CT Image Estimation for Calculation of Delivered Dose

描述（由申请人提供）：前列腺是一个可移动的器官，其位置和形状的瞬间和日常变化严重影响放射治疗的准确传递。因此，在现代技术要求很高的前列腺癌放射治疗中，标准做法是在大多数治疗中定位前列腺，以帮助实现向前列腺提供高剂量同时保护附近放射敏感的正常组织的目标。第一阶段的研究确立了一种新方法的可行性，该方法利用一类统计上可训练的可变形形状模型的独特数学特性，通过将参考CT图像数据映射到基于植入前列腺的标记物的测量位置的治疗空间来估计治疗图像数据。估计的图像捕获了治疗期间前列腺的姿势、位置和形状，适合于准确计算给前列腺和周围组织的剂量。这项II期研究的前两年将专注于基于这项创新技术的临床原型系统的设计和开发。该原型将在第三年的临床环境中进行评估。所设想的商业形式将接受来自多个设备的输入数据，包括治疗计划系统、治疗机和标记物跟踪系统，并将输出数据通信到治疗计划系统。在几秒钟内，所提出的原型将使用输入数据自动计算估计的治疗图像并计算交付的剂量。拟议的系统将利用目前正在进行的II期项目（R44 CA119571）所完成的工作，该项目将允许原型机与治疗内成像设备以及标记跟踪器一起使用。此外，所提出的方法可以实现一种称为同步动态剂量重建的剂量计算形式，这是最精确的ART形式所需要的。

DESCRIPTION (provided by applicant): The prostate is a mobile organ that demonstrates momentary and daily changes in position and shape that critically affect accurate delivery of radiation therapy. For this reason, in modern technically demanding radiation therapy for prostate cancer, it is standard practice to localize the prostate on most days treatment as an aid to achieve the goal of delivering a high dose to the prostate while protecting nearby radiosensitive normal tissues. The Phase I study established the feasibility of a novel approach that uses unique mathematical properties of a class of statistically trainable deformable shape models to estimate treatment image data by mapping reference CT image data into the treatment space based on measured positions of markers implanted in the prostate. The estimated images capture the pose, position and shape of the prostate during treatment and are suitable for accurate calculation of dose delivered to the prostate and immediately surrounding tissues. The first two years of this Phase II study will focus on design and development of a clinical prototype system based on this creative technology. The prototype will be evaluated in the clinical setting during the third year. The envisioned commercial form will accept input data from multiple devices including the treatment planning system, treatment machine, and marker tracking system, and will communicate output data to the treatment planning system. Within a few seconds the proposed prototype will use input data to automatically compute the estimated treatment image and calculate delivered dose. The proposed system will leverage work accomplished under a currently active Phase II project (R44 CA119571) that will allow the prototype to be used with intra-treatment imaging devices as well as marker-trackers. Moreover the proposed approach may enable a form of dose computation, called synchronized dynamic dose reconstruction, needed for the most accurate form of ART. PUBLIC HEALTH RELEVANCE: The prostate is a mobile organ that demonstrates momentary and daily changes in position and shape that critically affect accurate delivery of radiation therapy. For this reason, in modern technically demanding radiation therapy for prostate cancer, it is standard practice to localize the prostate on most days treatment as an aid to achieve the goal of delivering a high dose to the prostate while protecting nearby radiosensitive normal tissues. The Phase I study established the feasibility of a novel approach that uses unique mathematical properties of a class of statistically trainable deformable shape models to estimate treatment image data by mapping reference CT image data into the treatment space based on measured positions of markers implanted in the prostate. The estimated images capture the pose, position and shape of the prostate during treatment and are suitable for accurate calculation of dose delivered to the prostate and immediately surrounding tissues. The first two years of this Phase II study will focus on design and development of a clinical prototype system based on this creative technology. The prototype will be evaluated in the clinical setting during the third year. The envisioned commercial form will accept input data from multiple devices including the treatment planning system, treatment machine, and marker tracking system, and will communicate output data to the treatment planning system. Within a few seconds the proposed prototype will use input data to automatically compute the estimated treatment image and calculate delivered dose. The proposed system will leverage work accomplished under a currently active Phase II project (R44 CA119571) that will allow the prototype to be used with intra-treatment imaging devices as well as marker-trackers. Moreover the proposed approach may enable a form of dose computation, called synchronized dynamic dose reconstruction, needed for the most accurate form of ART. This project will contribute to public health in two ways: 1) It will allow an intra-treatment-image-guided form of treatment planning called Adaptive Radiation Planning that assures a close match between the planned and actual delivered doses; and 2) It will provide a needed but currently missing quality assurance step to confirm the safety of the delivered dose for each treatment session. This contribution is especially significant in light of a series of recent mistreatments that have led many to believe that new planning and delivery technologies have outpaced the development of clinically practical methods to independently validate safe planning and delivery.