(636g) Whole-Brain Computational Fluid Dynamic Analysis

Whole-Brain Computational Fluid Dynamic Analysis

The human brain is a large, metabolically sensitive organ requiring an uninterrupted supply of oxygen delivered through the blood stream. In order to properly perfuse cerebral tissue with oxygen, an extensive vascular network encompasses the brain. Cortical blood supply starts at the base of the brain with three major arteries, the left and right internal carotid artery (ICA) and the basilar artery (BA). These three arteries combine to form the Circle of Willis, which is a cyclic structure capable of redistributing blood throughout the brain in the case of occlusions or stenoses [1,2,3,4]. Six main cerebral arteries branch out from the Circle of Willis – left and right anterior, posterior, and middle cerebral arteries. Each of these arteries branch as they leave the Circle of Willis and travel along the gyrations of the cortical surface to create pial networks in independent territories of the brain. The pial vessels then penetrate the surface of the brain to form column-like microvascular structures that perfuse local regions with oxygen [5,6,7]. In a reverse process, blood is drained through the veins as they converge to the confluence of sinuses.

In order to detect morphologically consistent vascular architecture and blood flow properties of a specific patient, physicians commonly use non-invasive imaging modalities. These modalities produce three-dimensional images and structural reconstructions that can be visualized using volume and surface rendering respectively to facilitate interpretation by physicians. Various imaging modalities are used in clinical practice depending on what physiological properties are deemed relevant. In order to gather information regarding blood vessels, either Computed Tomography Angiography (CTA), Magnetic Resonance Angiography (MRA), or Digital Subtraction Angiography (DSA) is utilized. CTA is fast to acquire, MRA provides flow measurements, and DSA can be performed during surgeries. However, performing multiple scans on a patient is not feasible due to the limited resources in emergency centers and the time it takes to obtain a single medical image. Because of these limitations, developing a computational model of the vasculature that can accurately simulate physiological phenomena would be useful to physicians when investigating blood flow dynamics and devising a surgical plan.

While the expansive field of intracranial vasculature simulation has made great progress in the area of constructing computational fluid dynamics models capable of predicting hemodynamic patterns [8,9], many of these models do not address the flow dynamics in an entire, patient-specific brain. Models that treat the brain as a lump Windkessel or Balloon type model [10] are useful for predicting the aggregate behavior of the brain but are unable to address the growing need for quantitative predictions of cerebral blood flow, blood velocity, pressure distributions and angiographic transit times. In order to make these predictions, a complete cerebral vasculature model is needed that captures large arteries and veins as well as the microvasculature.

Based on images and reconstructions output from medical scanners, centerline extraction and radius detection can be performed on the blood vessel structure to represent a network of tubular objects [11]. We then perform data reduction and smoothing in order to limit the number of segments necessary to store accurate information about a blood vessel and to enforce G1 (positional and tangential) continuity between vessel segments. This technique yields a morphologically consistent patient-specific model of large cerebral vessels. However, hemodynamic models of the entire cerebral vasculature require an accurate reconstruction of not only the major vessels, but also require reconstruction of smaller vessels that exist beyond the scope of angiographic resolution. Detailed morphological studies have uncovered the hidden architecture of the microvasculature that bridges the gap between the major arterial and venous structures [5,6]. Vascular growth algorithms have been proven to be an accurate representation and have been thoroughly explored and validated against measured morphological and hemodynamic studies [12]. Once a complete tubular network of the major vessels and the microvasculature is constructed, it can be used to perform dynamic blood flow simulations.

The final, and most important, step in the entire process is analyzing the resulting data. In order to display the information so that fine details can be viewed along with the context of the entire vascular structure, our patient-specific vascular models are rendered in the 72 megapixel, stereoscopic, cylindrical CAVE2™ System. This immersive environment allows neurosurgeons, bioengineers, and researchers to walk through the unique three-dimensional structure of a specific patient. This affords unprecedented ability to assess morphology, spatial relationships, and simulation results of both macro and micro vessels.

References

[1]  S. Hirsch, et al., "Topology and Hemodynamics of the Cortical Cerebrovascular System," J Cereb Blood Flow Metab, vol. 32, pp. 952-67, Jun 2012.

[2]  P. Blinder, et al., "Topological Basis for the Robust Distribution of Blood to Rodent Neocortex," Proc Natl Acad Sci USA, vol. 107, pp. 12670-5, Jul 13 2010.

[3]  M. Brozici, et al., "Anatomy and Functionality of Leptomeningeal Anastomoses - A Review," Stroke, vol. 34, pp. 2750-2762, Nov 2003.

[4]  A. Shuaib, et al., "Collateral Blood Vessels in Acute Ischaemic Stroke: a Potential Therapeutic Target," Lancet Neurology, vol. 10, pp. 909-921, Oct 2011.

[5]  H. M. Duvernoy, et al., "Cortical Blood vessels of the Human Brain," Brain Res Bull, vol. 7, pp. 519-79, Nov 1981.

[6]  F. Cassot, et al., "Branching Patterns for Arterioles and Venules of the Human Cerebral Cortex," Brain Research, vol. 1313, pp. 62-78, Feb 8 2010.

[7]  F. Lauwers, et al., "Morphometry of the Human Cerebral Cortex Microcirculation: General Characteristics and Space-Related Profiles," NeuroImage, vol. 39, pp. 936-48, Feb 1 2008.

[8]  A. Bui, et al., “Dynamics of pulsatile flow in fractal models of vascular branching networks,” Med Biol Eng Comput, vol. 47, pp. 763-72, 2009.

[9]  L. Grinberg, et al., “Modeling Blood Flow Circulation in Intracranial Arterial Networks: A Comparative 3D/1D Simulation Study,” Annals of Biomedical Engineering, vol 39, pp. 297-309, 2011.

[10]  D.A. Boas, et al., “A Vascular Anatomical Network Model of the Spatio-Temporal Response to Brain Activation,” NeuroImage, vol. 40, pp. 1116-29, 2008.

[11]  F. Benmansour, et al., “Tubular Structure Segmentation Based on Minimal Path Method and Anisotropic Enhancement,” Int J Comput Vis, vol. 92, pp. 192-210, 2011.

[12]  R. Karch, et al., “Staged growth of optimized arterial model trees,” Annals of Biomedical Engineering, vol. 28, pp. 495-511, 2000.