VMAT QA: Measurement-guided 4D dose reconstruction on a patient
VMAT QA: Measurement-guided 4D dose reconstruction on a patient
复制标题
DOI：
10.1118/1.4729709
复制
发表时间：
2012-07-01
期刊：
影响因子：
3.8
通讯作者：
Feygelman, Vladimir
中科院分区：
文献类型：
--
作者：
Nelms, Benjamin E.;Opp, Daniel;Feygelman, Vladimir
目的：开发并验证容积调制弧形治疗（VMAT）质量保证（QA）工具，该工具将来自商用螺旋二极管阵列的时间分辨、低密度（类似于10 mm）圆柱形表面剂量图作为输入，并在任意患者数据集上输出高密度、容积、时间分辨剂量矩阵。第一次验证研究仅限于同质“患者”。“方法：将VMAT治疗输送到二极管阵列体模（ARcidae，Sun Nuclear Corp.，墨尔本，佛罗里达州）。3DVH软件（Sun Nuclear）使用测量引导剂量重建（MGDR）推导高密度体积剂量。MGDR圆柱体模的结果，然后用于扰动的三维（3D）治疗计划剂量的患者数据集，产生一个半经验的体积剂量网格。通过变形单独的子射束剂量而不是复合物，也可以对患者进行四维（4D）剂量重建。对于常规（3D）剂量比较，开发了两种方法，使用AAPM TG-119报告中的四种计划（多靶、C形、模拟前列腺和头颈部），包括其结构和目标。首先，将3DVH和治疗计划系统（TPS）累积点剂量与立方体水等效体模（“患者”）中的离子室进行比较。体模的形状与ARceptor不同，此外，靶点放置不对称。其次，将立方体中冠状和矢状绝对剂量分布与3DVH和TPS进行比较。对于时间分辨（4D）比较，进行了三次检验。首先，在输送结束时计算3D MGDR和累积时间分辨患者（4D MGDR）剂量之间的体积剂量差异，理想情况下它们应该相同。其次，将时间分辨（10 Hz采样率）离子室剂量与4D MGDR的累积点剂量与时间曲线进行比较。最后，加速器的输出是不同的，以评估的线性的4D MGDR与全球通量changes.Results：在四个TG-119 plans，平均PTV点剂量差异之间的立方体3DVH和离子室是0.1 +/-1.0%。对于1%/2 mm、2%/2 mm和3%/3 mm阈值组合，平均胶片与TPS伽马分析通过率分别为83.0%、91.1%和98.4%，而对于相同的相应标准，平均胶片与3DVH伽马分析通过率分别为88.6%、96.1%和99.5%。4D MGDR也足够准确。首先，对于每种情况下的99.5%体素，在递送结束时来自3D和4D MGDR的剂量在0.5%局部剂量误差/1 mm距离内一致。此外，所有失败的体素被限制到圆柱形重建体积的边缘。第二，在离子室和4D MGDR之间的剂量与时间曲线在1%内跟踪。最后，4D MGDR剂量随加速器输出线性变化：累积离子室和MGDR剂量之间的差异不超过1%（随机），输出变化范围为10%。“该方法不需要比标准QA更多的加速器时间，同时产生更多的临床相关信息。在一个异质性的胸部体模的验证正在进行中，作为最终的应用程序的4D MGDR虚拟运动的研究。(C)2012年美国医学物理学家协会。[http：//dx.doi.org/10.1118/1.4729709]
Purpose: To develop and validate a volume-modulated arc therapy (VMAT) quality assurance (QA) tool that takes as input a time-resolved, low-density (similar to 10 mm) cylindrical surface dose map from a commercial helical diode array, and outputs a high density, volumetric, time-resolved dose matrix on an arbitrary patient dataset. This first validation study is limited to a homogeneous "patient."Methods: A VMAT treatment is delivered to a diode array phantom (ARcCHECK, Sun Nuclear Corp., Melbourne, FL). 3DVH software (Sun Nuclear) derives the high-density volumetric dose using measurement-guided dose reconstruction (MGDR). MGDR cylindrical phantom results are then used to perturb the three-dimensional (3D) treatment planning dose on the patient dataset, producing a semiempirical volumetric dose grid. Four-dimensional (4D) dose reconstruction on the patient is also possible by morphing individual sub-beam doses instead of the composite. For conventional (3D) dose comparison two methods were developed, using the four plans (Multi-Target, C-shape, Mock Prostate, and Head and Neck), including their structures and objectives, from the AAPM TG-119 report. First, 3DVH and treatment planning system (TPS) cumulative point doses were compared to ion chamber in a cube water-equivalent phantom ("patient"). The shape of the phantom is different from the ARcCHECK and furthermore the targets were placed asymmetrically. Second, coronal and sagittal absolute film dose distributions in the cube were compared with 3DVH and TPS. For time-resolved (4D) comparisons, three tests were performed. First, volumetric dose differences were calculated between the 3D MGDR and cumulative time-resolved patient (4D MGDR) dose at the end of delivery, where they ideally should be identical. Second, time-resolved (10 Hz sampling rate) ion chamber doses were compared to cumulative point dose vs time curves from 4D MGDR. Finally, accelerator output was varied to assess the linearity of the 4D MGDR with global fluence change.Results: Across four TG-119 plans, the average PTV point dose difference in the cube between 3DVH and ion chamber is 0.1 +/- 1.0%. Average film vs TPS gamma-analysis passing rates are 83.0%, 91.1%, and 98.4% for 1%/2 mm, 2%/2 mm, and 3%/3 mm threshold combinations, respectively, while average film vs 3DVH gamma-analysis passing rates are 88.6%, 96.1%, and 99.5% for the same respective criteria. 4D MGDR was also sufficiently accurate. First, for 99.5% voxels in each case, the doses from 3D and 4D MGDR at the end of delivery agree within 0.5% local dose-error/l mm distance. Moreover, all failing voxels are confined to the edge of the cylindrical reconstruction volume. Second, dose vs time curves track between the ion chamber and 4D MGDR within 1%. Finally, 4D MGDR dose changes linearly with the accelerator output: the difference between cumulative ion chamber and MGDR dose changed by no more than 1% (randomly) with the output variation range of 10%.Conclusions: Even for a well-commissioned TPS, comparison metrics show better agreement on average to MGDR than to TPS on the arbitrary-shaped measurable "patient." The method requires no more accelerator time than standard QA, while producing more clinically relevant information. Validation in a heterogeneous thoracic phantom is under way, as is the ultimate application of 4D MGDR to virtual motion studies. (C) 2012 American Association of Physicists in Medicine. [http://dx.doi.org/10.1118/1.4729709]