ESTRO 38 Abstract book

S484 ESTRO 38

result in major uncertainties in the delivered dose. Therefore, presence of metal implants is often a determining criterion in deciding to contraindicate proton therapy. This work reports the implemented strategy at the Mediterranean Institute for Proton Therapy (Nice, France) that allowed a patient with lumbosacral chordoma to benefit from proton therapy with the Proteus®ONE IBA system despite large metallic spine implants. Material and Methods The titanium alloy materials implanted in the patient (screws, nuts, rods) were provided by the manufacturer. The purpose was to quantify the shift in proton range when implants were placed in the beam and compare it to the shift calculated in the TPS used for proton pencil beam scanning treatments (RayStation). EBT3 Gafchromic films were interspersed among 1 cm-thick solid water slabs. The phantom was placed along the beam direction to assess depth dose distributions. Metal implants were positioned at the films entrance. The phantom was irradiated using a 120 MeV-proton scanned field covering a 20 cm² area (see figure 1). CT images of the set-up were obtained by scanning the phantom with and without the use of Smart Metal Artifact Reduction (MAR) reconstruction algorithm. In RayStation, irradiation conditions were reproduced based on the CT images. Bragg peak positions measured in the films for each implant were compared to the calculated ones in several configurations: using both uncorrected and MAR-corrected images; overriding or not the density of the delineated implants in the TPS; using both Pencil Beam and Monte Carlo algorithms. Results In all cases, maximum error in TPS range calculation was 18.9 mm. In both CT series, the errors were minimized when implants density was overridden in RayStation. Although MAR-corrected images reduced beam hardening artifacts, they mis-reproduced implant dimensions. Hence, the following planning strategy was adopted: implants were delineated on uncorrected CT and contours were copied on MAR-corrected CT. The metal structure was overridden to a compound defined in RayStation based on Ti with a density of 4,42 g/cc. The range uncertainty parameter of robust optimization in the TPS was set to 5%. To account for maximal errors, 5 mm CTV-to-PTV margins were added to target and organs-at-risk near the implants. Conclusion In this work, evaluation of the maximum error in calculated proton range allowed to define adapted safety margins and create a robust treatment plan for a patient with lumbosacral chordoma with metal spine implants. The strategy implemented could be re-used in further proton therapy treatments. Besides, the use of a dual energy CT scan could improve the definition of high density structures CT-number, allowing reducing safety margins. Further investigations are ongoing, including absolute dose studies. PO-0911 Choose before you measure. Setting gamma parameters for different QA devices on the basis of ROC. M. Giżyńska 1,2 , D. Blatkiewicz 1 , M. Bukat 1 , M. Gil- Ulkowska 1 , S. Maluszczak 1 , A. Paciorkiewicz 1 , D.

2 Technical University of Denmark, Center for Nuclear Technologies, Roskilde, Denmark ; 3 Skåne University Hospital, Department of Radiation Physics, Lund, Sweden ; 4 Carleton University, Carleton Laboratory for Radiotherapy Physics, Ottawa, Canada Purpose or Objective Due to intra-fractional motion in the abdominal and thoracic region of cancer patients, there is a need for solutions to accurately calculate external x-ray beam radiotherapy dose in four dimensions with synchronization between the dynamic beam configuration and deforming anatomy. In order to facilitate determination of time- dependent accuracy of advanced radiotherapy techniques, there is furthermore a need for the advanced dose calculation solution to be time-resolved. The aim of this project was to develop and experimentally validate a user code for time-resolved Monte Carlo (MC) calculations of dose delivered to a dynamic thorax phantom. Material and Methods Time-resolved Monte Carlo simulations were based on the previously developed 4DdefDOSXYZnrc/EGSnrc MC user code, which scores dose in a time-varying deformable geometry. To improve efficiency of the simulations, photon cross-section enhancement (XCSE) was implemented in a region surrounding the voxel in which dose was scored. One three-dimensional conformal radiotherapy (3DCRT) plan and one volumetric modulated arc therapy (VMAT) plan (both 6 MV) were optimized on and delivered to a set of static and dynamic configurations of an in-house developed dynamic thorax phantom. Time- resolved MC calculations using linac logfile based input files were synchronized with and compared to measurements using a fiber-coupled organic plastic scintillator detector (PSD). Measurements and calculations were conducted in the center of a spherical tumor (PMMA) embedded in a motion controlled cylindrical lung insert (balsa wood), laterally positioned in the body (PMMA) of the thorax phantom. Results Using the time-resolved MC code with a XCSE factor of 8 and binning the data to a resolution of 100 ms, a statistical uncertainty of approximately 2% was achievable. Comparison with PSD measurements revealed cases with disagreements in low-dose regions and narrow dose peaks, which are attributed to uncertainty in the position of the PSD and currently under investigation. However, a majority of the results indicated good agreement; first of all in the accumulated dose (in most cases within 2-3%) and secondly in the dose as a function of time as indicated by agreements in temporally resolved dose gradients. This applies for both the 3DCRT and VMAT techniques applied to the set of phantom configurations investigated. Conclusion A novel user code for time-resolved MC calculations of dose delivered to a deforming anatomy was developed and initial validation included indications of good temporal agreements in detecting dose gradients compared to PSD dosimetry in a dynamic thorax phantom. The solution has high potential in assisting in the detection of underlying causes to deviations detected in the accumulated dose as delivered by advanced motion managed radiotherapy techniques. PO-0910 Lumbosacral proton therapy treatment of a patient with large spine metal implants C. Peucelle 1 , A. Gérard 1 , D. Maneval 1 , M. Vidal 1 , A. Falk 1 , J. Hérault 1 1 Centre Antoine Lacassagne, Radiotherapy, Nice, France Purpose or Objective Metal implants complicate radiotherapy delivery by creating artifacts in CT images. Artifacts, as well as CT- number inaccuracy of very high-density material, may affect range calculation in the TPS for proton therapy and