Index of content:
Volume 34, Issue 6, June 2007
- Therapy Scientific Session: Auditorium
TH‐E‐AUD‐01: Photon Energy Spectrum Emitted by a Novel Polymer‐Encapsulated 103Pd Source and Its Effect On the Dose Rate Constant34(2007); http://dx.doi.org/10.1118/1.2761765View Description Hide Description
Purpose: Although two independent groups have reported dosimetry parameters for the newly marketed OptiSeed103 source (made with a novel polymer encapsulation), there is currently no AAPM recommended consensus value for these parameters in clinical dosimetry. The aim of this work was to perform an independent determination of the dose‐rate constant (Λ) for the OptiSeed103 source using a technique based on photon spectrometry. Method and Materials: Three OptiSeed103 sources (Model 1032P) with known air‐kerma strength were used in this study. The photon energy spectrum emitted by each source along the radial direction in its bisector was measured with a high‐resolution Germanium spectrometer designed specifically for low‐energy photons. The dose‐rate constant of each source was calculated from its emitted relative energy spectrum with explicit consideration of the source geometry. Results: Unlike other sources made with traditional titanium encapsulation, the photons emitted by the OptiSeed103 source exhibited only slight spectral hardening; yielding a relative energy spectrum nearly identical to that of the bare . The dose‐rate constant calculated from the measured energy spectra was 0.664±0.025 cGyh−1U−1 in water. This value agreed, within experimental uncertainties, with the Monte‐Carlo‐calculated value (MCΛ) of 0.665±0.014 cGyh−1U−1 and the TLD‐measured value (with Monte‐Carlo‐calculated solid‐phantom‐to‐water conversion factor) of 0.675±0.051 cGyh−1U−1 reported by one group (Appl. Ratiat. Isotopes. 63, 311–321, 2005). However, it differed by −6.7% from the MCΛ of 0.712 ± 0.043 cGyh−1U−1 reported by the other group (Phys. Med. Biol. 50, 1493–1504, 2005). Conclusions: The results obtained in this work provide additional information needed for establishing a consensus value for the dose‐rate constant of OptiSeed103 source. It suggests that the eventual consensus value of Λ is likely to be closer to a value of 0.670 cGyh−1U−1 rather than 0.693 cGyh−1U−1 as currently recommended by the manufacturer based on the average of two previously reported values.
34(2007); http://dx.doi.org/10.1118/1.2761766View Description Hide Description
Purpose: Interstitially delivered gene therapy using GeneSeeds (small metallic containers similar in size to brachytherapy sources) may have many valuable properties for cancertreatment. We are currently pursuing both simulation and experimental work to more fully understand the dosimetryproperties of GeneSeeds and their potential use as both a direct tumoricidal agent and a co‐therapy for enhancing tumor radiation sensitivity. Method and Materials: Simulated GeneSeed dosimetry was based on modeling of the diffusion process using the physical properties of the GeneSeed and its potential biochemical payloads. Experimental data was drawn from pathology and immunohistochemistry performed on sectional slices of mouse xenograft tumors into which single GeneSeeds had been implanted. These data were combined into a computational model (implemented in Matlab) which computes the dose as a function of distance and orientation from the GeneSeed using cylindrical coordinates. We have also developed an interface for placing multiple GeneSeeds in a 3‐D space and computing their overall effect on a regular Cartesian coordinate grid, to facilitate integration with conventional treatment planning systems. Results: GeneSeed dosimetry differs from conventional brachytherapydosimetry by having a sigmoidally‐shaped falloff as opposed to an exponential falloff. This implies that positioning of the GeneSeeds may require a higher level of precision than what is expected for conventional brachytherapy with radioactive sources. However, homogeneous coverage of larger regions can still be achieved using controlled inter‐seed spacing. Image guided robotics may be of assistance in achieving the level of precision needed for this application. Conclusion: We have developed dosimetry models for single GeneSeeds based on a combination of experimental and simulation data. These models can be used to predict overall dosimetry of a larger distribution of GeneSeeds, similar to calculations for groups of brachytherapy seeds. GeneSeeds may have potential use in many organ targets.
34(2007); http://dx.doi.org/10.1118/1.2761767View Description Hide Description
Purpose: To understand the propagation of electrons through the acceleration gradient of the Xoft micro‐miniature x‐ray source via electrostaticmodeling of the fields and x‐ray pinhole camera imaging. Materials and Methods: 1) A 3D model of the source was built using OmniTrak3, and electron trajectories were traced from the hot cathode to the x‐ray producing anode. A finite element model of the filament temperature profile was input to a Richardson‐Dushman thermionic emission/extraction model to determine the electron emission density distribution. Assumptions made were that the field was graded linearly along the length, and that electrical connection between conductive and resistive components was ideal. 2) A ShadoCam4 x‐ray sensitive camera was used to acquire images of sources through a 30 μm pinhole mounted 15 mm from the source. The camera was 105 mm from the pinhole, providing a magnification of 7×. The camera was read out through a USB computer interface. Images were typically acquired over ten second integration times. Results: The actual images showed distinct patterns that were identifiable with model predictions. Structures associated with particular emission locations and launch angles on the filament were clearly observed. Conversely, patterns in the images were explainable in terms of parameters such as the absolute location of the filament. Image patterns which were relatively similar to the nominal had no correlation with measured spatial distribution patterns of the sources, but where there were notable variations in the image patterns it was possible to infer correlations with polar and/or azimuthal measurements. Conclusions: Electron trajectories within the accelerating structure were explored through pinhole imaging and computer modeling, and sensitivities to mechanical tolerances were established that were valuable in setting manufacturing tolerances.
This study was funded by Xoft, Inc.
3 Field Precision LLC, Albuquerque, NM.
4 Rad‐icon Imaging Corp, Santa Clara, CA.
34(2007); http://dx.doi.org/10.1118/1.2761768View Description Hide Description
Purpose: To accurately determine the exit skindose from MammoSite® Radiation Therapy System treatments using thermoluminescent dosimeters (TLDs) placed on the surface of the skin during each treatment fraction. Methods and Materials: Well‐characterized TLD‐100 chips were calibrated using to establish an energy response correction factor for measurements. TLD response vs. material thickness was measured and compared for thicknesses of Virtual Water™ and breast‐equivalent material from 2 to 10cm. Monte Carlo simulations were performed to determine the relationship between the dose to TLD and dose to basal skin layer for a range of treatment parameters (size and content of balloon, distance from skin, etc). Treatment planning system (TPS) predicted skindoses were also compared to the TLD measured skindose for 29 patients and two phantom setups. Results: The TLD‐100 energy response for relative to was determined to be 1.045. The TL vs. thickness curves for Virtual Water™ and breast‐equivalent material were found to be within 2.5% for all thicknesses studied for , and can be used interchangeably within the 2.5% limit, which falls within our measurement uncertainty of 3%. The Monte Carlo calculated dose to TLD agreed with the 1/r2 corrected basal skindose to within 0.5% for the entire range of treatment parameters studied when the dose to TLD vs. dose to water correction factor of 1.2 was applied. TPS predicted doses overestimated the TLD measured skindose by an average of 26% for the group of 29 patients studied, and by an average of 40% for the two phantom setups. Conclusions: TLDs placed on the surface of the breast during MammoSite RTS treatments can accurately measure the skindose when the proper correction factors are applied. The TLD measured dose is much more accurate than the TPS predicted dose.
TH‐E‐AUD‐05: Evaluation of Inter‐Fractional Dosimetric Variations in MammoSite Partial Breast HDR Irradiation34(2007); http://dx.doi.org/10.1118/1.2761769View Description Hide Description
Purpose: To evaluate the inter‐fractional dose variations due to changes in balloon shape and location in MammoSite partial breast HDR irradiation. Method and Materials: Eleven MammoSite HDR patients were treated with a dose of 34 Gy delivered in 10 fractions twice a day at our institution. For each of these patients, a plan was generated for the first fraction based on the CT scan acquired before treatment. The plan was then used for the remaining 9 fractions without any modification unless a significant change in balloon shape and/or location was observed on the CT scan acquired prior to each treatment fraction. To assess the inter‐fractional dose variations, we retrospectively contoured the target (including PTV_EVAL for planning evaluation) as well as critical structures and positioned the catheter on the CT datasets of fraction 2 – 10. Then, we generated two plans utilizing a multiple and single dwell position approaches while using the same dwell time distribution as for the clinical plan of the first fraction. A total of 220 plans generated on 110 CT datasets were evaluated using the following dosimetric metrics: PTV_EVAL coverage, target dose homogeneity index (DHI), target dose conformal index (COIN) as well as maximum dose to ipsilateral lung and skin.Results: For the multiple dwell position approach, the average (maximum) percent inter‐fractional variation relative to the first fraction was 1.5% (6.2%) for PTV_EVAL coverage, 1.1% (12.5%) for DHI, 3.4% (11.8%) for COIN, 10.2% (42.6%) for maximum ipsilateral dose, and 6.9% (29.4%) for maximum skin dose. No difference in the dosimetric values was observed using either a multiple or single dwell position approach. Conclusion: The inter‐fractional dose variation was patient‐specific and dependent upon the balloon deformity and location. While the average variation may not significantly impact the treatment, the maximum variation may be clinically significant.
TH‐E‐AUD‐06: HDR Brachytherapy of Prostate Patient in the Presence of Bi‐Lateral Hip Prostheses Using Megavoltage Cone‐Beam CT34(2007); http://dx.doi.org/10.1118/1.2761770View Description Hide Description
Only a rare concern a few years ago, an increasing number of patients with hip replacements are presenting for prostate radiotherapy.Image artifacts caused by the presence of hip replacements render CTimages useless for prostate delineation and catheter definition. We report on the use of a Mega‐Voltage Cone‐Beam CT (MVCBCT) system to identify the catheters and to complement the regular CT for target definition of High Dose Rate Brachytherapy of prostate cancer patients with hip replacements.
Following a routine trans‐rectal ultrasound guided catheter implant and recovery, MVCBCT and regular CT acquisitions were performed within few minutes. Metallic wires were inserted in the catheters during the MVCBCT image acquisition. A series of filters were applied on the MVCBCT image to maximise spatial resolution and image quality. The MVCBCT and CTimages were registered with the Anatomy Modelling module of the planning system using the catheters as landmarks. The fused images were used to delineate the target and organs at risk. Then, the MVCBCT and the volumes of interest were transferred to the brachytherapy planning system where catheters were identified solely from the MVCBCT image. The dose distribution was optimized by following our standard technique using the inverse planning IPSA.
Fused MVCBCT‐CT images greatly facilitated target delineation. The registration precision was better than 1.5 mm in the prostate and catheter areas. Catheters were easily visible and accurately defined on MVCBCT. The same dosimetric criteria used for regular HDR prostate planning were and achieved on this patient.
HDR treatment based only on CT is not possible for patients with bi‐lateral hip replacements. The use of MVCBCT for planning purpose allows to offer prostate and other pelvic patients with hip prosthesis the most advanced form of HDR brachytherapy.
Research supported in part by Siemens and Nucletron.
TH‐E‐AUD‐07: Measurement of the Effect of Shielding in a Tandem and Ovoid Applicator Used in HDR Brachytherapy34(2007); http://dx.doi.org/10.1118/1.2761771View Description Hide Description
Purpose: To analyze the effect of shielding present in a Fletcher Suit Delcos ovoid by comparing the dose distribution around it computed by treatment planning system (TPS) to the dose distribution measured using radiochromic film dosimetry for HDR remote afterloader. Methods and Materials: Gafchromic/EBT films were carefully wrapped around the ovoids [diameter 2.0 & 2.5cm] including its anterior and posterior aspects. Origin was marked at the center of one the sources on the colpostat and on the film for spatial coordination. The ovoid was radiated to a dose of 300 cGy. The films were scanned using Vidar VXR‐16 Scanner and analyzed using RIT software. The dose distribution in the planes above, below and on the sides of the ovoid was obtained. 3D analyses of the dose profiles were carried out and the results were compared with the dose distribution computed by TPS which does not account for the presence of shields. Results: The dose reduction in the anterior part of the colpostat affects maximally the dose to the bladder where a reduction up to 15 % [average 11.6%] was noted. The reduction of dose in the posterior part of the colpostat which is designed to shield rectum was found to be as high as 25%. The effect of shields can be seen in some part beyond the shields in a typical subtended solid angle. In the high dose region in the central plane where the shields are not present, insignificant difference in the measured & computed dose values was noticed. Conclusions: Shields were more effective in reducing dose to the rectum than for bladder primarily due to the applicator design. The present TPS substantially over‐estimates the dose to bladder and rectum including regions which lie in the shadow of the solid angle subtended by the shields.
- Clinical Measurements
MO‐D‐AUD‐01: A Comprehensive Study On the Heterogeneity Dose Calculation Accuracy in IMRT Using An Anthropomorphic Thorax Phantom34(2007); http://dx.doi.org/10.1118/1.2761232View Description Hide Description
Purpose: To provide a comprehensive study on the accuracy of many commonly used Intensity Modulated Radiation Therapy(IMRT)treatment planning systems using the Radiological Physics Center's (RPC) anthropomorphic thorax phantom. Method and Materials:Treatment planning systems (TPSs) from Corvus, Eclipse, Pinnacle, and Tomotherapy were evaluated using the RPC anthropomorphic phantom. Treatment plans were designed using the same clinical constraints and prescriptions so that 96% of the planning target volume (PTV) was covered by the prescription dose. The phantom is equipped with TLD located in the tumor,heart, and spinal cord and radiochromic film located in three anatomical planes intersecting the tumor center and extending into the lung.IMRT QA was performed to adjust the calculated dose distributions in order to isolate the effects of heterogeneity. Comparisons were made between each TPS calculation and measurement. In two instances, re‐calculations of the original correction based pencil beam (PB) plans were performed using the superposition convolution (SC) method. Results: TPSs employing superposition convolution algorithms predicted dose within 3.6% of the target TLD, while TPSs using correction based pencil beam algorithms predicted dose within 5.0% of the target TLD. Both algorithm types showed variations (2% to 38% in the cord and heart) in predicting low dose to normal structures. The dose distributions within the PTV and penumbra lung regions showed good agreement when using an SC algorithm. However, TPSs using the PB type algorithm overestimated dose in the PTV and underestimated the extent of penumbra broadening corresponding to the surrounding lung.Conclusion: This work demonstrated that superposition convolution algorithms found in widely used IMRTtreatment planning systems are able to calculate the dose accurately to the PTV and penumbra regions when low density heterogeneities are involved. Conflict of Interest: This work supported by PHS CA010953 and CA081647 awarded by NCI, DHHS.
MO‐D‐AUD‐02: Liquid Scintillation‐Based Proton Residual Range Measurement Using a Dynamic Biological Lung Phantom34(2007); http://dx.doi.org/10.1118/1.2761234View Description Hide Description
Purpose: To test and apply a scintillation‐based method to directly observe proton residual range and range variation in a dynamic biological lung phantom. Method and Materials: The dynamic biological lung phantom using a preserved swine lung has previously been compared to human lung and evaluated as an experimental platform for IGRT studies. Fiducial markers and a 5 cc artificial tumor were placed in the lung, and the phantom was imaged on a GE Lightspeed CT in 4D mode before and after seed and “tumor” implantation. The resulting 10‐phase DICOM datasets were evaluated in terms of anterior‐posterior (AP) water‐equivalent path length (WEL) variation throughout the ventilatory cycle. The preserved swine lung was irradiated in the AP direction with a 170 MeV proton beam through a 1 cm × 10 cm slit aperture. Proton residual range and range variation were directly observed using a custom‐built lucite chamber filled with scintillating fluid, monitored during irradiation by a video camera.Images were then analyzed using ImageJ to determine residual proton range. Results: WEL analysis on the imagedlung suggests a range of lung WEL values between 8 and 25 mm (average ∼16 mm). Peak calculated WEL difference at the tumor margin after implantation during ventilatory motion was approximately 5mm. Irradiation of the phantom demonstrated regional range variation across the lung on the order of 13 mm, with total WEL values equivalent to calculated values. Residual range variation due to ventilatory motion was less significant, at some points on the order of ±2.4 mm. Conclusion: The dynamic biological lung phantom in conjunction with the scintillation chamber has been shown to be a simple, inexpensive, and effective tool for the measurement of proton residual range and significant residual range variation in a complex biological system.
MO‐D‐AUD‐03: Verification of Lung Tumor Doses Calculated by the Eclipse AAA and Pinnacle CC Algorithms34(2007); http://dx.doi.org/10.1118/1.2761235View Description Hide Description
Purpose: The goal of this study was to investigate the dosimetric accuracy of the analytical anisotropic algorithm (AAA) for the treatment of lungtumors. This algorithm, recently implemented in Eclipse (Varian Inc., Palo Alto, CA), was also compared to the Pinnacle's (Philips Medical, Cleveland, OH) collapsed cone (CC) algorithm. Methods: Four phantoms were used in this study. Three lung‐, bone‐ and water‐equivalent slab phantoms were designed to validate the accuracy of the calculated doses in simple geometries. Doses were additionally measured in the CIRS (Norfolk, VA) thorax solid water phantom, which includes lung cavities and a cylindrical spine. All four phantoms were CT‐scanned with thermoluminescent dosimeters(TLD) in place. Plans were generated with the AAA algorithm using anterior‐posterior (AP), AP/PA, oblique, and intensity modulated (IMRT) beams with both 6MV and 18MV photons. The dose distributions for all plans, excluding the IMRT, were re‐computed with Pinnacle using monitor units matched to the AAA plans. Dose measurements were performed with TLDs and ion chambers at 6 different positions including three in tissue, one in spinal cord, and two in lung.Results: The AAA algorithm was found to calculate the dose in lung and tissue accurately to within 6% in all four phantoms. Ion chamber measurements showed 2–3% better agreement than the TLDs in the thorax phantom. For 6 MV, the differences between measured and calculated doses were less than 2% for both AAA and Pinnacle for all six points. However, for the 18 MV beam, the dose measured in the spinal cord was 5.9% greater than the calculated dose with both algorithms. Measured and calculated doses agreed to within the 2% for the IMRT plan. Conclusion: The AAA algorithm provides accurate dose calculations in and around heterogeneities, similar to that provided by the CC algorithm implemented in Pinnacle.
34(2007); http://dx.doi.org/10.1118/1.2761236View Description Hide Description
Purpose: To compare the measured dose distribution with the planned distribution over a PTV centrally located within the lung when heterogeneity corrections are taken into account. Method and Materials: The Radiological Physics Center's anthropomorphic thorax phantom includes a target (∼ 1 g/cm3) centrally located in the left lung (∼ 0.33 g/cm3). The phantom was sent to 25 institutions, each of which was instructed to design and deliver a stereotactic treatment plan. The plan was intended to deliver 20Gy (homogeneous calculation) to ⩾ 95% of the PTV and limit the lungdose at point 2 cm from the PTV edge to 11.7 Gy. The institutions were asked to recalculate the dose distribution with the heterogeneity correction using the monitor units determined from the homogeneous calculated plan. TLD and radiochromic films were used as dosimeters within the target region. Results: A total of 17 institutions met the phantom irradiation criteria: +/− 5% for DTLD/DInst, and +/−5mm DTA on all sides of the PTV, based on the heterogeneous calculated plan. For these irradiations, the delivered doses over the central 80% of the PTV were compared to the planned doses along 3 orthogonal profiles through the PTV. An average of 85% of the points in the profiles from the cases calculated with the superposition/convolution algorithm were within 5% of the calculation, while only 69% of the points from the plans using pencil beam and Clarkson were within the 5% of the plan. Conclusions: The superposition/convolution heterogeneity correction algorithm showed better agreement with the measured dose distribution across the PTV than the pencil beam and Clarkson algorithms because it more accurately accounted for the lack of lateral scatter.
Work supported by PHS grant CA10953 and CA081647 from the NCI, DHHS.
MO‐D‐AUD‐05: Characterization of Dose in Heterogeneous Situations: A Comparison of Treatment Planning System and Computer Aided Second Check Dosimetry QA Software Dose Evaluations34(2007); http://dx.doi.org/10.1118/1.2761237View Description Hide Description
Purpose: To quantify and compare the dosimetric predictions calculated by a treatment planning system, second check dosimetric computer QA software, and ion chamber measurements in heterogeneous situations for 6 and 18 MV photon beams. Method and Materials: An assortment of plastic tissue equivalent materials was used to compare the calculated dose predictions between the Pinnacle3treatment planning system, the RadCalc® QA computer software, and ion chamber measurements. The dosimetric accuracy of the plastic water, lung, and bone equivalent slab materials was assessed and validated through the use of simple geometries. After planning, doses for each slab arrangement were measured on a Varian 21EX accelerator with a second check performed by the RadCalc® computer software. Percentage differences between the computed and measured doses were then compared and quantified, providing information on the accuracy of the dose predictions. Results: Evaluation and comparison between the calculated dose values from the Pinnacle3, RadCalc®, and measured data indicate that discrepancies exist, even for simple geometric setups. Looking at percentage differences, the Pinnacle3 system (−3.82% – 4.33%) more accurately calculates the dose in the heterogeneous locations than does the RadCalc® software (−8.30% – 4.15%). Examination of all measured point locations show only about 4% of the Pinnacle3 system dose calculations, and almost 18% of the RadCalc® software dose calculations, have a percentage difference greater than ±3%. Conclusion: This work explores the clinical application and accuracy of using RadCalc® for dosimetric second checks. Even with the use of heterogeneity corrections, it is still not guaranteed that an accurate dose calculation will result when heterogeneous material is present. The CIRS Inc. IMRT Thorax and Pelvic 3D phantoms will be utilized for the continuing investigation of the accuracy of dose calculations performed by Pinnacle3 and RadCalc® involving additional complex geometries in a more anatomical construct.
MO‐D‐AUD‐06: Dependence of Total Scatter Factors of Small Beams On the Radial Distribution of the Electron Beam Incident On the Target: A Multi‐Detector and Monte Carlo Study34(2007); http://dx.doi.org/10.1118/1.2761238View Description Hide Description
Purpose: To investigate dependence of total scatter factors (sc,p) of small beams used in radiosurgery on the radial distribution of the electron beam incident on the target. The study was performed in order to clarify discrepancies observed in sc,p measurements performed by means of different detectors at the Cyberknife radiosurgery system and was aimed at obtaining a means to infer the electron radial distribution of a specific Cyberknife unit. Method and Materials: PTW PinPoint31014 and Exradin A16 microchambers, PTW30012 diode and TM60003 diamond were used to measure sc,p. The same detectors were simulated by means of the Monte Carlo code BEAMnrc to calculate correction factors for sc,p. BEAMnrc was also used to calculate theoretical values of sc,p. Accuracy of Monte Carlo simulation depends on the choice of energy, divergence and radial distribution of the electron beam incident on the target. The energy was determined by comparison to experimental TMRs. Radial distribution of the electron beam (expressed as FWHM, gaussian shape) was chosen to be optimal when correction factors of the 4 detectors were such that corrected sc,p converged to the same value within +/−1%. Results: measured sc,p of the 5mm collimator averaged 0.638 −4%+11% with the 4 detectors. E=7.2MeV best matched calculated to experimental TMRs. Optimal FWHM=2.3mm gave sc,p=0.673 −0.7%+0.3%. Correction factors decreased with increasing FWHM, while Monte Carlo‐calculated sc,p increased. Measured (corrected) and calculated sc,p matched at 0.673 only for FWHM=2.3mm. Conclusion: Variations of electron beam focussing can explain significant variations of sc,p. If one of the investigated detectors is used, it is possible to infer actual FWHM values and thus appropriate correction factors by comparison to pure Monte Carlo‐calculated sc,p values as a function of FWHM. This fact could be exploited by centres who do not have access to Monte Carlo codes to simulate their own system.
MO‐D‐AUD‐07: Determination of Output Factors for Stereotactic Radiosurgery Beams by Monte Carlo and Measurements34(2007); http://dx.doi.org/10.1118/1.2761239View Description Hide Description
Purpose: The Varian Trilogy accelerator provides high dose rate (1000MU/min) photon beams for stereotactic radiosurgery(SRS). Different output factors have been reported and their use may impair observations of dose response and optimization of prescribed dose. In this work, we investigated the output factors for the Trilogy 6 MV SRS beam using Mote Carlo simulations and measurements. Method and Materials: The Trilogy SRS cone collimators are 5 mm to 30 mm in diameter. Chamber measurement of output factors is difficult for small cone sizes. In this work, the measurement was carried out using a 0.015cc pinpoint chamber and Gafchromic EBT film. The Monte Carlo simulation was performed using the MCBEAM/MCSIM codes for linac head simulation and phantom dose calculation. The Monte Carlo calculations were validated using the measurement results and compared with the Varian recommended beam data. Results and Conclusions The agreement in percent depth doses and dose profiles between the simulation results, the measurement results and the Varian data was within 2%/1mm for the 10×10 cm2reference field and all the cones at a 100 cm source to surface distance. The output factors obtained from the Monte Carlo simulations were in excellent agreement with the values from the film measurements. Similar agreement was found between the Monte Carlo results and the pinpoint chamber values except for cones with diameter less than 20 mm where the 2 mm chamber diameter has become comparable to the field size. However, the Varian recommended output factors are consistently higher than the Monte Carlo results, especially for the 5 mm cone, where the difference reaches 10.9%. Therefore, great caution must be taken with the use of the Varian recommended beam data.
MO‐D‐AUD‐08: Dose Verification at the Surface of Air Cavities During Radiation Therapy Using the TomoTherapy Hi‐Art System34(2007); http://dx.doi.org/10.1118/1.2761240View Description Hide Description
Purpose: Air cavities are a significant inhomogeneity in radiation therapy. TomoTherapy, a helical dose delivery system, is dependent on a heterogeneity based treatment planning system. Plan optimization can be affected by changing the pitch of treatment delivery along with the modulation factor for the treatments. The system allows for calculations using a fine, normal, or coarse calculation grid. In this investigation, we assess the dose delivered to the surface and superficial regions of the cavity, the influence of the above parameters on dose delivery, and the accuracy of the planning system to represent the dose in these regions. Method and Materials: Four different 3×3 cm2 air cavities configurations in solid water were investigated‐with 9 and 1cm solid water above the cavity, with the cavity open at the surface, and with no cavity to measure the surface dose. Each cavity had a reference plan, two plans changing pitch, and two plans changing modulation. We compared the predicted dose to measureddose using both the Attix chamber and TLDpowder.Results: A large variation in the percent difference between measured and expected dose depends on the location of the air cavity below the phantom surface. The accuracy of the delivered dose varied with both modulation factor and pitch dependent on the cavity location in the phantom. The Attix chamber and TLDpowder showed essentially the same response, but the Attix chamber data was more precise. Conclusions: The accuracy of the delivered to expected dose on the cavity surface was greater when the air cavity was far below the surface (0–5%). As the cavity moved closer to the surface, the deviations in measured to expected dose increased to minus 40–50% of the measureddose at the surface.
34(2007); http://dx.doi.org/10.1118/1.2761241View Description Hide Description
A method based on dose‐response function was developed in the past, which verifies the beamlet weight of intensity modulated radiation therapy(IMRT) from doseimage in electronic portal imaging device(EPID) and reconstructsdose in a patient. The establishment of a linear relationship between beamlets and dose responses in patient and in electronic portal imaging device(EPID) was the key to this methodology. The responses are to be predetermined by a full‐scope Monte Carlo calculation in patient and EPID structures. In this study, the method was validated through measurement using in‐phantom and exit film dosimetry which simulated in‐patient and EPID measurements, respectively. The in‐phantom film was inserted at 100 cm from the target and at 12.5 cm depth within a 25‐cm thick phantom and the exit film was placed at 139.5 cm from the target and at 2 cm depth within a 4‐cm thick EPID phantom. For the validation, a 6MV X‐ray beam with the size of 6 × 6 cm2 was perpendicularly exposed to the phantom. Responses to each beamlet (0.2 × 5 mm2) within a phantom and an EPID phantom were then calculated. Using the calculated responses, the exit film dose was used to inversely reconstruct the in‐phantom dose, which was then compared with the measured in‐phantom dose. In a second study, an IMRT beam intensity reconstruction was investigated computationally. The dose comparison in patient showed a difference of less than 3 %. Some propagated noise was found in the reconstructed intensity distribution, suggesting the need for noise filtration prior to reconstruction. The reconstruction took less than 10 seconds of calculation time and 10 MB of memory. The method is accurate as well as effective for the dosereconstruction of IMRT. A follow‐up study will include detailed modeling of a therapeutic beam and EPID and experiments.
- IMRT: Delivery
34(2007); http://dx.doi.org/10.1118/1.2761660View Description Hide Description
Purpose: To develop an effective and efficient leaf sequencing algorithm for intensity‐modulated arc therapy (IMAT). Methods: The input to our sequencing algorithm includes: (a) A set of (continuous) intensity patterns optimized by a treatment planning system for a sequence of equally spaced beam angles (10 degrees apart); (b) the IMAT maximum leaf motion constraint; (c) the number of arcs, k, that the user specifies based on the complexity of the problem. The output is a set of k treatment arcs that best approximates the set of input intensity patterns at all beam angles. The MLC shapes for each output arc are interconnected to guarantee a smooth delivery without violating the IMAT maximum leaf motion constraint. Our sequencing algorithm consists of two key steps. First, the intensity profiles aligned with each MLC leaf pair at all beam angles are converted into k MLC leaf openings using a k‐link shortest path algorithm, where k is the specified number of arcs for the delivery. The delivered photon flux using these leaf openings best approximates the desired intensity distribution. Second, the leaf openings are connected into k IMAT treatment arcs under the maximum leaf motion constraint using the minimum‐cost matching and shortest path algorithms. Results: The performance of our leaf sequencingsoftware has been tested for four treatment sites (prostate, breast, head‐and‐neck, and lung). In all cases, our leaf sequencing algorithm provides efficient and highly conformal IMAT plans that rival the counterpart tomotherapy plans and significantly improve the IMRT plans. Execution times of our software that range from a few seconds to 2 minutes are observed on a laptop computer equipped with a Pentium M Processor of 2.0 GHz. Conclusion: This research provided strong evidence that IMAT equipped with an effective leaf sequencing algorithm can provide a feasible and high quality implementation of IMRT.
34(2007); http://dx.doi.org/10.1118/1.2761661View Description Hide Description
Purpose: To determine the effect of organ motion and gating on high dose rate IMRT step and shoot delivery and to investigate the feasibility to high dose rate IMRT delivery with a gating system. Method and Materials: A 6‐field IMRT plan for livercancer with different monitor units (77 to 225 MUs) was delivered with an in‐house made motor driven phantom, which allows sinusoidal movement with a changeable motion period between 1 second and 10 second. The motion of phantom was set along patient superior‐inferior direction with a 1‐cm amplitude and a 5‐second cycle. A 0.6 c.c. ion chamber was put at the isocenter inside the moving solid water phantom. A gating system was applied with a variety of gating settings, including the phase range and the length of the breathing period. Repeated measurements were taken for gated and non‐gated delivery with different gating settings and two dose rates, 500 and 1000 MU/Min. Results: The separate segment doses and total doses of the IMRT plan were compared crosswise with 30∼70%, 30∼60% phase gating, and without gating with 1000 and 500 MU/Min. A reproducible 3.1% difference of the total dose between 1000 MU/Min and 500MU/Min at 30∼70% phase gating setting was observed. The value was verified with the XiO treatment planning system, indicating that the delivery at 1000MU/Min dose rate with gating is closer to the original plan calculation. Our results also show that there was no significant effect from the gating system on small MUs or large fields. Conclusions: Our results using 1000 to 500 MU/Min dose rate suggests that higher dose rate is feasible and beneficial for IMRT treatment delivery with gating; there is no significant effect for small MU numbers or large field size on dose output.
TH‐C‐AUD‐03: Impact of Respiratory Motion On Helical Tomotherapy Dose Delivery: A Phantom Study with 3‐D Volumetric Dose Measurements34(2007); http://dx.doi.org/10.1118/1.2761662View Description Hide Description
Purpose: To investigate the impact of respiratory motion on helical tomotherapy dose delivery with 3D volumetric dose measurements. Method and Materials: Two helical tomotherapy plans were generated for a polystyrene box phantom with simulated target and critical organs using different filed width. Respiratory motion was simulated by a sinuosoidal motion of 15mm longitudinal and 5mm lateral ranges at 15cycle/min. Tight longitudinal and transverse motion margins of 7mm and 3mm, respectively, were used to expand CTV to PTV. Each plan was delivered to the phantom with and without motion for various numbers of fractions for a total of 16 factions. 3D volumetric doses were measured with films in the transverse planes every 6mm throughout the phantom. These volumetric doses were imported to a treatment planning system for analysis. Dose distributions and Dose‐volume‐histograms were compared between the plans and the measurements as well as between measurements with and without motion. V98, and V100 were also examined for the CTV and PTV. Results: The doses delivered to the stationary phantom agreed with the treatment plan. The dose delivered to the moving phantom maintained the general shape of the dose distribution in the treatment plan but was affected by the motion. At some spots, the dose discrepancy between deliveries with and without phantom motion could be up to 20%. Dose distributions and DVHs showed that dose coverage to the PTV was noticeably compromised. In some cases, V98 was reduced by over 7% and V100 over 9% compared to those of stationary phantom measurements. However, the coverage of CTV remained as good as the plan and the stationary measurement, with practically no change in V98 and V100. Conclusion: Respiratory motion can cause significant dose error in helical tomotherapy. However, adequate motion margins on the target can substantially reduce the motion effect on target dose coverage.
34(2007); http://dx.doi.org/10.1118/1.2761663View Description Hide Description
Purpose: The current height (and corresponding transmission) of multileaf collimator(MLC) leaves may be suboptimal for the efficient delivery of IMRT. In this study, we have examined the degree to which the efficiency of step‐&‐shoot IMRT delivery can be improved with a multi‐layer MLC design. Materials and Methods: The multi‐layer MLC design tested in this study is based on the concept of dividing the leaves into multiple layers of equal thickness along the planes perpendicular to central axis. The combined leaf height and transmission would match that of current MLC designs. We modified our in‐house leaf‐sequencing program (CIMO) to accommodate the multi‐layer MLC. CIMO uses a simulated annealing to minimize discrepancies between the optimized and sequenced fluence maps. We assume each layer of leaves can move independently and the leaves within the same layer follow the current constraints imposed by each type of MLC. Using various number of apertures, we performed leaf sequencing studies for both Varian and Elekta MLCs using a total of 20 optimized fluence maps extracted from Pinnacle3. The 1‐norm error obtained between the sequenced and optimized fluence maps using different number of MLC layers were then compared. Results: Fewer aperture shapes per beam are required to achieve the same level of 1‐norm error for the sequenced fluence maps when a multi‐layer MLC was used. On average, using a 2‐layer MLC can reduce the number of apertures by 28% for both Varian and Elekta MLCs. This value increases to 50% when a 3‐layers MLC is used. If the same number of apertures is used, increasing the number of MLC layers to 3 can reduce the 1‐norm error, on average, by 42%. Conclusions: The multi‐layer MLC can dramatically increase the delivery efficiency for step‐&‐shoot IMRT. Future dose verification using Monte‐Carlo simulation will also be performed.