Assessing Placental Structure and Function by Unified Fluid Mechanical Modelling and in-vivo MRI

我们的计划是通过创造一种革命性的方法来评估和分析怀孕环境，以减少不良的妊娠结局。具体地说，我们将建立专门的核磁共振技术(这种技术是非侵入性的，不会释放电离辐射)，专注于高风险怀孕的妇女，这将使母亲和婴儿能够进行个性化护理和个性化治疗。胎盘(或胎盘)对成功妊娠至关重要，大多数重大妊娠并发症都与胎盘问题有关。超声波是目前监测怀孕的主要方法，但它有局限性：不能一次看到整个胎盘，结果根据执行者的不同而不同，它不能告诉我们胎盘功能如何，只能告诉我们它看起来是什么样子。由于婴儿的位置和母亲的体重等因素，超声图像也可能很差。我们将使用MRI的主要功能来实现对胎盘健康的更详细的评估-关于胎盘如何工作的信息以及对其结构的更详细的描述。尽管我们可以在胎盘MRI扫描中看到一些重要的妊娠并发症的明显变化，如胎儿生长受限和先兆子痫，但影响MRI信号的胎盘具体变化尚未确定，尤其是因为胎盘的结构和功能之间的关系是复杂的。检测胎盘细微变化的技术将使怀孕并发症的早期和准确诊断成为可能。在这个项目中，我们将开发这样一种技术，将磁共振成像与计算建模相结合，以提供对胎盘工作情况的有价值的见解。首先，我们将建立整个胎盘的高度详细的计算模型，包括组织和血管尺寸等精细结构细节。接下来，我们将开发可以预测这些计算胎盘中的血流量和氧气水平的工具。然后，我们使用模拟来预测计算胎盘的MRI扫描结果。最后，使用机器学习技术，我们将指示一个算法从数以千计的计算胎盘模拟中学习，以便提供一个成像工具，给定胎盘MRI扫描作为输入，可以显示胎盘的结构、血流和氧气水平。通过使用这个基于MRI的工具来调查有并发症的怀孕，我们将测试它识别胎盘不正常工作的能力。我们将通过将我们的工具与专业病理学家对分娩的胎盘进行的孕后评估进行比较来实现这一点(许多胎盘状况，如结构异常或血液供应不良，目前只能通过这种方式诊断)。对于多名孕妇，我们将在怀孕期间使用我们基于MRI的工具来计算与胎盘结构和血流相关的值，然后进行产后胎盘评估。这将使我们能够测试我们的基于MRI的工具是否可以在怀孕期间判断胎盘是否不工作，而不必等待专家病理学家的产后评估。我们将首先探索确定胎盘问题是否由母体或胎儿血液供应问题引起的方法。这在临床实践中将是极其重要的，未来可能会允许对母亲进行重点治疗，以减少不良结局。显示孕妇和胎儿血液供应问题之间的差异只是目前不能在怀孕期间进行诊断的一个例子，我们在这里使用它来展示我们的模型驱动的胎盘成像工具在直接影响的应用中的潜力。通过开发的工具可以确定的其他潜在情况包括胎儿绒毛树(这是胎盘的详细结构)的损害，以及氧气供应的水平。

Our plan is to reduce poor pregnancy outcomes by creating a revolutionary way to assess and analyse the pregnancy environment. Specifically, we will establish specialised MRI (which is non-invasive and emits no ionising radiation) techniques focusing on women with high-risk pregnancies, which will enable personalised care and individualised treatments for mothers and babies. The placenta (or afterbirth) is vital for a successful pregnancy, and most major pregnancy complications are associated with placental problems. Ultrasound is currently the main way to monitor pregnancies, however it has limitations: the whole placenta can't be seen at once, findings vary depending on who is performing it, and it cannot tell us how the placenta is functioning, only what it looks like. Ultrasound images can also be poor due to the baby's position and factors such as the mother's weight. We will use key features of MRI to enable a more detailed assessment of placental health - information on how the placenta is working and a more detailed picture of its structure. Although we can see clear changes in placental MRI scans for a number of important pregnancy complications, such as fetal growth restriction and pre-eclampsia, the specific changes in the placenta that affect the MRI signal are yet to be determined, not least because the relationship between the placenta's structure and how well it functions is complex. Techniques for detecting subtle placental changes would enable early and accurate diagnosis of pregnancy complications. In this project, we will develop such a technique by combining MRI with computational modelling to give valuable insight into how well the placenta is working. First of all, we will build highly detailed computational models of whole placentas, incorporating fine structural details such as tissue and blood vessel dimensions. Next we will develop tools that can predict blood flow and oxygen levels in these computational placentas. We then use simulations to predict what MRI scans would look like for our computational placentas. Finally, using machine learning techniques, we will instruct an algorithm to learn from thousands of simulations on computational placentas, in order to deliver an imaging tool that, given a placental MRI scan as input, can show the structure, blood flow, and oxygen levels of the placenta. By using this MRI-based tool to investigate pregnancies with complications, we will test its ability to identify cases where the placenta is not working properly. We will do this by comparing our tool with a post-pregnancy assessment of the delivered placenta by a specialised pathologist (many placental conditions, such as abnormal structure or poor blood supply, can currently only be diagnosed in this way). For multiple pregnant participants, we will use our MRI-based tool during the pregnancy to calculate values relating to placental structure and blood flow, then undertake an after delivery assessment of the placenta. This will allow us to test if our MRI-based tool can tell if the placenta is not working during pregnancy without having to wait for an after delivery assessment by a specialist pathologist. We will first explore ways of determining if problems with the placenta are caused by issues with the supply of maternal or fetal blood. This would be extremely important in clinical practice and in the future may allow focused treatment of the mother to reduce poor outcomes. Showing differences between maternal and fetal blood supply problems is just one example of a diagnosis that cannot be during the pregnancy at present, and we use it here to demonstrate the potential of our model-driven placenta imaging tools in a directly impactful application. Further potential conditions that could be identified with the developed tools include damage to fetal villous trees (this is the detailed structure of the placenta), and levels of oxygen supply.