A Daily QA Phantom for Linear Accelerator with Image-Guided Radiosurgery Capability

Purpose/Objective(s): Image guided radiosurgery and stereotactic body radiotherapy (IG-SRS/SBRT) have been widely adopted on linear accelerators. Ideally, a daily quality assurance program on such machines should provide the targeting accuracy from image guidance quantitatively and efficiently. A suitable phantom for such test requires sufficient landmarks for robust online image registration and a high contrast ball bearing (BB) to provide clear images for Winston Lutz (WL) test. Such phantoms, however, are not widely available. In this study, we investigated a prototype daily QA phantom that allows all major components of the IG-SRS/SBRT process to be tested. Materials/Methods: The phantom is a 12 cm cube with 3 pairs of AL rods at 3 corners, and 4 embedded AL ceramic BBs. The BB in the center of phantom is 5/16 inch in diameter, while the others are ¼ inch in size. On the treatment table, the phantom sits atop a holder with build-in pitch, roll, and pitch. The holder can be index to the couch at a preset starting position. A high resolution planning CT was acquired of the phantom on a level surface. A plan was then generated with isocenter placed precisely in the central BB of the phantom. The plan was delivered on a medical linear accelerator and a radiosurgery system, using cone beam CT (CBCT) and stereoscopic x-ray system for image guidance, respectively. Both imaging systems provide 6DoF deviations that were subsequently applied to the treatment couch. Afterwards, verification imaging was acquired with an on-board imaging (OBI) KV/MV pair on a medical linear accelerator and a 2nd pair of stereotactic x-ray images on a radiosurgery system, and WL tests were performed to measure the agreement between central BB and machine isocenter. Results: The plan was delivered 12 times each on the two machines. The initial correction, deviation from verification imaging, and WL test results (μ ± σ) are provided in the table below, where R represents 3D vector length.The verification imaging indicated that all 6DoF corrections were applied correctly, and WL test showed that both systems are able to localize well within 1 mm accuracy on a phantom. Conclusion: The phantom incorporates necessary features to enable daily QA with the goal of simulating IG-SRS/SBRT workflow and quantitatively evaluating localization accuracy. Accuracy determined from this daily QA procedure should represent the upper limit of targeting accuracy for ideal conditions. [Figure presented