Feedback Control and Inferential Modeling for Radiotherapy Treatment Couch

描述（由申请人提供）：呼吸诱导的肿瘤运动导致计算机断层扫描图像上的伪影，并导致放射治疗治疗计划和输送的不准确性。传统的解决方案包括在肿瘤周围增加“安全裕度”，这通常会导致正常组织的不必要的照射，从而增加正常组织的毒性。目前市售的线性加速器方法包括在放射输送期间屏气和门控。前者不能被患者很好地耐受，而后者增加了治疗递送时间，从而增加了伪患者运动的风险，甚至降低了递送的辐射的生物有效性。本建议的目标是开发一个使用反馈控制的真实的时间运动同步治疗床。本提案的具体目的是：1）设计和测试用于RT输送的原型呼吸诱导运动同步治疗床的可行性; 2）使用皮肤标记开发推理模型，该模型将使用偏最小二乘（PLS）模型将皮肤标记的位置与肿瘤运动相关联。为了开发这种治疗床，将对治疗床动力学和控制系统进行建模，并在患者特定的肿瘤运动轨迹上进行测试。一旦原型控制器和治疗床的开发完成，将进行完整的系统测试。将开发使用偏最小二乘回归的推断模型，以使用放置在患者皮肤上的运动替代物来预测肿瘤的位置。将使用多个治疗日采集的呼吸数据测试其稳健性。我们提案的结果有可能对肺癌和上腹部癌症的放射治疗产生重大影响。拟议的探索性研究工作将提供一个实时的运动管理框架，使用实时运动同步治疗床的进一步发展的基础。其在临床实践中的最终实施将通过使用更小的肿瘤边缘和准确的肿瘤跟踪来降低正常组织的毒性。这将使得能够以可靠的方式更广泛地使用高度适形的治疗，例如调强放射治疗（IMRT）和立体定向体放射治疗（SBRT）（其对肿瘤运动高度不宽容）。

DESCRIPTION (provided by applicant): Respiration induced tumor motion causes both artifacts on computed tomography images and leads to inaccuracies in the planning and delivery of radiation therapy treatments. Conventional solutions included adding a "safety margin" around the tumor which can often result in unnecessary irradiation of normal tissue thereby increasing normal tissue toxicity. Current commercially available methods with linear accelerators include breath-holds and gating during radiation delivery. The former is not tolerated well by patients while the later increases treatment delivery time, thereby increasing the risk of spurious patient motion and even reducing the biological effectiveness of the delivered radiation. The goal of this proposal is to develop a real- time motion-synchronized treatment couch using feedback control. The specific aims of this proposal are to 1) design and test the feasibility of a prototype respiration induced motion-synchronized treatment couch for RT delivery and 2) develop an inferential model using skin markers that will correlate their position with tumor motion using partial-least squares (PLS) models. In order to develop this treatment couch, the couch dynamics and control system will be modeled and tested on patient-specific tumor motion trajectories. Once the development of the prototype controller and couch has been completed, a full system test will be performed. Inferential models using partial least squares regression will be developed to predict the position of the tumor using motion surrogates placed on the skin of the patient. Its robustness will be tested with respiration data acquired on multiple treatment days. The results from our proposal have the potential to significantly impact radiation therapy of lung and upper abdominal cancers. The proposed exploratory research effort will provide the foundation for further development of a real-time motion management framework using a real-time motion-synchronized treatment couch. Its eventual implementation in clinical practice will lead to reduced normal tissue toxicity through the use of smaller tumor margins and accurate tumor tracking. This will enable more widespread use of highly conformal treatments such as intensity- modulated radiation therapy (IMRT) and stereotactic body radiation therapy (SBRT) (which are highly unforgiving of tumor motion) in a reliable manner.