腹部仿真变形体模的制作及其在变形配准验证中的应用
发布时间：2019-01-18 16:35
【摘要】：多年来,变形图像配准(Deformation Image Registration,DIR)方法在自适应放射治疗(Adaptive Radiation Therapy,ART)中的应用发展已经进行了很多努力。然而,我们需要回答一个更为紧迫的问题就是如何评估在ART背景下的这些DIR方法的不确定性和验证其准确性。本研究希望制作一个腹部仿真变形体模,和提供一种可定量评价DIR算法几何精度和剂量叠加精度的腹部变形体模制作方法,该体模可对人体器官变形造成的解剖结构的复杂变化进行模拟。本研究提出的体模是一个腹部仿真变形体模,体模包含的模拟器官有软组织、肝脏、肾脏、脾脏、胃、椎骨和两个转移肿瘤。所有的器官模具都使用来自一个卵巢癌病人的轮廓线进行3D打印成阴模所得到。器官模具塑造了由PVC树脂粉和对苯二甲酸二辛酯增塑剂不同混合比例的混合物制成的可变形材料。不同密度的软聚氯乙烯(Polyvinylchlorid,PVC)可以通过多项式拟合曲线得到,该多项式描述了 CT值与PVC树脂粉-增塑剂混合比例之间的关系。刚性的椎骨通过模具塑形白水泥和纤维素纸浆的混合物制成。将所有模拟器官按照它们的解剖位置放置在一个中空的假人容器里,用模拟肌肉和脂肪平均CT值的变形材料进行密封,最后得到一个腹部体模。195块1mm3的橡胶立方体作为验证DIR几何精度和剂量叠加精度的标记点均匀地嵌入到体模里。为了促进剂量测定验证,在平行体模上下方向的位置上挖掘两通道用于电离室插入。使用来自一个开源DIR工具包(DIRART)的三种DIR算法测试了该体模用于基于器官标记点的DIR精度验证和体内剂量测定验证的可行性。制作的变形材料展示了很好的弹性特性和CT值随时间推移是稳定的,证明这是变形体模的一种满意材料。该变形体模的内部模拟器官具有与真实人体器官相似的CT值,三维解剖结构和空间位置关系,体模的CT图像跟真实病人的CT图像高度相似且对比度明显,在图像上可清晰区分各器官的轮廓和标记点的位置。测量的剂量跟治疗计划系统(treatment planning system,TPS)上计算的剂量显示了一个很好的一致性。应用该体模进行基于标记点的精度分析验证了使用体模评估器官方面的DIR精度的可行性。在这项工作中,我们设计和制作了一个3D腹部仿真变形体模,描述了关于体模制作的详细步骤,包括变形材料配方、器官模塑、标记点嵌入等等,测试了体模的性能,例如变形材料的弹性和CT值随时间推移的稳定性,还证明这体模验证DIR精度的可行性。正如实验结果显示,体模制作过程简单,体模材料便宜和不需要专门工具制作。任何机构可以根据这些步骤制作一个相同或者相似的体模,作为DIR质量保证的一个常规工具。
[Abstract]:Over the years, many efforts have been made in the application of deformable image registration (Deformation Image Registration,DIR) in adaptive radiation therapy (Adaptive Radiation Therapy,ART). However, we need to answer a more urgent question: how to evaluate the uncertainty and verify the accuracy of these DIR methods in the context of ART. The purpose of this study is to make an abdominal simulated deformable model and to provide a method to quantitatively evaluate the geometric accuracy and dose superposition accuracy of the DIR algorithm. The phantom can simulate the complex changes of anatomical structure caused by the deformation of human organs. The phantom proposed in this study is an abdominal phantom containing soft tissue, liver, kidney, spleen, stomach, vertebrae and two metastatic tumors. All organ molds are obtained by 3D printing negative models using contour lines from a ovarian cancer patient. A deformable material was molded by a mixture of PVC resin powder and dioctyl terephthalate plasticizer. Soft polyvinyl chloride (Polyvinylchlorid,PVC) with different densities can be obtained by polynomial fitting curve. The polynomial describes the relationship between CT value and the ratio of PVC resin powder to plasticizer. Rigid vertebrae are made from a mixture of molded white cement and cellulose pulp. All the simulated organs were placed in a hollow dummy container in accordance with their anatomical position and sealed with a deformable material that mimicked the average CT value of muscle and fat. Finally, an abdominal phantom and a rubber cube of 1mm3 are obtained as marks to verify the geometric accuracy of DIR and the accuracy of dose superposition, which are uniformly embedded into the phantom. In order to facilitate dosimetry verification, two channels were excavated in the position parallel to the upper and lower directions of the phantom for the insertion of the ionization chamber. Three DIR algorithms from an open source DIR toolkit (DIRART) were used to test the feasibility of using the phantom to verify the accuracy of DIR based on organ marker points and in vivo dosimetry. The deformable materials show good elastic properties and the CT value is stable over time. It is proved that the deformable material is a satisfactory material for the deformable model. The internal simulated organ of the model has the CT value similar to that of the real human organ, the three-dimensional anatomical structure and the spatial position relation. The CT image of the phantom is highly similar to the CT image of the real patient and the contrast is obvious. The contour of each organ and the location of the marking point can be clearly distinguished on the image. The measured dose shows a good consistency with the dose calculated on the treatment planning system (treatment planning system,TPS. The feasibility of using the phantom to evaluate the DIR accuracy of organs is verified by using the mark-based accuracy analysis. In this work, we designed and made a 3D abdominal simulation deformable model, and described the detailed steps for the fabrication of the phantom, including the formulation of the deformable material, the molding of the organ, the embedding of the marking points, and so on, and tested the performance of the phantom. For example, the elasticity of deformed materials and the stability of CT values over time, and the feasibility of verifying the accuracy of DIR by this model is also proved. As the experimental results show, the fabrication process is simple, the material is cheap and no special tools are needed. Any organization can make an identical or similar phantom based on these steps as a general tool for DIR quality assurance.
【学位授予单位】：南方医科大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;R815
本文编号：2410899
[Abstract]:Over the years, many efforts have been made in the application of deformable image registration (Deformation Image Registration,DIR) in adaptive radiation therapy (Adaptive Radiation Therapy,ART). However, we need to answer a more urgent question: how to evaluate the uncertainty and verify the accuracy of these DIR methods in the context of ART. The purpose of this study is to make an abdominal simulated deformable model and to provide a method to quantitatively evaluate the geometric accuracy and dose superposition accuracy of the DIR algorithm. The phantom can simulate the complex changes of anatomical structure caused by the deformation of human organs. The phantom proposed in this study is an abdominal phantom containing soft tissue, liver, kidney, spleen, stomach, vertebrae and two metastatic tumors. All organ molds are obtained by 3D printing negative models using contour lines from a ovarian cancer patient. A deformable material was molded by a mixture of PVC resin powder and dioctyl terephthalate plasticizer. Soft polyvinyl chloride (Polyvinylchlorid,PVC) with different densities can be obtained by polynomial fitting curve. The polynomial describes the relationship between CT value and the ratio of PVC resin powder to plasticizer. Rigid vertebrae are made from a mixture of molded white cement and cellulose pulp. All the simulated organs were placed in a hollow dummy container in accordance with their anatomical position and sealed with a deformable material that mimicked the average CT value of muscle and fat. Finally, an abdominal phantom and a rubber cube of 1mm3 are obtained as marks to verify the geometric accuracy of DIR and the accuracy of dose superposition, which are uniformly embedded into the phantom. In order to facilitate dosimetry verification, two channels were excavated in the position parallel to the upper and lower directions of the phantom for the insertion of the ionization chamber. Three DIR algorithms from an open source DIR toolkit (DIRART) were used to test the feasibility of using the phantom to verify the accuracy of DIR based on organ marker points and in vivo dosimetry. The deformable materials show good elastic properties and the CT value is stable over time. It is proved that the deformable material is a satisfactory material for the deformable model. The internal simulated organ of the model has the CT value similar to that of the real human organ, the three-dimensional anatomical structure and the spatial position relation. The CT image of the phantom is highly similar to the CT image of the real patient and the contrast is obvious. The contour of each organ and the location of the marking point can be clearly distinguished on the image. The measured dose shows a good consistency with the dose calculated on the treatment planning system (treatment planning system,TPS. The feasibility of using the phantom to evaluate the DIR accuracy of organs is verified by using the mark-based accuracy analysis. In this work, we designed and made a 3D abdominal simulation deformable model, and described the detailed steps for the fabrication of the phantom, including the formulation of the deformable material, the molding of the organ, the embedding of the marking points, and so on, and tested the performance of the phantom. For example, the elasticity of deformed materials and the stability of CT values over time, and the feasibility of verifying the accuracy of DIR by this model is also proved. As the experimental results show, the fabrication process is simple, the material is cheap and no special tools are needed. Any organization can make an identical or similar phantom based on these steps as a general tool for DIR quality assurance.
【学位授予单位】：南方医科大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;R815
【参考文献】
相关期刊论文 前5条
1 王锐濠;张书旭;林生趣;;肿瘤精确放疗新进展——图像引导自适应放疗[J];中国医学物理学杂志;2012年04期
2 甄鑫;周凌宏;卢文婷;张书旭;周露;;改进Demons算法的验证及其在4D-CT轮廓线推衍中的应用[J];南方医科大学学报;2010年12期
3 周付根;段卓镭;;Demons算法在四维CT图像配准中的应用[J];CT理论与应用研究;2009年01期
4 刘咏梅;李平;;图像引导的自适应放疗研究进展[J];中国肿瘤临床;2008年22期
5 戴建荣;胡逸民;;图像引导放疗的实现方式[J];中华放射肿瘤学杂志;2006年02期
相关博士学位论文 前1条
1 甄鑫;自适应放射治疗中的图像变形配准关键技术研究[D];南方医科大学;2013年
相关硕士学位论文 前1条
1 王建华;结合线性二次放射生物模型的图像配准技术的研究[D];清华大学;2010年
，本文编号：2410899
本文链接：https://www.wllwen.com/yixuelunwen/yundongyixue/2410899.html
最近更新
教材专著
