W Cho

The Experimental Verification of Collapsed Cone Convolution Algorithm for Dose Calculation in Radiation Treatment Planning System

The aim of this study is to evaluate the accuracy the collapsed cone convolution (CCC) algorithm for dose calculation in a radiation treatment planning system (TPS), CorePLANTM. We implemented beam models for various setup conditions in TPS, we calculated dose using CCC algorithm for 6 MV and 15 MV photon beam in 50 × 50 × 50 cm3 water phantom of the assessment criteria for the dose distribution. Generated beam models from a specific depth were evaluated by comparing PDDs and profile of measured data and calculated data to fit in other depths. Field sizes were 4 × 4 cm2, 6 × 6 cm2, 10 × 10 cm2, 20 × 20 cm2, 30 × 30 cm2 and 40 × 40 cm2, classified as open beam cases and wedged beam cases, respectively. The values were calculated using CCC algorithm, the discrepancies between measurement data and calculation data was found. The delivered dose in small field size was underestimated in field but overestimated out of field. In PDD, the Dmax depth of small field was reduced. But, for over 4 × 4 cm2, the beam profile and PDD showed good agreement between measured and calculated data. We verified the accuracy of CCC algorithm in CorePLAN.

SU-E-T-316: The Design of a Risk Index Method for 3D Patient Specific QA

PURPOSE To suggest a new guidance for the evaluation of 3D patient specific QA, a structure-specific risk-index (RI) method was designed and implemented. METHODS A new algorithm was designed to assign the score of Pass, Fail or Pass with Risk to all 3D voxels in each structure by improving a conventional Gamma Index (GI) algorithm, which implied the degree of the risk of under-dose to the treatment target or over-dose to the organ at risks (OAR). Structure-specific distance to agreement (DTOA), dose difference and minimum checkable dose were applied to the GI algorithm, and additional parameters such as dose gradient factor and dose limit of structures were used to the RI method. Maximum passing rate (PR) and minimum PR were designed and calculated for each structure with the RI method. 3D doses were acquired from a spine SBRT plan by simulating the shift of beam iso-center, and tested to show the feasibility of the suggested method. RESULTS When the iso-center was shifted by 1 mm, 2 mm, and 3 mm, the PR of conventional GI method between shifted and non-shifted 3D doses were 99.9%, 97.4%, and 89.7% for PTV, 99.8%, 84.8%, and 63.2% for spinal cord, and 100%, 99.5%, 91.7% for right lung. The minimum PRs from the RI method were 98.9%, 96.9%, and 89.5% for PTV, and 96.1%, 79.3%, 57.5% for spinal cord, and 92.5%, 92.0%, 84.4% for right lung, respectively. The maximum PRs from the RI method were equal or less than the PRs from the conventional GI evaluation. CONCLUSION Designed 3D RI method showed more strict acceptance level than the conventional GI method, especially for OARs. The RI method is expected to give the degrees of risks in the delivered doses, as well as the degrees of agreements between calculated 3D doses and measured (or simulated) 3D doses.

SU‐E‐J‐37: The Validation Tool for Compensation of Patient Positioning Error Using DRR Images

PURPOSE The present study was designed to develop the validation tool for compensation of patient positioning error using digitally reconstructed radiograph (DRR) extracted from three-dimensional computed tomography (3DCT) and two orthogonal kilo-voltage x-ray images. METHODS To generate DRR image from 3DCT, the ray casting which is most straightforward method was applied in this study. The traditional ray casting algorithm finds the intersections of a ray with all objects, voxels of the 3DCT volume in the scene, with nearest-neighbor interpolation method. Similarity between extracted DRR and orthogonal image was measured by using normalized mutual information method. All process was done by using Matlab. Two orthogonal image was acquired from Cyber-knife system from anterior-posterior view and right lateral view. 3DCT and two orthogonal image of an anthropomorphic Alderson-Rando phantom and head and neck cancer patient were applied in this study. Finally, we designed graphic user interface (GUI) for easy use. RESULTS Registration accuracy with average errors of 2.12 mm ± 0.5 mm for transformation and 1.23° ± 0.4° for rotation using an anthropomorphic Alderson-Rando phantom has been acquired. CONCLUSION We demonstrated that this validation tool could compensate the patient positioning error. For further study, with the developed validation GUI tool for compensation of patient positioning error, we will add the registration tool by manual/auto using cone-beam CT and kilo-voltage CT image to utilize clinically in heavy-ion radiation treatment center in Korea which scheduled for completion in 2016.

SU‐E‐T‐509: Photon Spectrum Modeling of Flattening Filter Free (FFF) Beam and the Optimization of Model Parameters

PURPOSE To determine the distribution of photon spectrum on flattening filter free (FFF) beams, novel and fast optimization methods that were applicable on a convolution/superposition dose calculation algorithm were implemented. METHODS Two-step optimization method was designed to model the virtual photon spectrums for FFF beams. At first, simple functional form of photon spectrums proposed by E. S. M. Ali was modified and used to make rough shapes of photon spectrum. The distributions of photon spectrums were defined at various field sizes (FSs) to consider the changes of the contribution for scattered photons. Percent depth doses (PDDs) at various FSs were used, and collapsed cone convolution (CCC) algorithm was used to calculate PDDs by considering cone-shaped photon fluence in fields. At next, an arbitrary functional form of photon spectrums where the values of photon intensity itself were free parameters was designed. Line search method was used for optimization and gradient terms at each free parameter were derived from CCC algorithm to enhance the speed of iterations. RESULTS The mean energies of the optimized spectrums were decreased from 1.40 to 1.21 MeV for 6 MV FFF beams and from 2.45 to 1.27 MeV for 10MV FFF beams as FSs were increased from 3×3 to 40×40 cm2 because of the contributions of scattered photons. The shape of the spectrums were not greatly changed with field sizes, but root mean squared differences (RMSDs) between the measured PDDs and the calculated PDDs using optimized spectrums were increased upto 0.87% as the FSs were decreased to 3×3 cm2 . CONCLUSIONS Developed method for spectrum modeling showed good agreements when the PDDs were calculated with the optimized results. Suggested method is proper to the radiation treatment planning systems because it only requires measured PDDs, and based on the analytic dose calculation algorithm. This work was supported by the National Research Foundation of Korea(NRFK)⌉ grant funded by the Korean government (MEST)⌉ (Grant No. K20901000001-09E0100-00110⌉), Varian Medical Systems, NCI Grant No.1R01 CA98523⌉, and NSF Grant No. 0854492⌉.

SU‐E‐T‐566: A Dosimetric Comparison of Partial Breast Irradiation Technique Depending on the Tumor Locations in Patient's Breast

PURPOSE The present study was designed to suggest the optimal treatment modalities among 3D-conformal radiation therapy, electron beam therapy and helical-tomotherapy depending on classified tumor locations in patients breast using Partial Breast Irradiation (PBI) technique. METHODS The CT dataset of six patients who had received lumpectomy treatments was used to treatment plans of 3D-conformal radiation therapy, electron beam therapy and helical-tomotherapy. The tumor locations were classified into eight sections according to the quadrants of the breast and to the superficial or deep positions. To evaluate the dosimetric results from the suggested treatment modalities and tumor locations, conformation number, radical dose-homogeneity and delivered doses to normal tissue were calculated. Kruskall-Wallis, Mann-Whitney U and Bonferroni method was used as statistical analysis. RESULTS Helical-tomotherapy is not recommendable method in PBI technique because the dosimetric results from radical dose- homogeneity and the delivered dose to organ at risk showed worse results at all tumor locations compared to other modalities. For helical-tomotherapy, the volume of heart received more than 2.5 Gy was maximized more than 90% of total heart volume at all tumor locations. Electron beam therapy showed good sparing effect to normal tissues and acceptable target coverage in Lower Inner Quadrant-Superficial (LIQ-S) and Lower Inner Quadrant- Deep (LIQ-D) cases. CONCLUSIONS Electron beam therapy could be recommendable method to treat tumor at LIQ-S, LIQ-D locations rather than 3D-conformal radiation therapy, helical-tomotherapy at PBI technique because electron beam therapy is considered to provide the acceptable target coverage and the greatly lower dose to surrounding tissue.

SU‐E‐T‐296: Optimization of the Energy Selection System with Varying Magnetic Field for Laser‐Accelerated Proton Beams

PURPOSE Laser-based accelerated proton beams are unsuitable for clinical use because of their broad energy spectra. For this reason, it is essential to employ an energy selection system (ESS). The purpose of this study is to determine optimum parameters of the ESS which uses a varying magnetic field to generate Bragg-peak. METHODS We simulated an accelerated proton beams using radiation pressure acceleration mechanism with carbon-proton mixture target. The density ratio (n = 6) between the protons and the carbon ions is one of optimization parameters in determining the accelerating mechanisms. The ESS was implemented by the Geant4 Monte Carlo toolkit. In order to optimize the hole size and position of the energy selection collimator, and magnetic field at ESS, these parameters were simulated for acquiring energy and dose distributions by changing each values. RESULTS The proton energy distributions had a poly-energetic distribution after passing through the ESS. As the magnetic field was increased, the mean energy of the proton beams also was increased. Also as the hole size was increased, the energy bandwidth of proton passed through the ESS was increased. The hole size and position of the energy selection collimator were effectively optimized to 2 cm and 5 cm from the z-axis, respectively. CONCLUSIONS We simulated laser-accelerated proton beams using ESS for generation of Bragg-peak. Our results suggest that the ESS with magnetic field variation can effectively generate a Bragg-peak suitable for use in proton radiation therapy. Our ESS can be applied to pencil beam scanning proton therapy.

Development of a Beam Source Modeling Technique for a Flattening Filter Free (FFF) Beam

This study was focused on a new beam source modeling technique for a flattening filter free (FFF) beam. The model was based on a previous three source model, and improved by introducing off axis ratio (OAR) of photon fluence to the primary and scattered photon sources to generate cone shaped dose profiles. The model parameters and the OAR were optimized from measured head scatter factors and a dose profile with 40 x 40 cm2 field size by using line search optimization algorithm. The model was validated by comparing various dose profiles on 6 and 10 MV FFF beam from a True Beam STx linear accelerator. Planar dose distributions for clinically used radiation fields were also calculated and compared with measured data. All calculated dose profiles were agreed with the measured data within 1.5% for 6 MV FFF beam, and within 1% for 10 MV FFF beam. The calculated planar doses showed good passing rates (> 94%) at 3%/3 mm of gamma indexing criteria. This model expected to be easily applicable to any FFF beams for treatment planning systems because it only required measured PDD, dose profiles and output factors which were easily acquired during conventional beam commissioning process.

SU‐GG‐T‐614: The Implementation of Auto‐Optimization Method to Determine Photon Energy Spectrums and Dose Profiles with Various Field Sizes for Collapsed Cone Convolution Algorithm

Purpose: As a part of developing treatment planning system, auto‐optimization method was designed to determine photon energy spectrum and the shape of dose profile with various field sizes.Method and Materials: The initial distribution of energy spectrum was designed using approximate formulation. Multi‐spectrum kernels were constructed with the determine photon energy fluence using published mono‐energy kernel, and collapse cone convolution algorithm was implemented to calculate PDDs and dose profiles by considering the effects of beam hardening with depth, kernel tilting, beam softening with off axis distance, off axis ratio, and finite beam source size. Objective function was defined with the sums of differences between measured PDD or dose profiles and calculated PDD or dose profiles. Auto‐optimization method based on steepest decent method was used to minimize objective function by calculating gradient values at each iteration.Results: Determined energy spectrums were not similar to rear photon spectrums, but calculated PDDs showed good agreements with measured data. The dose error beyond build‐up depth was less than 2 % at any depth. Some under‐estimated results (>5%) were showed at the buildup regions ( 30 x30 cm2) because not considering electron contamination. Calculated dose profiles were also well agreed with measured dose profiles (<3% errors) up to 10×10 cm2, but some discrepancies at penumbra region were appeared at large field. Conclusion: Determined energy spectrums of photon beam with various field sizes were useful to simulate the variation of PDDs. Multi‐spectrum model with field sizes based on auto‐optimization was good approximated method to simulate measured dose distribution, and it seemed to compensate the limitation of classical convolution/superposition algorithm using finite radius of kernel.

SU-GG-T-386: Monte Carlo Study of Absorbed Dose to Solid Water for External Auditing

Purpose: An external audit for radiotherapy has been performed by the third parties to maintain a uniform quality of patient care among different facilities. Among a lot of auditing items, we developed a method to determine absorbed dose to water from Farmer‐type ion‐chamber measurements in solid water within the context of AAPM TG‐51 protocol. The AAPM T‐51 protocol has liquid water as a phantom material for clinical reference dosimetry. Instead of liquid water, the use of a solid phantom is convenient for an external auditing of busy clinics. Method and Materials: Due to different compositions and densities of solid water, correction factor for irradiation geometry and chamber responses are needed to convert measurements in solid water into absorbed dose to water. To avoid the complexity of measurement we kept the reference condition of solid water identical to that of water. The compositions and homogeneity of commercial solid water are varied among different manufactures and not consistent with the compositions provided by manufacturers. Therefore, the compositions of solid water used in this study were experimentally determined by using an electron probe micro‐analyzer (EPMA). The absorbed dose conversion factors for the solid water phantom were measured and, calculated by using the EGSnrc Monte Carlo code system. Results:Measured and calculated conversion factors under the reference condition were in a range of 1.00 – 1.005 for 6–15 MV photon beams, and 1.001 – 1.021 for 6–20 MeV electron beams. The total uncertainty of TG‐51 protocol measurement using solid water was determined to be ±1.5%. Conclusion: The measurement time (including setup time of a solid water phantom and a chamber for several photon and electron beams) was typically less than 30 minutes for external auditing. Conflict of Interest: Research sponsored by Korea Institution of Nuclear Safety Corporation

Development of IMRT Treatment Planning System

Development experience of IMRT treatment planning system focused on the objective function and optimization method was presented. Beamlet based fast dose calculation algorithm was implemented to calculate Dij matrix and a quadratic objective function based on DVH volume constraint was implemented. It was solved with a steepest descent method and the optimized intensity map could be obtained. For the verification of the algorithm, several test phantoms and sample plan was tested. As a result, all the dose volume histogram satisfied the treatment prescriptions and the high dose regions maintained the shape of PTV and did not violated OAR. In conclusion, it could incorporate clinical treatment prescriptions into optimization module and generate corresponding dose distributions within reasonable time.