ESTRO 38 Abstract book

S977 ESTRO 38

terms of technique (VMAT or Tomotherapy), anatomic region, prescription and ordered from 1 to 6 to better filter the data. Results As expected, results based on previous treatments shows that as higher the modulation factor gets, lower is the number of pixels passing the gamma criteria. For VMAT, minimum and máximum values of 94,6% y 99,9% in gamma were found to be related to plans having modulation factors of 3,56 y 1,24 respectively. For Tomotherapy the same behaviour was observed. Applying a class solution is also important to take into account prescription and volume of the targets. For Tomotherapy reducing the expected value for modulation factor during calculation also impacts plan quality, so a good bias must be found. Whilst for VMAT, rising the segment size during planning was determinant to reduce modulation factor. Conclusion Currently, modern machines are capable of delivering modulated plans in better accordance with the planning theoretical data. Performing PSQA for every single patient might be evaluated depending on internal statistics for each department. More effective and representative plans yielding better QAs results could be achieved if there is a better understanding of the factors provided by treatment planning systems. EP-1804 Experimental validation of a novel technique to derive stopping power ratio from MRI in soft tissue J. Scholey 1 , D. Chandramohan 1 , T. McClave 1 , A. Sudhyadhom 1 1 The University of California- San Francisco, Radiation Oncology, San Francisco, USA Purpose or Objective to experimentally validate the Unified Composition (UC) method and compare to “ground truth” measurements for determining mean ionization potential (I m ) and stopping power ratio (SPR) using MRI for dose calculation in proton Porcine soft tissue types (muscle, liver, brain, and spleen) were cryoground to produce homogeneous samples. Tissues were divided into two groups for determination of I m and SPR using 1) Bragg’s additivity rule (BAR), considered the ground truth (GT) method, and 2) the Unified Compositions (UC) method. For the GT method, I m values were calculated directly using BAR with elemental composition determined from chemical combustion analysis of each tissue, as shown in equation 1. SPR values were determined by the Bethe-Bloch equation through measurements of physical density and atomic composition. For the UC method, phantoms were created containing each tissue type and pure water, isopropyl alcohol, acetone, and propargyl alcohol, which were used for proton density and water fraction calibration curves for MRI. Phantom images were acquired using kVCT, MVCT, and 3T MRI. I m values were calculated from percent water/organic material by mass and hydrogen content of organic material computed using water/organic 1 H separation and proton-density weighted MRI scans, respectively, as shown in equation 2. For calculation of SPR, relative electron density (P e rel ) was determined from MVCT using calibration of an electron density phantom. For comparison, SPR was also calculated for each tissue using kVCT and the stoichiometric calibration method. therapy for soft tissues. Material and Methods

Results Of the tissues evaluated, SPR values computed with the UC method agreed on average to within 0.7% of the GT method, with results shown in table 1. SPR values computed from the stoichiometric calibration saw greater disagreement to GT values, with percent differences of 5.3, 5.2, 3.8, and 4.5% for muscle, liver, brain, and spleen, respectively.

Conclusion It is possible to achieve sub-percent accuracy using the UC method for determination of I m and SPR from MRI and MVCT imaging as compared to direct computation using GT principles for soft tissues evaluated in this study. All SPR values determined by the UC method were within experimental error of GT determination. Of note, P e rel is a predominant source of error in SPR calculations (as SPR varies linearly with P e rel and logarithmically with I m ) and ongoing investigations are looking to improve this accuracy. SPRs computed via the UC method saw closer agreement with those computed by GT than the widely- used stoichiometric calibration method. EP-1805 The Effect of Material Heterogeneity in Endorectal Brachytherapy with 192Ir, 75Se and 169Yb Sources T. Shoemaker 1,2 , T. Vuong 3,4 , H. Glickman 1 , S. Kaifi 3 , G. Famulari 1 , S.A. Enger 1,2,4 1 McGill University, Medical Physics Unit, Montreal, Canada ; 2 McGill University Health Centre, Research Institute, Montreal, Canada ; 3 Jewish General Hospital- McGill University, Department of Radiation Oncology, Montreal, Canada ; 4 McGill University, Department of Oncology, Montreal, Canada Purpose or Objective Differences between prescribed and administered dose in high dose rate endorectal brachytherapy (HDR-EBT) were investigated by evaluating dose to clinical target volume (CTV) and organs at risk (OARs) calculated with a Monte Carlo based dose calculation tool, RapidBrachyMC. In addition, dose coverage, conformity and homogeneity were compared between the radionuclides 192 Ir, 75 Se and 169 Yb for use in HDR-EBT. Material and Methods RapidBrachyMC, which builds on the Geant4 MC-toolkit, was used to perform postimplant dosimetry using computed tomography (CT) images for 8 patients, each treated with 3 fractions of HDR-EBT at the Jewish General Hospital in Montreal, Canada. In the HDR-EBT treatment a 192 Ir MicroSelectron v2 source was used with a flexible intracavitary applicator capable of fitting a tungsten rod in its central lumen for OAR shielding. In combination with this applicator, two balloons filled with iodine solution were used, one used to fix the applicator to the rectum and the other to displace the contralateral rectal wall. Four segmentation schemes were simulated. 1) According