A novel-integrated quality assurance phantom for radiographic and nonradiographic radiotherapy localization and positioning systems

Transcription

1 A novel-integrated quality assurance phantom for radiographic and nonradiographic radiotherapy localization and positioning systems Amy S. Yu a) Tyler L. Fowler, and Piotr Dubrowski Department of Radiation Oncology, Stanford University School of Medicine, Stanford, CA, USA (Received 5 January 2018; revised 26 March 2018; accepted for publication 25 April 2018; published 23 May 2018) Purpose: Various localization and positioning systems utilizing radiographic or nonradiographic methods have been developed to improve the accuracy of radiation treatment. Each quality assurance (QA) procedure requires its own phantom and is independent from each other, so the deviation between each system is unavailable. The purpose of this work is to develop and evaluate a single-integrated QA phantom for different localization and positioning systems. Methods: The integrated phantom was designed in three-dimensional (3D) CAD software and 3D printed. The phantom was designed with laser alignment marks, a raised letter S on the anterior surface for optical surface monitoring system registration, a core for radiofrequency (RF) tracking system alignment, eight internal fiducials for image alignment, and an isocentric bearing for Winston Lutz test. Tilt legs and rotational stage were designed for rotational verification of optical surface mapping system and RF tracking system, respectively. The phantom was scanned using a CT scanner and a QA plan was created. This prototype phantom was evaluated against established QA techniques. Results: The QA result between the proposed procedure and established QA technique are and mm, respectively, for RF tracking system and and mm for Winston Lutz test. There is no significant difference for the QA results between the established QA and proposed procedure (P > 0.05, t test). The accuracy of rotational verification for surface mapping system and RF tracking system are less than 0.5 and 1 compared the predefined value. The isocenter deviation of each location system is around l mm. Conclusion: We have designed and evaluated a novel-integrated phantom for radiographic and nonradiographic localization and positioning systems for radiotherapy. With this phantom, we will reduce the variation in measurements and simplify the QA procedures American Association of Physicists in Medicine [ Key words: localization positioning system, nonradiographic and radiographic QA, quality assurance, radiotherapy 1. INTRODUCTION With escalating doses and decreasing margins for modern radiation therapy treatments, a positioning system with high accuracy is essential. In order to deliver the dose precisely, several localization techniques have been developed for positioning patients, so treatment can be delivered with precision and efficiency. These techniques can be used before and/or during the treatment to setup and monitor patient position and can be classified into two broad categories: radiographic and nonradiographic systems. A radiographic system includes kv/mv portal images and cone-beam computed tomography (CBCT). A nonradiographic system includes an optical surface mapping system and a radiofrequency (RF) tracking system. An optical surface mapping system (e.g., OSMS, AlignRT, C-RAD, and humediq) is a 3D imaging technology that provides a high resolution and accurate 3D surface data referenced to the treatment isocenter. The RF tracking system (e.g., Calypso) is a tracking device using RF waves to localize beacons within or near the target. Each system requires its own periodic measurement of specified parameters to ensure that hardware and software function safely and reliably with its own procedure and phantom. 1,2 It is time consuming to setup the QA phantom and to perform the procedure for each system. Many independent 3 5 and manufacture-provided phantoms and analysis tools are currently available but often result in excessive duplication of effort in performing routine QA on treatment systems. A single phantom can provide a more efficient procedure and reduce setup variation when one test object is replaced at isocenter with another. Therefore, in this study, a novel phantom was designed and evaluated to integrate QA procedures for RF tracking system, optical surface mapping system, MV isocenter (Winston Lutz test), and imaging system isocenter. By combining these QA tests into a single phantom, we seek to have an independent QA method from the vendor, to reduce the number of phantoms required for QA, to reduce the time required of a medical physicist to perform these QA tasks, and to increase the reproducibility of phantom placement. Unique to our design is the inclusion of rotational accuracy testing of RF tracking and optical surface mapping systems. Most importantly, it is feasible to evaluate the deviation of isocenter between the individual systems Med. Phys. 45 (7), July /2018/45(7)/2857/ American Association of Physicists in Medicine 2857

2 2858 Yu et al.: Integrated QA phantom and procedure METHODS AND MATERIALS 2.A. Phantom design The integrated phantom ( cm 3 ) was designed in SolidWorks 3D CAD software (Dassault Systemes, Velizy-Villacoublay, France) and 3D printed using a Ultimaker 2 + printer (Ultimaker B.V., Geldermalsen, the Netherlands) with a 0.4 mm printing nozzle, 0.1 mm layer resolution and using polylactic acid (PLA) material. The reason PLA was chosen for printing material is its abundance, ease of use, and cost-effectiveness for rapid prototyping. The wall thickness of the phantom is 2 mm. The phantom was 3D printed in seven separate parts and assembled using nylon hardware. The phantom allows the evaluation of the radiation isocenter (Winston Lutz test), megavoltage portal imaging localization, kilovoltage onboard imaging localization, optical surface mapping system, RF tracking system, and the accuracy of room lasers. Additionally, there is an internal rotational stage that can rotate implanted RF beacons to a predefined angle (10 ) for testing the accuracy of the RF tracking system to detect target rotation while maintaining the precise position of the isocentric bearing. The rotational core can be adjusted from outside the phantom with a knob and set to the predefined angle. This rotational core was designed to place RF beacons outside the radiation field to avoid confusion when performing the image analysis for the Winston Lutz test. The placement of beacons is showed in Fig. 1. There are eight internal radiopaque fiducials for imaging system alignment and a 6.35 mm isocentric bearing for Winston Lutz test. The phantom can also be tilted to a predefined angle (10 ) using the extendable legs to test the accuracy of the surface mapping system to detect rotation. Additionally, the phantom features 1 mm wide external grooves to check room laser alignment. The surface of the phantom was designed with a raised logo and shape to create a unique surface topography for tracking with the optical surface mapping system, that is, for this prototype, a raised letter S is on the anterior optical surface for surface mapping (S-Phantom). The detail of phantom design is shown in Fig. 2. The angle of the RF beacon rotational stage was evaluated via CT imaging. Three sets of CT images were acquired with the knob at center (no rotation), switching the knob to the left and to the right without moving the phantom. By blending FIG. 2. Integrated quality assurance phantom featuring (a) laser alignment marks, (b) a raised letter S on the anterior surface for optical surface mapping system registration feature, (c) rotational radiofrequency beacon stage for rotational verification with a knob (arrow) to rotate the stage from outside the phantom, (d) isocentric bearing for Winston Lutz test, and (e) the tilt legs for the rotational verification of the optical surface mapping system. The dots are RF beacons placed on the rotational stage and outside radiation fields to avoid confusion when performing the image analysis for the Winston Lutz test. [Color figure can be viewed at wileyonlinelibrary.com] the two images, the angle of rotation can be measured in the treatment planning system (Fig. 3). The tilted angle of the physical phantom with extended legs was calculated by trigonometry. The dimension of the phantom is 15 cm and when the phantom is tilted, the phantom was raised by 2.6 cm. 2.B. Proposed QA workflow The S-Phantom was scanned using a CT scanner (Discovery 710, GE Healthcare) with 1.25 mm slice thickness. A QA plan was created including fields for setup using the onboard imaging and fields for the Winston Lutz test in the treatment planning system (Eclipse, Varian, Palo Alto, CA, USA). The optical surface mapping system (Varian, Palo Alto, CA, USA) uses a 3D point cloud to represent the surface of subject s body obtained from a 3D camera in-room monitoring system and compares it with a CT-derived surface as a reference, imported via a DICOM file from treatment planning FIG. 1. The placement of beacons inside the integrated phantom, (a) the transverse plane, (b) sagittal plane, and (c) coronal plane of the CT image. This rotational stage (arrow) and the placement of the beacons (cyan cross) were designed, so beacons were placed outside the radiation field for Winston Lutz test. The orange line is the radiation field border for Winston Lutz test. The cyan crosses are the beacons which are located outside the radiation fields. [Color figure can be viewed at wileyonlinelibrary.com]

3 2859 Yu et al.: Integrated QA phantom and procedure 2859 (a) (b) FIG. 3. The angle of rotational stage for the RF beacon was evaluated by the CT images. The two CT images were blended, and the angle of rotational stage was measured in the treatment planning system; (a) the rotational stage was rotated to the left and (b) the rotational stage was rotated to the right. [Color figure can be viewed at wileyonlinelibrary.com] system. Therefore, a body contour of the S-Phantom was created and carefully reviewed in the treatment planning system and exported to the optical surface mapping system. For RF tracking system, two beacons were placed on one rotational stage and the third beacon was placed on the other rotational stage (as seen in Fig. 2). This arrangement ensures the center of mass of the three beacons falls at the isocenter. Beacons were identified and marked/localized in the treatment planning system and imported into the RF tracking workstation. Individual QA tests for each system were performed using the S-Phantom and subsequently compared to the vendor-provided QA method. After the S-Phantom was validated against each vendor s method, an integrated QA procedure was performed to quantify and evaluate the deviation of the isocenter position of each system. Truth) on the couch and the distance (D) between QA phantom isocenter to the calibrated isocenter was calculated by Eq. (1) 6 : qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi D ¼ XQA 2 þ Y2 QA þ Z2 QA (1) where X QA, Y QA, and Z QA are the distances of the QA isocenter to the calibrated RF tracking system isocenter in the X, Y, and Z axes, respectively. The X, Y, and Z axes are in the lateral, longitudinal, and vertical directions, respectively. After the displacement between the calibrated isocenter and phantom isocenter was recorded, the rotational stage was switched to the predefined angle (10 ) with a designed knob to verify the accuracy of rotation for the RF tracking system (Figure 4). 2.C. Comparison between S-Phantom and vendor s phantoms For each system, QA was performed first using the vendor s phantom with the recommended method, followed by QA using the S-Phantom with methods as close as to the way we performed vendor s QA to get fair comparisons. Twosided t test was used to evaluate the difference of the QA results between the S-Phantom and vendor s phantom, and differences with P > 0.05 were considered statistically insignificant (n = 10). The calibrated isocenter is the one defined during the calibration of each system. The QA isocenter is the isocenter revealed as a result of performing the specific QA test; indicating any discrepancy from the ideal calibrated isocenter. 2.C.1. Radiofrequency beacon-tracking system The vendor s QA phantom (Calypso â, Varian, Palo Alto, CA, USA) and the S-Phantom both contain three beacon transponders used to verify system performance by comparing their current position against one established during calibration. The phantom was aligned with the lasers (Ground 2.C.2. Optical surface mapping system The optical surface mapping system (OSMS â, Varian, Palo Alto, CA, USA) QA phantom provided by the vendor consists of a large calibration plate with a grid of printed circles. It was placed on the couch with the center of the plate aligned to the isocenter. The mapping system acquired a pair of images of the plate and a software routine registered the centers of the circles. For the S-Phantom, once the S-Phantom was setup with onboard imaging (Ground Truth), the deviation between the DICOM reference surface and the S-Phantom surface on the couch was recorded. After the deviation between two isocenters was recorded, the tilt legs were extended to verify the pitch angle for the optical surface mapping system (Fig. 4). 2.C.3. Winston Lutz test A Winston Lutz phantom provided by the vendor (Varian, Palo Alto, CA, USA) was aligned with the aid of digital graticule to the center of Winston Lutz target via a pair of MV images. Next, images of various radiation

4 2860 Yu et al.: Integrated QA phantom and procedure 2860 FIG. 4. The accuracy of position and rotation verification for (a b) the radiofrequency tracking system and (d e) the optical surface mapping system. (c) The rotational stage was switched to the predefined angle (10o) with a small knob (arrow) and (f) the tilt legs were extended (arrow) to verify the pitch angle for the optical surface mapping system. (b) The box indicates the rotation of the target detected by the radiofrequency tracking system. The boxes indicate the pitch (d) before and (e) after the tilt legs got extended out from the phantom. [Color figure can be viewed at wileyonlinelibrary.com] fields shaped with MLC were acquired by the electronic portal imaging device (EPID) with the gantry angles: 0, 90, 180, and 270 ; couch angles: 0, 45, 90, 270, and 315 ; and collimator angles: 0, 90, and 270. EPID images were analyzed by RIT V software (Radiological Imaging Technology, Colorado Springs, CO). The 3D displacement, the distance between the center of the radiation field and the sphere, was recorded and defined as Eq (1): where XQA, YQA, and ZQA minimize the overall setup displacement on all EPID images in the X, Y, and Z axes, respectively.7 The X, Y, and Z axes are in the lateral, longitudinal, and vertical directions. For the S-Phantom, the S-Phantom was also aligned with the aid of digital graticule to the central of ball bearing with a pair of MV images. EPID images of radiation field shaped with MLC with the bearing ball at the isocenter were acquired at different angles of couch, gantry, and collimator as described above. 2.D. Deviation between the Isocenter of each System and the Imaging Isocenter This test is to streamline the QA procedure after the phantom is validated against vendor s method. The S-Phantom was initially positioned on the treatment couch using the room lasers. A kv image pair and CBCT were taken for the fine adjustment, so the isocenter of each system will be linked to the imaging isocenter which is the Ground Truth for the integrated QA. The uncertainty and the stability of the onboard imaging system were evaluated by the results of monthly isocal test (Varian, Palo Alto, CA, USA) that shows to be very stable (standard deviation = mm) over 12 months. All the tests were repeated 10 times (n = 10). After the S-Phantom was accurately positioned on the couch, the optical surface mapping system was turned on and the deviation between the exported DICOM phantom surface and its current position on the couch was recorded for the lateral, longitudinal, and

5 2861 Yu et al.: Integrated QA phantom and procedure 2861 TABLE I. The displacements of the DICOM reference surface and the S-Phantom centered at the radiographic isocenter as measured by the optical surface mapping system. Couch angles (Mean SD) Vertical (mm) Lateral (mm) Longitudinal (mm) Rotation ( ) Roll ( ) Pitch ( ) TABLE II. Comparison between vendor s phantom and the S-Phantom. Vendor phantom S-Phantom (n = 10) Optical surface mapping system (n = 10) Isocenter location N/A mm Rotational verification N/A <0.3 Radiofrequency tracking system (n = 10) Isocenter location mm mm Rotational verification N/A <1 Winston Lutz test (n = 10) Isocenter location mm mm Total QA time 60 min 20 min vertical directions as well as the rotation, pitch, and roll for couch positions 0, 45, 90, 270, and 315. Next, the RF tracking system was deployed, and the deviation of QA isocenter and calibrated isocenter of RF tracking system were recorded. Finally, the Winston Lutz test was then performed and analyzed as described above. The projected location of the room lasers on the phantom with respect to the radiographic isocenter was verified visually at the end of the procedure. 3. RESULTS 3.A. Comparison between S-Phantom and vendor s phantom The deviation between the QA isocenter and calibrated isocenter locations with the use of S-Phantom and vendor QA phantom are and mm (P = 0.91, n = 10) for RF tracking system, and and mm (P = 0.87, n = 10) for Winston Lutz test. For the optical surface mapping system, however, the vendor-provided daily QA procedure does not report an isocenter deviation between calibrated and phantom isocenters but only a relative shift between two camera pods. Therefore, the largest relative shift between two camera pods, mm (n = 10), was reported here. Correspondingly, the result using the S-phantom is mm (n = 10) while couch is at 0 o. Room laser alignment agrees well within 0.5 mm with respect to the radiographic isocenter. Finally, the accuracy of rotational verification for optical surface mapping system with extendable legs and RF tracking system with the rotational stage is less than 0.5 and 1 degrees compared to the designed value (Fig. 4), respectively. 3.B. Deviation between the Isocenter of each System and the Imaging Isocenter The result of the deviation between the imaging isocenter and RF tracking system isocenter is mm (n = 10). The displacement of the DICOM reference surface and the surface of the S-Phantom for the lateral, longitudinal, and vertical directions as well as the rotation, pitch, and roll are listed in Table I for each couch angle (n = 10). The deviation between the imaging isocenter and optical surface mapping system isocenter is mm while couch is at zero. The entire procedure takes about 20 min from setting up the S-Phantom to finishing the data acquisition. The summary of comparison between the S-Phantom and vendor s phantom are listed in Table II. 4. DISCUSSION The optical surface mapping system is often used for head and neck and brain stereotactic radiosurgery treatment, 8,9 RF tracking system can be used for stereotactic ablative body radiotherapy for low-risk prostate cancer, 10 and kv/mv/ CBCT imaging is widely used for the routine patient setup. Radiation isocenter verification (Winston Lutz test) is essential for brain stereotactic radiosurgery treatment. These technologies have been developed to improve the accuracy and the precision of treatments, and thereby, enable clinicians to deliver higher doses of radiation directly to the tumor while potentially avoiding healthy surrounding tissues with the increased precision and decreased margin. However, each positioning system requires an independent calibration and periodic measurement of specified parameters. QA phantoms can take up a significant amount of storage space within a radiation therapy department (Fig. 5) and require medical physicist staff to setup several different phantoms which increase the total time required for QA and reduce the positional reproducibility. A phantom was designed and evaluated to streamline the QA procedure for different routinely used localization and positioning systems. As suggested by TG-142, localization systems should align with the radiological isocenter within 1 mm or 2 mm

6 2862 Yu et al.: Integrated QA phantom and procedure 2862 FIG. 5. A comparison of process maps illustrates the reduction in number of phantoms. (a) Uncertainty introduced by the reproducibility of quality assurance (QA) phantom setup (left) and the reduction of uncertainty with the single isocentric QA phantom and procedure (right). With the use of this integrated phantom, it is possible to reveal the deviation of the isocenters from each system. (b) The collection of phantoms required for the QA procedures (left) compared to one single phantom can do it all (right). Note: RF, radiofrequency tracking system; Surface, optical surface mapping system; Radiation, Winston Lutz test. [Color figure can be viewed at wileyonlinelibrary.com] for IGRT depending on what radiation treatment procedures will be done on the particular machine. 1 In previous studies, optical surface mapping systems have demonstrated that the accuracy of calibration is within 1 mm and 1 which agrees with our study. 11, 12 Another study evaluated the optical surface mapping system with respect to CBCT imaging and showed similar accuracy for positioning with differences less than 1 mm for the linear vector displacement and 0.5 for rotational inaccuracy. 4 Likewise, our results of the RF tracking system accuracy are comparable to the previous results evaluating the phantom provided by the vendor; namely about 1 mm linearly. 13 Comparisons for the rotation test were limited by the number of decimal places displayed by the RF tracking system (none) relative to the optical surface mapping system (one). Therefore, we were not able to report a higher resolution value in this study. In addition to our current study confirming the accuracy of each system independently, by using a single phantom for a number of QA tests, an intercomparison may be made that is independent of the uncertainty introduced by placement and replacement of multiple phantoms and allows one to quantify the deviation between each positioning system relative to a single system (Fig. 5). This transparent evaluation and inspection of intersystem variability can give confidence between different positioning systems, cross validate system QA results, and potentially uncover miscalibrations or system drift over time. 5. CONCLUSIONS It is essential to effectively and reliably perform routine QA checks independent from the vendor s method. In this study, we developed and evaluated an integrated phantom for radiographic and nonradiographic localization and positioning systems. In addition to routine QA, the rotational accuracy of the optical surface mapping system and the RF tracking systems can be verified. With this integrated QA phantom, it is feasible to quantify the physical distance from radiographic isocenter for each system. ACKNOWLEDGMENT No funding sources were used to support this study. CONFLICT OF INTEREST The authors have no relevant conflicts of interest to disclose. a) Author to whom correspondence should be addressed. Electronic mail: amysyu@stanford.edu. REFERENCES 1. Klein EE, Hanley J, Bayouth J, et al. Task Group AAoPiM: Task Group 142 report: quality assurance of medical accelerators. Med Phys. 2009;36: Willoughby T, Lehmann J, Bencomo JA, et al. Quality assurance for nonradiographic radiotherapy localization and positioning systems: report of Task Group 147. Med Phys. 2012;39: Peng JL, Kahler D, Li JG, Amdur RJ, Vanek KN, Liu C. Feasibility study of performing IGRT system daily QA using a commercial QA device. J Appl Clin Med Phys. 2011;12:3535.

7 2863 Yu et al.: Integrated QA phantom and procedure Mancosu P, Fogliata A, Stravato A, Tomatis S, Cozzi L, Scorsetti M. Accuracy evaluation of the optical surface monitoring system on EDGE linear accelerator in a phantom study. Med Dosim. 2016;41: Wooten HO, Klein EE, Gokhroo G, Santanam L. A monthly quality assurance procedure for 3D surface imaging. J Appl Clin Med Phys. 2010;12: Muralidhar KR, Komanduri K, Rout BK, Ramesh KK. Commissioning and quality assurance of Calypso four-dimensional target localization system in linear accelerator facility. JMedPhys. 2013;38: Low DA, Li Z, Drzymala RE. Minimization of target positioning error in accelerator-based radiosurgery. Med Phys. 1995;22: Pan H, Cervino LI, Pawlicki T, et al. Frameless, real-time, surface imaging-guided radiosurgery: clinical outcomes for brain metastases. Neurosurgery. 2012;71: Cervino LI, Pawlicki T, Lawson JD, Jiang SB. Frame-less and mask-less cranial stereotactic radiosurgery: a feasibility study. Phys Med Biol. 2010;55: Mantz C. A phase II trial of stereotactic ablative body radiotherapy for low-risk prostate cancer using a non-robotic linear accelerator and realtime target tracking: report of toxicity, quality of life, and disease control outcomes with 5-year minimum follow-up. Front Oncol. 2014;4: Bert C, Metheany KG, Doppke K, Chen GT. A phantom evaluation of a stereo-vision surface imaging system for radiotherapy patient setup. Med Phys. 2005;32: Peng JL, Kahler D, Li JG, et al. Characterization of a real-time surface image-guided stereotactic positioning system. Med Phys. 2010;37: Santanam L, Noel C, Willoughby TR, et al. Quality assurance for clinical implementation of an electromagnetic tracking system. Med Phys. 2009;36: