Quantitative MRI/PET bimodal pharmacokinetic modeling to improve diagnostic accuracy in medical imaging

在临床上，各种成像技术被用来查看患者的内部，找出哪里出了问题。PET/MRI是一种将磁共振成像(MRI)和正电子发射断层扫描(PET)这两种成熟的成像技术相结合的新技术。PET/MRI扫描仪正在进入加拿大的医院，几项研究已经展示了PET/MRI相对于单独的PET和MRI扫描仪的一些优势。然而，PET/MRI扫描仪非常昂贵，为了确保加拿大人从这项投资中获得高回报，我们必须充分挖掘这项技术的潜力。 我们的研究重点是分子成像，这是成像的一个特定分支，它量化了生理过程，并非常详细地表征了组织。为了做到这一点，我们注入了可以在成像过程中进行监测的分子“探测器”。例如，PET可以使用可追踪的糖来检测细胞的糖消耗，MRI可以使用对比剂产生探针来检测流向肿瘤的血液。这类信息可用于临床计划和监测抗癌治疗，以及诊断病理。 然而，分子成像并非易事。一个困难是，注射的探针在体内的分布既取决于它的预定目标(我们正在寻找什么)，也取决于它与身体的任何其他相互作用。想象一下，一栋有许多房间的建筑，病人需要治疗。注射探头就像是把护士和医生送进去，他们的目标是照顾这些病人。成像就像透过窗户观看一样。观察者如何知道医生在房间里是因为他/她找到了病人，还是因为他/她正在寻找病人？如果医生停止活动，他/她可能会因为房间很忙而被困在房间里。他/她也可能正在与健康的人讨论。分子成像是类似的：我们注射一个探针来检测肿瘤，然后我们在身体的某个地方检测到这个探针。然而，我们永远不能确定探测器是否真的找到了它想要的目标，或者它是否仅仅因为它在探索身体而找到了它的目标。 这个问题的解决方案相当简单：我们派出另一个团队，他们唯一的目标是走进大楼，探索房间，而不是停下来照顾病人。最后，我们比较两个团队的位置，差异表明患者在哪里。这就是PET/MRI派上用场的地方：我们有两个观察者(成像技术)，每个观察者能够探测一个探针。第一个探头是有目标的(例如，针对肿瘤)，第二个是无目标的，并指示如果第一个探头只是在“探索”身体的话它会在哪里。两者之间的差异表明目标所在的位置。 到目前为止，由于这里描述的问题，分子成像在临床上的领域是有限的。我们的工作将消除这一限制。利用PET/MRI，我们将开发监测两个探针的方法。除其他外，我们将能够在图像中区分非靶标(非特定)和靶标(特定)信号，这将使我们能够更准确地描述组织，并在分子水平上提供真实、可靠和定量的信息，临床医生可以使用这些信息来诊断和监测疾病。

In the clinic, various imaging technologies are used to look inside patients and find out “what’s wrong”. PET/MRI is a newly available technology which combines two established imaging technologies, Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET). PET/MRI scanners are making an entry into Canadian hospitals, and several studies have presented some of the advantages of PET/MRI over separate PET and MRI scanners. However, PET/MRI scanners are very expensive, and to make sure Canadians get a high return from this investment, we must explore the full potential of this technology. Our research focuses on molecular imaging, a specific branch of imaging that quantifies physiological processes and characterizes tissues in very fine detail. To do this, we inject molecular “probes” that can be monitored during imaging. For example, PET can be used to detect the sugar consumption of cells using a traceable sugar, and MRI can be used to detect blood flow to a tumor using a contrast generating probe. This kind of information can be used in the clinic to plan and monitor anti-cancer therapy, and diagnose pathologies. However, molecular imaging is not easy. A difficulty is that distribution of an injected probe inside the body depends on both its intended target (what we are looking for) and any other interaction it has with the body. Imagine a building with many rooms where patients require treatment. Injecting a probe is like sending nurses and physicians inside whose goal is to attend to these patients. Imaging is like watching through the windows. How can an observer know whether a physician is inside a room because he/she has found a patient, or because he/she is looking for one? If a physician stops moving, he/she might simply be stuck in a room because the room is very busy. He/She might also be discussing with someone healthy. Molecular imaging is similar: We inject a probe to detect a tumor, and we detect the probe somewhere in the body. Yet, we can never be certain if the probe has indeed found its intended target, or if it is there simply because it is exploring the body “searching” for its target. The solution to this problem is fairly simple: We send another team, whose only goal is to walk inside the building and explore the rooms without stopping to attend to patients. At the end, we compare the location of both teams, and the difference indicates where the patients are. This is where PET/MRI comes in handy: we have two observers (imaging technologies) capable of detecting one probe each. The first probe is targeted (e.g., toward a tumor), and the second one is untargeted, and indicates where the first probe would be if it was just “exploring” the body. The difference between the two indicates where the target is located. As of now, the field of molecular imaging in the clinic is limited because of the problem described here. Our work will remove this limitation. Using PET/MRI, we will develop methods to monitor two probes. We will, among other things, be able to make the difference between untargeted (unspecific) and targeted (specific) signal in images, which will enable us to characterize tissues more accurately, and provide real, solid and quantitative information at the molecular level that clinicians can use to diagnose and monitor diseases.