Development of A High Throughput Image-Guided IMRT System for Preclinical Research

项目概要/摘要 小动物临床前放射生物学实验对于在人体之前测试安全性和有效性至关重要 临床试验。然而，受限于现有技术，临床前动物研究存在很大差异 来自最先进的人体治疗剂量一致性。因此，动物研究很难模拟 人类治疗的放射生物学、放射免疫学和毒性环境。差距产生不利影响 我们有能力有意义地测试用于人工翻译的假设。凭借数十年的 随着人类放射治疗技术的进步，人类放射治疗已经实现了基于 技术突破，包括调强放射治疗（IMRT），这是小鼠无法实现的 实验。一种针对小动物尺度调制 X 射线强度的实用装置和算法是 弥合差距的第一步。在 NIH R21 拨款的支持下，我们设计了一种新型小动物 IMRT 剂量调节器称为稀疏正交准直器（SOC）。与硬件同样重要的是，我们创建了 使数学工具能够提供比理论上可实现的分辨率更高的 SOC IMRT 计划 基于 MLC 的小型 IMRT。我们委托并测试了原型 SOC，以提供高度调制的 计算机和模型上的剂量。尽管如此，强度调制装置之间仍然存在很大差距 以及适合广泛采用和影响的小动物 IMRT 系统。所需的时间、资源和 创建复杂的 SOC-IMRT 计划的培训与临床前设置不兼容。此外，没有 自动化，现有的图像引导小动物 IMRT 治疗对于治疗活体动物来说速度慢得令人望而却步 在麻醉下。最后，当前在成像和治疗模式之间切换的手动方法导致 剂量输送中棘手的不确定性。我们建议使用自动化、机器人技术和系统来填补这些空白 优化。我们提出以下具体目标。具体目标 1 (SA1)。自动器官分割 使用深度学习神经网络的小鼠。具体目标 2 (SA2)。开发一个功能齐全、自动化、 高效的 IMRT 系统。具体目标 3 (SA3)。机器人多鼠标自动化的开发和验证 用于自动成像和治疗的治疗环境（Multi-MATE）。除了剂量测定之外，我们还将量化 时间性能，这对于小动物 IMRT 系统至关重要。因此，除了改善 硬件准确性和可靠性，拟议项目将提供完全自动化的规划和交付 系统，从而消除了广泛采用小动物 IMRT 的最后障碍。的成功 拟议的项目将帮助现有研究充分发挥人工翻译的潜力，并使未来成为可能 假设检验准确的复杂剂量分布至关重要。

Project Summary/Abstract Preclinical radiobiology experiments on small animals are crucial to test the safety and efficacy before human clinical trials. However, limited by currently available technologies, preclinical animal studies substantially differ from state-of-the-art human treatments in dose conformity. Consequently, the animal studies poorly mimic the radiobiological, radioimmunological, and toxicity environment of human therapies. The disparity adversely affects our ability to meaningfully test hypotheses that are intended for human translation. With decades of advancement, human radiotherapy has achieved high targeting accuracy and dose conformality based on technological breakthroughs, including intensity-modulated radiotherapy (IMRT), which is unavailable for mouse experiments. A practical device and algorithm to modulate the x-ray intensity for the scale of small animals is the first step to bridge the gap. With the support of an NIH R21 grant, we engineered a novel small animal IMRT dose modulator termed sparse orthogonal collimator (SOC). Equally important as the hardware, we created the enabling mathematical tools to deliver SOC IMRT plans with higher achievable resolution than a theoretically miniaturized MLC-based IMRT. We commissioned and tested prototypical SOCs to deliver highly modulated doses in silico and on phantoms. Nonetheless, there are still large gaps between an intensity modulation device and a small animal IMRT system suitable for broad adoption and impact. The required time, resources, and training to create sophisticated SOC-IMRT plans are incompatible with preclinical settings. Furthermore, without automation, the existing image-guided small animal IMRT treatment is prohibitively slow for treating live animals under anesthesia. Lastly, the current manual method to switch between imaging and therapy modes results in intractable uncertainties in dose delivery. We propose to fill these gaps using automation, robotics, and system optimization. We propose the following specific aims. Specific Aim 1 (SA1). Automated organ segmentation for mice using deep learning neural networks. Specific Aim 2 (SA2). Development of a fully functional, automated, and efficient IMRT system. Specific Aim 3 (SA3). Development and validation of a robotic Multi Mouse Automated Treatment Environment (Multi-MATE) for automated imaging and treatment. Besides dosimetry, we will quantify the time performance, which is critical to small animal IMRT system. As a result, in addition to improving the hardware accuracy and reliability, the proposed project will provide a fully automated planning and delivery system, thus removing the last barriers towards the broad adoption of small animal IMRT. The success of the proposed project will help existing research to achieve the full potential for human translation and enable future hypotheses testing where accurate complex dose distribution is critical.