Assessing deformable image registration in the lung using hyperpolarized-gas MRI

摘要 放射治疗期间患者解剖特征的位置和方向的变化，如果不正确的话 管理不当，可能会导致目标组织剂量不足和/或邻近健康组织剂量过大。 可变形图像配准(DIR)由于具有几何对齐两幅图像的能力，正在成为 对于处理这些解剖变异在放射治疗中日益重要。DIR的精度直接关系到DIR的精度 影响其临床应用的成败。因此，仔细评估DIR算法是至关重要的 在它们可能被用于指导临床决策之前是必要的。当前的DIR评估方法 注重形态结构，而不是整个变形的生理有效性。最近，我们 已经展示了一种新的超极化3He标记磁共振技术，它能够直接、在体内和 在区域基础上非侵入性测量生理性肺变形。这种独特的成像技术 为评估、验证和改进DIR算法在肺中的使用带来了巨大的希望。我们的龙 学期目标是将超极化气体标记磁共振应用于肺生物力学研究，开发更多生理学 完善的肺部DIR算法，最终改善肺癌的放射治疗。总体目标 这一应用的目的是优化超极化3He标记磁共振技术，并建立其在 DIR评估。目标1是发展和优化一种基于三维超极化3He标记的方法 直接测量肺变形的MRI。目标2是开发生理上健全的数字胸腔幻影 基于HP3He标记的MRI，并演示其在肺DIR评估中的应用。成功 这些目标的实现将产生一种基于超极化~3He标记磁共振成像用于DIR的新方法 在肺部进行放射治疗的评估。它还将产生许多新的磁共振成像技术和新的 一系列数字胸腔幻影。测量真实的生理性肺变形的能力使我们的 技术是评估、验证和改进肺DIR算法的一个很有前途的工具。本研究 将给研究和临床带来重大变化。在短期内，它可能会为我们提供新的见解 肺生物力学的复杂性，丰富了我们对DIR的理解，产生了黄金标准的数据集 可能有益于研究界的肺变形，并为临床实施 目录从长远来看，这可能会导致开发更复杂的DIR工具来改善放射治疗 肺癌，导致更精确地向肺肿瘤提供放射治疗，并减轻辐射- 对周围正常组织造成损伤。

ABSTRACT Changes in the position and orientation of a patient’s anatomical features during radiotherapy, if not properly managed, can lead to underdosage of the target and/or overdosage of neighboring healthy tissues. Deformable image registration (DIR), owing to its ability to geometrically align two images, is becoming increasingly important in radiotherapy for managing these anatomy variations. The accuracy of the DIR directly impacts the success of its clinical applications. Careful assessment of DIR algorithms is therefore a critical necessity before they may be used to inform clinical decision making. Current methods of DIR assessment focus on morphological structures but not on the physiological validity of the entire deformation. Recently, we have demonstrated a novel hyperpolarized 3He tagging MRI technique that is capable of directly, in vivo, and non-invasively measuring physiological lung deformation on a regional basis. This unique imaging technique holds great promise for assessing, validating, and improving the use of DIR algorithms in the lung. Our long term goal is to apply hyperpolarized gas tagging MRI to study lung biomechanics, develop more physiologically sound DIR algorithms for the lungs, and eventually improve radiotherapy of lung cancer. The overall objective of this application is to optimize the hyperpolarized 3He tagging MRI technology and establish its usefulness for DIR assessment. Aim 1 is to develop and optimize a methodology based on 3D hyperpolarized 3He tagging MRI for directly measuring lung deformation. Aim 2 is to develop physiologically sound digital thorax phantoms based on HP 3He tagging MRI and demonstrate their use for DIR assessment in the lung. Successful completion of these aims will yield a novel methodology based on hyperpolarized 3He tagging MRI for DIR assessment in the lung for radiotherapy. It will also yield a number of novel MR imaging techniques and a new series of digital thorax phantoms. The ability to measure true physiological lung deformation makes our technique a promising tool for the assessment, validation, and improvement of lung DIR algorithms. This study will bring important changes to research and the clinic. In the short term, it may lend new insights into the complexities of pulmonary biomechanics, enrich our understanding of DIR, generate gold-standard datasets of lung deformation that may benefit the research community, and provide guidance for clinical implementation of DIR. In the long term, it may lead to development of more sophisticated DIR tools for improving radiotherapy of lung cancer, resulting in more precisely delivered radiation treatment to lung tumors and mitigating radiation- induced injury to surrounding normal tissues.