Making sense of highly-disturbed blood flow dynamics

血液在体内流动的方式（血液动力学）被广泛认为对动脉粥样硬化和动脉瘤等血管疾病有很大影响，这些疾病是心脏病发作和中风的主要原因。不幸的是，与这些疾病相关的复杂血液动力学很难直接测量，这导致越来越多地使用与患者医学图像相关的计算机模拟血流，我们称之为基于图像的计算流体动力学（CFD）。在过去的五年里，我和我的团队发现了强有力的证据，表明血液动力学可能比许多医生和工程师认为的更复杂。教科书告诉我们，血液流动大多是有序的（层流），只有很少混乱（湍流），这使我们能够采取方便的捷径来加速CFD模拟。通过不走这些捷径，我们发现，在心跳的部分过程中可能会出现高频流动波动，并且在下一次心跳中可能会或可能不会看起来相同。虽然湍流的血液流动会加剧构成血管壁和血液的细胞，但目前还不清楚这些不太湍流的脉动流（我们称之为“高度扰动”）是否同样令人不安。为了理解这一点，我们必须首先阐明这些流动的性质，这些流动没有得到很好的研究，因为它们介于工程师通常处理的层流和湍流之间。这是我提出的NSERC发现研究的主要目标。首先，我们将弄清楚如何可视化这些高度时变的三维（因此，四维）流，并量化它们的相对复杂性。因为我们预计这些流动在视觉上是如此复杂，我们还将使用声音来利用人类的能力，即通过耳朵比眼睛更好地区分频率。其次，由于高度扰动的流动可能对驱动基于图像的CFD的患者数据中的不确定性敏感，因此我们将使用巧妙的技术来确定我们对CFD预测的信心。第三，由于血管的形状在决定血液动力学方面起着重要作用，我们将看到，鉴于这些不确定性，我们是否可以至少以相同的置信度预测仅从血管形状高度干扰血流的可能性，这一基础工程研究将与我在多伦多西部医院的临床研究同时进行，预测脑动脉瘤破裂的血流动力学因素，为我们提供了满足我们目标所需的各种患者数据。最终，我们获得的知识将应用于其他心血管疾病，其中高度干扰的流动现在也被发现，并将为医生和工程师提供更清晰的指导方针，如何充分模拟和/或预测它们。

The way blood flows in the body (hemodynamics) is widely thought to have a big impact on vascular diseases like atherosclerosis and aneurysms, the leading causes of heart attacks and strokes. Unfortunately, the complex hemodynamics associated with these diseases are difficult to measure directly, which has led to increasing use of computer simulations of blood flow tied to medical images of the patient, something we call image-based computational fluid dynamics (CFD).In the last five years my team and I have uncovered strong evidence that hemodynamics may be more complex that many doctors and engineers think. Textbooks tell us that blood flow is mostly orderly (laminar), and only rarely chaotic (turbulent), which has allowed us to take convenient shortcuts to speed up CFD simulations. By not taking these shortcuts, we have found that high-frequency flow fluctuations can arise during parts of the heartbeat, and may or may not look the same in the next beat. While turbulent blood flow is known to aggravate the cells that make up vessel walls and blood, it remains unclear whether these not-quite-turbulent pulsatile flows, which we term “highly disturbed”, are equally troublesome. To understand that, we must first shed some light on the nature of these flows, which are not well studied because they fall between the laminar and turbulent flows that engineers usually deal with. This is the main goal of my proposed NSERC Discovery research.First, we will figure out how to visualize these highly time-varying, three-dimensional (so, four dimensional) flows, and to quantify their relative complexity. Because we expect these flows to be so visually complicated, we will also use sound to take advantage of humans' ability to discriminate frequencies better by ear than by eye. Second, because highly-disturbed flows are likely to be sensitive to uncertainties in the patient data that drives image-based CFD, we will use clever techniques to determine just how confident we can be in our CFD predictions. Third, because the shape of blood vessels plays a major role in determining the hemodynamics within, we will see if, in light of those uncertainties, we can predict with at least the same confidence the likelihood of highly disturbed blood flow from vessel shape alone, potentially avoiding the need for CFD in the clinic.This basic engineering research will be carried out in parallel to my clinical research with Toronto Western Hospital on hemodynamic factors for predicting brain aneurysm rupture, providing us with the wide variety of patient data needed to meet our objectives. Ultimately, the knowledge we gain will have application to other cardiovascular diseases where highly-disturbed flows are now also being uncovered, and will provide doctors and engineers with clearer guidelines on how to adequately simulate and/or anticipate them.