Diffusion-Weighted Imaging-Based Adaptive Replanning for the MR-Linac

项目摘要 接受放射治疗(RT)的口腔和颅面部癌症患者，如人乳头瘤病毒- 人乳头状瘤病毒阳性口咽癌(HPV OPC)放射治疗的副作用较多 对健康组织的伤害。尽管RT对HPV OPC有很高的疗效，但辐射引起的后遗症可能会持续存在 几十年的生存，大大降低了患者的口腔健康和生活质量。对健康的毒性 组织可以通过自适应重新规划来减少，在该规划中，辐射束的几何形状被重新优化 在为期数周的放疗过程中定期进行，以考虑肿瘤缩小和正常组织变形。 适应性再计划目前在口腔和颅面肿瘤的临床上是可行的，因为最近发展了一种 核磁共振/直线加速器混合装置(MR-LINAC)。适应性治疗只使用了基本的解剖学核磁共振 用于监测肿瘤体积的脉冲序列。然而，我们提出了一种适应性治疗策略，使用 一种名为扩散加权成像(DWI)的功能磁共振技术，可以评估正常组织功能， 确定肿瘤内抗辐射的亚体积，并预测患者对RT的反应。这是一个假设 研究表明，来自DWI的功能信息可以被实施到自适应重新规划过程中 口腔和颅面部癌症在临床上是可行的(使用新的MR-LINAC设备)，并将减少 副作用。为了检验这一假设，我们将首先开发一个多元回归模型，该模型将 肿瘤和健康组织的ADC值对HPV OPC患者预后的影响。这一信息将是 整合到MR-LINAC基于DWI的适应性重新规划工作流程中(具体目标1)。接下来，酒后驾车- 基于自适应再规划的方法将在日常患者图像上进行回溯建模。一剂 将开发与MR-Linac基于MRI的剂量计算方法兼容的累积算法 并被用来测量危险器官的累积剂量。累积剂量将与正常剂量相关 组织并发症概率模型以确定这种方法是否降低了副作用的风险 (具体目标2)。这些特定目标的预期结果是开发自适应RT方法 使用来自DWI的函数数据。这一治疗方案的临床可行性和益处将是 通过计算机和统计建模进行了论证，使其最终可以应用于临床。这 Project将通过提供个性化的、有针对性的RT来对HPV OPC患者产生积极影响 在保留正常组织和减少副作用的同时治疗肿瘤。此外，这项工作将产生更广泛的影响 在口腔、牙齿和颅面健康领域，通过引入一种新的治疗范例，直接 在不影响肿瘤控制的情况下监控并对正常组织损伤做出反应。

Project Summary Patients undergoing radiation therapy (RT) for oral and craniofacial cancers such as human papillomavirus- positive oropharyngeal cancer (HPV+ OPC) experience a host of side effects caused by radiation-induced injury to healthy tissues. Although RT is highly curative for HPV+ OPC, radiation-induced sequelae can persist for decades of survivorship, significantly degrading a patient's oral health and quality of life. Toxicity to healthy tissues can be reduced by adaptive replanning, in which the geometry of the radiation beams is re-optimized periodically during a multi-week course of RT to account for tumor shrinkage and normal tissue deformation. Adaptive replanning is now clinically feasible for oral and craniofacial cancers due the recent development of a hybrid MRI/linear accelerator device (MR-Linac). Adaptive treatments have used only basic anatomical MRI pulse sequences to monitor the tumor volume. However, we propose an adaptive treatment strategy that uses a functional MRI technique called diffusion-weighted imaging (DWI), which can assess normal tissue function, identify radioresistant sub-volumes within tumors, and predict patient response to RT. The hypothesis of this study is that the functional information from DWI can be implemented into the adaptive replanning process for oral and craniofacial cancers such that it is clinically feasible (with the new MR-Linac device) and will reduce side effects. To test this hypothesis, we will first develop a multivariate regression model relating changes in ADC values of the tumor and healthy tissues to HPV+ OPC patient outcomes. This information will be integrated into a DWI-based adaptive replanning workflow for the MR-Linac (Specific Aim 1). Next, the DWI- based adaptive replanning approach will be modeled retrospectively on daily patient images. A dose accumulation algorithm compatible with the MR-Linac's MRI-based dose calculation method will be developed and employed to measure cumulative doses to organs at risk. Cumulative doses will be related to normal tissue complication probability models to determine whether this approach lowers the risk of side effects (Specific Aim 2). The expected outcome of these specific aims is the development of an adaptive RT approach that uses functional data from DWI. The clinical feasibility and benefit of this treatment scheme will be demonstrated through in silico and statistical modeling so that it may eventually be used in the clinic. This project will positively impact patients with HPV+ OPC by enabling the delivery of personalized, targeted RT to the tumor while sparing normal tissues and reducing side effects. Further, this work will have a broader impact on the field of oral, dental, and craniofacial health by introducing a novel treatment paradigm that directly monitors and reacts to normal tissue injury without compromising tumor control.