Abstract：
OBJECTIVE: To demonstrate a framework for calculating daily dose distributions for proton therapy in a timeframe amenable to online evaluation using CT-on-Rails.
METHODS: Tasks associated with calculation of daily dose are fully automated. A rigid registration between daily and planning images is used to propagate beams and targets for calculation of daily dose; additionally, risk structures are propagated using deformable registration to facilitate online evaluation. An end-to-end constancy test was carried out using a pelvis phantom containing a simulated target and bladder contour. 97 Daily fan-beam CT data sets associated with 10 clinical patients were processed to demonstrate feasibility and utility of online evaluation. Computing times and dosimetric differences are reported.
RESULTS: The phantom constancy test took 62 s to complete with no notable discrepancies in the registrations or calculated dose. Max doses were identical for target and bladder contours on initial and repeat scans (359 and 310 cGy (RBE) respectively). Total processing time for 97 daily patient images averaged 154.6 s (73.0 - 222.0 s; SD = 31.8 s). On average, dose calculation accounted for 35 % of total processing time. Average differences in D95 for target contours was 1.5 % (SD = 1.6 %) with a max decrease of 5.9 % on a single daily image.
CONCLUSIONS: Daily dose can be automatically calculated in a timeframe amenable to online evaluation using scanner utilities in conjunction with the scripting API of a commercial treatment planning system. Online evaluation of dose in proton therapy is useful to detect clinically relevant changes, guide setup, and facilitate treatment or replanning decisions.
METHODS: Tasks associated with calculation of daily dose are fully automated. A rigid registration between daily and planning images is used to propagate beams and targets for calculation of daily dose; additionally, risk structures are propagated using deformable registration to facilitate online evaluation. An end-to-end constancy test was carried out using a pelvis phantom containing a simulated target and bladder contour. 97 Daily fan-beam CT data sets associated with 10 clinical patients were processed to demonstrate feasibility and utility of online evaluation. Computing times and dosimetric differences are reported.
RESULTS: The phantom constancy test took 62 s to complete with no notable discrepancies in the registrations or calculated dose. Max doses were identical for target and bladder contours on initial and repeat scans (359 and 310 cGy (RBE) respectively). Total processing time for 97 daily patient images averaged 154.6 s (73.0 - 222.0 s; SD = 31.8 s). On average, dose calculation accounted for 35 % of total processing time. Average differences in D95 for target contours was 1.5 % (SD = 1.6 %) with a max decrease of 5.9 % on a single daily image.
CONCLUSIONS: Daily dose can be automatically calculated in a timeframe amenable to online evaluation using scanner utilities in conjunction with the scripting API of a commercial treatment planning system. Online evaluation of dose in proton therapy is useful to detect clinically relevant changes, guide setup, and facilitate treatment or replanning decisions.
摘要：
目的：展示了一个框架，用于计算质子治疗的日剂量分布，该框架适用于使用RailsCT进行在线评估的时间范围。
方法：与日剂量计算相关的任务是完全自动化的。每日和计划图像之间的刚性配准用于传播光束和目标以计算每日剂量；此外，使用可变形配准来传播风险结构，以促进在线评估。使用包含模拟目标和膀胱轮廓的骨盆体模进行端到端恒定测试。处理了与10名临床患者相关的97个每日扇形束CT数据集，以证明在线评估的可行性和实用性。报告计算时间和剂量测定差异。
结果：体模恒定性测试需要62秒才能完成，注册或计算剂量没有明显差异。在初始和重复扫描中，目标和膀胱轮廓的最大剂量相同（分别为359和310cGy（RBE））。每天97张患者图像的总处理时间平均为154.6s（73.0-222.0s；SD=31.8s）。平均而言,剂量计算占总处理时间的35%。目标轮廓的D95平均差异为1.5%(SD=1.6%),在单个每日图像上最大减少5.9%。
结论：每日剂量可以在一个时间范围内自动计算，可使用扫描仪实用程序结合商业治疗计划系统的脚本API进行在线评估。质子治疗中剂量的在线评估有助于检测临床相关变化，指南设置,并促进治疗或重新规划决策。
方法：与日剂量计算相关的任务是完全自动化的。每日和计划图像之间的刚性配准用于传播光束和目标以计算每日剂量；此外，使用可变形配准来传播风险结构，以促进在线评估。使用包含模拟目标和膀胱轮廓的骨盆体模进行端到端恒定测试。处理了与10名临床患者相关的97个每日扇形束CT数据集，以证明在线评估的可行性和实用性。报告计算时间和剂量测定差异。
结果：体模恒定性测试需要62秒才能完成，注册或计算剂量没有明显差异。在初始和重复扫描中，目标和膀胱轮廓的最大剂量相同（分别为359和310cGy（RBE））。每天97张患者图像的总处理时间平均为154.6s（73.0-222.0s；SD=31.8s）。平均而言,剂量计算占总处理时间的35%。目标轮廓的D95平均差异为1.5%(SD=1.6%),在单个每日图像上最大减少5.9%。
结论：每日剂量可以在一个时间范围内自动计算，可使用扫描仪实用程序结合商业治疗计划系统的脚本API进行在线评估。质子治疗中剂量的在线评估有助于检测临床相关变化，指南设置,并促进治疗或重新规划决策。
