New tools and method for monitoring ionizing radiation delivery in medical physics

这项研究的目的是开发提高放射治疗效率的工具。我们将通过以下两种不同的策略来做到这一点。第一项战略是开发新的剂量测量装置，能够快速、准确和精确地测量放射外科等最先进的放射治疗技术所提供的复杂辐射剂量。第一个装置是一个探测器，能够在三维空间和时间的函数中测量辐射剂量。放射治疗的成功需要精确的放射剂量，与肿瘤的形状紧密匹配。因此，三维剂量测量是保证放疗确实能按计划达到目标的最佳方法。第二个3D探测器将用于研究放射治疗期间的冲击解剖变化。解剖改变会降低放射治疗的效果。因此，重要的是定量评价它们的影响。虽然这些影响可以通过计算机模拟来估计，但目前还没有仪器可以直接测量它们作为时间的函数。这些新的检测设备和我们将开发的分析软件来解释它们的结果，将有可能帮助医学物理学家和其他专业人员更好地控制癌症治疗的准确性。我们第二个策略的目标是用智能算法建立一个虚拟的安全网，自动分析放射治疗过程中产生的图像，以保证治疗的准确性。大多数放射治疗输送机都配备了x射线成像功能。原则上，这些图像可用于监测患者，但这样做需要耗费大量时间和资源。此外，以人为基础的监测往往不一致。因此，这种成像能力往往没有得到充分利用，特别是在繁忙的加拿大放射治疗诊所。通过设想一个预测模型，我们期望在潜在的问题有机会改变治疗质量之前发现它们。这将构成敏捷放射治疗的基础，这是一种新的治疗范式，旨在通过称为放射组学的图像处理技术预测放射治疗患者的变化，然后自动建议可能的行动过程，以尽早有效地纠正问题。这些智能算法最终可能会被纳入放射治疗工作流程，并作为决策支持系统，帮助指导医学物理学家甚至医生。

The goal of this research is to develop tools to improve the efficiency of radiation treatments. We will do this by following two distinct strategies. The first strategy is to develop new dose measuring devices capable of rapidly, accurately and precisely measuring the complex radiation dose delivered by state-of-the-art radiotherapy techniques such as radiosurgery. The first device is a detector capable of measuring radiation doses in three dimensions and as a function of time. The success of radiotherapy requires an accurate delivery of radiation dose that tightly match the shape of a tumor. Thus, 3D dose measurement is the best possible way of guaranteeing that radiotherapy can indeed reaches its target as planned. A second 3D detector will be built to study the impact anatomical changes during radiotherapy. Anatomical changes can reduces the efficacy of radiation treatments. It is therefore important to quantitatively evaluate their impact. While these impact can be estimated by computer simulations, there is currently no instrument that can directly measure them as a function of time. These new detector devices and the analysis software we will develop to interpret their results will have the potential to help medical physicists and other professionals to better control the accuracy of cancer treatments. The goal of our second strategy is to build a virtual safety net with smart algorithms that automatically analyze images produced during radiation treatments to guarantee that treatments are delivered accurately. Most radiotherapy delivery machines are equipped with X-ray imaging capability. In principle, these images can be used to monitor patients, but doing so is time and resource intensive. Furthermore, human-based monitoring is often inconsistent. Therefore, such imaging capability are often underused, especially in the context of busy Canadian radiotherapy clinics. By conceiving a prediction model, we expect to detect potential problems before they have a chance of altering treatment quality. This will form the basis of agile radiotherapy, a new treatment paradigm that aims to anticipate changes in radiotherapy patients through an image processing technique known as radiomics and then automatically suggest possible course of actions to efficiently correct problems as early as possible. These smart algorithms could eventually be included in the radiotherapy workflow and act as a decision support system that will help guide medical physicists and even physicians.