ESTRO 37 Abstract book

S38

ESTRO 37

Table 1 shows that good agreement was found between the MR to MV alignment values found with the plumb-line and the official Elekta phantom MR to MV alignment (mm) L/R S/I A/P Plumb line +0.93 -1.43 +0.5 Elekta phantom +0.79 -1.09 +0.44 Conclusion The prototype components of a new MR-Linac QA phantom have been developed and tested. It has been demonstrated that they provide a good check of gantry angle, EPID position and MR to MV alignment. The main strength of this phantom is that these measurements are solely dependent on an absolute metric, gravity, rather than any phantom or linac manufacturing tolerances. OC-0080 Dosimetric evaluation of a 3D printed phantom for patient-specific pre-treatment plan verification D.N. Makris 1 , E. Zoros 2 , T. Boursianis 3 , E. Pappas 4 , T.G. Maris 3 , E.P. Efstathopoulos 1 1 Medical School-National and Kapodistrian University of Athens, Second Department of Radiology, Athens, Greece 2 Medical School-National and Kapodistrian University of Athens, Medical Physics Laboratory, Athens, Greece 3 University of Crete- Heraklion- Crete- Greece, Department of Medical Physics, Heraklion, Greece 4 Technological Educational Institute of Athens- Greece, Department of Radiology and Radiotherapy, Athens, Greece Purpose or Objective Stereotactic radiosurgery (SRS) is a well-established treatment approach for the management of a wide variety of lesions, mainly in the brain. Patient-specific dose verification becomes paramount in SRS treatments where small photon beams and steep dose gradients are employed. Although current patient-specific quality assurance techniques could detect critical dose errors, the dose distribution is not directly measured but is instead reconstructed. The scope of this work is to study the dosimetric characteristics of a 3D printed phantom based on anonymized real patient’s CT scan and pave the way towards a truly patient-specific plan verification methodology. Material and Methods Using the treatment planning CT scan of a real patient with 6 brain metastases (target volumes of 0.023 - 0.989cc), a 3D model of the patient’s external contour and bone structures was constructed. Patient skin was modelled by applying a 3mm thickness to the external contour. The model was cropped till the lower jaw, while patient immobilization apparatus was removed. A commercially available 3D printer was selected for 3D printing of the model, based on studied radiologic characteristics of the 3D-printed material samples. The hollow 3D-printed phantom was filled with polymer gel (tissue equivalent) and CT scanned using the same immobilization apparatus. In order to dosimetrically evaluate the phantom, the real patient’s treatment plan was applied to the phantom’s CT scan and corresponding calculated 3D dose distribution was exported. Following an anatomic based co-registration of the two CT scans, patient and phantom dose distributions were compared in terms of 1D profiles, 2D isolines, 3D gamma index (passing criteria: 1%/1mm), target and critical organs Dose Volume Histograms (DVHs) and plan quality metrics. Results The 3D printed material is bone equivalent in terms of HU (mean HU = 1037±57) which also applies to the printed “skin”. All 1D profiles evaluated showed very good agreement in both soft tissue and bone structures, while a mean discrepancy of up to 8% was observed for the skin dose, as expected. An additional dose offset was

observed in and around sinuses, which were not modelled. 3D gamma index comparison for the entire geometry resulted in a passing rate of 97.98% which was also reflected in DVH comparison. The mean dose for the smallest target was found to deviate by 1.4%, while for all remaining targets discrepancies did not exceed 0.6%.

Figure: (a) Profile on an axial slice through two targets. (b) Isolines comparison on a coronal slice including two targets. (c) Indicative DVHs for one metastasis. Conclusion An anatomical replica of the patient was constructed using commercially available 3D printer. Dosimetric evaluation revealed that the phantom can be regarded as patient equivalent (except for the skin and sinuses areas), enabling truly patient-specific plan verification if coupled with a comprehensive dosimetric system. OC-0081 From range measurements to dose: in-vivo dosimetry using the prompt gamma camera S. Toscano 1 , K. Souris 2 , E. Sterpin 1 , G. Janssens 3 , J. Petzoldt 3 , F. Vander Stappen 3 , J. Smeets 3 , X. Geets 4 , K. Teo 1 KU Leuven, Laboratory of Experimental Radiotherapy- Department of Oncology-, Leuven, Belgium 2 Université catholique de Louvain, Institute of Experimental and Clinical Research IREC, Louvain-la- Neuve, Belgium 3 IBA - Ion Beam Applications, R&D, Louvain-la-Neuve, Belgium 4 Clinique Universitaire Saint Luc- UCL, Radiation oncology, Brussels, Belgium 5 University of Pennsylvania School of Medicine, Radiation oncology, Philadelphia, USA Purpose or Objective In this contribution we present a new methodology to reconstruct the delivered dose during a pencil beam scanning proton therapy treatment, using prompt gamma imaging for range verification of the proton beam. Material and Methods A prototype prompt gamma camera was used to record the signal emitted along the proton tracks during the delivery of proton therapy for a brain cancer patient. By comparing the recorded prompt gamma depth profiles with simulation, a verification of the proton range is possible. Using the range shift calculation we developed a new method based on Monte Carlo calculations to reconstruct the delivered dose. Our Monte Carlo simulation tool, MCsquare, is a fast and accurate method of simulating heavy charged particles inside voxelized geometry, such that a full treatment can be simulated in few minutes. In our approach the shift measured from the prompt gamma imaging can be simulated either as a density scaling of the voxels along the proton trajectory or a spot-specific energy correction. Two methodologies have been implemented and compared: 1. Density correction: A ray-tracing inside the CT image geometry is performed to calculate the water equivalent path length (WEPL) for every proton beamlet, and a density correction factor is calculated as the ratio between the 'shift-corrected” WEPL and the expected one. In MCsquare calculations the stopping powers are scaled by this spot-specific scaling factor and the dose is recalculated accordingly. 2. Energy correction: For every beamlet the range in water is calculated from the 'shift-