University of Alberta, Department of Oncology, Division of Medical Physics, 11560 University Ave, Edmonton, Alberta T6G 1Z2, Canada.
Alberta Health Services, Department of Medical Physics, Cross Cancer Institute, 11560 University Ave, Edmonton, Alberta T6G 1Z2, Canada.
Biomed Phys Eng Express. 2022 Sep 7;8(6). doi: 10.1088/2057-1976/ac8e72.
. To present and share an open-source system (phantom and software) for verifying the targeting accuracy of linac-based, single-isocenter, multi-target radiotherapy. This quality assurance test extends the traditional Winston-Lutz test, which considers a single target located at isocentre.. Plans for a 3D-printed phantom are provided, which can be customized to accommodate various target (BB) positions. Given BB positions and gantry/collimator/couch combinations, the software generates multi-leaf collimator positions to facilitate multi-target Winston-Lutz (MTWL) plan creation. The software determines deviations between detected and expected BB positions on MV images resulting from MTWL plan delivery. BBs are located using a Hough circle detection algorithm, which is modified to favour the detection of circles: (1) having a reasonable size, (2) that are contained within the radiation field, and (3) having reasonable pixel intensities. Validation was performed in two ways: (1) using synthetic data with zero targeting errors and (2) by measuring real linac targeting errors and comparing against results obtained using a commercial system.. Validation using the synthetic data yielded a mean (maximum) absolute discrepancy of 0.11 mm (0.21 mm), which is comparable to the synthetic phantom resolution (0.2 mm). The mean (maximum) absolute discrepancy compared to the commercial system is 0.13 mm (0.43 mm). These values are similar to results obtained with repeated deliveries of the same MTWL plan with the same phantom setup. Both validation tests yield reasonable results and are therefore considered successful. The MTWL test was performed independently by three physicists on two linacs to investigate repeatability, resulting in a mean (maximum) absolute discrepancy of 0.14 mm (0.51 mm) among the various attempts.. Successful completion of this quality assurance test, using our customizable and open-source system, provides confidence that multi-target, single isocentre radiotherapy treatments can be delivered with sufficient geometric accuracy according to the chosen tolerance level.
. 介绍并分享一个用于验证基于直线加速器的单等中心、多靶区放射治疗靶区定位准确性的开源系统（包括体模和软件）。该质量保证测试扩展了传统的 Winston-Lutz 测试，后者仅考虑位于等中心的单个靶区。... 提供了 3D 打印体模的计划，可以根据各种靶区（BB）位置进行定制。根据 BB 位置和机架/准直器/治疗床组合，软件会生成多叶准直器位置，以方便创建多靶区 Winston-Lutz（MTWL）计划。软件会根据 MTWL 计划的实施情况，确定 MV 图像上检测到的和预期的 BB 位置之间的偏差。使用 Hough 圆检测算法来定位 BB，该算法经过修改，以利于检测具有合理大小、位于射束范围内且具有合理像素强度的圆。验证是通过两种方式进行的：（1）使用零靶区误差的合成数据，以及（2）通过测量真实直线加速器的靶区误差，并与使用商业系统获得的结果进行比较。... 使用合成数据进行验证，得出的平均（最大）绝对偏差为 0.11 毫米（0.21 毫米），与合成体模分辨率（0.2 毫米）相当。与商业系统相比，平均（最大）绝对偏差为 0.13 毫米（0.43 毫米）。这些值与使用相同体模设置重复实施相同 MTWL 计划的结果相似。两种验证测试均得出合理的结果，因此被认为是成功的。三位物理学家在两台直线加速器上独立进行了 MTWL 测试，以研究重复性，结果显示，在不同的尝试中，平均（最大）绝对偏差为 0.14 毫米（0.51 毫米）。... 使用我们的可定制和开源系统成功完成这项质量保证测试，使我们有信心根据所选容差水平，能够以足够的几何精度实施多靶区、单等中心放射治疗。
