DOS 542 - Week 3 Discussion
Initial Post: QA of IGRT Systems
Ensuring the accuracy and precision of imaging components of radiation delivery systems is critical to maintaining the safety and efficacy of the system. A robust quality assurance (QA) process tests the various parts of the system at daily, weekly, monthly, and annual intervals. Loyola's physics faculty has designed a local QA program based largely on TG-142, which is an extensive set of recommendations for QA of all of the components of a linear accelerator.1 Each of Loyola's four linear accelerators have different onboard imaging (OBI) capabilities to provide image guided radiation therapy (IGRT), and patients are matched to appropriate machines based on the needs of their treatments.
The two Varian 21EX machines, known locally as "2163" and "2164", and the TrueBeam all have megavolt (MV) imaging capabilities. The TrueBeam and 2163 also have kilovolt (kV) and cone beam CT (CBCT) capabilities. Each of these imaging systems is tested daily with a simple set of procedures. A small block phantom is placed on the treatment table, and the room's lasers are aligned to a set of surface marks on the phantom.2 Images of this phantom are acquired with MV, kV, and CBCT as available on each machine. In each case, a set of radiopaque markers inside the phantom are used to measure the shifts necessary to drive the treatment couch to the place the treatment isocenter at a secondary location in the phantom. The couch shifts are applied and the therapists check that the lasers have moved to the set of marks indicating the second point of interest inside the phantom. The magnitude of the shifts should be approximately 5 mm, and the the laser alignment on the phantom should be within 1 mm of the markings after the shifts have been applied.
The TrueBeam has a more extensive set of imaging tests, since it is used for stereotactic procedures. In addition to the basic phantom test performed with the linac gantry at 0 degrees and the table at 0 degrees, and rotating set of gantry and table angles are imaged each day of the week.3 Each day, images are taken with the table at 0 and the gantry at 90, 180, and 270 degrees. Two more images are taken with some combination of the gantry at 0 or 180 and the table at 90, 45, 0, 315, or 270 degrees depending on the day of the week.
Both the 2163 and TrueBeam are equipped with optical tracking systems for respiratory gating. Each day, the cameras are calibrated by placing a block on the treatment table at the isocenter position. Reflective markers on the blocks are picked up by the cameras to set the calibration.
The fourth linac, a Novalis system, uses an entirely different OBI system. The ExacTrac system included R-ray imagers mounted in the floor at 45 degree superior/posterior oblique angles, such that they are pointed through the isocenter in the inferior/anterior direction. Daily warmup includes taking a series of images at each available energy of the imagers.4 Once the X-ray imagers are ready, they need to be tested using a phantom. Before this can happen, the optical tracking portion of the ExacTrac system must be calibrated to set the isocenter location. Each day, a small phantom is placed on the treatment table and aligned to the room lasers. Reflective markers on the phantom are detected by stereoscopic cameras, and the calibration is tested by moving the phantom slightly and ensuring that the tracking system is able to automatically move the couch such that the phantom is back at the isocenter. These reflective markers are the basis of the stereotactic positioning system, so it is critical that this is done correctly. Once the isocenter has been calibrated and confirmed based on room lasers, a different phantom is placed on the table. This phantom also has reflective markings that are picked up by the cameras. The table automatically moves the phantom's measurement point to the laser isocenter. Once in position, the X-ray imagers shoot a pair of obliquely orthogonal images which reveal internal details of the phantom. Shifts to a second set of marks are then calculated, applied, and then verified on the phantom. Once this is complete, a Winston Lutz pointer is attached to the stereotactic frame mounting point on the couch, and it is aligned to the room lasers. It is imaged with the X-ray system, and minuscule errors in agreement can be detected. The tolerance of this setup is 0.4 mm.
Monthly QA of each system is more involved. In addition to running the same isocenter verification tests as the daily QA, the monthly checks also evaluate the quality of the imaging and the health of the imaging components.5 The MV imaging panels are tested for defective pixels using a dark field test. This looks for pixels that are registering a signal significantly higher than other pixels in their local region of the panel. The MV panels are also tested for contrast resolution using a Las Vegas Phantom. The phantom has circles milled out at various depths and sizes, and the MV images are examined to determine how many of the circles are visible when captured at each available MV energy.
Monthly image quality testing of the kV imaging systems is performed with a Leeds Phantom, which is a device with a series of discs of different sizes embedded inside it to test low contrast resolution, as well as arrays of parallel bars mounted at different separation distances to test high contrast resolution. The test involves counting how many discs are visible on the kV image, and determining the closest spacing of the parallel line arrays that is still discernible as separate lines. The Varian specification suggests 12 of the 18 discs must be visible, but Loyola wants to see 13. The baseline for line separation is 2.24 line pairs per millimeter.
Monthly image quality testing of the CBCT systems uses a Catphan Phantom, which houses a wide assortment of internal features including cylindrical plugs of materials with known densities, such as air, acrylic, and low density polyethylene. A CBCT image is acquired and a 7x7 pixel region of interest (ROI) is placed inside each plug. The mean Hounsfield Unit (HU) reading from this ROI must be within 50 HU of the reference value for each material type. The uniformity of the HU readings is also tested by sampling a 20x20 pixel ROI in the center of the phantom and near the edge in each of the cardinal directions. These readings must agree within 30 HU according to Varian's specifications. High contrast resolution is tested by scrolling within the phantom image to an array of differently spaced parallel bars. The baseline for the smallest discernible group is 8 line pairs per cm for a head scanning protocol image and 5 line pairs per cm for a pelvis scanning protocol image. Low contrast is tested by scrolling inside the phantom image to a series of discs of different sizes. The smallest disc that is discernible as a circle should be no larger than 6 mm based on acceptance testing.
Annual QA of the MV imaging panels includes a test of their full range of motion and a test of the collision detection system.6 Annual QA of the kV and CBCTsystems mostly deals with energy output of the X-ray sources. The kV imaging system is tested by a physics consultant to ensure that the kV beam quality and dose output are in line with the specifications required by the state of Illinois. The CBCT is also tested by a physics consultant to ensure that the dose output is in line with specifications required by the state of Illinois.
- Klein EE, Hanley J, Bayouth J, et al. Task Group 142 report: quality assurance of medical accelerators. Med Phys. 2009;36(9):4197-4212. http://dx.doi.org/10.1118/1.3190392
- Roeske J. Daily quality assurance procedure for the Varian OBI. [Departmental Memorandum]. 2015.
- Rusu I, Roeske J. TrueBeam warmup procedures. [Departmental Memorandum]. 2015.
- Sethi A, Roeske J, ExacTrac quality assurance. [Departmental Memorandum]. 2015.
- Rusu I. TrueBeam monthly OBI QA procedures. [Departmental Memorandum]. 2015.
- Roeske J. Annual linac quality assurance procedures. [Departmental Memorandum]. 2015.