J Hanley

Task Group 142 report: Quality assurance of medical acceleratorsa)

The task group (TG) for quality assurance of medical accelerators was constituted by the American Association of Physicists in Medicines Science Council under the direction of the Radiation Therapy Committee and the Quality Assurance and Outcome Improvement Subcommittee. The task group (TG-142) had two main charges. First to update, as needed, recommendations of Table II of the AAPM TG-40 report on quality assurance and second, to add recommendations for asymmetric jaws, multileaf collimation (MLC), and dynamic/virtual wedges. The TG accomplished the update to TG-40, specifying new test and tolerances, and has added recommendations for not only the new ancillary delivery technologies but also for imaging devices that are part of the linear accelerator. The imaging devices include x-ray imaging, photon portal imaging, and cone-beam CT. The TG report was designed to account for the types of treatments delivered with the particular machine. For example, machines that are used for radiosurgery treatments or intensity-modulated radiotherapy (IMRT) require different tests and/or tolerances. There are specific recommendations for MLC quality assurance for machines performing IMRT. The report also gives recommendations as to action levels for the physicists to implement particular actions, whether they are inspection, scheduled action, or immediate and corrective action. The report is geared to be flexible for the physicist to customize the QA program depending on clinical utility. There are specific tables according to daily, monthly, and annual reviews, along with unique tables for wedge systems, MLC, and imaging checks. The report also gives specific recommendations regarding setup of a QA program by the physicist in regards to building a QA team, establishing procedures, training of personnel, documentation, and end-to-end system checks. The tabulated items of this report have been considerably expanded as compared with the original TG-40 report and the recommended tolerances accommodate differences in the intended use of the machine functionality (non-IMRT, IMRT, and stereotactic delivery).

Concise Matrix Analysis of Point-Based Prostate Targeting for Intensity Modulated Radiation Therapy

Intensity Modulated Radiation Therapy (IMRT) has recently emerged as an effective clinical treatment tool to treat various types of cancers by limiting the external beam dose to the surrounding normal tissue. However, the process of limiting external radiation dose to the tissue surrounding the tumor volume is not a trivial task. Several parameters including tumor volume and inhomogeneity, position and shape of the tumor volume, and the geometrical distribution of the radiation beams directly affect the determination of the external radiation dose. In addition, a major variable in effective delivery of the radiation dose is “set-up error” caused by the changes in patient position. Any changes in the position of the patient affect the geometrical location of the tumor volume and, therefore, need to be accommodated in the delivery of radiation beams during the treatment. This work presents a complete matrix representation required to calculate the three-dimensional rigid body homogeneous transformation matrices corresponding to external beam radiotherapy setup error and subsequent corrections in treatment beam parameters. A new concise orthogonal rotation solution is presented for use with clinical noisy data. Monte Carlo simulations prove the new matrix results are consistently better than the standard inverse solution. The required corrections in beam table, gantry, and collimator angles as function of the planned beam gantry angle are derived. For transformations that include a rotation on the sagittal plane, the required offsets to beam parameters are complex functions of the planned gantry angle but are clearly documented graphically for clinical use. A case study is presented with an error analysis that supports the use of the presented method in a clinical environment. Clinical implementation and evaluation of the presented method with patient data is also included in the paper.

SU‐FF‐T‐147: Determination of TG43 Parameters for Cs‐131 Model CS‐1R2 Seed Using Radiochromic EBT Film Dosimetry

Purpose: To measure the 2D dose distributions for Cs‐131 seed model CS‐1R2 for distances from 0.06cm to 4cm using radiochromic EBT film dosimetry. TG43 dosimetric parameters were generated. Method and Materials: Each radiochromic film (GAFCHROMIC® EBT lot ♯35076) was in contact with a model CS‐1R2 seed (IsoRay Medical) at the center of a solid water phantom, 30×30×20cm. A multiple film technique was employed. More than 50 films were separately exposed to 10 seeds, with the product of air kerma strength and exposure time between 14 and 800 Uhr. The seed strength ranged from 10 to 4U (NIST traceable) during the experimental runs. For calibration, 15 films (of the same lot) were exposed to 50kV x‐ray (M50) in air at 100cm SSD with doses ranging from 0.2 to 20Gy. Since the EBT film response is almost identical for M50 and M40 x‐rays, the energy response can be considered to be flat in this region and thus also the same for the Cs‐131 energy. All experimental, calibration and background films were scanned (pixel resolution 0.2mm) using a CCD camera‐based densitometer, with red and green light sources. Conversion from net optical density readings to doses were achieved based on the calibration curve established for each light source used in scanning. The 2D dose values in cylindrical coordinates were converted to polar coordinates, and the TG43 parameters were generated. Results: The dose rate constant, radial dose function and 2D anisotropy function were determined, and compared with those reported by other authors. General agreement was found. Conclusion: It is feasible to determine TG43 parameters for Cs‐131 model CS‐1R2 seed in solid water phantom using radiochromic EBT film. This method is superior to TLDdosimetry (1) in providing data with high spatial resolution for distances down to 0.06 cm and (2) within reasonably achievable time frame.

A clinical procedure for case-specific analytical validation of mono-modality image fusion in image guided radiotherapy.

Image guided radiotherapy can be performed by fusing the daily treatment and reference planning computed tomography scans. Decreased errors in patient setup can lead to smaller target margins that significantly improve treatment efficacy and outcome. The purpose of this work is to present a clinical procedure to analytically compute the accuracy of the registration. Accepted techniques such as normalized mutual information intensity based three-dimensional image registration can be validated using a large automated point sample. Without a user independent metric it is not possible to determine effect of the fusion error on the calculated correction in patient setup

SU‐GG‐T‐569: A Real‐Time Graphical User Interface‐Driven Radiotherapy Treatment Planning Simulator

Purpose: To present a stand‐alone RadiotherapyTreatment Planning simulator (RTPsim) that physicists can use for interactive classroom instruction and provide to students for independent study. Method and Materials: A PC with MATLAB programming language and the imaging toolbox is required. The RTPsim components are (1) a set of grayscale images for dose computation with regions of interest for dose area histogram analysis; (2) a double Gaussian pencil beam algorithm to create treatment fields; (3) dose distribution calculation routines for individual open or wedged beams and for a composite of weighted beams; and (4) a Graphical User Interface (GUI) to run the system with such user friendly features as command buttons to run algorithms, text or cursor variable input, list boxes for patient selection, and radio buttons for beam modifiers such as wedges. Results: The percent depth dose (%DD) for a typical 10×10 cm 6MV beam created by pencil beams closely approximates the actual clinical data. By computing and storing a correction factor table, the %DD values match exactly. The incorporation of divergence and a low dose region beyond the geometrical field edge result in excellent agreement in calculated versus clinical data for open beam profiles. Wedged field profiles show reasonable agreement with our clinical system. Several test cases from texts were reproduced accurately for interactive teaching purposes. As an example, an RTPsim wedged tangent field breast plan comparable to a clinical plan was produced. Variations in wedge, weighting, and normalization point can be demonstrated in real time. Use of a MATLAB GUI is suitable for planning demonstration; physicists and dosimetrists found the GUI convenient and easy to use. Conclusion: The user‐friendly RadiotherapyTreatment Planning simulator can simulate a clinical treatment planningcomputer with sufficient accuracy and speed for interactive educational purposes and independent learning.

TU‐E‐BRB‐07: Optimal Treatment Beam Fluence Generation for Volumetric Arc Therapy Using Dose Image Backprojection with Initial Corona Calculation

Purpose: Derive a novel solution to the initial premise that IMRT fluence can be easily calculated using dose image backprojection. Method and Materials: Calculations were performed using MATLAB with the imaging toolbox using parallel beam geometry. A phantom matrix P was constructed with a concave target, simultaneous integrated boost, and organ at risk with respective desired percent doses of 100, 110, and 40. Let R:SP be the backprojection of sinogram S of P. Iterative reconstruction using standard Ratio Method converges on a solution by computing Si+1 = Si * (P / R:SiP). For the novel Corona Method, we compute the corona using iterative backprojection reconstruction using ratios on only the target volumes T, T’ = R:S0T, and then superimpose the organ at risk, P’ = T’ + OAR. Iterative backprojection on P’ is then performed by successive addition correction, Si+1 = Si‐1 + (SO − Si), where SO is the sinogram of P’. Any negative intensities are set to zero during the iteration process. Target coverage is improved using the Compensated Phantom Method by computing a revised target based on the results Z of the process described thus far. We then define a new target T” = T’/Z and superimpose the OAR, such that P” = T” + OAR. The final procedure is to perform additive iteration constrained to contain only positive intensities. Results: Ratio Method: Target and boost areas receive their respective dose goals but the organ at risk unacceptable. Corona Method : OAR acceptable but poor target coverage. Compensated Phantom Method : Acceptable target coverage and the OAR goal is achieved. Conclusion: Dose image backprojection with initial corona calculation significantly reduces the dose to the organ at risk while maintaining acceptable target coverage.

SU‐FF‐T‐51: Robotically‐Assisted Minimally Invasive Brachytherapy: Dosimetric Aspects

Purpose: To investigate the dosimetric aspects of robotically‐assisted minimally invasive brachytherapy.Materials and Methods: Using the Intuitive Surgical DaVinci S System, the following were inserted into a porcine abdominopelvic cavity through a 12 mm auxiliary port and sutured directly to the pelvic sidewall: three 6‐French catheters; the Xoft Axxent delivery system; a prototype minimally invasive applicator (MIA). Plans were generated in Nucletrons treatment planning software for an Ir‐192 source and the Xoft 50 kV source. The ability of the prototype MIA to reproducibly position catheters was tested using filmdosimetry. Other applicators were designed to allow an en face application of the source, and various catheter arrangements were investigated. For the en face application, the sensitivity of the delivered dose to random deviations from normal incidence of the catheters (⩽ 5 degrees) and regular catheter spacing were tested. Results: All three trials for introducing the catheters were performed with ease by the surgeon. After the apparatus was sutured into place, it maintained a fixed geometry. The dosimetric measurements using GafChromic EBT film showed a uniform distribution of dose. The treatment plan for the en‐face arrangement using the Xoft source with 16 catheters at random angles showed a deviation from the planned dose of ⩽ 0.39%. Conclusions: The feasibility of performing robotically‐assisted minimally invasive brachytherapy was demonstrated for techniques utilizing both parallel and en face geometries. A customized applicator was designed to deliver a uniform dose to the target. The reproducibility of the device was confirmed with filmdosimetry. For the geometries tested, the en face application was found to be insensitive to random deviations of the catheters of up to 5 degrees from normal incidence. These techniques may be used for minimally invasive intra‐operative or post‐operative brachytherapy procedures using real‐time or atlas‐based planning.

A rapid and robust iterative closest point algorithm for image guided radiotherapy

Our work presents a rapid and robust process that can analytically evaluate and correct patient setup error for head and neck radiotherapy by comparing orthogonal megavoltage portal images with digitally reconstructed radiographs. For robust data Photoshop is used to interactively segment images and registering reference contours to the transformed PI. MatLab is used for matrix computations and image analysis. The closest point distance for each PI point to a DRR point forms a set of homologous points. The translation that aligns the PI to the DRR is equal to the difference in centers of mass. The original PI points are transformed and the process repeated with an Iterative Closest Point algorithm until the transformation change becomes negligible. Using a 3.00 GHz processor the calculation of the 2500x1750 CPD matrix takes about 150 sec per iteration. Standard down sampling to about 1000 DRR and 250 PI points significantly reduces that time. We introduce a local neighborhood matrix consisting of a small subset of the DRR points in the vicinity of each PI point to further reduce the CPD matrix size. Our results demonstrate the effects of down sampling on accuracy. For validation, analytical detailed results are displayed as a histogram.

SU‐GG‐T‐07: I‐125 Seed Dosimetry in the Near Field Using Gafchromic® EBT Film

Purpose: To perform radiochromic EBT film dosimetry in the near field of model MED3631A/M I‐125 seed and to determine the TG‐43 dosimetric parameters. Method and Materials: Gafchromic® model EBT films (lot♯35076) were horizontally positioned, one at a time, above and in contact with a NAS model MED3631A/M I‐125 seed horizontally placed at the center of a solid water phantom (RMI457, 30×30×20cm3). A multiple‐film technique was employed. To cover the distance range of 0.06–5cm, 50 EBT films of different sizes (4×4–12×12cm2) were irradiated by 10 individual seeds (NIST traceable air kerma strengths ∼10U, based on current NAS calibration) using different exposure times (1–295hr). A separate set of 19 calibration films was exposed to 50kV x‐rays at 100 cm SSD for doses of 0.2–40Gy at the University of Wisconsin. All experimental, calibration and background films were scanned using a CCD100 densitometer using a green light source with fine spatial resolution (0.2mm). Based on the established calibration curve, dose conversion for the experimental films and generation of the TG‐43 dosimetric parameters (for L=4.2mm) were achieved using IDL v6.0. Results: The dose rate constant was 1.004 cGy/Uh. The radial dose function was obtained for the radial distances from 0.06 to 5cm. The 2D anisotropy functions for distances from 0.1 to 5cm were determined. General agreement with published values and those recommended by TG‐43U1 report for r⩾0.5cm was found. The near field data for 0.25cm is close to Rivards Monte Carlo results for the “ideal capsule orientation”, but different from the TG43U1 consensus values based on Rivards weighted “diagonal” and “vertical” capsule orientations. Conclusion: Using EBT film dosimetry with the multiple‐film technique, TG‐43 dosimetric parameters were determined for model MED3631A/M I‐125 seed in the near field and out to 5cm, allowing for accurate treatment planning calculations. Research supported partially by North American Scientific Inc.

SU‐GG‐T‐117: Uniform Treatment of a Moving Target with Un‐Gated Stationary Fields

Purpose: To describe the mathematical formalism necessary to compute the stationary treatment field fluence for the entire respiratory cycle that delivers a uniform dose to a moving target. Method and Materials: The desired doseD to the central plane of a moving target perpendicular to a stationary radiation beam with fluence F can be mathematically described as D = M*F where M is a binary coefficient matrix. Our algorithm computes M as function of the periodic motion and determines F by iterative gradient descent search or linear programming. The theory was tested by analyzing a model in detail and simulating the motion on a treatment planning system by shifting the beam relative to a stationary target. A clinical moving target described in the literature as suitable for gated treatment was also analyzed. The motion was divided into phases and M determined, F optimized for a single stationary field encompassing one phase near the center of the motion, and then dose to target computed. Results: We have developed a matrix mathematical formalism to compute a stationary non‐gated field fluence that can be optimized to treat a moving target and spare normal tissue. For the model case with comparable coverage using an integrated target, the use of a single field with the optimized fluence decreased the 40% normal tissue dose from 79% to 52% of the volume. The clinical target computed fluence gave a mean dose of 100% with 92.4%, 110.1%, and 5.2 as the minimum, maximum, and standard deviation respectively. Conclusion: A technique has been developed to compute the stationary treatment field fluence for the entire respiratory cycle that delivers a uniform dose to a moving target without gating. Results indicate the potential for improved normal tissue sparing while delivering a relatively uniform target dose.