G Gill

TH-CD-304-02: Clinical Uncertainty of in Vivo Dosimetry for Intensity-Modulated Radiation Therapy Using Optically-Stimulated Luminescent Dosimeters

Purpose: Several studies have reported the physical properties of optically-stimulated luminescent dosimeters (OSLDs) and suggest their efficacy for clinical in vivo dosimetry, but few publications have assessed the clinical uncertainties associated with OSLD-based in vivo dosimetry for conformal and intensity-modulated treatment. The purpose of the current work is to identify and characterize clinical uncertainties for OSLD-based in vivo dosimetry. Methods: OSLDs are placed in dosimetrically-appropriate locations on a weekly basis, covered with small 5 mm bolus squares, exposed during patient treatment, read, and compared with doses predicted from the treatment planning system. Six (6) parameters were identified as significant contributors to uncertainty in the process: Inherent physical OSLD uncertainty (σ_ OSLD), OSLD reader uncertainty (σ_reader), dose calculation uncertainty in the build-up region (σ _calc), uncertainty of depth for planned dose (σ_d), dosimetric uncertainty due to daily image-guided shifts (σ_ s), and placement uncertainty (σ_p). σ_d, σ_ s, and σ_p were estimated by analyzing clinical OSLD dosimetry for inverse-planned intensity-modulated treatment plans (prostate, lung, and head-and-neck) and field-in-field intensity-modulated treatment plans (breast). Total uncertainty was estimated by summing the 6 components in quadrature. Results: σ_OSLD was defined by the manufacturer (±3.0%). σ_calc was assumed to be approximately ±5.0% at 5 mm depth from other works in the literature. σ_reader and σ_s were measured to be ±1.0 and ±5.8% at 5 mm depth respectively. σ_p was found to be ±3.6% for breast, ±4.2% for prostate, ±4.4% for lung, and ±10.6% for head-and-neck. Total uncertainty was ±9.0% for breast, ±11.5% for prostate, ±13.2% for lung, and ±16.1% for head-and-neck. Conclusion: Site-specific clinical uncertainty for a limited selection of sites ranged from ±9.0% to ±16.1%. The largest components were image-guided shifts and estimated placement uncertainty. A wider selection of anatomical sites, site-specific correction factors, and clinical tolerance/action level guidelines will be presented.

SU‐F‐T‐52: Study of Energy Dependent Effect of Dosimetry Systems Used in Therapeutic Soft X‐Ray Energy Range

PURPOSE To investigate energy dependent effects of different dosimetry systems which can be used as in vivo dosimetry monitoring for intraoperative radiotherapy in therapeutic soft x-ray energy range. METHODS Three dosimetry systems were evaluated in therapeutic soft x-ray energy range: optically stimulated luminescent dosimeter (OSLD) nanoDots, radiochromic EBT2 and EBT3 films. The x-ray photons were produced by a Zeiss Intrabeam 50 kV x-ray radiotherapy system. Solid water and bolus slabs with different thicknesses were used in the process of irradiation. An aluminum filter set was used to measure HVLs of X-rays. Calibration curves were made at different depth of boluses. RESULTS Half Value Layers at depths of 0, 3, 10, and 20 mm of solid water were measured to represent the energy change versus depth, yielding 0.306, 0.482, 0.865 and 0.901 respectively and indicating nearly unchanged HVL beyond 1 cm depth. The responses of each system at different depths were normalized to the response at 2 cm depth. In film dosimetry, the response is calculated as optical density (OD). The results show that there is nearly the same energy dependence for EBT2 and EBT3. At a HVL of 0.482 mm Al, the relative responses of nanoDots and EBT3 are 0.85 ± 0.04 and 0.89 ± 0.03 compared to those at 0.901 mm Al HVL, respectively, indicating no obvious difference between those two systems within the measurement uncertainty. CONCLUSION It was observed that the studied dosimeter response increases about 13% from the x-ray energy of 0.48 mm Al to 0.90 mm Al. Therefore, caution should be exercised in using an appropriate calibration curve, and x-ray beam hardening effect has to be taken into account.

SU-F-T-321: The Effect of an Electromagnetic Array Used for Patient Localization and Tumor Tracking On OSLD in Vivo Dosimetry

PURPOSE The purpose of this study was to observe the effect of an electromagnetic array used for patient localization and tumor tracking on optically-stimulated luminescent in-vivo dosimetry. METHODS A linear accelerator equipped with four photon energies was used to irradiate optically stimulated luminescent dosimeters (OSLDs) at the respective dmax depths and in the buildup region, with and without the presence of an electromagnetic array used for tumor tracking and patient localization. The OSLDs were placed on solid water slabs under 5 mm bolus and on each face of an octagonal phantom, and irradiated using both static beam and arc geometry, with and without the electromagnetic array under our setup. The electromagnetic array was placed 6 cm above the phantom to coincide with similar distances used during patient treatment. Ionization chamber measurements in a water phantom were also taken initially for comparison with the simple geometry OSLD measurements and published data. RESULTS Under simple geometry, a negligible change was observed at dmax for all energies when the electromagnetic array was placed over the setup. When measuring at five millimeter depth, increases of 1.3/3.1/16/18% were observed for energies 4X/6X/10X/15X respectively when the electromagnetic array was in place. Measurements using the octagonal phantom yielded scattered results for the lateral and posterior oblique fields, and showed increases in dose to the OSLDs placed on the anterior and lateral anterior faces of the phantom. CONCLUSION Placing the electromagnetic array very close to the patients surface acts as a beam spoiler in the buildup region (at 5 mm depth), which in turn causes an increase in the measured dose reading of the OSLD. This increase in dose is more pronounced when the OSLD is placed directly underneath the electromagnetic array than off to one side or the other.

SU‐E‐T‐738: Verification of an Automated Weighted Sector‐Integration Algorithm for Determining Output Factors for Electron Cutouts

Purpose: To verify the accuracy of an automated weighted‐sector‐integration electron output factor calculation algorithm for a variety of cone and cutout sizes, beam energies and clinical cutout shapes in a multi‐site radiation medicine department. Methods: A new weighted sector‐integration algorithm for electron output factor estimation has been implemented on a custom database structure. To automate the cutout shape determination, the cutout contour points are imported electronically from the treatment planning system, or digitized on a computer monitor using a digitally reconstructedradiographimage from an Electronic Medical Records system. The contour points are converted to polar coordinates and a sector‐integration summation performed with a weighting scheme corresponding to the angular size of each contour sector. Sector output factors used in the summation process are taken from a database of circular cutout measurements for a range of cone sizes, radii and energies. The algorithm was validated over a range of cutout sizes (2–20 cm), electron cones (10×10 cm2 to 20×20 cm2), cutout shapes (rectangular, circular and irregular including narrow fields and fields with concavities or convexities) and nominal beam energies (6–18 MeV) used in clinical electron beam therapy.Results: Differences between calculated and measured output factors for the patient cutouts studied were less than 2%, except in the case of long and narrow cutout shapes measured at low energies, which remained within 2.5% Conclusions: In a busy multi‐hospital environment, measurement of output factors for electron cutouts places a considerable time burden on physics staff. An estimation method that provides sufficiently accurate output factors for the wide variety of cutouts is a valuable clinical tool. Further work with handling unusually concave and long, narrow shapes at lower energies will be of value to extend the extreme limits of the algorithms accuracy.

SU‐E‐T‐102: Factors That Affect the Accuracy of IMRT Quality Assurance Measurements and Their Clinical Significance

Purpose: The purpose of this work was to perform error analysis on various processes that directly affect the accuracy of quality assurance (QA) measurements for intensity modulated radiation therapy(IMRT)delivery and to gauge their overall clinical impact for various clinical sites. Methods: In this study, we developed a numerical model based on the propagation of errors to quantify the complexity of treatment plans by the level of the intensity of modulation, and correlated the level of the modulation directly to the agreement of the IMRT QA results. We also provide detailed analysis of the uncertainties from IMRT QA results stratified by treatment site specific plans and fidelity in plan delivery. Results: Using a 5%/2mm criterion, it was found that the level of agreement of IMRT QA measurements was better for pelvic and chest plans than for head and neck, breast and brain plans. Errors tended to first increase with an increase in the number of Multi‐leaf Collimator(MLC) segments from 5 to 20 and then decrease with further increase in the number of segments. It was determined that other factors that impact the accuracy of the measurements include measurement device limitations, machine output variations, and phantom setup errors. Similarly, uncertainties in the beam data modeling and the calculation algorithm, as well as the initial beam commissioning data are main contributing factors to the accuracies of the treatment planning systems. Conclusions: The overall agreement of IMRT QA results is affected by both ‐ the accuracy of the measurement and that of the treatment planning system. We will demonstrate that the IMRT QA analysis criteria currently used in many clinics can be modified and improved to reflect the uncertainties of the process using the error propagation model developed in this study.

SU‐E‐T‐115: Dose Rate Dependence of OSLDs in Flattened and Flattening‐Filter Free Photon Beams

Purpose: Optically stimulated luminescent dosimeters (OSLDs) are increasingly utilized for in vivo patient dosimetry. Current characterizations of OSLDs are limited to conventional dose rates and flat beams. The TrueBeam linac (Varian, Palo Alto, CA) is capable of producing flattening‐filter‐free (FFF) 6 MV beams with extremely high dose rates. The purpose of this work is to characterize OSLD dependencies on dose rate and presence/absence of flattening filters.Methods: OSLDs were placed at central axis of the TrueBeam accelerator on 5 cm solid water, covered with 1.5 cm bolus, and exposed to a range of doses (0, 100, 200, 300, 500, 800, 1100, and 2100 cGy) and dose rates (10, 200, 400, 600 cGy/min for conventional 6 MV, 400, 600, 1000, and 1400 cGy/min for 6 MV FFF). Twelve OSLDs were spaced 1 cm apart from left‐to‐right through the central axis to capture beam profiles of the flattened and FFF beam. OSLDs were read and corrected for dosimeter sensitivity and read‐out depletion. Linearity of counts with dose was assessed with regression. Profiles were compared with measurements from ion chamber array. Results: Comparable linearity up to 500 cGy and supralinearity from 500–2100 cGy was observed across all dose rates for both flattened and FFF beams. Slopes for all dose rates were within 2% for flattened beams and within 1% for FFF beams. Slopes for FFF beams were, on average, 3.7% higher than flattened beams. OSLD beam profiles approximated both flattened and FFF array profiles, but OSLDs read 5% greater than the array profile at the “horns” of the Purpose: In stereotactic radiosurgery(SRS), targets in the brain are located in CT or MR images and localized relative to external fiducial. Physicians need accurate assessments of the accuracy with which they can expect to hit targets in order to prescribe appropriate margins. We have developed a new tool which is set up for irradiation in a manner similar to a patient demonstrates the clinical accuracy of the imaging device. No commercial device currently exists which allows measurement of accuracy using both CT and MR planning images. Methods: The device contains radiosurgical targets visible by both CT and MR, with radiochromic film to capture the delivered dose distribution. The device consists of several thin acrylic plates holding non‐ferrous CT fiducial targets which pierce a piece of film. Adjacent to the film are channels filled with copper sulfate solution to provide MR fiducials. The assembly fits inside a water‐filled head phantom which can be secured to a SRS system. The phantom is imaged using CT or MR. Targets are determined using the treatment planning software and methods appropriate to the SRS device. Targeting accuracy is determined by measuring distances on the film between dose distribution centers, relative to the pinhole made by the CT fiducials. Results: Measurements of GK SRS accuracy with 4 mm and 8 mm collimation sizes show sub‐millimeter agreement for planned and delivered dose distributions with CT or MR imaging, in agreement with GK SRS manufacturer specifications. Conclusions: Assessing accuracy of dose delivery for GK or linacSRS can be limited by the use of one imaging modality for patient treatment. This phantom design enables the use of MR‐only, CT‐only, or MR‐CT‐combined image‐based assessment for GK or linacSRS approaches, providing physicists and radiation oncologists with basic accuracy information that is relevant to patient treatment.

SU‐GG‐T‐94: An Automated Tool for Determining Output Factor for Electron CutOuts

Purpose: To provide streamlined methods for quickly and accurately determining the output factor for irregularly shaped electron cutouts. Measurement on a linear accelerator of output factor for electron cutouts is a time consuming physics task. Existing methods that calculate output factor for shaped apertures are approximations based on cutout length and width, or require a printed version of a cutout shape for either manual measurement or scanning and digitizing. Methods and Materials: An integrated spreadsheet and database provides several options for output factor determination: a measured value or an approximation based on cutout length and width may be used, an image based database containing previously measured cutouts can be searched, or the electron cutout shape contour can be imported electronically from the Philips Pinnacle3treatment planning system, and the output factor calculated without further user interaction. For the last option, a weighted sector‐integration based algorithm is used that employs polar coordinates and does not require either the calculation of the intersection of equispaced vectors with the block contour or the repeated rotation of the contour. A set of measurements taken for a variety of energies, electron cone sizes, and circular cut out diameters is used to generate a standard output factor table for the sector integration approach. Results: The output factor calculations were tested on a set of cutouts for electron boost treatments. Comparison of measured and calculated output factors are presented. Conclusion: In a busy clinic environment, measurement of output factors for electron cutouts is time burden on physics staff, so automated methods of calculation that involve a minimum of staff time while preserving accuracy have proved valuable in our clinics.

SU‐GG‐T‐352: A Tracking Database for Clinical Implementation of Microdosimeters

Purpose: To develop a database tracking tool to facilitate clinical implementation of a patient microdosimetry system with Optically Stimulated Luminescent Dosimeters. The database allows day to day tracking of treatment dose verification and long term analysis of microdosimeter efficacy and accuracy for all patients treated in a multi‐site department. Methods and Materials: The nanodot microdosimeters (Landauer, Inc., Glenwood, IL) are irradiated for one field per patient, once a week, at three hospital sites. The volume of data produced creates the need for an organized and systematic way of monitoring weekly dose reading results. It also provides a substantial base of information for evaluating the processes involved in the planning, placement, and reading of the nanodots. A nanodots tracking database package was created for this purpose using Microsoft Access. For each nanodot irradiated, parameters regarding the individual dosimeter, patient, staff involved and treatment site are recorded, along with the expected and measured doses. Results:Nanodot data has been collected in a tracking database composed of thousands of readings. This data is followed by the physicist and approved by the physician to verify dose delivery for a given patients treatment. Examples are given illustrating how the nanodot data can presented graphically to evaluate results over time, and filtered for quality assurance evaluation of patient treatments within a multi‐machine multi‐site department. Conclusion: An organized database is essential for the management of a large scale implementation of patient microdosimetry use. The study of long term trends in microdosimeter data can be used to point to areas in which processes can be optimized and quality of care improved.