Ultrasound-based sensors for the fusion and motion correction of MRI and PET/CT data

项目摘要 放射科配备了许多不同类型的医学成像扫描仪，如磁 共振成像（MRI）、正电子发射断层扫描（PET）、计算机断层扫描（CT）、X射线， 超声成像（USI）和单光子发射计算机断层扫描（SPECT）等。 显然，这些模式都不是理想的;否则，任何人都没有理由购买任何模式。 他人每种模态都有其自身的对比机制，并提供其自身类型的信息。 一个理想的扫描仪可以在这里定义为一个可以提供任何这些对比度，其 硬件将包括所有主要扫描技术所需的硬件， 期望的扫描类型，例如，MRI+CT+USI，继续。显然，在实践中，这种理想的扫描仪将是 相当不切实际。与每种模态相关的硬件组件将竞争少量的 患者周围的可用空间，这些组件会以可能严重阻碍 他们的功能。即使这些技术问题可以解决，复杂性也可能会导致成本上升， 令人望而却步的高度。一种更现实的多模态策略是在不同的设备上顺序地执行扫描。 物理扫描仪，然后融合图像，主要是通过软件。包括混合扫描仪 例如广泛使用PET/CT和可用但不常见的PET/MRI。的事实 涉及PET不是巧合，而是证明PET可能是目前最好的 这是一种用于诊断/分期癌症的可用工具，并且它还需要良好的解剖图像，例如那些 用于计算适当的衰减校正图。 特别是PET/MRI扫描仪是真正的工程杰作。两种合并模式 证明在癌症检测/分期和神经科学领域中特别有用。可以说，主要问题是， 与这些扫描仪可能是，他们是罕见的，非常昂贵，两个问题， 当然了工程杰作可能需要尊重，但它们也往往需要高昂的价格; 相比之下，本项目代表了一种实用且低成本的多模态成像方法。为 同样的原因也促使供应商开发PET/MRI扫描仪，因为这两种模式都非常强大， 他们各自的权利，但如此互补，以及，我们在这里专注于结合PET和MRI收购。但 拟议的工作并不限于这两种扫描类型，将来它很可能被证明可以推广到任何扫描类型。 这可以是来自任何成像模态和/或放射治疗设备的顺序扫描的组合。它是基于 基于超声波的传感器的发展，可以附着在病人的皮肤上， 他/她通过顺序扫描，作为一个共同的线程跨模态。安排了54名患者 临床PET/CT扫描将被招募，以验证拟议的技术，其中一半将是 核磁共振扫描也是如此。

Project summary Radiology departments are equipped with many different types of medical imaging scanners, such as Magnetic Resonance Imaging (MRI), Positron Emission Tomography (PET), Computed Tomography (CT), X-ray, ultrasound imaging (USI) and Single-Photon Emission Computed Tomography (SPECT), among others. Clearly, none of these modalities is ideal; otherwise, there would be no rationale for anybody to purchase any others. Each modality is characterized by its own contrast mechanism and provides its own type of information. An ideal scanner might be defined here as one that could provide any of these contrasts, whose hardware would include that required for all of the main scanning techniques, whereby one could simply select the desired scan type, e.g., MRI+CT+USI, and proceed. Clearly, in practice, such an ideal scanner would be quite impractical. Hardware components associated with each modality would compete for the small amount of space available around the patient and these components would interfere in ways that might seriously impede their functioning. Even if such technical problems could be solved, complexity would likely drive cost to prohibitive heights. A more realistic multi-modality strategy is to perform scans sequentially, on different physical scanners, and to fuse images afterward, mostly through software. Exceptions include hybrid scanners such as PET/CT, which are widespread, and PET/MRI, which are available but uncommon. The fact that both involve PET is not a coincidence, but rather a testimony to the fact that PET may well currently be the best available tool for diagnosing/staging cancer, and that it further requires good anatomical images such as those provided by CT or MRI for proper attenuation correction maps to be calculated. In particular, PET/MRI scanners are true masterpieces of engineering. The two combined modalities prove especially useful in the fields of cancer detection/staging and neuroscience. Arguably, the main problem with these scanners may be that they are rare and very expensive, two problems that are not unrelated of course. Masterpieces of engineering may demand respect but they also tend to demand a high price; in contrast, the present project represents a practical and low-cost approach to multi-modality imaging. For the same reasons that drove vendors to develop PET/MRI scanners, because both modalities are so powerful in their own rights yet so complementary as well, we focus here on combining PET and MRI acquisitions. But the proposed work is not limited to these two scan types, and in the future it might well prove generalizable to any combination of sequential scans from any imaging modalities and/or radiation therapy devices. It is based on the developments of ultrasound-based sensors that can be attached to the patient's skin and accompany him/her through sequential scans, acting as a common thread across modalities. Fifty-four patients scheduled for a clinical PET/CT scan will be recruited to validate the proposed technology, and half of them will be scanned by MRI as well.