Index of content:
Volume 33, Issue 6, June 2006
- Therapy Moderated Poster Session: Exhibit Hall F
- Moderated Poster ‐ Area 1 (Therapy): IMRT Planning and Dosimetry
33(2006); http://dx.doi.org/10.1118/1.2240129View Description Hide Description
Purpose: To evaluate the potential for Co‐60 based tomotherapy including dose delivery and mega‐voltage CT (MVCT). Tomotherapy is a rotational implementation of IMRT that provides highly conformal doses and patient setup verification using MVCT. Current tomotherapy is limited to linear accelerators. This poster presents advances in our investigation of Cobalt‐60 based tomotherapy, including MVCT. Method and Materials: The fundamental components for the Co‐60 tomotherapy dose delivery and MVCT imaging experiments are a benchtop motion stage and a clinical Co‐ 60 MDS Nordion T‐780 unit. Film and polymer‐gel dosimetry are used to validate the tomotherapy dose delivery planned using in‐house software.Imaging is provided by a Varian Portal Vision LC250 EPID. MVCT imaging is demonstrated using a variety of phantoms, including an anthropomorphic head phantom, and various contrast phantoms. EGS Monte Carlo simulation is used to model different beam delivery approaches such as source design for increased radiation output. Results: The computer simulations, filmdose measurements, and three‐dimensional polymer gel dosimetry all demonstrate that Co‐60 tomotherapy provides conformal dose delivery required of modern IMRT techniques. Film measurements show that dose delivery corresponds excellently with treatment plans, validating our in‐house planning system. Treatment planning studies show that Co‐60 tomotherapy delivery compares favourably with that from linac based 6MV tomotherapy. Dose volume histograms show identical coverage and avoidance of target critical organs. Imaging results show that Co‐60 CT provides sufficient contrast and resolution for image guidance. Results from Monte Carlo studies show that it is possible to increase beam output for a dedicated Co‐60 tomotherapy unit by modifying the source geometry. Conclusion: Co‐60 is well suited to tomotherapy and imaging applications; the development of clinical implementations of Co‐60 tomotherapy is warranted and work continues in our centre along these lines.
SU‐DD‐A1‐02: Variations of Energy Spectra and Water‐To‐Material Stopping‐Power Ratios in Three‐Dimensional Conformal and IMRT Photon Fields33(2006); http://dx.doi.org/10.1118/1.2240130View Description Hide Description
Purpose: Complex dose distributions and dose gradients in IMRT may cause spatial variations in photon‐ and electron‐energy spectra. This study examined the change of photon‐ and electron‐energy spectra, and their effects on dosimeter response and water‐to‐material stopping power ratios (SPR) for 3D and IMRT beams. The later term is an important factor for dosimetry protocols and obtaining dose‐to‐water conversion in Monte Carlo dose calculations. Method and Materials: The Monte‐Carlo BEAM‐EGSnrc system was used to simulate external‐beam photon fields with 3D or IMRT features. Electron and photon energy fluences and spectra were calculated on a voxel‐by‐voxel basis using track‐length estimation for 3D and IMRT treatment plans. The water‐to‐material SPR were averaged over the voxel of interest with the electron spectra using the Spencer‐Attix theory. The relative response of ion chambers, films, and TLDs were modeled using the photon and electron spectra. Results: There was a strong spatial dependence of photon‐energy spectra in both the 3D and IMRT fields. The low‐energy (<100 keV) component of the photon spectra increased inversely with doses because of the contribution of the scatteredphotons. A similar effect was observed for electrons but to a much smaller extent. As a result, the response of film could increase by more than 10% in the low‐dose region, whiles the changes of ion chamber and TLD response were within 3%. On the other hand, the variation of the water‐to‐material SPR with energy spectra and spatial locations was not clinical significant (< 1%) for soft tissue, cortical bone, and lung, and was less than 2% for dry air. Conclusion: Photon‐ and electron‐spectra are spatial‐ and dose‐dependent in 3D and IMRTphoton fields. The spectra variation should be considered for certain dosimeters whose responses are energy dependent. For patient‐like materials, the water‐to‐material SPR was relative stable in spite of the spectral variation.
33(2006); http://dx.doi.org/10.1118/1.2240131View Description Hide Description
Purpose: To compare traditional IMRTquality assurance using film dosimetry and small volume ionization chambermeasurements with two new commercial products, the Wellhofer® MatriXX ionization chamber array and Varian® Portal Dosimetry. Available analysissoftware,hardware requirements and approximate operator times for data preparation, measurement and analysis will also be examined. Method and Materials: Fluence patterns from several 6X and 18X IMRT treatment plans for pelvis and head and neck radiotherapy patients were measured using radiographic film, the MatriXX array, and the Portal Dosimetry array. In each case the fluence pattern predicted by the treatment planning system was compared to the measured fluence pattern using ordinary γ‐analysis. Absolute dose at a point in a low‐gradient region of the fluence was also measured in the solid water phantom with an ionization chamber and compared to the dose prediction of the TPS. The absolute dosemeasured at the same point by the MatriXX array was also compared. Results: The absolute dosemeasurements made in a region of low‐gradient using an ionization chamber were, on the average, within 3% of the TPS predicted dose. The absolute dosemeasurements made using the MatriXX were, on the average, within 5% of the predicted dose. The ion chamber and MatriXX agreed to within 3%. An average of about 4% of pixels failed an ordinary γ‐analysis using 5% dose agreement and 3mm DTA criteria for both film and MatriXX measurements. A smaller percentage of pixels measured using Portal Dosimetry failed. The time spent preparing the data was comparable for all methods. Data measurement and analysis times were significantly reduced using the MatriXX and Portal Dosimeter procedures. Conclusions: This work indicates significant time savings for the new methods. In addition, the MatriXX system measures absolute dose at each chamber position.
33(2006); http://dx.doi.org/10.1118/1.2240132View Description Hide Description
Purpose: To develop a general formalism with various correction factors to predict dmax entrance dose with the new hemispherical brass buildup caps to be used with MOSFET detectors in anterior prostate IMRT fields and thereby integrate in vivo IMRTdose measurement as part of routine QA process in IMRTradiotherapyMethod and Materials: We have used the new wide energy hemispherical build‐up caps for this study. Due to its high density and high atomic number it provides the minimal amount of metal needed to achieve full build‐up at Dmax for a range of photon energies. We have developed a general formalism to predict Dmax entrance dose by applying necessary correction factors after studying the response of MOSFET with brass build up caps for energy, dose rate, dose reproducibility, SSD and patient specific IMRT correction factor. Results: In vivo Prostate IMRTdose measurements with MOSFET detectors using brass buildup caps was performed and compared against dose predicted by two different treatment planning systems. We used both 6 MV and 10 MV for this study and compared the in vivo MOSFET detector reading with dose predicted by Philips Pinnacle ( 6 MV) and CMS XiO ( 10 MV ) treatment planning systems respectively. We achieved a overall accuracy of better than ± 5% on measured patient doses.Conclusion: Routine IMRT QA in most institutions today only involves verifying the optimized fluence map delivered to the patient in a test phantom at a certain preset depth. Based on our work here, we believe adding in vivo IMRTdosimetry with MOSFET detectors using the new brass build up caps along with routine fluence map verification in phantoms and MLCquality assurance offers greater accuracy and confidence in actual dose delivered to the patient.
33(2006); http://dx.doi.org/10.1118/1.2240133View Description Hide Description
Purpose: To investigate the utility of using ray tracing to extract intrinsic information from CT, contour and primary dose data in order to determine initial conditions (number of arcs, arc weights, arc ranges and leaf positions) that can be input into an Intensity Modulated Arc Therapy (IMAT) optimization routine. Methods and Materials: Patient CT and contour data was ray‐traced to determine PTV and PTV‐OAR arcs. An additional arc was determined by the calculation of a ray importance factor (RIF) through ray tracing of the primary dose ray‐tracing of the PTV. All three sets of arcs were then input into a previously described leaf position optimization algorithm. This method was tested on two geometries by ray tracing 27 equi‐spaced beams. The optimized arc deliveries (number of arcs, arc weights, arc ranges and leaf positions) were then input into a fast dose calculation algorithm, NXEGS (NumeriX LLC) for dose calculation and comparison with primary dose as calculated by ray tracing.Results: RIF arc addition reduced the objective function by 20% for geometry 1 and 8% for geometry 2. Leaf position optimization further reduced the objective function by 27% for geometry 1 and 29% for geometry 2. Calculation of dose using NXEGS provides accurate dose distributions for IMAT. Conclusions:Ray tracing can quickly provide information about number of arcs, arc ranges, arc weights and leaf positions with very little user input. Leaf position optimization can improve leaf positions once the initial number of arcs and arc ranges are determined. Together these two steps can produce intensity modulated arcs for further optimization with a more accurate dose calculation algorithm.
33(2006); http://dx.doi.org/10.1118/1.2240134View Description Hide Description
Purpose: With IGRT, the geometric uncertainty in treatment can be reduced, which makes it feasible to implement IMRT dose painting with a reasonable resolution. In this cancer center, an on‐line realignment protocol is utilized for prostate cancer patients. This IGRT protocol is based on use of implanted gold fiducial markers and EPI. In this study, dose escalations with urethra sparing have been tested by using IMRT dose painting. Method and Materials:CT scans of three patients were chosen from the IGRT group. The original 3D‐CRT plan (74Gy/37fr, 10mm PTV margin) was used as a reference. In test IMRT plans two PTVs were generated. PTV1 was defined as 5mm extension of prostate. PTV2 was generated from PTV1 with 5mm margin subtracted for bladder, rectum, and urethra. Two raw plans were generated. Plan 1 was 74Gy/37fr to PTV1, and Plan 2 was 74Gy/37fr to PTV2. Then, the urethra sparing IMRT boost plan was generated as a weighted sum of the two raw plans, e.g. . Different combinations of weighting factors were tested: w1∈[0.6, 1], w2∈. [0.1, 0.5]. The dose to each organ was calculated with organ motion simulated based on actually recorded EPI image mismatches. The tumor control probability (TCP) and effective dose were used to evaluate the plans. Results: To achieve the same urethra D50 (minimum dose to 50% volume) as the reference plan, the highest weighting combination was w1=0.7, w2=0.5 . This yields significant dose reduction in bladder and rectum. For the considered patient the TCP increases from ∼74% to ∼95%. Conclusion: With IGRT, the urethra sparing IMRT dose painting is superior to the 3D‐CRT plan. The total prescription dose can be as high as 88Gy, with TCP of ∼95% and lower GI complication. Since urethra has been spared, the GU complication will be less.
- Moderated Poster ‐ Area 1 (Therapy): Stereotactic, Single and Hypofractionated Treatment I
SU‐EE‐A1‐01: Calibration of a Cobalt‐60 Irradiator for Stereotactic Radiosurgery Following the AAPM TG51 Protocol33(2006); http://dx.doi.org/10.1118/1.2240155View Description Hide Description
Purpose: Compare calibration of the Leksell Gamma Knife according to the American Association of Physicists in Medicine Task Groups 21 and 51 protocols. Materials and Methods: The Gamma Knife calibration phantom (The Phantom Laboratory, Inc., Salem, NY) is designed to fill with water and support an Exradin (Standard Imaging, Inc., Middleton,WI), model A‐16, ionization chamber positioned at its center. The phantom and chamber assembly was mounted in a Leksell stereotactic ring. The location of chamber's sensitive volume was determined using computed tomography and the Leksell fiducial frame. The chamber‐phantom assembly was attached to the 18 mm helmet in the Gamma Knife by the stereotactic ring. The phantom's geometry allowed radiation beams from each of the 201 Gamma Knife cobalt‐60 sources to converge along an 8 cm path to the ionization chamber's sensitive volume. This is equivalent to the arrangement by which one calibrates the Gamma Knife using the manufacturer‐supplied polystyrene phantom. Results: The phantom could be attached to the Gamma Knife so that the ionization chamber was reproducibly positioned at the convergence of the beams. Because of the phantom's design, either trunnions or automatic patient positioning system could attach the phantom. Comparisons using different phantoms and protocols resulted in the following calibration ratios for TG‐21 in the polystyrene sphere phantom, TG‐21 in the water phantom and TG‐51 in the water phantom, respectively: 1.00, 1.010, 0.996. Transmission measurements using a block of identical material indicate that the phantom's 2mm plastic shell would result in an error of approximately 0.6% if ignored. Conclusions:Calibration of the Gamma Knife can be performed in liquid water using the AAPM TG‐51 protocol, thereby eliminating any uncertainties with respect to the composition of the polystyrene from. Calibration values for the Gamma Knife that were obtained using the three methods for our phantoms agree to within 1.4%.
SU‐EE‐A1‐02: Analysis of Photon Beam Data From Multiple Institutions: An Argument for Reference Data33(2006); http://dx.doi.org/10.1118/1.2240156View Description Hide Description
Purpose: Beam data requirements to support sophisticated treatment planning and delivery techniques are increasingly rigorous. Small field photon measurements are particularly challenging for many centers and practitioners. The purpose of this work is to compare measured beam data characteristics from identical linear accelerators contributed by multiple institutions. Methods and Materials: Measured beam data from 43 “identical” 6 MV linear accelerators were collected from 43 different institutions. A common treatment planning system was used by all participating institutions, standardizing the data collected and simplifying the analysis. Beam data consisted of percent depth dose (PDD), crossbeam profiles and relative scatter factors (SF) as a function of field size. Beam data for field sizes less than 1 × 1 cm2 were contributed by the majority of institutions. A dose‐to‐monitor unit conversion factor was also obtained. All data were normalized in a consistent manner for direct comparison. Data were analyzed using a commercial analysis package. Mean, standard deviation, minimum and maximum deviation were calculated for the PDD data. A one‐population t‐test was applied to PDD, scatter factors and dose‐to‐MU factors to identify statistically significant differences. Results: PDD data for a 10×10 cm2field size were remarkably consistent among institutions, with 1σ variation of less than 1% at all depths beyond dmax. In contrast, significant variation was observed in small field PDD data; at 0.6×0.6 cm2, the PDD at 10 cm fell outside the 95% confidence level at 63.2% of institutions. Measurement of small field output factors proved to be equally variable. Several significant outliers were noted in dose‐to‐monitor unit conversion factors. Conclusions: Significant differences exist in beam data collected by multiple institutions for identical linear accelerators. Uniform procedures are needed to increase the quality and consistency of measured beam data. Use of a reference set of beam data may help to eliminate fundamental errors.
SU‐EE‐A1‐03: A Novel Three‐Dimensional Radiochromic Film Phantom for Use with Stereotactic Radiosurgery Units33(2006); http://dx.doi.org/10.1118/1.2240157View Description Hide Description
Purpose: To create a three‐dimensional (3D) film dosimeter capable of simultaneously measuring the entire relative dose distribution of the volume fields of an Elekta Gamma Knife (GK) unit.
Methods and Materials: A spherical head phantom was constructed out of Virtual Water™ (VW™). This phantom was constructed with a bored hole allowing the insertion of a stack of film 2.5cm in thickness and 5cm in diameter. The hole is fitted with two unique fiducial rods that prevent both rotation and inversion of the film. Radiochromic film with a thickness of 105 microns is used, allowing approximately 240 layers of film to be inserted into the phantom. One layer of film is assumed to be water/tissue equivalent; however, the water/tissue equivalency of a thick stack of film has not been determined. Monte Carlo MCNP5 methods were used to determine the water/tissue equivalency of a thick stack of radiochromic film. Results: Using MCNP5 simulations, the water/tissue equivalency of a stack of film 2.5cm thick was determined. For a simplified model of the film phantom, the dose distribution in the active layer of the pieces of film was found to be within −1.7% of the dose distribution in similar layers of VW™, demonstrating that a stack of radiochromic film may be used as 3D dosimeter. Conclusion: Using a stack of film as a 3D dosimeter limits the resolution of the determination of the relative dose distribution only by the resolution of the scanner and by the thickness of the film layers. This allows resolutions of 50×50×105 microns3 to be achieved. Most 3D dosimeters require advanced imaging equipment to read out the data but a 3D film dosimeter allows any institution which has a flat bed scanner to obtain 3D dose distribution information.
SU‐EE‐A1‐04: Verifying Internal Target Volume Using Cone‐Beam CT for Stereotactic Body Radiotherapy Treatment33(2006); http://dx.doi.org/10.1118/1.2240160View Description Hide Description
Purpose: The internal target volume (ITV) could be determined using 4D CT simulation images and be verified in the treatment room using on‐board cone‐beam CT(CBCT) since the CBCT projection images are acquired over approximately 10 breathing cycles. This study used a 4D dynamic phantom to verify the accuracy of this technique and also to develop a procedure for using CBCT to clinically verify ITV in stereotactic body radiotherapy(SBRT) treatment. Method and Materials: A CIRS 4D dynamic phantom, with a target ball and precisely controlled motion, was imaged using a 4D CT scanner. A Varian RPM system was used for respiratory gating. Ten 3‐D image sets were generated corresponding to 10 breathing phases. The ITV was determined based on the phase images. To assess concordance, on‐board CBCTimages of the target ball were compared with the 4D‐CT defined ITV. SBRT patient with tumor targets located in the thorax and upper abdomen were similarly scanned using phase gating 4D CT. The ITVs were compared between simulation CT and CBCT scans to identify localization error. Results: The dynamic phantom motion was 20mm along the inferior‐superior direction, 5mm along the anterior‐posterior direction, and 2mm along the left‐right direction with a cycle time of 4 seconds. The concordance of the CBCT and ITV matching was within 1 mm. For the lungSBRT patient, the target volume based on the CTimages without respiratory gating was 0.7 cc. The ITV was 2 cc. The ITV matches well with the CBCTimages. The localization errors between free‐breathing CT and CBCT were 2 mm to the right, 1 mm to the anterior, and 2 mm to the superior. Conclusion:CBCT provides an accurate assessment of the ITV for targets affected by respiratory motion.
SU‐EE‐A1‐05: Determination of Beam Margins for SRT/IMRT of Small Lung Cancers Based On Monte Carlo Simulations33(2006); http://dx.doi.org/10.1118/1.2240168View Description Hide Description
Purpose: This work investigates the beam margins used in treatment planning for stereotactic radiotherapy (SRT) and intensity‐modulated radiotherapy(IMRT) of small lung lesions based on Monte Carlo(MC) simulations. Method and Materials: Ninety SRT/IMRT treatment plans generated in a commercial treatment planning system were recalculated using MC simulations with different combinations of beam margin (0 to 18 mm), lung density (0.1 to 0.5 g/cm3) and planning target volume (PTV) (10 to 50 cc) based on the patient geometries built from CTimages. Each plan was normalized at D95, of the dose‐volume histogram (DVH) so that the comparison between different plans could be made quantitatively in terms of minimum dose (D99) and maximum dose (D1) in the PTV. The relationship between the beam margin and lung density/tumor size was fitted into modeled functions. The beam margin needed for a particular plan with certain lung density and PTV size can be determined based on the clinical acceptance criteria based on maximum/minimum doses and other normal tissue constraints. Results: The maximum and minimum doses were found to vary with beam margins, the volumes of PTV and lung densities. The relationships between them have been quantitatively generalized into functions from the simulation data. It was found that the maximum dose decreased with increasing beam margin while the minimum dose increased with beam margin when the beam margin was less than 1.5 cm. The trends were reversed with the increasing PTV volume. Conclusion: The generalized formulas for maximum and minimum doses can be used for the estimation of the minimal beam margin required in SRT/IMRT for adequate dose coverage for small lungtumors.
33(2006); http://dx.doi.org/10.1118/1.2240174View Description Hide Description
Purpose: To compare IMRT treatment techniques for a simulated para‐spinal mass located in the thoracic spine of an anthropomorphic phantom and to measure the accuracy of Megavoltage‐CT (MVCT) images for localizing spinal anatomy in the T‐Spine region. Methods and Materials:Treatment planningCTimages were acquired on a kilovoltage CT simulator of a whole body anthropomorphic RANDO phantom and used to create a planning target volume (PTV) covering the T7 to T9 vertebral bodies. The fixed gantry IMRT cases were planned using the Pinnacle treatment planning system and delivered on a Varian 21EX with a 120‐leaf multileaf collimator. Inverse treatment plans were created with 9 and 12 equally spaced fixed fields starting at 0‐degrees (IEC Scale). Inverse planning was performed using Direct Machine Parameter Optimization (DMPO), and gradient decent optimization with sliding window leaf sequencing. Helical tomotherapy cases were planned using the TomoTherapy HI‐ART treatment planning system. Relative dose measurements were made using calibrated film placed in the RANDO phantom. MVCT images of the RANDO phantom were acquired with a tomotherapy system and fused with the treatment planningCTimages. The phantom was then correctly positioned, and the fusion error was measured by imaging the T7 and T9 phantom vertebrae. A principal component analysis was used to determine the largest factors in image registration. Results: The 9‐Field DMPO and helical tomotherapy cases had PTV uniformities of 10% and maintained a large dose gradient. Conclusions: Helical tomotherapy and 9‐Field DMPO treatments yielded similar dose gradients (10%/mm) and PTV dose uniformity indices (10%). The sliding window treatment deliveries were consistently worse in cord sparing and dose uniformity. Anthropomorphic phantom studies indicated that megavoltage CTimages were capable of imaging the spine for placement at isocenter within 1‐mm of the desired position.
- Moderated Poster ‐ Area 1 (Therapy): Brachytherapy I
TU‐EE‐A1‐01: Capabilities of a CT‐Suitable, Patient‐Adaptive HDR/PDR Intracavitary Brachytherapy Applicator for the Treatment of Cervical Cancer33(2006); http://dx.doi.org/10.1118/1.2241586View Description Hide Description
Purpose: To demonstrate the capability of a prototype, high‐dose‐rate/pulsed‐dose‐rate (HDR/PDR) intracavitary brachytherapy cervical applicator to be (a) imaged using computed tomography(CT) with minimal artifacts as compared to current, clinically‐utilized applicators and (b) deliver dose to a simulated disease plane that is at least equivalent to that delivered by the FW applicator while delivering less dose to multiple, simulated rectal planes for equivalent loadings. Capability (b) demonstrates the prototype applicator's ability to adapt to varying patient anatomies utilizing a remotely adjustable shield contained within the colpostat. Method and Materials:CTimage sets were acquired of the Fletcher‐Suite‐Delclos (FSD), Fletcher‐Williamson (FW), and prototype applicators (utilizing a step‐and‐shoot technique) and artifacts generated by each were compared qualitatively. Images were acquired of the applicators positioned parallel to the table and in positions that simulated applicator placement during treatment. Using film validated MCNPXMonte Carlo(MC) models of the prototype and FW applicators, dose comparisons (min, max, average, and dose‐surface histograms) were made in simulated disease (1cm medial to colpostat) and multiple rectal (1cm distal and 0, 0.5, and 1cm medial to the colpostat) planes. Results: Preliminary results indicate that the prototype applicator is CT‐friendly; qualitatively minimizing anatomy‐obscuring artifacts compared to equivalent FW and FSD image sets. Additionally, the prototype applicator is able to deliver comparable dose levels (within +/− 5%) to a simulated disease plane while reducing dose (32% average) to varying simulated rectal planes when compared to equivalent FW treatments. Conclusion: The prototype applicator is able to be CTimaged with minimal artifacts and substantially reduce dose delivered to the rectum while maintaining dose delivered to disease when compared to current ICBT technologies. Currently, the model is being extended to include two prototype applicators and a 15° tandem so that dose‐volume histograms can be generated using specific‐patient cases of varying anatomical geometries.
TU‐EE‐A1‐02: Quality Assurance of Partial Breast Irradiation Using Permanent Breast 103Pd Seed Implant (PBSI)33(2006); http://dx.doi.org/10.1118/1.2241587View Description Hide Description
Purpose ‐ For early stage breast cancer, accelerated partial breast irradiation using High Dose Rate brachytherapy appears as effective as whole breast radiation. A permanent breast seed implant (PBSI) technique has been developed that realizes the implantation of 103Pd seeds under ultrasound guidance in a single one‐hour session. The objective of this study is to compare early and delayed post‐implant dose distribution to assess the Quality of the procedure. Material and Methods ‐ A Phase I/II clinical trial has been activated in May 2004 and as of February 2006 fourty seven patients have received PBSI. A minimal peripheral dose of 90Gy was prescribed to a volume corresponding to the CTV plus a margin of 1.5 cm. Each patient has a CT scan immediately following seed implantation and another one at two months. After identification of the seeds, new plans were calculated using the MMS TPS. Results: ‐ On average, 70 seeds are used per patient, with an activity ranging from 1.59U to 2.7U per seed. The V100 values demonstrate a satisfactory coverage and minimal variations over the course of 2 months. V100 were improving over time demonstrating a learning curve. The V200 for both the PTV and CTV shows significant and dramatic increases over time (65% and 87% respectively, p<0.001). The mean PTV and CTV volumes were small and do not vary significantly between the “immediate” and two month post‐implant. Conclusions ‐ Post‐implant PBSI Quality Assurance data shows adequate dose coverage of the target volumes that remains stable over time suggesting that there is no significant seed motion. Significant changes in the hot‐dose sleeves are seen two months from seed implantation. This could be due to a breast oedema at the time of implant that disappears with time and/or to the development of a retractile fibrosis over time.
33(2006); http://dx.doi.org/10.1118/1.2241588View Description Hide Description
Purpose: To show the ability of a large‐volume free‐air chamber to experimentally determine the air‐kerma strength of a new 1cm 103Pd coiled brachytherapy sources and show its potential for determining the air‐kerma strength of longer (2–6cm) coiled sources. Method and Materials: A Variable Aperture Free‐Air Chamber (VAFAC) has been constructed for making air‐kerma rate measurements of low‐energy photon‐emitting brachytherapy sources with photon energies up to 70 keV. The VAFAC has been used to determine the air‐kerma strength of 1cm coiled brachytherapy sources. The present US air‐kerma strength standard for low dose rate (LDR) brachytherapy sources is the National Institute of Standards and Technology Wide‐Angle Free‐Air Chamber (NIST WAFAC). The results obtained with the VAFAC were compared to air‐kerma strengths determined with an Accredited DosimetryCalibration Laboratory (ADCL) well ionization chamber (traceable to the NIST WAFAC) and direct measurement in the NIST WAFAC. At present, the WAFAC is unable to measure the air‐kerma rate of the coiled sources longer than 1cm due to its geometric limitations. The VAFAC does not have the same geometric limitations as the NIST WAFAC, which extends its capabilities to measure sources up to 6cm in length. Results: It has been shown that air‐kerma strengths determined for three 1cm length coils with the VAFAC are within 1% agreement with both an ADCL well ionization chamber and NIST WAFAC values. Conclusion: This work shows that the VAFAC can accurately determine the air‐kerma strength of 1cm 103Pd coiled brachytherapy sources. It also shows the ability of the VAFAC to measure sources up to 6cm in length, a feat which was previously impossible due to the limitations of the NIST WAFAC. Conflict of Interest: Research sponsored by RadioMed Corporation.
33(2006); http://dx.doi.org/10.1118/1.2241589View Description Hide Description
Purpose: To use the EGSnrc Monte Carlo (MC) code for calculations of photon energy spectra and TG‐43 dosimetry parameters for Xoft, Inc's miniature x‐ray brachytherapy source. The importance of MCtreatment planning for brachytherapy is also investigated. Method and Materials: Calculations of in‐air photon energy spectra and the dose distribution around the source in water were performed. The radial dose function, anisotropy function, and the absolute dose rate were calculated and compared with measurements made by Rivard et al (submitted to Medical Physics). Calculations were done to investigate how parameters ignored by TG‐43 affect dose delivered to the medium. The effects of realistic breast tissue and a finite irradiated volume were investigated. Results: Calculated in air photon spectra show excellent agreement with measurements in the energy range of ∼10–50kV. TG‐43 dosimetry parameters agree well with measurements but show a significant dependence on incident electron angles. Comparison of the dose in water to breast tissue show that calculations done in water may overestimate dose to breast tissue. The difference in dose to breast and water varies greatly with distance from the source and differences as large as 18% occur near the source. Calculations done in an infinite medium overestimate dose at the surface by 7% when compared with the case of a source placed 2cm from the surface of a phantom. Conclusion:MC calculations of in‐air photon energy spectra and TG‐43 dosimetry parameters have been performed and agree well with measurements. Calculations show that by ignoring the effects of realistic tissues and finite irradiated volumes, the TG‐43 dosimetry protocol may significantly overestimate dose delivered to a patient. Using MC would improve treatment planning accuracy allowing for better correlation of treatment outcome to dose delivered.
TU‐EE‐A1‐05: The Use of Directional Interstitial Sources to Reduce Skin Dose in Breast Brachytherapy33(2006); http://dx.doi.org/10.1118/1.2241590View Description Hide Description
Purpose: To investigate the feasibility of reducing the skindose with temporary LDR multicatheter breast implants with the use of directional 125I interstitial sources in comparison to conventional HDR interstitial breast brachytherapy.Method and Materials: The treatment plan for a patient treated with HDR interstitial brachytherapy with 192Ir was compared to a directional 125I treatment plan in the same dataset. Directional sources contain an internal radiation shield that greatly reduces the intensity of radiation in the shielded direction. They have a similar dose distribution to non‐directional sources on the unshielded side. Several dosimetric parameters are compared including target volume coverage, dose homogeneity index, and the skin surface areas receiving 30%, 50% and 80% of the prescription dose (S30, S50 and S80, respectively). The HDR prescription dose was 34 Gy in 10 fractions. Results: Similar excellent target coverage was achieved by both directional LDR and HDR (99.2% and 97.5%, respectively). Moreover, for a 170‐cc target volume, the dose homogeneity index was 0.82 for both LDR and HDR (V100 was 211.4 cc or 225.7 cc, and V150 was 39.1 cc or 40.4 cc, respectively). However, with directional LDR, the following reductions in skindose may be achieved: S30 is reduced from 100.6 cm2 to 62.6 cm2, S50 from 50.6 cm2 to 16.1 cm2, and S80 at 2 cm2 to null. The reduction in V50 for the whole breast is more than 100 cm3 (386.1 cc vs. 489.2 cc). Conclusion: As compared to HDR, directional interstitial 125I sources allow similar dose coverage to the subcutaneous target, while significantly lowering the skindose due to a quicker fall‐off beyond the target. Directional LDR sources can produce a similar dose homogeneity index, but the biological characteristics are more tolerable to the patient and can potentially reduce the risk of late skin and subcutaneous toxicity.
TU‐EE‐A1‐06: Application of TG_43U1 Formalisms to Calculate Dose Distribution Around a 5 Cm Long RadioCoil Source33(2006); http://dx.doi.org/10.1118/1.2241591View Description Hide Description
Purpose: To evaluate application of TG‐43U1 parameters for the determination of dose distribution around a 5cm long RadioCoilTM™ source. In addition TG‐43U1 recommended linear interpolation technique for 2D anisotropy function (F(r,θ) has been evaluated for elongated sources. Materials and Methods:Dosimetriccharacteristics of 5cm long RadioCoil™ brachytherapy source have been determined following TG‐43U1 recommendations using MCNP5 Monte Carlo code. In addition dose profiles along the longitudinal axis of the 5cm long source have been determined using MCNP5 simulation and theoretical technique. Monte Carlo calculated dose profiles were compared with theoretically values to evaluate accuracy of TG‐43U1 recommended formalisms for elongated brachytherapy sources. Results: TG‐43U1 recommended dosimetriccharacteristic were utilized to calculate dose distribution around an elongated brachytherapy source. Fifth order polynomial fit to the F(r,θ) was applied to extend F(r,θ) from 0° to 90° for the points falling on the source. Dose profiles at radial distances 0.5, 1.0, 1.5, and 2.0cm away from the central axis of the source has been determined using Monte Carlo simulation technique and TG‐43U1 recommended formalisms. Results of theoretically calculated dose profile were compared with Monte Carlo simulation data. Application of TG‐43U1 recommended minimal radial distances for 2D anisotropy function for dose calculation indicated that F(r,θ) for additional radial distances are required for good agreement between the two methodologies for calculating dose distribution. In addition, linear interpolation technique recommended by TG‐43U1 has also been investigated to extract F(r,θ) for the points falling in between the TG‐43U1 recommended radii. Conclusion: Results of these investigations indicate that TG‐43U1 formalisms can be extended for elongated brachytherapy sources, if F(r,θ) is tabulated for radial distances of 0.5 to 5.0cm with 0.5cm increment L/2 ± 0.2cm. Moreover, with the addition of recommended radial distances for 2D anisotropy functions, the linear interpolation technique more closely replicates Monte Carlo simulated data.
- Moderated Poster ‐ Area 2 (Therapy): IMRT Planning and Delivery
33(2006); http://dx.doi.org/10.1118/1.2241593View Description Hide Description
Purpose: To include all noncoplanar beam directions in the beam orientation optimization (BOO) problem in intensity‐modulated radiation therapy(IMRT)treatment planning, and to demonstrate that high‐quality treatment plans can be obtained using fewer beams than are typically used in equi‐spaced plans. Method and Materials: Because the data storage requirements for each beam restrict the number of beams that can be considered in BOO, the majority of previous BOO research has focused on considering just coplanar angles and/or a handful of pre‐selected noncoplanar directions, which comprise only a small subset of all solutions. In contrast, our approach allows for the generation of beam data for promising directions thus avoiding the data storage restriction and significantly increasing the size of the solution space, possibly leading to improved treatment plans. We use a response surface (RS) method that allows us to generate the beam data on‐the‐fly only as necessary. We consider the problem of adding a single (noncoplanar) beam to a locally optimal 3‐beam solution, thus yielding a 4‐beam plan. Several varying implementations of the RS algorithm were tested on six head‐and‐neck cases using gantry and couch rotations, each on a 10° grid. Results: The 4‐beam treatment plans obtained using the RS method were comparable to locally optimal 4‐beam solutions, and were also comparable to the 5‐ and 7‐beam equi‐spaced plans typically used in head‐and‐neck treatment plans. Conclusion: For head‐and‐neck cases, quality plans with fewer beams than standard 5–7 beam treatment plans can be obtained if BOO is applied. While the inclusion of noncoplanar orientations in BOO is useful in terms of improving the FMO objective function, the resulting improvements in the treatment plan are not always clinically significant.
This work supported in part by NSF DMI‐0457394 and the NSF Alliances for Graduation Education and the Professoriate and Graduate Research Fellowship programs.
33(2006); http://dx.doi.org/10.1118/1.2241594View Description Hide Description
Purpose: To develop new algorithms/softwares that optimally split any intensity‐modulated fields of large widths into multiple subfields under the MLC maximum leaf spread constraint such that the total beam‐on time for delivering the resulting subfields is minimized. Methods and Material: Due to the maximum leaf spread (MLS) constraint of MLCs, intensity‐modulated fields used in IMRT whose widths exceed a given threshold must be split into multiple subfields. This results in increased beam‐on time of the treatment. We studied two versions of the field splitting problems: 1) Splitting a large field into non‐overlapping subfields along paths that are orthogonal to the MLC leaf motion direction (this is a generalization of field splitting along straight lines); (2) splitting with overlapping, allowing adjacent subfields to overlap with each other. We developed two new field splitting algorithms (called FSMP and FSO) for these two problem versions, which mathematically guarantee to minimize the total beam‐on time of the splitting. Our algorithms are based on graph algorithmic techniques in computer science and linear programming tools in operations research. Results: We implemented our new algorithms, and experimented with them on 58 large intensity‐modulated fields for 11 clinical cases obtained from the Department of Radiation Oncology, University of Maryland Medical School. We conducted comparisons with CORVUS 5.0, and with a recent field splitting algorithm (denoted by FSSL), which splits along straight lines. For every tested field, the total beam‐on times of the four methods, CORVUS 5.0, FSSL, FSMP, and FSO, are always in decreasing order. Comparing with CORVUS 5.0 and FSSL, our new algorithms showed considerable improvements (on average, 21% and 12%, respectively) in the total beam‐on time. Conclusion: We developed two new field splitting algorithms under the MLS constraint of MLCs to minimize the total beam‐on time. Our algorithms improved the previous field splitting approaches considerably.