Real-time Volumetric Imaging for Motion Management and Dose Delivery Verification

用于运动管理和剂量输送验证的实时体积成像

• 批准号: 10659842
• 负责人: Zhen Tian
• 金额: $ 52.64万
• 依托单位: EMORY UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-04-24 至 2028-03-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10659842
• 关键词: 3-Dimensional; Abdomen; Address; Adopted; Algorithms; Anatomy; Area; Articular Range of Motion; Artificial Intelligence; Biometry; Clinical; Clinical Data; Collimator; Compensation; Data; Development; Dose; Ensure; Geometry; Image; Implant; Inferior; Inhalation; Interdisciplinary Study; Linear Accelerator Radiotherapy Systems; Liver; Location; Lung; Malignant neoplasm of lung; Masks; Medical; Methods; Modernization; Monitor; Monte Carlo Method; Motion; Normal tissue morphology; Organ; Pancreas; Patients; Performance; Photons; Physicians; Physics; Plant Leaves; Radiation; Radiation Dose Unit; Radiation Oncology; Radiation therapy; Reproducibility; Respiration; Risk; Site; Supervision; System; Techniques; Technology; Testing; Three-Dimensional Image; Time; Tissues; Treatment outcome; Treatment-related toxicity; Tumor Volume; Uncertainty; X-Ray Computed Tomography; X-Ray Medical Imaging; automated segmentation; cancer radiation therapy; cancer site; cancer therapy; clinical application; clinical practice; convolutional neural network; cost; cost effective treatment; deep learning; dosimetry; falls; generative adversarial network; imaging Segmentation; improved; innovation; long short term memory; migration; novel; patient safety; radiation delivery; respiratory; simulation; success; tumor; tumor eradication

Project Summary/Abstract Stereotactic body radiation therapy (SBRT) is one of the most effective, well-tolerated, and cost-effective treatments. The success of SBRT relies heavily on the precision of dose delivery, due to the typically small tumor size, the very high radiation dose per fraction, and the sharp dose fall-off outside the target. For those sites where the tumor moves due to respiration, motion management is indispensable to ensure the high-precision dose delivery of SBRT. Current motion management strategies are either treating a large area encompassing the tumor motion range, or only delivering radiation dose within a small window (e.g., a gating window or at the end of inhale) of tumor motion cycle via indirect and inferior tumor motion monitoring (such as external surrogates or implanted fiducial markers). In-treatment real-time volumetric imaging is highly desired to enable direct, accurate, and markerless 3D tumor tracking for better motion management and capture unexpected large tumor motion for patient safety. The availability and accuracy of in-treatment real-time patient 3D anatomy information is also essential to the development of more active and advanced motion management technologies, such as multileaf collimator tracking and 4D treatment delivery. The unpredictable motion change during treatment can lead to substantial deviation of the delivered dose from the planned dose. Adaptive radiotherapy can compensate for the dosimetric errors by adapting the subsequent fractions. However, due to the notable changes of respiration, the pre-treatment imaging cannot provide the patient’s actual in-treatment anatomy to assess the actual delivered dose for adaptive radiotherapy. In-treatment real-time volumetric imaging is needed to enable dose-guided adaptive SBRT. Despite these strong needs, real-time volumetric imaging is not currently available due to the big challenge of reconstructing an instantaneous 3D image from very few 2D projections to meet the real-time requirement. To fill this clinical gap, we plan to develop a real-time volumetric imaging-based tumor tracking and dose verification (RITD) system using novel techniques in deep learning, imaging, Monte Carlo simulation and high-performance computation, and use lung SBRT treatment as a testbed. We will accomplish the following specific aims: 1) To develop and refine a real-time on-board volumetric imaging and tumor tracking method; 2) To develop an image correction method and a tumor/multi-organ segmentation method on the volumetric images; 3) To evaluate the performance of the proposed RITD system and assess its clinical benefit. The innovation of this study lies in developing new deep-learning approaches to enable real-time on-board volumetric imaging and build accurate tumor tracking and dose verification capability into cancer radiotherapy. It has substantial potential to improve lung SBRT treatment outcomes by reducing targeting uncertainty, improving treatment accuracy and precision, and enabling dose-guided adaptive lung SBRT. It paves the way for more active and advanced motion management (e.g., truly 4D radiotherapy). The proposed RITD system may be adapted for other cancer sites; thus, it has far-reaching clinical potential.

项目总结/摘要 立体定向放射治疗（SBRT）是最有效，耐受性好，成本效益高的治疗方法之一。 治疗。SBRT的成功在很大程度上依赖于剂量输送的精确性，因为通常肿瘤很小 尺寸、每部分非常高的辐射剂量以及靶外的急剧剂量下降。对于这些网站 在肿瘤由于呼吸而移动的情况下，运动管理是必不可少的，以确保高精度 SBRT的剂量递送。当前的运动管理策略要么是处理包括 肿瘤运动范围，或仅在小窗口内输送放射剂量（例如，门控窗口或在 吸气结束）的肿瘤运动周期，通过间接和下肿瘤运动监测（如外部替代物 或植入的基准标记）。非常需要治疗中实时体积成像， 精确、无标记的3D肿瘤跟踪，可更好地进行运动管理并捕获意外的大肿瘤 为了患者安全的动议。治疗中实时患者3D解剖信息的可用性和准确性 对于开发更主动和更先进的运动管理技术也至关重要，例如 多叶准直器跟踪和4D治疗输送。治疗期间不可预测的运动变化可能 导致输送剂量与计划剂量的实质性偏差。适应性放疗可以补偿 对于剂量测定误差，通过调整随后的分数。然而，由于 呼吸，治疗前成像不能提供患者的实际治疗中解剖结构来评估 适应性放射治疗的实际输送剂量。需要治疗中实时体积成像， 剂量引导的自适应SBRT。尽管有这些强烈的需求，但实时体积成像目前还不可用 由于从非常少的2D投影重建瞬时3D图像以满足 实时要求。为了填补这一临床空白，我们计划开发一种基于实时体积成像的肿瘤 跟踪和剂量验证（RITD）系统，使用深度学习，成像，蒙特卡罗 模拟和高性能计算，并使用肺SBRT治疗作为试验平台。要全面完成 具体目标如下：1）开发和完善实时机载体积成像和肿瘤跟踪 方法; 2）开发了图像校正方法和肿瘤/多器官分割方法， 3）评价所提出的RITD系统的性能并评估其临床益处。 本研究的创新之处在于开发新的深度学习方法，以实现实时机载 体积成像和建立精确的肿瘤跟踪和剂量验证能力到癌症放射治疗。 它具有通过降低靶向不确定性来改善肺SBRT治疗结局的巨大潜力， 提高治疗准确性和精确度，并实现剂量引导的自适应肺SBRT。它铺平了道路 对于更主动和高级的运动管理（例如，真正的4D放射治疗）。拟议的RITD系统 可能适用于其他癌症部位;因此，它具有深远的临床潜力。