Volume 35, Issue 6, June 2008
Index of content:
- Therapy Moderated Poster Session: Exhibit Hall E
- Moderated Poster — Area 1 (Therapy): IMRT Planning
35(2008); http://dx.doi.org/10.1118/1.2961350View Description Hide Description
Purpose:Treatment planning system/delivery system combinations affect IMRT optimization. Delivery method (step and shoot vs. dynamic), leakage radiation variations (lack of secondary collimators on the Elekta SL‐S MLC head), hardware constraints and leaf edge shape all factor in when a radiation treatment planning system (RTPS) optimizes MLC segments. Different optimization algorithms within treatment planning systems influence IMRT planning results as well. In our study we evaluated delivery efficiencies from Elekta, Varian and Siemens machines coupling with two different treatment planning systems. Method and Materials: 16 IMRT plans were created for prostate and headneck (HN) sites (5 and 6 co‐planar beams respectively) using CMS Xio and Philips ADAC Pinnacle. Plans were optimized using the same dose and DVH objectives and modulated using clinical thresholds. The monitor units needed to deliver the weight point dose for each field were tabulated for each site and machine source data. Results: All plans have similar quality meeting the objectives. HN plans have an average MU efficiency of 5.02 +/− 1.09 MU/cGy. The plans created on the Pinnacle RTPS had a lower mean efficiency but was not statistically significant due to the high standard deviation. Prostate plans were more efficient than the HN plans with an average value of 2.19 +/− 0.47 MU/cGy. Pinnacle prostate plans had a mean MU efficiency of 1.69 +/− 0.18 MU/cGy compared to 2.51 +/− 0.22 MU/cGy from Xio. Conclusion: There is no statistical difference in MU efficiency for HN plans from different planning/delivery system combinations. Prostate plans have higher delivery MU efficiencies than those from HN plans, most likely due to less modulation. For the prostate plans there exist slight differences in MU efficiencies with unknown clinical significance between different planning/delivery systems.
35(2008); http://dx.doi.org/10.1118/1.2961351View Description Hide Description
Purpose: Planning and delivery in HN‐IMRT is challenging for Elekta linacs because of numerous constraints on beam delivery systems. The purpose of this study is to find a set of planning parameters that are applicable to most patients and optimal in terms of plan quality, delivery efficiency and dosimetric accuracy. Method and Materials: Four types of plans were created for each of 12 patients: ideal fluence optimization (FO), conventional two‐step optimization (TS) consisting of FO followed by MLC conversion, segment weight optimization (SW) and direct machine parameter optimization (DMPO). Maximum number of segments (NS) and minimum segment area (MSA) were varied in DMPO. Plan quality was evaluated based on score, dose distributions and dosimetric indices. Delivery efficiency was evaluated by irradiation time, and dosimetric accuracy by Mapcheck. Results: Plan quality deviates most from ideal FO for TS, with slight improvement for SW. DMPO is the closest to FO with the least variation among patients. NS of 80–160 in DMPO yield optimal plans. At larger NS (⩾80), plan quality decreases with MSA as expected, except for MSA <8cm2, which suggests presence of local minima in the DMPO algorithm. The irradiation time is strongly dependent on the plan segments (NSactual), weakly dependent on MUs, and independent of MSA. Typical plans with 79 segments and 8cm2 MSA have ∼747 MU and take ∼ 8 minutes to deliver; this increases to ∼13 minutes for 158 segments. Dosimetric accuracy is independent of DMPO parameters. Conclusion: The superior quality of DMPO plans makes it ideal for planning HN‐IMRT on Elekta linacs and its consistency allows development of class solutions. However, the vulnerability of local minima warrants such a study to systematically evaluate the effect of parameters in new planning techniques. The optimal set of parameters should be chosen to balance plan quality and delivery efficiency.
SU‐DD‐A1‐03: On the Quantification of the Dosimetric Accuracy of Collapsed Cone Convolution Superposition Algorithm for Small Lung Volumes Using IMRT35(2008); http://dx.doi.org/10.1118/1.2961352View Description Hide Description
Purpose: To quantify the accuracy of a collapsed cone convolution superposition (CCCS) algorithm against Monte Carlo(MC) simulations for small lung lesions subject to electronic disequilibrium when very small segments are used during the IMRT optimization process. Method and Materials: IMRT plans for eleven (n=11) lung patients were created using Pinnacle3 7.6 planning system (TPS). Lung lesions measuring <3 cm in max. diameter and <27 cm3 were previously treated in our institution with SBRT techniques. The optimized intensity maps of each plan were then used to calculate the dose distributions using the CCCS algorithm. For each patient, seven optimized plans were created with varying MLC minimum segment sizes—0.25cm2 to 6cm2. The linear accelerator was modeled using MC code EGSnrc\BEAMnrc and verified against commissioned measured data. Intensity maps for each plan and the patient CT dataset from the TPS were exported to our MCsoftware. All patients were planned using a 5‐field IMRT plan (Millennium 120‐leaf MLC and Varian 2100C 6MV beam). Dose distributions were calculated and normalized so that the isocenter receives 45.0Gy. Isodose distributions, DVH, and ROI statistics were used for comparison between the two calculation methods. Results: Comparison of the DVHs from Pinnacle3 and MC show similar target coverage between CCCS algorithm and MC. Differences in the minimum, maximum and mean dose of the PTV were less than ±5% while doses to critical structures agreed within ±6% for all cases investigated. Discrepancies in the dose to very small structures have been observed and due to volume effects between the two systems. Conclusion: Good agreement exists in the dose distributions predicted by the CCCS algorithm and MC method. The CCCS algorithm can accurately predict doses to small lungtumors.
35(2008); http://dx.doi.org/10.1118/1.2961353View Description Hide Description
Purpose: Continuous prescription gradients may not be practical or necessary in the implementation of dose painting. This study optimized dose distributions to various discrete prescriptions and tested them for conformity to continuous, non‐uniform prescriptions. Method and Materials: The planning geometry consisted of a virtual cylindrical water phantom, containing a concentric heterogeneous tumor volume. The tumor was prescribed a uniform dose with boost subvolumes receiving escalated doses of 20, 50, or 100 % relative to background. Variable dose gradients extended up to 2 cm from the edge of 1 cm radius boost regions. Prescriptions were defined over gradient regions either continuously or in discrete steps. An iterative linear least squares method was employed to optimize helical tomotherapy dose distributions. Tumor, normal tissue, and subvolume importances were fixed relative to one another. Dosimetric evaluation included quantification of the volumetric percentage of boost and gradient regions whose plans conformed to prescriptions within 5 %. Results: Dose conformity to prescriptions in high dose contrast boost regions increased from 47 % of the volume for an infinite gradient to 100 % of the volume for a continuous gradient over 2 cm. In high contrast gradient regions of 1 cm radial distance, conformity ascended from 22 % of the volume for a single step to 91 % of the volume for a continuous gradient. To achieve dose conformity in 90 % of a 1 cm gradient region volume at low contrast, the prescription step size must be no larger than 0.5 cm. Conclusion: Results demonstrate that finite thresholds in prescription discretization exist to ensure planned dose conformity to continuous prescriptions. Adequate prescription gradient step size is dependent on beamlet dimensions, boost region size, and local gradient magnitude. Evaluation of prescription discretization may impact dose painting treatment planning protocols in its application to clinical cases.
SU‐DD‐A1‐05: Treatment Planning of Prone Position Accelerated Partial Breast Irradiation Customized for Magnetic Resonance Imaging Guidance35(2008); http://dx.doi.org/10.1118/1.2961354View Description Hide Description
Purpose: It is important to deliver an optimized dose distribution to an accurately delineated target volume for a successful accelerated partial breast irradiation (APBI) treatment. Breast magnetic resonance imaging(MRI) could improve lumpectomy delineation. We investigated treatment plans for CyberKnife and conventional C‐arm LINAC delivery of MRI‐guided APBI in prone position. Method and Materials: We acquired a CT scan of a patient positioned on a mechanical assembly with the same geometry as the dedicated MR coil to simulate the position during breast MRI acquisition. Ellipsoidal virtual PTV's of 4 cm were defined in the anterior part and in the lower outer quadrant of the right breast. Treatment plans were generated by inverse planning techniques for each system. For CyberKnife, we minimized beam directions from posterior to anterior to protect critical normal tissues(heart,lung and contralateral breast) and normalized the maximum dose to 120% of the prescription dose (38.5 Gy). For the conventional LINAC, we excluded any beams directed towards the critical normal tissues and normalized the plan to achieve 99% of the prescription coverage for 99% of PTV. Results: LINAC‐based plans resulted in maximum doses of 41.6 Gy in PTV and ⩽ 0.5 Gy in the critical normal structures. CyberKnife‐based plans led to 90% of the prescription dose coverage for 95% of PTV. Maximum doses were ⩽ 0.5 Gy in contralateral breast, ⩽ 1.9 Gy in heart, and ⩽ 4.2 Gy in both lungs. The fraction of normal ipsilateral breast tissue receiving ⩾ 37 Gy was 1% and 5% for CyberKnife and LINAC plans respectively. Conclusion: CyberKnife and LINAC generate treatment plans for MRI‐guided prone position APBI with acceptable dose distributions. The standard LINAC requires careful beam configuration to secure clearances between equipments. CyberKnife results in better conformity and slightly higher doses in the critical normal tissues.
SU‐DD‐A1‐06: Correlation Between Number of Beams and Monitor Units, in the Context of VMAT and RapidArc, in IMRT35(2008); http://dx.doi.org/10.1118/1.2961355View Description Hide Description
Purpose: Reduction in treatment time and monitor units (MU) are desired for an optimum intensity modulated radiotherapy(IMRT) plan. We investigated the relationship between increasing field numbers and the total number of monitor units and segments in IMRT plans of prostate cancertumors, and its impact on arc techniques such as intensity modulated arc therapy (IMAT) and volumetric modulated arc therapy (VMAT). Method and Materials: Ten prostate cancer patients treated to the same clinical protocol of 79.2 Gy were investigated. Individual plans that consisted of 5, 7, 9, 18 and 36 equally spaced, 15 MV beams were created for each patient. The effect of beam numbers on each plan was analyzed for 1) PTV Uniformity Index, 2) Segments, 3) Monitor Units (MU) and 4) DVH of all organs.Results: MU changes were most significant when number of fields increased from 5 to 9. Number of segments showed similar trends, where maximum increase in segment numbers occurred when field numbers increased from 5 to 9. The UI results showed both increasing and decreasing trends when field numbers increased between different patients, but were not significant. DVH comparison of organs at risk showed maximum changes in volume dose when the number of fields increased from 5 to 7. Conclusion: Increasing the field numbers from 9 to 36 fields did not result in dramatic changes in MU, segments, uniformity or DVH. Our study showed that for traditional IMRT step and shoot arrangements, to achieve similar homogeneous coverage of deep seated tumors, significant increases in the number of fields did not result in similar increases in monitor units or segments. Provided the continuous beam‐on and segment changes do not affect the overall treatmentdelivery speed, techniques based on modulated arc therapy should deliverdose distributions similar to IMRT techniques in considerably less time.
- Moderated Poster — Area 2 (Therapy): Brachytherapy
SU‐DD‐A2‐01: The Metal Artifact Reduction Technique for Accurate Prostate Seed Localization Using Limited EPID Projections and Deformable Registration35(2008); http://dx.doi.org/10.1118/1.2961356View Description Hide Description
Purpose: Scatter radiation can degrade the quality of a kilovoltage cone‐beam CT (KV CBCT)image due to the use of a larger flat‐panel detector. Especially, if KV CBCT is employed for seed localization, the effect of streak metal artifacts of the seeds can be magnified in CBCTimages. In this study, we describe the preliminary application of MV cone‐beam reconstruction (CBR) image from EPID projections and deformable registration for metal artifact reduction in a test phantom. Method and Materials: To perform the CBR and deformable registration based technique, the following three steps were used. (1) Extract a seeds only image and an image of seeds with bright metal streak regions by using a user‐defined threshold value in the CBR image and KV CBCTimages. (2) A derived intensity based 3D ‘demon’ deformable registration algorithm was applied to both seeds images, and the use of the algorithm generated the deformation field. (3) Resample the original CTimage according to the deformation field. Results: The intensities of the metal artifact region were decreased with the use of the deformation grid, as the seeds in KV CBCTimages were deformed as compared to the original seeds in CBR images. Although image quality was not dramatically enhanced, the metal artifact removed volume which causes the dose calculation error was 2.795 mm3 over 61 CT slices of the test phantom. Conclusion: Preliminary results with a test phantom demonstrated a bright streak metal artifact reduction effect. We believe that if dose calculation associated with KV CBCT yields reliable results, our study could provide an efficient method for post‐implant dosimetry. Conflict of Interest: This research sponsored by the Seoul R&BD Program (10550) and the Mid‐ and Long‐Term Nuclear R&D Program of the Ministry of Science and Technology (M20706000007‐07M0600‐00710).
35(2008); http://dx.doi.org/10.1118/1.2961357View Description Hide Description
Purpose: Eye plaques provide a good alternative to enucleation for ophthalmic tumors. The majority of eye plaques in the U.S. use brachytherapy seeds in a Collaborative Ocular Melanoma Study (COMS) standardized design. The plaques are 3 mm thick and uncomfortable to the patients, dosimetrically having localized hot spots. Implant Sciences developed 1 mm thick eye plaques using ion‐implantation techniques. Dosimetric characterization of prototype plaques using radiochromic film and a silicon diode is reported. Method and Materials: Prototype 16 mm eye plaques consisted of titanium‐encapsulated hemispheric quartz substrates with an layer at the inner surface, and a gold backing. Dosimetry was performed using GAFCHROMIC XR‐T and EBT film in a specially designed Solid Water “eye” phantom inserted into a 30 cm × 30 cm × 23 cm full scatter Solid Water phantom, as well as using a Scanditronix stereotactic diode in a water tank. Films were sandwiched between 1 mm thick phantom inserts perpendicular to the plaque's central axis. The diode was used for point measurements along the central axis and scans across the plaques. The films and diode were calibrated using a calibrated I‐Plant 3500 seed in Solid Water and water respectively, applying the TG‐43 formalism. Results:Dose distributions in planes perpendicular to the central axis of the plaques were radially isotropic and uniform. Film results, reproducible within 4%, agree well with silicon‐diode measurements. At 5 mm depth, measured dose rates are between 50 and 200 cGy/hour. Conclusion:Implant Sciences' ion‐implantation technique enables manufacturing of thin eye plaques with optimal dose distributions and clinically‐useful dose rates. GAFCHROMIC film in a Solid Water phantom and the stereotactic diode in water are convenient and reproducible dosimeters for eye plaque dosimetry.Conflict of Interest: Supported in part by SBIR NIH Grant 2R44CA092913‐02.
SU‐DD‐A2‐03: Measurement of Skin Dose When Using FlexiShield® with the Axxent® Electronic Brachytherapy System35(2008); http://dx.doi.org/10.1118/1.2961358View Description Hide Description
Purpose: To study the skinsurfacedose delivered when FlexiShield, a flexible tungsten silicone sheet, is placed on the skin to shield personnel in the treatment room. This surfacedose is compared to the surfacedose delivered when no FlexiShield is present and when a silicone and tungsten‐only shielding is used. Method and Materials: Axxent FlexiShield is designed to shield the operator and other personnel from x‐ray radiation from the Axxent® Model S700 X‐ray Source. In order to attenuate fluorescencephotons from tungsten in the shield, a barrier layer containing a tuned‐Z material was added to the silicone tungsten. A source holder was fabricated to hold FlexiShield against the surface of a PTW 34013 soft x‐ray ionization chamber while allowing for the placement of the x‐ray source at varying distances from the shield surface with water‐equivalent material filing the intervening space. The skindose is defined as the dose measured by the ionization chamber whose active face is recessed ∼0.3mm from the FlexiShield inner surface.Results: For a source to skin distance of 3cm, the increased skindose was measured to be 6.9% ± 0.5% for FlexiShield compared to no shielding. Silicon and tungsten‐only shielding produced an increased skindose of 14.4% % ± 1.9%. These measurements compare well with MCNP5 simulations of the dose to the basal layer of the skin for a 3cm source to skin distance, which predict an increase of 7% for FlexiShield and 16% for a tungsten‐only shield for a simulated Axxent x‐ray source spectrum. Conclusion: The use of FlexiShield produces a slightly increased dose near the surface of the skin when compared no shielding used. The tuned‐Z layer added to FlexiShield decreases the measured dose near the skinsurface by about a factor of two. Conflict of Interest: Research sponsored by Xoft, Inc.
35(2008); http://dx.doi.org/10.1118/1.2961359View Description Hide Description
Purpose: To verify the deviation of dose calculation based on commonly used TG43U1‐based algorithms from Monte Carlo simulations in eyeplaque implants, and to evaluate the dose perturbation introduced by eyeplaque. Method and Materials: Monte Carlo technique was used to generate 3D dose distributions of single IsoAid™ I‐125 brachytherapy sources. A novel computer code EyeMC was developed to utilize the single‐source 3D dose distribution to generate dose distributions of multiple‐source implants for COMS eye‐plaques, through rotating, superposing, displacing and interpolating methods. In addition, Monte Carlo simulations were performed for the 8‐seed implant to verify the accuracy of the EyeMC and treatment planning system, and to verify the magnitude seed‐eyeplaque effects. Eight seeds were placed as two isocentric squares at the same plane, with the inner square width of 0.6 cm and outside square width of 1.1 cm. The treatment planning system Brachyvision™ was also used to calculate the doses at points of interest with point‐ and line‐source approximations. Results: The point‐source approximation used in the treatment planning system agreed with line‐source approximation within 10%. Compared to the Monte Carlo simulation, the treatment planning system using line‐source approximation significantly underestimated the dose values at the eyeplaque axis, by 13% at 2 mm, 18% at 1.0 cm, 20% at 2.0 cm, and 24% at 3 cm. The eyeplauqe‐seed effect was small but noticeable. The presence of the eyeplaque could reduce dose up to 10% at the distance of interest up to 3 cm. The data from EyeMC agreed with multiple‐seed full Monte Carlo simulation within 3%. Conclusion: Our comparison of commonly used TG43U1‐based algorithm shows that neither point‐source nor line‐source approximations provides accurate dose calculations for multiple‐seed implant eyeplaques at short distances. Our study suggests that a Monte Carlo‐aided code, such as EyeMC, is needed to increase the accuracy of dose calculation.
SU‐DD‐A2‐05: Uncertanity of Real Time in Vivo Dosimetry with MOSFET Linear Array in I‐125 Prostate Permanent Implant Brachytherapy35(2008); http://dx.doi.org/10.1118/1.2961360View Description Hide Description
Real‐time dose monitoring is important to improve the accuracy of prostate permanent implant brachytherapy (PI). However, few data about in vivo dosimetry in PI have been reported because dosimetry of low energy photons from I‐125 sources is challenging. The purpose of this study is to estimate the dosimetric uncertainty in I‐125 permanent implant brachytherapy using 5‐linear array MOSFET. We evaluated the physical characteristics including angular response variability, linearity, calibration factor, and detector response variability for photon beams from I‐125 seed. All experiments were performed using I‐125 seeds to eliminate the energy dependence. Furthermore phantom study was performed to estimate the dosimetric uncertainty in preclinical settings. For isotropic analysis, the response of 270 and 90 degrees were 8 to 15 % lower than 0 degree and in another angle, the differences of responses were within 5%. However, the deviation of each angle was within 6%. The linearity of the readings is excellent (R2=0.098). The fading effect of MOSFET was not observed. Calibration factor was obtained with the standard deviation within 4%. We found that calibration factor was changed by seed‐detector distance (SDD) by 10% in shorter distance (5mm and 7mm) and 4% in greater distance compared with SDD 10mm. The same tendency was observed by Monte Carlo Simulation.Monte Carlo simulation showed that the beam hardening effect occurred as the SDD increased, which can increase additional deviation. In phantom study, treatment planning was performed with 47 seeds in a modified peripheral loading manner. We measured the urethral dose at base, midland, and apex positions during 20 minutes. In mid grand and base regions, measure dose agreed within 6% and in apex region less than 15% differences were observed. These data suggest that in vivo dosimetry in permanent implant brachytherapy with the appropriate correction can be performed within 15%.
35(2008); http://dx.doi.org/10.1118/1.2961361View Description Hide Description
Purpose: Applications of I‐125 and Pd‐103 sources in brachytherapy treatment of prostate cancers, with long and short doubling time, respectively, are well established. However, for treatment of prostate cancer composed of cells with different doubling time, a brachytherapy source with mixture of I‐125 and Pd‐103 radioisotopes might be more desirable. In this project, dosimetric characteristics of a novel brachytherapy source, composed of a mixture of I‐125 and Pd‐103 radioisotopes, has been determined following AAPM TG‐43U1 protocol. Method and Materials: A new brachytherapy source model, ADVANTAGE™ HYBRID PdI, has been designed by IsoAid™, LLC for treatment of prostate cancer. A dose calculation model was developed to determine the dosimetric characteristics of this source design. In this model, the Monte Carlo simulated parameters from each isotope of the source were individually determined assuming 50% dose contribution from each isotope. These values were then utilized for determination of the parameters of the mixed source. The outcomes of these investigations have been incorporated in the VariSeed planning system for the evaluation of an actual clinical case. Results: The results of these investigations indicated that the dose rate constant of a source with 50% dose contribution from each isotope is 0.699 cGy/hr/U. The radial dose function, 2D‐ and 1D‐ anisotropy function of the new source design have been tabulated for clinical applications. The tolerance of each parameter was evaluated for isotope variation within the source. The percent variation of V100 and D90 were found to be less than 1% and 4%, respectively, which are within the acceptable range. Conclusion: The TG‐43 dosimetric parameters of a multi‐isotope source have been determined for clinical applications in the treatment of prostate cancer with tumors composed of cells with variety of potential doubling times.
This project is partially sponsored by IsoAid™, LLC.
- Moderated Poster — Area 1 (Therapy): IMRT Delivery
SU‐EE‐A1‐01: Acceleration of Tomotherapy Treatment Delivery by Increasing Pitch and Decreasing Modulation35(2008); http://dx.doi.org/10.1118/1.2961374View Description Hide Description
Purpose: To determine the feasibility of reducing Tomotherapy treatment delivery time by planning with increased pitch and reduced modulation factor (MF). Method and Materials: Twelve patients with Tomotherapy treatment sites (head/neck, prostate, prostate/nodes, lung) and fraction size of 1.8–2.5 Gy originally planned with a low pitch (< 0.3) and fast gantry period (< 20 s) were replanned. The first replan increased the pitch to 0.43 and used the original MF (MF0). If the first replan was acceptable and the gantry period increased, the patient was replanned again with pitch = 0.43 and a lower MF to reduce treatment time. This was repeated until the plan (DVH) became unacceptable or the gantry period reached its minimum (15 s). If the first replan was not satisfactory or if the gantry period did not increase, the MF was increased by 0.5 to improve the plan and to test if the treatment time was still reduced. Results: For all patients, we were able to produce a DVH comparable to the original DVH using pitch = 0.43 and MF0. When comparing the original DVH to the accelerated plan DVH that most closely matched it, treatment times were reduced by 13–66%. Comparing these accelerated plans to the original plans showed the PTV median dose increase 0.4±0.5 Gy, insignificant changes in target coverage and homogeneity index, an average change in critical structure median dose of −0.8±1.3 Gy, and a maximum critical structure median dose increase of 1.3 Gy. Conclusion: Faster treatment times can be achieved by planning with increased pitch and reduced modulation. Treatments planned with pitch = 0.43 and MF0 showed little or no plan degradation and most had faster treatment times. Further reduction in treatment time was possible by reducing the MF. Conflict of Interest: Some of the authors have financial interest in TomoTherapy, Inc.
SU‐EE‐A1‐02: Image Guided Hypofractionated IMRT for Multiple Intracranial Metastases — An Alternative to Arc‐Based SRS35(2008); http://dx.doi.org/10.1118/1.2961375View Description Hide Description
Purpose:Stereotactic radiosurgery(SRS) uses an invasive head‐frame and involves long treatment times when multiple isocenter plans are treated, which becomes intolerable for some patients. This study evaluates intensity‐modulated radiotherapy(IMRT) for treating multiple (⩾4) intracranial tumors in comparison to arc‐based SRS.Method and Materials: This study involved 3 patients treated with image‐guided hypofractionated IMRT (Eclipse) and 3 treated with SRS using arc therapy (XKnifeRT). There were 28 intracranial lesions in total with planning target volumes (PTVs) ranging from 0.195cc to 5.64cc. For this study, 3 IMRT patients were re‐planned for arc therapy, while 3 SRS patients were re‐planned for PTVs (=GTVs+2mm) for both IMRT and arc therapy. By keeping the minimum surface dose of the PTVs the same between the IMRT and arc plans, we have compared maximum doses and PITV (the ratio between the volume of Prescribed Isodoseline and the Target Volume) for the two methods. We also compared the doses received by 10%, 20% and 50% of the whole brain, as well as the maximum doses to the critical structures (eyes and brainstem). Results: The average maximum doses to the 28 targets were 125.85%+/− 18.4% for arc plans and 115.2%+/−4.6% for IMRT plans. The average PITVs were 2.65+/−1.22 and 1.93+/−0.52 for the arc and IMRT plans, respectively. For critical structures, average doses received by 10%, 20% and 50% of the brains, as well as the maximum doses to other structures, were higher for IMRT plans. However, the treatment time for IMRT is significantly reduced (e.g. 15 min) compared to that for multiple isocenter SRS (up to hours). Conclusion:IMRT plans appeared to achieve comparable conformality compared to arc based SRS. This provides a good alternative for poor performance status patients. However, further study needs to be conducted to ensure the PITV calculation accuracy for small targets.
35(2008); http://dx.doi.org/10.1118/1.2961376View Description Hide Description
Purpose: The purpose of this study was to investigate the adequacy of full‐field non‐flat megavoltage beams for planning and delivery of step‐and‐shoot Intensity Modulated Radiation Therapy for different body sites. Method and Materials:IMRTtreatment optimization creates non‐flat fluence maps to accomplish prescribed treatment objectives. With “Stereotactic” mode available in commercial linacs with the flattening filter removed, non‐flat beams are now available that produce dose rates much higher than the flattened beam modes in current clinical use, thus offering the potential of deliveringtreatments much faster. We have recently installed and commissioned such a photon beam. This beam is currently restricted in its clinical use to 5×5cm2. Full field beam data was acquired for this beam in research mode. The full‐field stereotactic mode beam data was modeled in Eclipse TPS. IMRT plans previously delivered for different target sizes and complexities ranging from prostate, head and neck and abdomen were re‐planned using the same target and normal tissue constraints. Plans comparable to the clinically accepted treatment plans were obtained. These plans were compared using several plan comparison parameters, along with the total number of MU's and segments required to deliver a comparable plan, and correlated to the size of the target. Results: Our results show that clinically acceptable treatment plans are easily accomplished with the non‐flat beams, with comparable MU's for smaller targets, but proportionally larger MU's for larger targets. The number of segments required to deliver comparable plans with the non‐flat beam were very similar, but slightly higher. Conclusion: Our study indicates that treatmentdelivery with highly modulated fields do not require flattened beams for clinical treatmentdelivery. The benefit of three times higher dose rates than the flattened beam will make treatmentdelivery significantly faster for non‐flat beams. This benefit will be much more significant for hypofractionated treatments.
35(2008); http://dx.doi.org/10.1118/1.2961377View Description Hide Description
Purpose: Increasing the temporal precision of IMRT delivery is hampered by current mechanical limitations. We are studying the feasibility of a novel system that would electronically modulate a radiation beam and deliver a time dependent IMRT beam by incorporating pencil beam scanning with a dual focus collimator. This design requires a sharply focused photon pencil beam (FWHM ∼ 1 cm at 100 cm SSD), comparable to beams achievable with current LINAC technology. Collimator specifics were investigated using Monte Carlo (MC) techniques to evaluate the capability of producing photon pencil beams. Method and Materials: The BEAMnrc code was used to develop a geometry incorporating a novel 200 hexagon channel tungstencollimator. Simulations were performed varying collimator thickness, size, and spacing of the channels. Dose in water phantoms was calculated at depth of 10 cm using DOSXYZ. Similar MLC generated pencil beams were measured at equivalent depth using Computed Radiography(CR). Script files were written in MATLAB® to calculate peak output, FWHM, and off‐axis dose/dmax ratios. Results: Pencil beam FWHM and output decreased with increasing collimator thickness (2 cm collimator: FWHM = 1.5 cm, 9 cm collimator: FWHM = 0.6 cm), keeping all other factors constant, and compared favorably to the MLC pencil beam FWHM of 0.9 cm. Off‐axis dose leakage ratios decreased 20% between 2 and 9 cm thickness. These ratios also decreased 20% with increasing hexagon spacing (0.25 – 0.75 mm). An 8‐cm collimated (0.5 mm channel spacing, 0.43 mm radius) pencil beam (FWHM = 1.1 cm) matched closely the MLC pencil beam. Conclusion: A dual focus collimator was designed using Monte Carlo modeling to produce photon pencil beams comparing favorably with current LINAC technology. Future work will focus on comparing theoretical isodose distributions to those of our clinical IMRT system. Conflict of Interest: Research sponsored in part by Varian.
SU‐EE‐A1‐05: A Real‐Time Feedback Control Algorithm to Compensate 2D Target Motion with a Dynamic Multileaf Collimator35(2008); http://dx.doi.org/10.1118/1.2961378View Description Hide Description
Purpose: The aim of the work is to develop a real‐time control algorithm for a dynamic dose delivery with a multileaf collimator(MLC) that accounts for 2D target motion and deformation. Method and Materials: The multithreaded real‐time feedback control receives the target's position and deformation in real‐time from an independent localization device and approximates the future target position to calculate the next leaf positions. For lateral shifts, the leaves either include the entire target, exclude any healthy tissue, or utilize a percental trade‐off between surrounding tissue and target. Furthermore, the leaves can be placed so that an adjacent organ‐at‐risk (OAR) is not irradiated. For leaf positions and velocities, the physical constraints of the 160 MLC, Siemens, were obeyed. Theoretical simulations with different motion models were first performed to test the feedback algorithm. Experiments with a modified MLC demonstrate the feasibility of the method and the delay times within the feedback control loop were assessed. Results: The adaptive sequencer allows to compensate in real‐time for anatomically feasible target motion by adapting the radiation field to the new target position. Simulations indicate a significant improvement in conformity. Depending on the selected tracking mode, either an underdosage of the target edges, an overdosage of the adjacent tissue or a percental trade‐off occurs. First experiments proved the feasibility of the tracking algorithm combined with the dynamically adapted MLC though further developments to reduce the MLCcontrol system latency are required to improve the feedback control loop. Conclusion: The simulations indicate an accurate application of the dose to moving targets if the real‐time feedback control is applied. The developed sequencer provides in real‐time the required new leaf positions to the MLCcontrol system which then adapts the treatment field to the target motion. Conflict of Interest: Research supported by Siemens Medical Solutions.
SU‐EE‐A1‐06: A Real‐Time Dynamic‐MLC Control Algorithm for Delivering IMRT to Targets Exhibiting 2D Rigid Motion in the Beam's Eye View35(2008); http://dx.doi.org/10.1118/1.2961379View Description Hide Description
Purpose: We have developed a real‐time MLC control algorithm that allows for delivery of dynamic‐MLC (DMLC) IMRT to targets exhibiting 2D rigid motion in the beam's eye view (BEV). Method and Materials: The control algorithm consists of two components: 1) Construction of baseline DMLC leaf trajectories, and 2) real‐time control loop. The synchronized baseline leaf trajectories are constructed using target motion data that is collected prior to delivery. Only target motion that is aligned with leaf travel is included in this step. To account for target motion in the BEV that is not aligned with leaf travel, we have implemented a real‐time leaf‐pair switching mechanism, which allows the MLC to track motion along this axis in discrete increments of the leaf width. Using patient data, 36 target trajectories were constructed. One of these trajectories was used to construct the baseline leaf trajectories, and the others were used to simulate a 35 fraction IMRT treatment. Errors were analyzed using difference maps and a distance‐to‐agreement analysis.Results: The results indicate that 2D tracking resulted in deliveries that were superior to both no tracking and 1D tracking. A 160ms system lag time produced errors that were approximately equal those that resulted from ignoring one component of motion altogether. Additional results show that the algorithm's performance is very insensitive to the level of agreement between the target motion collected prior to delivery, and the motion observed during delivery. Conclusion: Over the course of a fractionated IMRT treatment, the MLC tracking algorithm is able to accurately compensate for 2D rigid target motion in the BEV. The performance of the algorithm is insensitive to the difference between the target motion measured during planning and the motion that actually occurs during delivery.
- Moderated Poster — Area 2 (Therapy): Stereotaxy and High Precision Techniques
SU‐EE‐A2‐01: A General Formula Predicting Near‐Target Dose Fall‐Off Characteristics for Different Isocentric and Non‐Isocentric Delivery Modalities35(2008); http://dx.doi.org/10.1118/1.2961380View Description Hide Description
Purpose: To investigate whether a common relationship exists governing the near‐target dose fall‐off for Gamma Knife, Cyberknife and MLC‐based Novalis linac systems. Method and Materials: Three groups of intracranial patient cases treated with Gamma Knife, Cyberknife, Novalis MLC‐based system were selected for the study. Each group contains 10–20 delivered treatment plans. In each case, the target was covered conformally by a peripheral isodose line ranging from 45% (Gamma Knife) to 95% (Novalis) of the maximum dose inside the target. Based on the divergence theorem, the dose fall‐off near the target was derived as , where V0 is volume of the prescribed isodose line covering at least 99% of the target volume, γ is a constant fitting parameter, which measures the steepness of the dose fall‐off near the target. The goodness of the fit and the average γ‐values were obtained for all cases of each modality. Results: The formula fitted excellently for all cases with the linear correlation coefficient R2 exceeded 0.99 in the log‐log plot for each case. No obvious dependence was found for the size of the targets, the number of the beams or the isocenters used for each group. An average γ‐value of 1.49± 0.13 was obtained for Gamma Knife, 1.48± 0.41 for Cyberknife, and 1.46 ± 0.19 for Novalis group of cases. No statistical significance was found in the mean value differences among the groups. However, Gamma Knife exhibited the smallest range of deviations in the mean value while Cyberknife exhibited the largest range of deviations in the mean value. Conclusion: A general formula and a common parameter was demonstrated for relating the average dose fall‐off near the target for Gamma Knife, Cyberknife and Novalis MLC‐based Linac systems. Despite large physical differences, nearly identical dose fall‐off was found for these modalities for treating intracranial lesions.
SU‐EE‐A2‐02: Verification of Source and Collimator Configuration for Gamma‐Knife® Perfexion TM using Panormaric Imaging35(2008); http://dx.doi.org/10.1118/1.2961381View Description Hide Description
Purpose: To develop a method of verifying the source and collimator configuration of Leksell Gamma Knife® Perfexion™. Method and Materials: The new model of stereotactic radiosurgery machine, Perfexion™, with modified source configuration allows extended reach of targets located in the cranial, neck and cervical regions. The control system allows automatic selection of appropriate built‐in collimator modules eliminating the need of time consuming manual installation of collimator helmets as in the older model of Gamma Knife system. However, the geometric configuration of collimator modules cannot be easily verified. The conventional method of exposing a film at the isocenter plane provides only a composited dose image, which is difficult to interpret in terms of the integrity of each individual source and corresponding collimator system. A method has been developed to capture a panoramic view of 192 Cobalt sources and corresponding images are utilized to verify the integrity and configuration of 192 sources. The images were acquired by exposing Gafchromic films wrapped around the surface of a specially designed 16 cm diameter cylindrical phantom. The phantom was mounted at the isocenter, with its axis aligned along the longitudinal axis of the couch. Depending upon the azimuthal angle of the source location, the shape and size of the source images were calculated and compared with the acquired images. Results: The images allowed clear identification of each of the 192 sources, verifying their integrity and selected collimator sizes. In this presentation the results of the source image alignment with respect to the expected collimator geometry will be described. In addition, the feasibility of relative dosimetric evaluation of the individual sources by the panoramic images will be presented. Conclusion: This method of source/collimator verification by panoramic imaging can provide an enhancement of commissioning and routine quality assurance of the Gamma Knife systems.