Department of Medical Physics, Liverpool and Macarthur Cancer Therapy Centres and Ingham Institute, Sydney, NSW 2170, Australia.
Centre for Medical Radiation Physics, University of Wollongong, Wollongong, NSW 2522, Australia.
Med Phys. 2017 Oct;44(10):5457-5466. doi: 10.1002/mp.12484. Epub 2017 Aug 22.
The aim of this study was to validate the accuracy of an exit detector-based dose reconstruction tool for helical tomotherapy (HT) delivery quality assurance (DQA).
Exit detector-based DQA tool was developed for patient-specific HT treatment verification. The tool performs a dose reconstruction on the planning image using the sinogram measured by the HT exit detector with no objects in the beam (i.e., static couch), and compares the reconstructed dose to the planned dose. Vendor supplied (three "TomoPhant") plans with a cylindrical solid water ("cheese") phantom were used for validation. Each "TomoPhant" plan was modified with intentional multileaf collimator leaf open time (MLC LOT) errors to assess the sensitivity and robustness of this tool. Four scenarios were tested; leaf 32 was "stuck open," leaf 42 was "stuck open," random leaf LOT was closed first by mean values of 2% and then 4%. A static couch DQA procedure was then run five times (once with the unmodified sinogram and four times with modified sinograms) for each of the three "TomoPhant" treatment plans. First, the original optimized delivery plan was compared with the original machine agnostic delivery plan, then the original optimized plans with a known modification applied (intentional MLC LOT error) were compared to the corresponding error plan exit detector measurements. An absolute dose comparison between calculated and ion chamber (A1SL, Standard Imaging, Inc., WI, USA) measured dose was performed for the unmodified "TomoPhant" plans. A 3D gamma evaluation (2%/2 mm global) was performed by comparing the planned dose ("original planned dose" for unmodified plans and "adjusted planned dose" for each intentional error) to exit detector-reconstructed dose for all three "Tomophant" plans. Finally, DQA for 119 clinical (treatment length <25 cm) and three cranio-spinal irradiation (CSI) plans were measured with both the ArcCHECK phantom (Sun Nuclear Corp., Melbourne, FL, USA) and the exit detector DQA tool to assess the time required for DQA and similarity between two methods.
The measured ion chamber dose agreed to within 1.5% of the reconstructed dose computed by the exit detector DQA tool on a cheese phantom for all unmodified "Tomophant" plans. Excellent agreement in gamma pass rate (>95%) was observed between the planned and reconstructed dose for all "Tomophant" plans considered using the tool. The gamma pass rate from 119 clinical plan DQA measurements was 94.9% ± 1.5% and 91.9% ± 4.37% for the exit detector DQA tool and ArcCHECK phantom measurements (P = 0.81), respectively. For the clinical plans (treatment length <25 cm), the average time required to perform DQA was 24.7 ± 3.5 and 39.5 ± 4.5 min using the exit detector QA tool and ArcCHECK phantom, respectively, whereas the average time required for the 3 CSI treatments was 35 ± 3.5 and 90 ± 5.2 min, respectively.
The exit detector tool has been demonstrated to be faster for performing the DQA with equivalent sensitivity for detecting MLC LOT errors relative to a conventional phantom-based QA method. In addition, comprehensive MLC performance evaluation and features of reconstructed dose provide additional insight into understanding DQA failures and the clinical relevance of DQA results.
本研究旨在验证基于出口探测器的剂量重建工具在螺旋断层放疗（HT）质量保证（DQA）中的准确性。
开发了基于出口探测器的 DQA 工具，用于患者特异性 HT 治疗验证。该工具使用 HT 出口探测器测量的射束中无物体（即静态治疗床）的正弦图，对计划图像进行剂量重建，并将重建剂量与计划剂量进行比较。使用带有圆柱形固体水（“奶酪”）模体的供应商提供（三个“TomoPhant”）计划进行验证。每个“TomoPhant”计划都通过有意修改多叶准直器叶片打开时间（MLC LOT）误差进行了修改，以评估该工具的灵敏度和稳健性。测试了四个场景；叶片 32 被“卡住打开”，叶片 42 被“卡住打开”，随机叶片 LOT 首先以平均值关闭 2%，然后关闭 4%。然后，对于三个“TomoPhant”治疗计划中的每一个，五次运行静态治疗床 DQA 程序（一次使用未经修改的射束，四次使用修改后的射束）。首先，将原始优化的输送计划与原始机器不可知的输送计划进行比较，然后将具有已知修改的原始优化计划（有意的 MLC LOT 误差）与相应的误差计划出口探测器测量结果进行比较。对于未经修改的“TomoPhant”计划，对计算剂量和离子室（A1SL，标准成像公司，威斯康星州，美国）测量的剂量进行了绝对剂量比较。对于所有三个“Tomophant”计划，通过比较计划剂量（未经修改计划的“原始计划剂量”和每个有意误差的“调整计划剂量”）与出口探测器重建剂量，进行了 3D 伽马评估（2%/2mm 全局）。最后，使用 ArcCHECK 体模（Sun Nuclear 公司，佛罗里达州墨尔本）和出口探测器 DQA 工具对 119 个临床（治疗长度<25cm）和三个颅脊柱照射（CSI）计划进行了 DQA 测量，以评估 DQA 所需的时间和两种方法之间的相似性。
对于所有未经修改的“Tomophant”计划，在奶酪模体上，出口探测器 DQA 工具计算的重建剂量与离子室测量的剂量相差 1.5%以内。对于所有考虑使用该工具的“Tomophant”计划，计划剂量与重建剂量之间观察到极好的伽马通过率（>95%）一致。从 119 个临床计划 DQA 测量中，出口探测器 DQA 工具和 ArcCHECK 体模测量的伽马通过率分别为 94.9%±1.5%和 91.9%±4.37%（P=0.81）。对于临床计划（治疗长度<25cm），使用出口探测器 QA 工具和 ArcCHECK 体模分别进行 DQA 的平均时间为 24.7±3.5 和 39.5±4.5 分钟，而 3 个 CSI 治疗的平均时间分别为 35±3.5 和 90±5.2 分钟。
与传统的基于体模的 QA 方法相比，出口探测器工具在执行 DQA 方面速度更快，并且对检测 MLC LOT 误差的灵敏度相当。此外，综合的 MLC 性能评估和重建剂量的特点提供了对 DQA 失败和 DQA 结果临床相关性的额外深入了解。
