρ = 10 -3 g/cm 3 ). Secondly, taking as reference the Carleton Laboratory for Radiotherapy Physics Database, different radioactive seeds have been implemented in the code. Moreover, several simulations based on CT scan of prostate cancer have been performed. The M 1 model is validated with a comparison with a standard, accurate but time consuming, statistical simulation tools as PENELOPE. Results The M1 code is capable of calculating 3D dose distribution with 1mm 3 voxels without statistical uncertainties in few seconds instead of several minutes as PENELOPE. Thanks to its capability to take into account the presence of inhomogeneities and strong density gradients, the dose distributions significantly differ from those calculated with the TG-43 approximations. More in detail: inter-seed attenuation is treated, the real chemical composition of the different tissues can be taken into account and the effects of patient dimensions are considered. Conclusion In the comparison with the MC results the excellent accuracy of the M 1 model is demonstrated. In general, M 1 , as the MC codes, overcomes the approximations that are formalised in TG-43 in order to decrease the complexity of the calculations. Thanks to its reduced computational time and its accuracy M 1 is a promising candidate to become a real-time decision support tool for brachytherapists. PO-0947 Image-guided brachytherapy with 106Ru eye plaques for uveal melanomas using post implantation MRI G. Heilemann 1 , N. Nesvacil 2 , M. Blaickner 3 , L. Fetty 1 , R. Dunavoelgyi 4 , D. Georg 2 1 Medical University of Vienna/ AKH Vienna, Department for Radiotherapy, Vienna, Austria 2 Medical University of Vienna/ AKH Vienna, Department for Radiotherapy/ Christian Doppler Laboratory for Medical Radiation Research for Radiation Oncology, Vienna, Austria 3 Austrian Institute of Technology GmbH, Health and Environment Department Biomedical Systems, Vienna, Austria 4 Medical University of Vienna/ AKH Vienna, Department for Ophthalmology and Optometry, Vienna, Austria Purpose or Objective In radiation oncology magnetic-resonance imaging (MRI) is an important modality for tissue characterization, target delineation and allows image-guidance due to its high soft tissue contrast as a tool for better cancer treatment. In 106 Ru-brachytherapy of uveal melanomas MRI is mainly used for pre-treatment planning scans to assess tumor size and location. However, post-implantation MR scans yield additional information on the plaque position in relation to the target volume and critical structures. Together with

Results 3D Cartesian coordinates can be obtained with 0.2 mm accuracy using a single EPID panel. The panel can clearly identify dwell positions 2 mm apart even with the catheter at 24 cm distance (Figure 1c) from the panel. Absolute coordinates can be obtained by adding reference points (representing the corners of the water phantom) in the treatment plan that can be related with the position of the water phantom over the panel during the experiments. An in-house developed software compares all dwell positions/times against the treatment plan. The software can also monitor the sequence of the treatment identifying the afterloader channel connected to each catheter. Therefore, it is possible to detect catheter misplacements, swapped transfer tube connections, wrong dwell times and/or positions and also verify the source activity. Conclusion This work describes an experimental system that can be implemented in the clinic providing experimental pre- treatment verification that is not currently available. This method provides several advantages when compared against other dosimeters such as films or MOSFETs as it combines a 2D dosimeter, which has an online response. Our system can detect several problems that would be unnoticed during the treatment if only traditional QA is performed. PO-0946 Entropic model for real-time dose calculation: I-125 prostate brachytherapy application. G. Birindelli 1 , J.L. Feugeas 1 , B. Dubroca 1 , J. Caron 1,2 , J. Page 1 , T. Pichard 1 , V. Tikhonchuk 1 , P. Nicolaï 1 1 Centre Lasers Intenses et Applications, Interaction- Fusion par Confinement Inertiel- Astrophysique, Talence, France 2 Institut Bergonié Comprehensive Cancer Center, Department of radiotherapy, Bordeaux, France Purpose or Objective This work proposes a completely new Grid Based Boltzmann Solver (GBBS) conceived for the description of the transport and energy deposition by energetic particles for brachytherapy purposes. Its entropic closure and mathematical formulation allow our code (M 1 ) to calculate the delivered dose with an accuracy comparable to the Monte Carlo (MC) codes with a computational time that is reduced to the order of few seconds without any special processing power requirement.