4D Robust Optimization in Intensity-Modulated Proton Therapy

描述(申请人提供)：申请人的直接职业目标是过渡到高素质的独立研究人员，并在紧张和支持性的研究环境中建立一个小型实验室。从长远来看，申请者希望发展一份扎实的学术生涯，专注于医学辐射物理领域的转化性研究，该领域影响最先进的放射治疗模式的各个方面。这位候选人于2007年在普林斯顿大学获得博士学位，后来在洛斯阿拉莫斯国家实验室担任博士后研究员，在计算物理、数学和算法以及软件开发方面拥有丰富的经验，特别是大规模并行高性能计算(HPC)的代码开发。应聘者还具有扎实的分析和数学技能，能够解决复杂的物理问题，并具有跨学科背景。2010年7月，应聘者进入德克萨斯大学MD安德森癌症中心辐射物理系教职，担任助理教授(研究轨道)。MD安德森是一家以研究为导向的综合性癌症医院，也是美国领先的癌症中心。辐射物理系提供必要的临床、研究和教育资源，以支持与癌症治疗相关的物理和剂量学研究。该部门的质子治疗中心-休斯顿(PTC-H)于2006年5月开始为患者治疗。PTC-H是世界上少数几个有能力提供扫描质子束强度和能量调制治疗的质子治疗中心之一，这是拟议研究的主要重点。在获奖期间，候选人将专注于新的先进放射治疗方法的开发和验证。在K25基金的支持下，申请者计划1)获得放射治疗临床和研究方面的深入知识和实践经验；2)获得深入的解剖学知识；3)获得深入的医学成像知识；4)获得中等水平的生物统计学知识；5)获得中等水平的放射生物学知识。为了实现这些目标，申请者将与一群经验丰富的导师和合作研究人员合作，开展专注于癌症放射治疗的联合项目，在莱斯大学和德克萨斯大学休斯顿健康科学中心参加学术课程，接受放射治疗的临床培训，并参加研讨会和会议。强度调制质子疗法(IMPT)的四维(4D)稳健优化被选为K25培训计划的研究主题，以帮助申请者获得成为独立研究人员所需的经验。使用IMPT治疗肺癌带来了许多挑战，需要通过研究来解决这些挑战，以最大限度地发挥这种有希望的治疗方式的治疗效益。候选人在这项研究中学到的东西将广泛适用于该领域的许多研究领域。原则上，IMPT 有最大的潜力提供高度适形的肿瘤靶点覆盖，同时保留邻近的健康器官。然而，质子的特性(例如，剂量突然下降超出范围和散射)使IMPT非常容易受到不确定因素的影响。不确定的来源包括肿瘤缩小、体重减轻、患者设置的变化、呼吸运动、CT数量和停止功率比的不确定，以及质子剂量计算算法中的近似。目前在IMPT中管理不确定性的做法类似于调强放射治疗(IMRT)，即在临床靶区周围分配安全裕度以产生计划靶区。一般而言，在面对不确定性的情况下，所产生的剂量分布是不稳健的，即，提供给患者的可能与计算机设计的治疗计划上看到的显著不同，并可能导致不可预见的临床后果。因此，研究开发合适的4D稳健优化方法，以提高IMPT计划对包括规则运动和不规则运动在内的不确定性的最优性和稳健性，是至关重要的。我们的假设是，4D稳健优化可以使IMPT计划对不确定性不那么敏感，并实现比基于边际优化的传统计划更好地保存正常组织(两者至少都是两倍)。我们建议在以下具体目标上验证我们的假设：(1)量化解剖运动及其不确定性；(2)开发和实施4D IMPT优化；(3)增强IMPT计划的健壮性；(4)验证IMPT 4D稳健优化。与以前的IMRT中的4D稳健优化方法相比，申请人提出的研究具有几个创新方面，包括用于处理不确定和规则运动的自适应的、根据需要自动优化的边界、使用在量子力学中广泛应用的微扰理论来求解薛定谔方程、用于处理不规则运动的存储器分布式并行化以及用于解决具有挑战性的高计算机内存要求的问题。我们预计，我们在IMPT中的开创性4D稳健优化研究将填补我们在适当方法方面的空白，以最大限度地减少IMPT中不确定性的影响，并为癌症患者，特别是胸腹癌患者带来显著好处。该项目不涉及美国以外的活动，也不涉及与国际合作者的伙伴关系。

DESCRIPTION (provided by applicant): The applicant's immediate career goal is to make the transition to a high-caliber independent researcher and to establish a small laboratory in an intense and supportive research environment. In the long term, the applicant hopes to develop a solid academic career focusing on translational research in the area of medical radiation physics that impacts various aspects of the state-of-the-art radiotherapy modalities. The candidate, who obtained his PhD from Princeton University in 2007 and later worked at the Los Alamos National Laboratory as a postdoctoral researcher, has extensive prior experience in computational physics, mathematics and algorithms, and software development, especially code development in massive parallel high-performance computing (HPC). The candidate also has solid analytical and mathematical skills to solve complicated physics problems, along with an interdisciplinary background. In July 2010, the candidate joined the faculty of the Department of Radiation Physics of The University of Texas MD Anderson Cancer Center as an assistant professor (research track). MD Anderson is a research-driven, comprehensive cancer hospital and is the leading cancer center in the United States. The Department of Radiation Physics provides the clinical, research, and educational resources necessary to support physics and dosimetry research related to cancer therapy. The department's Proton Therapy Center-Houston (PTC-H) began patient treatments in May 2006. PTC-H is one of few proton treatment centers in the world to have the capability to deliver intensity- and energy-modulated treatments with scanned proton beams, which is the major focus of the proposed research. During the award period, the candidate will focus on the development and validation of novel advanced radiation therapy methodologies. With the support of this K25 grant, the applicant plans to 1) obtain in-depth knowledge and hands-on experience in the clinical and research aspects of radiation therapy; 2) obtain in- depth knowledge of anatomy; 3) obtain in-depth knowledge of medical imaging; 4) obtain moderate knowledge of biostatistics; and 5) obtain moderate knowledge of radiobiology. To achieve these objectives, the applicant will work with a group of experienced mentors and collaborating researchers on joint projects focusing on the radiotherapy of cancers, take academic courses at Rice University and The University of Texas-Health Science Center at Houston, undergo clinical training in radiation therapy, and attend seminars and conferences. Four-dimensional (4D) robust optimization of intensity-modulated proton therapy (IMPT) has been chosen as the research topic for this K25 training program to help the applicant gain the experience necessary to become an independent investigator. The use of IMPT to treat lung cancers presents numerous challenges that need to be addressed through research to maximize the therapeutic benefits of this promising modality. What the candidate learns during this research will be widely applicable to many areas of research in this field. In principle, IMPT has the greatest potential to provide highly conformal tumor target coverage, while sparing adjacent healthy organs. However, characteristics of protons (e.g., the abrupt drop-off of dose beyond the range and scattering) make IMPT highly vulnerable to uncertainties. Sources of uncertainty include tumor shrinkage, weight loss, variation in patient setup, respiratory motion, uncertainties in CT numbers and stopping power ratios, and approximations in proton dose calculation algorithms. The current practice for managing uncertainties in IMPT is similar to that for intensity-modulated radiation therapy (IMRT), i.e., assigning a safety margin around the clinical target volume to produce a planning target volume. The resulting dose distributions are, in general, not robust in the face of uncertainties, i.e., what is delivered to the patient may be significantly different from what is seen on the computer-designed treatment plan and may lead to unforeseen clinical consequences. Therefore, investigations leading to the development of suitable 4D robust optimization methods to improve the optimality and robustness of IMPT plans to uncertainties, including regular and irregular motion, are vital. Our hypothesis is that 4D robust optimization can render IMPT plans less sensitive to uncertainties and achieve better sparing of normal tissues (both by at least a factor of two) than conventional plans optimized on the basis of margins. We propose to test our hypothesis in the following specific aims: (1) to quantify anatomy motion and its uncertainty; (2) to develop and implement 4D IMPT optimization; (3) to enhance the IMPT plan robustness; and (4) to validate IMPT 4D robust optimization. Compared to previous 4D robust optimization approaches in IMRT, the research proposed by the applicant has several innovative aspects, including a "customized", as-small-as-necessary margin optimized spontaneously to handle uncertainties and regular motion, the use of perturbation theory, widely used in quantum mechanics to solve the Schr¿dinger equation, to handle irregular motion, and memory-distributed parallelization to solve the challenging high-computer-memory requirement problem. We expect that our pioneering 4D robust optimization research in IMPT will fill gaps in our knowledge about appropriate ways to minimize the influence of uncertainties in IMPT and lead to significant benefits for cancer patients, especially those with thoracic and abdominal cancers. This project doesn't involve activities outside of the United States or partnerships with international collaborators.