Paddick Ian, Mott Judith, Bedford James, Filatov Petr, Grishchuk Diana, Orchin Gavin, Houston Peter, Eaton David J
Queen Square Radiosurgery Centre, National Hospital for Neurology and Neurosurgery, London, United Kingdom.
Northern Centre for Cancer Care, Department of Radiotherapy Physics, Newcastle upon Tyne Hospital NHS Foundation Trust, Newcastle upon Tyne, United Kingdom.
Pract Radiat Oncol. 2023 Sep-Oct;13(5):e451-e459. doi: 10.1016/j.prro.2023.05.005. Epub 2023 Jun 7.
Stereotactic radiosurgery treatment delivery can be performed with a range of devices, each of which have evolved over recent years. We sought to evaluate the differences in performance of contemporary stereotactic radiosurgery platforms and also to compare them with earlier platform iterations from a previous benchmarking study.
The following platforms were selected as "state of the art" in 2022: Gamma Knife Icon (GK), CyberKnife S7 (CK), Brainlab Elements (Elekta VersaHD and Varian TrueBeam), Varian Edge with HyperArc (HA), and Zap-X. Six benchmarking cases were used from a 2016 study. To reflect the evolution of increasing numbers of metastases treated per patient, a 14-target case was added. The 28 targets among the 7 patients ranged from 0.02 to 7.2 cc in volume. Participating centers were sent images and contours for each patient and asked to plan them to the best of their ability. Although some variation in local practice was allowed (eg, margins), groups were asked to prescribe a specified dose to each target and tolerance doses to organs at risk were agreed upon. Parameters compared included coverage, selectivity, Paddick conformity index, gradient index (GI), R50%, efficiency index, doses to organs at risk, and planning and treatment times.
Mean coverage for all targets ranged from 98.2% (Brainlab/Elekta) to 99.7% (HA-6X). Paddick conformity index values ranged from 0.722 (Zap-X) to 0.894 (CK). GI ranged from a mean of 3.52 (GK), representing the steepest dose gradient, to 5.08 (HA-10X). The GI appeared to follow a trend with beam energy, with the lowest values from the lower energy platforms (GK, 1.25 MeV; Zap-X, 3 MV) and the highest value from the highest energy (HA-10X). Mean R50% values ranged from 4.48 (GK) to 5.98 (HA-10X). Treatment times were lowest for C-arm linear accelerators.
Compared with earlier studies, newer equipment appears to deliver higher quality treatments. CyberKnife and linear accelerator platforms appear to give higher conformity whereas lower energy platforms yield a steeper dose gradient.
立体定向放射外科治疗可以通过一系列设备进行，近年来每种设备都有所发展。我们试图评估当代立体定向放射外科平台的性能差异，并将它们与之前一项基准研究中早期平台的迭代版本进行比较。
以下平台被选为2022年的“先进技术”：伽玛刀Icon（GK）、射波刀S7（CK）、Brainlab Elements（医科达VersaHD和瓦里安TrueBeam）、配备HyperArc（HA）的瓦里安Edge以及Zap-X。使用了2016年一项研究中的6个基准病例。为反映每位患者治疗转移灶数量增加的趋势，增加了一个14靶点病例。7名患者中的28个靶点体积从0.02到7.2立方厘米不等。向参与的中心发送了每位患者的图像和轮廓，并要求他们尽最大能力进行计划。虽然允许局部操作存在一些差异（例如，边界），但要求各小组为每个靶点规定特定剂量，并就危及器官的耐受剂量达成一致。比较的参数包括覆盖率、选择性、帕迪克适形指数、梯度指数（GI）、R50%、效率指数、危及器官的剂量以及计划和治疗时间。
所有靶点的平均覆盖率从98.2%（Brainlab/医科达）到99.7%（HA - 6X）不等。帕迪克适形指数值从0.722（Zap-X）到0.894（CK）。GI的平均值范围从3.52（GK）（代表最陡的剂量梯度）到5.08（HA - 10X）。GI似乎随束流能量呈现一种趋势，能量较低的平台（GK，1.25 MeV；Zap-X，3 MV）的值最低，能量最高的平台（HA - 10X）的值最高。平均R50%值从4.48（GK）到5.98（HA - 10X）不等。C型臂直线加速器的治疗时间最短。
与早期研究相比，更新的设备似乎能提供更高质量的治疗。射波刀和直线加速器平台似乎具有更高的适形性而能量较低的平台产生更陡的剂量梯度。
