S28 Abstracts / Brachytherapy 13 (2014) S15eS126

seeds or ISA in multiple seed implants are neglected up to day. The goal ofthis project is to determine a novel method to solve the impact of the doseperturbations (P-value) and/or inter-seed attenuation (ISA-value) effect inBTPS based on MC simulation and Artificial Neural Networks (ANN).This method was validated for LDR I-125 to Ir-192.Materials and Methods: In this model, perturbation and ISA of dosedistribution from one source by the surroundings sources are calculatedusing the MC pre-calculations 3D kernels of P-values for a set of binarygroups of one active and one inactive source. These kernels were obtainedfor different radial distances and angular positions. The total P-valuemodel is equal to the product of the P-value of the binary groups of theone active and inactive source. The accuracy of this model have beenexamined by a comparison of the calculated P-value and ISA-value withthis model by the values, which are directly obtained using full MonteCarlo water simulations (FMCWS) for some multiseed implants. Theseimplant configurations are: single active source in the center and (1)inactive sources on four different sides, (2) inactive sources are parallel toeach other on the along source axis ‘‘r,’’ and (3) more inactive sources areparallel to each other on the away axis ‘‘z.’’ The P-value MC simulationsfrom binary groups of active-inactive sources was used to train theArtificial Neural Networks (ANN) in MATLAB software. Base on thisapproach, once trained the network generalizes, to produce ISAcorrection response for any unknown binary group source combination.Then the ANN ISA correction data for any unknown combinations of anactive and inactive source which it has not been obtain directly by MC,imported to the BTPS base on the total ISA-value formulation model.Results: Figure 1 shows comparison betweenModel and FMCWS for I-125(Isoaid) source. The active source are placed in center and the inactivesources are in (0.5, 0.5), (1, 0.5), (2, 2), and (3, 1) for angles of 20, 40,60, and 90 degrees, respectively. For all cases, the total perturbation andISA formulisms model agree with FMCWS. These new P and ISAformulism have better accuracy for Ir-192 than the I-125 due to Comptonscattering. The differences between the trained ANN and MC were up to1%. The I-125 ISA formalism accuracy for case one, two, and three are0.8%, 5%, and 5%, respectively. These corresponding values for Ir-192are less than 0.5%, 3.5%, and 3%, respectively.Conclusions: This new model provide inputs for brachytherapy planningsoftware to correct the ISA effect in dose calculations for multi-seedimplants based on TG-43U1 algorithm using ANN and MC methods.

OR20 Presentation Time: 4:24 PM

Monte Carlo Simulation of HDR Ir-192 Brachytherapy Cancer

Treatments

Gabriel P. Fonseca, MSc1,2, Brigitte Reniers, PhD2, Josef Nilsson, PhD3,

Maria Persson, MSc3, �Asa Carlsson Tedgren, PhD3, H�elio Yoriyaz, PhD1,

Frank Verhaegen, PhD2,4. 1Centro de Engenharia Nuclear, Instituto de

Pesquisas Energ�eticas e Nucleares, S~ao Paulo, Brazil; 2Radiation

Oncology, Maastro Clinic, Maastricht, Netherlands; 3Section of

Radiotherapy Physics and Engineering, Dept of Medical Physics,

Karolinska University Hospital, Stockholm, Sweden; 4Medical Physics

Unit, Department of Oncology, McGill University, Montr�eal, QC, Canada.

Purpose: Modern treatment planning systems (TPS) for brachytherapyare now available that are based on model-based dose calculationalgorithms (MBDCA). They enable heterogeneity corrections whichare needed to replace the TG-43U1 water dose formalism with a moreaccurate approach. With the aim of evaluating the impact of the transitdose, applicators, body boundaries and tissue composition in clinicalcases, several HDR Ir-192 brachytherapy treatment cases includingprostate, gynecological (Gyn), arm and head and neck cases wereevaluated.Materials and Methods: AMBDCA, AMIGOBrachy based on the MonteCarlo code MCNP6, was used to import brachytherapy treatment plansprovided by four different hospitals planned using commercial TPS,BrachyVision� and Oncentra�. These plans were then edited through aninteractive graphical interface. CT images were segmented into tissuesand the applicators (needles or Gyn cylinders) were included to evaluatetheir impact on the dose distributions. The treatment plans were selectedto provide an overview of brachytherapy treatments since there is aconsiderable range in the number of dwell positions, needles and type ofapplication. Some treatments were performed with dwell positions withinsoft tissue at a few millimeters from bone tissue. The transit dose impactwas evaluated using a continuous source distribution defined through thesource trajectory with probability distribution weighted by theinstantaneous source speed at each position.Results: The differences between the results obtained using a MBDCAbased on the TG-186 formalism and TG-43U1 water dose formalism arecase dependent with dose differences ranging from negligible (!0.5%) upto 20% even within the target volume. For the majority of the cases thisimpact cannot be fully represented by clinical parameters as DVH orD90, e.g., inter needle attenuation can cause up to 6% underdose atshadow regions that do not impact significantly when averaging the doseover the target. Figure 1 shows a dose ratio for a head and neck case withdifferences up 10% behind the esophagus. The transit dose wasindependently evaluated since it could be as relevant as the tissuecomposition. Particularly for interstitial implants the proximity of thedwell positions and the target volume reduces the dwell time making thetransit dose more significant. The results obtained for two prostate caseswith similar target volume and prescribed dose can lead to a 10%overdose to the urethra for one case, with only half the overdose for thesecond case. This highlights the importance of an adequate catheterdistribution and of the transit dose.Conclusions: The effect of tissue heterogeneities and applicators can besignificant, while the transit dose effect depends of the catheters anddwell position distribution. Those aspects are relevant due to the varioustypes of applicators commercially available and due the wide variety ofbrachytherapy treatments.

OR21 Presentation Time: 4:33 PM

Dynamic Modulated Brachytherapy for Cervical Cancer

Dae Yup Han, MSc, Matthew J. Webster, MSc, Daniel J. Scanderbeg, PhD,

Catheryn M. Yashar, MD, William Y. Song, PhD. Radiation Medicine and

Applied Sciences, University of California, San Diego, San Diego, CA.

S29Abstracts / Brachytherapy 13 (2014) S15eS126

Purpose: We propose a new intra-uterine tandem design that are capable ofcreating non-isotropic 192-Ir dose distributions and give unprecedenteddose conformality for treatment of cervical cancer.Materials and Methods: The MRI-compatible DMBT tandem design has6 peripheral holes of 1mm diameter, on a non-magnetic tungsten-alloy rod(18.0 g/cc), enclosed in a polyoxymethylene tube (1.41 g/cc), with a totaldiameter of 6mm. The tungsten-alloy shield has a diameter of 5mm andthe grooves are evenly distributed (60�) on the surface of tungsten-alloyshield. A VariSource iX HDR source was used for delivering radiationdose. MCNPX Monte Carlo was used to simulate the resulting non-isotropic dose distributions. An in-house developed HDR brachytherapyplanning platform, with intensity modulated planning capability usingSimulated Annealing and Constrained-Gradient Optimization algorithms,was used to re-plan 75 plans (5 fractions (fx) of 15 patient cases) treatedwith a conventional tandem & ovoids (T&O) applicator. For the proposedtandem designs, the plans were optimized with the same ovoids in place,as the conventional T&O plans. All DMBT plans were normalized tomatch the HRCTV V100 coverage of the original T&O plans.Results: Generally, the plan qualities were markedly better using DMBT.Mean 2cc doses to the bladder were 4.0�0.2Gy and 3.37�0.19Gy, forT&O and DMBT, respectively. For the rectum, they were 2.30�0.17Gyand 1.82�0.1Gy. For the sigmoid, they were 1.58�0.19Gy and

1.48�0.17Gy. The best single plan reduction was in the bladder dose,0.32Gy (40.8%), and this was due to the horseshoe-like wrapping of thebladder around the CTV, thus benefitting the most with the doseconformation enhancements achieved by the proposed designs. The bestsigmoid improvement, 0.13Gy (27.5%), was shown for endophyticgrowth type cervical cancer. The HRCTV dose heterogeneity index (DHI)was 2.56�0.24 and 2.36�0.22 for T&O and DMBT, respectively.Conclusions: We have shown the new tandem designs that advance theconformality of image-guided cervix HDR, in congruence with thecurrent trend of 3D image based planning to maximize the therapeutic ratio.

OR22 Presentation Time: 4:42 PM

Measurement and Comparison of Surface Dose outside Treatment

Area for Different Radiation Treatment Modalities in Breast Cancer

Patients Using Thermoluminescence Dosimeters

Suraj Prasad Khanal, Master in Medical Physics1, Zoubir Ouhib, MS,

DABR2, Theodora Leventouri, PhD1. 1Physics, Florida Atlantic

University, Boca Raton, FL; 2Lynn Cancer Institute, Boca Raton, FL.

Purpose: There is an increasing concern regarding radiation-relatedsecondary cancer risks in long term radiotherapy survivors and acorresponding need to be able to predict cancer risks at significantradiation doses.Purpose of this study is to measure and compare surface dose outside thetreatment area for different breast cancer radiation treatment modalities.Measured absorbed doses outside treatment volume are the degree ofscattered and leakage radiation for the treatment modalities.Materials and Methods: The TLD-100 chips were calibrated according tothe types of beam that were going to be used for the particular patient.Patient’s consensus has been made to participate in the study. The TLDswere placed at six different locations on the patient. Those surface pointswere selected based on the critical organ surrounding the target area.TLDs were then exposed during patient treatment and read for evaluationaccording to a protocol of the University of Wisconsin RadiationCalibration Laboratory (UWRCL). Energy response factors for differentbeams are applied for dose calculation.Results: Average absorbed doses in cGy for different modalities are: ForAccuboost [Sternum (6.51�2.93), Shoulder (5.18�2.21), Eye(1.74�0.84), Thyroid (5.50�2.75), Contralateral Breast (8.52�3.86),Lower Abdomen (4.50�2.63)], For SAVI [Sternum (3.06�1.28),Shoulder (2.26 �1.11), Eye (1.51�0.52), Thyroid (2.00�0.73),Contralateral Breast (2.74�1.49), Lower Abdomen (1.67�1.22)], ForMammosite ML [Sternum (5.27�2.12), Shoulder (5.58�2.77), Eye(2.65�0.68), Thyroid (3.38�1.03), Contralateral Breast (2.80�1.22),Lower Abdomen (2.82�2.25)], For Electron boost [Sternum (0.52�0.19),Shoulder (0.66 �0.16), Eye (0.52�0.29), Thyroid (0.52�0.13),Contralateral Breast (0.41�0.24), Lower Abdomen (0.32�0.18)], ForPhoton boost [Sternum (2.62�1.82), Shoulder (1.03 � 0.42), Eye(0.64�0.20), Thyroid (0.96�0.55), Contralateral Breast (1.84�0.69),Lower Abdomen (0.76�0.11)].Conclusions: Statistically there are no significant differences in terms of outof field surface doses among three studied APBImodalities. However, out offield surface dose measured with APBI modalities at different POIs aresignificantly different with electron boost and photon boost. Our studyshows that the maximum possible doses out of normalized nominaldelivered dose (2 Gy/fraction) at POIs are: 7.90% in contralateral breast,6.32% in sternum, 5.70% in thyroid, 5.50% in shoulder, 1.93% in eye and4.3% in lower abdomen during radiation treatments. However, anabsorbed surface doses are strongly tumor site quadrant depended. This isnot a high dose as compare to target dose but deposit to a large part ofbody which have potential to induce secondary malignancy. It will beuseful considering various influences of out of field surface dose whiletreating a breast cancer woman with radiation. Special care should begiven in delivering parameters such as patient set up, catheterarrangement and field set up to minimize out of field surface doses duringtreatment.