Research

 

Objective

 

Our broad research objectives are:

 

·         Improve our understanding of the mechanisms responsible for the initiation, progression and outcome of cerebrovascular diseases

·         Develop new techniques for assessing the risk of stroke of a given patient

·         Develop new techniques to personalize and optimize minimally invasive endovascular treatments of these diseases

 

Research Team

 

The research team combines experts in image-based vascular hemodynamics modeling (Dr. Cebral), computational fluid dynamics and grid generation (Dr. Löhner), interventional neuroradiology (Dr. Putman), biostatistics and epidemiology (Dr. Sheridan), and vascular segmentation and medical image analysis (Dr. Frangi). The members of this team have actively collaborated in the past few years on computational fluid dynamics studies of cerebrovascular diseases, and have published several joint papers [1-13].

 

Stroke

 

Cerebrovascular diseases, such as vascular dementia or stroke, are the leading causes of morbidity and mortality in the United States. Stroke is the single leading cause of nursing home admission, and one of the fastest growing health care costs. Stroke can either be ischemic (due to lack of blood flow or oxygen to the brain) or hemorrhagic (bleeding into the brain). Ischemic strokes are caused by an interruption of blood flow to a portion of the brain typically due to blockages of blood vessels. Hemorrhagic strokes are most commonly due to the rupture of a cerebral aneurysm.

 

Cerebral aneurysms

 

Cerebral aneurysm rupture is a leading cause of hemorrhagic strokes. Cerebral aneurysms are pathological dilatations of the arterial wall frequently located near arterial bifurcations in the circle of Willis [14-16]. The most serious consequence is their rupture and intracranial hemorrhage into the subarachnoid space, with an associated high mortality and morbidity rate [17-20]. Intracranial aneurysms are particularly difficult to treat, and often do not produce symptoms before they rupture [21]. Greater availability and improvement of neuroradiological techniques have resulted in more frequent detection of unruptured aneurysms. Because prognosis of subarachnoid hemorrhage is still poor, preventive surgery is increasingly considered as a therapeutic option. But every treatment carries a risk, which sometimes matches or exceeds the yearly risk of aneurysm rupture. Therefore, the best patient care would be to treat only those patients who are likely to rupture [22-24]. Planning elective surgery requires a better understanding of the process of aneurysm formation, progression, and rupture so that a sound judgment between the risks and benefits of possible therapies can be made. These processes are not well understood. Previous studies [14, 25-34] have identified the major factors involved in these processes: a) hemodynamics, b) wall biomechanics and mechanobiology, and c) peri-aneurysmal environment.

 

Image-Based Computational Fluid Dynamics Simulation Environments

 

Since to date there are no reliable non-invasive techniques for in vivo quantification of aneurysmal blood flow patterns, we use a patient-specific image-based computational modeling and simulation approach. Over the past several years, we have developed state of the art computational tools for patient-specific modeling of hemodynamics in human arterial systems. We have implemented these tools in efficient and novel combinations [11, 35] into a software that has allowed us to conduct numerous simulations for studying biological fluid dynamics problems [2, 3, 5, 36-44]. This simulation software includes tools for: a) image processing and segmentation, b) geometry modeling, c) unstructured grid generation, d) incompressible flow solution, and e) visualization.

 

 

Figure 1 shows the graphical user interface of the software environment used for constructing image-based vascular models (left panel). An example of a patient specific model of a cerebral aneurysm constructed from 3D rotational angiography images is shown in the right panel.

 

Validation of Vascular CFD Techniques with In Vitro Models

 

In a recent collaborative study, we compared the numerical solutions obtained with the GMU software against numerical results obtained at UCLA using the commercial software Fidap (Fluent Inc.) and available experimental measurements performed with PIV on a square duct bent 90^o  [45]. This study showed that both solvers provided accurate and consistent solutions (see Figure 2).

 

 

Figure 2: Bent square duct: comparison of velocity profiles computed with GMU software (black lines) and Fidap (blue lines) with experimental results (diamonds), along lines at 0^o, 30^o, 60^o and 90^o (left to right).

 

Deformable model segmentations have been validated in a variety of applications. The reconstruction of a synthetic numerical phantom closely matched the prescribed degree of stenosis while providing a smooth surface model [46] (Figure 3a). A vascular model reconstructed from a CTA image of a glass phantom of a stenosed carotid bifurcation was within less than 5% of the vessel diameters of the glass phantom [47, 48] (Figure 3c). The reconstruction of a stenosed renal artery from a contrast-enhanced MRA image was in excellent agreement with measurements of the vessel dimensions performed directly on the DSA images [41, 46] (Figure 3b). Subsequently, in a study designed to validate the CFD methodology [37, 43, 47], we measured velocity profiles using phase-contrast MR under steady flow conditions below and above the bifurcation of the glass phantom of the carotid artery with stenosis (Figure 3c, left). A comparison of a CFD simulation and the PC-MR measurements showed close agreement between the computed and measured velocity profiles (Figure 3c, right). In another study [41], we constructed a flow through glass phantom of a stenosed renal artery with dimensions derived from a contrast-enhanced MRA image, and found a very strong correlation between the computed and measured differential pressures across the stenosis (Figure 3d).

 

 

Figure 3: a) validation of iso-surface deformable model using synthetic numerical phantom of stenosed vessel, b) validation of iso-surface deformable model with angiographic data of stenosed renal artery, c) validation of CFD velocity profiles with phase-contrast MR measurements in vitro model of stenosed carotid artery, and d) validation of unsteady CFD pressure drops with in vitro model of stenosed renal artery.

 

Validation of Subject-Specific Vascular CFD Models with In Vivo Multi-Modality Image Data

Validation of computational models with in vivo data is problematic since there is no gold standard for measuring cerebrovascular blood flow in humans. Although we have shown good agreement between numerical models and in vivo measurements obtained with PC-MR (Figure 4a) and Doppler ultrasound (DUS)  (Figure 4b) in carotid arteries [3, 40, 42, 49], validating intraaneurysmal flow patterns is far more difficult. Currently, it is only possible to validate these models indirectly. One approach is to predict the signal intensity of time-of-flight (ToF) MRA images of cerebral aneurysms. The fact that these images suffer a signal loss in regions of disturbed flows has been used by Sato et al. [50] to visualize the high inflow regions by plotting iso-intensity surfaces. A computational model created from the image data of Sato et al. [50] showed that, as expected, regions of high signal intensity coincided with regions of high flow velocity [4] (Figure 4c).

 

 

Figure 4: Validation of patient-specific vascular CFD models with in vivo multi-modality image data: a) validation of velocity profiles with phase-contrast MR measurements of a normal carotid artery, b) validation of CFD peak velocity at a stenosed internal carotid artery with Doppler ultrasound, and c) validation of inflow regions in a cerebral aneurysm using time-of-flight MRI image intensity maps.

 

In a recent study [51], we show that patient-specific CFD models can correctly predict the location and shape of the major intra-aneurysmal flow structures that can be identified by conventional angiography. For this purpose, patient-specific models of three cerebral aneurysms were constructed from 3D rotational angiography images and CFD simulations performed. Using the resulting velocity fields, contrast transport was simulated and visualizations created to provide a “virtual” angiogram.  These models were then compared to images from high frame rate conventional angiography to compare flow structures. The CFD simulations showed three distinct flow types ranging from simple to complex. Virtual angiographic images showed good agreement with images from conventional angiography for all three aneurysms with analogous size and orientation of the inflow jet, regions of impaction, and flow type. Large intra-aneurysmal vortices and regions of outflow also corresponded between the images. An example for an ICA aneurysm is presented in Figure 5.

 

 

Figure 5: Model of internal carotid artery aneurysm from 3DRA images, visualizations of intraaneurysmal flow patterns and wall shear stress distribution at three instants during the cardiac cycle, and comparison of high frame rate conventional angiograms and “virtual” angiograms obtained from the CFD simulations.

 

Models of the Circle of Willis from MRA Data

 

We have constructed realistic, patient specific models of the blood flow in the circle of Willis from MRI image data [52, 53]. The anatomical models were constructed from MRA images of normal volunteers using deformable models and surface merging algorithms [54]. The physiologic flow conditions were derived from PC-MR measurements in all the vessels of the Circle of Willis (Figure 6a). Visualizations of the distributions of flow among the vessels comprising the circle of Willis were performed using “virtual angiograms”, a novel technique that simulates the injection and passage of a contrast agent into the blood vessel [36, 52] (Figure 6c). Regions of increased and decreased wall shear stress were observed at arterial bifurcations and regions of high vessel curvature, which are preferred locations for aneurysm development (Figure 6b).

 

Study of Aneurysm Hemodynamics and Rupture

 

Recently, we conducted a pilot study of the association between intraaneurysmal hemodynamic characteristics from CFD models and the rupture of cerebral aneurysms [8]. A total of 62 patient-specific models of cerebral aneurysms were constructed from 3D angiography images.

 

 

Figure 7: Intra-aneurysmal hemodynamic characteristics and rupture: a) Examples of small (left) and large (right) aneurysms with intraaneurysmal flow patterns ranging from simple stable flow patterns (top) to complex unstable flow patterns (bottom); b) examples of aneurysms with small (top) and large (bottom) flow impaction regions; and c) relationship between rupture and flow pattern type (left) and size of the flow impingement region (right).

 

Computational fluid dynamics simulations were performed under pulsatile flow conditions measured on a normal subject. The aneurysms were classified into different categories depending on the complexity and stability of the flow pattern (Figure 7a), the location and size of the flow impingement region (Figure 7b), and the size of the inflow jet. A large variety of flow patterns was observed. Interesting trends in the distribution of ruptured and unruptured aneurysms among these categories were found (Figure 7c). Specifically, 72% of ruptured aneurysms had complex or unstable flow patterns, 80% had small impingement regions and 76% had small jet sizes. Conversely, unruptured aneurysms accounted for 73%, 82% and 75% of aneurysms with simple stable flow patterns, large impingement regions and large jet sizes, respectively.  Aneurysms with small impingement sizes were 6.3 times more likely to have experienced rupture than those with large impingement sizes (p=0.01).

 

Sensitivity Analysis

 

In order to characterize the sensitivity of our aneurysm classification with respect to several modeling parameters, we performed a preliminary sensitivity study using four aneurysms, one of each of the flow types described in the previous section. The results are shown in Figure 8. The left panel shows, from left to right, the effects of changing the viscosity model, increasing and decreasing the total flow on the distribution of wall shear stresses for the four studied aneurysms [11]. The right panel shows the corresponding effects on the intraaneurysmal flow pattern on one of them. In this study we found that the intraaneurysmal hemodynamic characteristics were not significantly affected by changes in the total flow rates, the flow divisions, and the non-Newtonian viscosity model. We also found that the most important factor affecting the intraaneurysmal flow pattern is the geometry of the aneurysm and the parent vessel. In the absence of flow information, sensitivity analyses similar to the one presented in this paper can be conducted in order to ensure a proper characterization of the intraaneurysmal hemodynamics. Special attention must be paid to obtain accurate geometrical models since the flow patterns strongly depend on the shape of the aneurysm sac.

 

 

In another study [55] we showed that models using a truncated parent vessel underestimated the intraaneurysmal WSS and shifted the impaction zone to the neck when compared with the native geometry (Figure 9). These effects were more pronounced in aneurysms where upstream curvature was substantial. We concluded that failure to properly model the inflow stream contributed by the upstream parent artery can significantly influence the results of intra-aneurysmal hemodynamic models.

 

Figure 9: Examples of an ICA (top row) and an MCA aneurysm (bottom row) models including the geometry of the parent vessel (left models) and truncated near the neck (right models). Visualizations of WSS and intraaneurysmal flow patterns illustrate the effects of neglecting the secondary flows induced by the geometry of the upstream parent vessel.

 

Recently, we investigated the effects of vessel wall motion on the hemodynamics of cerebral aneurysms [56, 57]. The motion of the arterial walls was quantified from biplane digital subtraction angiography (DSA) images. Landmark points were manually selected in the first frame and tracked on subsequent frames using non-rigid registration procedures [58-62] (Figure 10c). In one patient, it was found that the amplitude of the bleb deformation was larger (by a factor of about 2) than the rest of the aneurysm sac (Figure 10b and d). This observation confirms similar findings recently reported using 4D-CTA [63] and suggest that the vessel wall may be weaker in the bleb than in the rest of the sac (i.e. there was a localized damage of the wall at the site of the bleb). Computational hemodynamic models were constructed from 3DRA images and wall motion was directly imposed to the CFD models. The CFD calculation revealed a region of elevated WSS in the dome of the aneurysm (Figure 10e), close to the neck of the lobulation, and relatively high WSS inside the lobulation at peak systole (second frame from left of Figure 10f). These characteristics were not significantly affected if a rigid-walled model was used (Figure 10f – bottom row). Therefore, this study suggests that rigid wall models yield a reasonable approximation of the in vivo hemodynamic patterns.

 

 

Figure 10: Study of effects of vessel wall motion on aneurysmal hemodynamics: a) model from 3DRA image, b) regions defined to measure wall motion on angiogram (vessel, sac, lobulation), c) selection and propagation of landmark points using image registration algorithms to measure wall motion, d) motion amplitudes at different time frames for each aneurysm region, e) visualization of regions of elevated WSS at five instants during the cardiac cycle, and f) visualization of WSS distribution using compliant (upper row) and rigid (bottom row) models.

 

In addition, we performed a comparison of the numerical results obtained with CFD models constructed from 3DRA and CTA images of a patient with a basilar tip aneurysm. It was found that the distribution of WSS and intraaneurysmal flow patterns obtained with both models were qualitatively similar (Figure 11). Due to the lower resolution of the CTA images, some of the arteries were of slightly larger caliber in the CTA model, and therefore this model exhibited lower absolute velocities than the 3DRA model for the same inflow rates. This resulted in a difference in the magnitudes of the WSS between the two models. However, similar differences are obtained by the daily variations of the physiologic flow conditions. This study was performed in collaboration between the GMU and the UCLA groups, using UCLA’s image data and GMU’s software.

 

 

Figure 11: Comparison of models constructed from CTA and 3DRA images of a patient with a basilar tip aneurysm: a) 3DRA image, b) 3DRA model, c) WSS in 3DRA model, d) and e) flow pattern in 3DRA model, f) CTA image, g) CTA model, h) WSS in CTA model, i) and j) flow pattern in CTA model. Visualizations are presented at peak systole.

 

Extension of Methodology to Aneurysms with Multiple Feeding Vessels

 

Computational hemodynamics studies of cerebral aneurysms based on 3DRA images have been limited to aneurysms with a single avenue of flow [1, 2, 6, 8, 64]. However, many aneurysms accept blood from two feeding vessels. For instance, aneurysms located in the anterior communicating artery (ACoA) receive blood from the left and right A1 segments of the anterior cerebral arteries, or aneurysms located in the basilar artery (BA) receive blood from the left and right vertebral arteries (VAs). Recently, we developed a novel technique to construct models of such aneurysms from multiple 3DRA images, overcoming this limitation [13, 65]. The idea is to obtain a 3DRA image by contrast injection in each of the feeding vessels, construct corresponding vascular models and then co-register and fuse them into a single watertight model of the aneurysm including all avenues of flow. The methodology is illustrated in Figure 12. This figure shows an aneurysm in the ACoA (a), another in the BA (b), and a model of the entire Circle of Willis of a patient with five cerebral aneurysms (c). This latter model was constructed from three 3DRA images, and illustrates the flexibility of this approach to construct complex arterial networks harboring aneurysms. With this approach, high fidelity patient-specific computational models of cerebral aneurysms accepting blood from more than one feeding vessel can now be constructed. These models are very important in order to understand the effects of unbalanced flows on the intraaneurysmal flow pattern, and may help understand why these aneurysms tend to rupture at a higher rate than aneurysms with a single route of flow.

 

Virtual Interventions

 

Surgical treatment for cerebrovascular diseases (endarterectomy or bypass surgery for atherosclerosis and clipping for cerebral aneurysms) is no longer the first line of therapy, especially for patients at high surgical risk. Minimally invasive endovascular interventions (catheterization) are increasingly used for these patients. These endovascular therapies aim at restoring the normal hemodynamic conditions in the affected arteries. The effectiveness of these interventions depends on the fulfillment of this goal. Therefore, understanding the hemodynamics after endovascular interventions is important for improving and personalizing these procedures. This includes design of better endovascular devices such as stents and coils, and selection of the best treatment option for a given patient (for example selection of the best available stent to divert the blood flow away from the cerebral aneurysm of a given individual).

 

Figure 13: Example of a patient-specific model of a cerebral aneurysm “virtually” stented with two different stent designs.

 

 References

 

1.   Cebral, J.R., M. Hernandez, and A.F. Frangi. Computational analysis of blood flow dynamics in cerebral aneurysms from CTA and 3D rotational angiography image data. in International Congress on Computational Bioengineering. 2003,Vol.1, 191-198, Zaragoza, Spain,Sept. 24-26.

2.   Cebral, J.R., M. Hernandez, A.F. Frangi, C.M. Putman, R. Pergolizzi, and J.E. Burgess. Subject-specific modeling of intracranial aneurysms. in SPIE Medical Imaging. 2004,Vol.5369, 319-327, San Diego, California,Feb. 14-19.

3.   Cebral, J.R., C.M. Putman, R. Pergolizzi, J.E. Burgess, and P.J. Yim. Multi-modality image-based models of carotid artery hemodynamics. in SPIE Medical Imaging. 2004,Vol.5369, 529-538, San Diego, California.

4.   Cebral, J.R., M.A. Castro, T. Satoh, J.E. Burgess, and C.M. Putman. Evaluation of image-based CFD models of cerebral aneurysms using MRI. in ISMRM Flow Motion Workshop. 2004 Zurich, Switzerland,July 11-13.

5.   Cebral, J.R., M.A. Castro, R. Pergolizzi, C.M. Putman, and J.E. Burgess. Clinical applications of computational fluid dynamics in analysis and treatment of cerebral aneurysms. in Bioengineering Consortium Symposium (BECON). 2004 Bethesda, Maryland.

6.   Cebral, J.R., M.A. Castro, J.E. Burgess, and C.M. Putman. Cerebral aneurysm hemodynamics modeling from 3D rotational angiography. in IEEE International Symposium on Biomedical Imaging. 2004 944-947, Arlington, Virginia,April 15-18.

7.   Cebral, J.R., M.A. Castro, J.E. Burgess, R. Pergolizzi, and C.M. Putman, Characterization of Cerebral Aneurysms for Assessing Risk of Rupture Using Patient-Specific Computational Hemodynamics Models. Proc. Eastern Neuroradiological Society (ENRS), Boston, Massachusetts, November 2004, 2004.

8.   Cebral, J.R., M.A. Castro, J.E. Burgess, R. Pergolizzi, M.J. Sheridan, and C.M. Putman, Characterization of cerebral aneurysm for assessing risk of rupture using patient-specific computational hemodynamics models. AJNR American Journal of Neuroradiology, 2005, 26: 2550-2559.

9.   Cebral, J.R., M.A. Castro, D. Millan, A.F. Frangi, and C.M. Putman. Pilot Clinical Investigation of Aneurysm Rupture Using Image-Based Computational Fluid Dynamics Models. in SPIE Medical Imaging. 2005,Vol.5746, 245-256, San Diego, CA,Feb. 12-17.

10.  Cebral, J.R., M.A. Castro, and C.P. Putman, Numerical Simulation of Flow Alterations after Carotid Artery Stenting From Multi-Modality Image Data. Third MIT Conference on Computational Fluid and Solid Mechanics, Boston, MA, 2005.

11.  Cebral, J.R., M.A. Castro, S. Appanaboyina, C.M. Putman, D. Millan, and A.F. Frangi, Efficient pipeline for image-based patient-specific analysis of cerebral aneurysm hemodynamics: Technique and sensitivity. IEEE Transactions in Medical Imaging, 2005, 24(1): 457-467.

12.  Castro, M.A., C.P. Putman, and J.R. Cebral, Application of Vascular CFD for Clinical Evaluation of Cerebral Aneurysms. Third MIT Conference on Computational Fluid and Solid Mechanics, Boston, MA, 2005.

13.  Castro, M.A., C.M. Putman, and J.R. Cebral, Patient-Specific Computational Modeling of Cerebral Aneurysms with Multiple Avenues of Flow from 3D Rotational Angiography Images. IEEE Trans on Medical Imaging, 2005, submitted.

14.  Stehbens, W.E., Intracranial aneurysms, in Pathology of the Cerebral Blood Vessels, 1972, 351-470.

15.  Weir, B., Unruptured intracranial aneurysms: a review. Journal of Neurosurgery, 2002, 96: 3-42.

16.  Foutrakis, G.N., H. Yonas, and R.J. Sclabassi, Saccular aneurysm formation in curved and bifurcation arteries. AJNR American Journal of Neuroradiology, 1999, 20: 1309-1317.

17.  Tomasello, F., D. D'Avella, F.M. Salpietro, and M. Longo, Asymptomatic aneurysms. Literature meta-analysis and indications for treatment. Journal of Neurosurgical Science, 1998, 42(1): 47-51.

18.  Winn, H.R., J.A. Jane, J. Taylor, D. Kaiser, and G.W. Britz, of asymptomatic incidental aneurysms: re-view of 4568 arteriograms. Journal of Neurosurgery, 2002, 96(1): 43-49.

19.  Linn, F.H., G.J. Rinkel, A. Algra, and J. van Gijn, Incidence of subarachnoid hemorrhage: role of region, year, and rate of computed Tomography: a meta-analysis. Stroke, 1996, 27(4): 625-629.

20.  Kaminogo, M., M. Yonekura, and S. Shibata, Incidence and outcome of multiple intracranial aneurysms in a defined population. Stroke, 2003, 34(1): 16-21.

21.  Kim, C., J. Cervos-Navarro, C. Patzold, Y. Tokuriki, Y. Takebe, and K. Hori, In vivo study of flow pattern at human carotid bifurcation with regard to aneurysm development. Acta Neurochir (Wien), 1992, 115(3-4): 112-117.

22.  Kassell, N.F., J.C. Torner, E.C.J. Haley, J.A. Jane, H.P. Adams, and G.L. Kongable, The International Cooperative Study on the Timing of Aneurysm Surgery. Part 1: Overall management results. J Neurosurg, 1990, 73(1): 18-36.

23.  Nishioka, H., J.C. Torner, C.J. Graf, N.F. Kassell, A.L. Sahs, and L.C. Goettler, Cooperative study of intracranial aneurysms and subarachnoid hemorrhage: a long-term prognostic study. II. Ruptured intracranial aneurysms managed conservatively. Arch Neurol, 1984, 41(11): 1142-1146.

24.  White, P.M. and J.M. Wardlaw, Unruptured intracranial aneurysms. J Neuroradiol, 2003, 30(5): 336-350.

25.  Kerber, C.W. and C.B. Heilman, Flow in experimental berry aneurysms: method and model. AJNR Am J Neuroradiol, 1983, 4(3): 374-377.

26.  Kerber, C.W., S.G. Imbesi, and K. Knox, Flow dynamics in a lethal anterior communicating artery aneurysm. AJNR Am J Neuroradiol, 1999, 20(10): 2000-2003.

27.  Gobin, Y.P., J.L. Counard, and P. Flaud, In vitro study of haemodynamics in a giant saccular aneurysm model: influence of flow dynamics in the parent vessel and effects of coil embolization. Neuroradiology, 1994, 36(7): 530-536.

28.  Kyriacou, S.K. and J.D. Humphrey, Influence of size, shape and properties on the mechanics of axisymmetric saccular aneurysms. J Biomech, 1996, 29(8): 1015-1022.

29.  Ortega, H.V., Computer simulation helps predict cerebral aneurysms. J Med Eng Technol, 1998, 22(4): 179-181.

30.  Buonocore, M.H., Visualizing blood flow patterns using streamlines, arrows, and particle paths. Magn Reson Med, 1998, 40(2): 210-226.

31.  Boecher-Schwarz, H.G., K. Ringel, L. Kopacz, A. Heimann, and O. Kempski, Ex vivo study of the physical effect of coils on pressure and flow dynamics in experimental aneurysms. AJNR Am J Neuroradiol, 2000, 21(8): 1532-1536.

32.  Groden, C., J. Laudan, S. Gatchell, and H. Zeumer, Three-dimensional pulsatile flow simulation before and after endovascular coil embolization of a terminal cerebral aneurysm. J Cereb Blood Flow Metab, 2001, 21(12): 1464-1471.

33.  Metcalfe, R.W., The promise of computational fluid dynamics as a tool for delineating therapeutic options in the treatment of aneurysms. AJNR Am J Neuroradiol, 2003, 24(4): 553-554.

34.  Kayembe, K.N.T., M. Sasahara, and F. Hazama, Cerebral aneurysms and variations of the circle of Willis. Stroke, 1984, 15: 846-850.

35.  Cebral, J.R. and R. Löhner, Efficient simulation of blood flow past complex endovascular devices using an adaptive embedding technique. IEEE Transactions in Medical Imaging, 2005, 24(4): 468-477.

36.  Calamante, F., P.J. Yim, and J.R. Cebral, Estimation of bolus dispersion effects in perfusion MRI using image-based computational fluid dynamics. NeuroImage, 2003, 19: 342-352.

37.  Cebral, J.R., P.J. Yim, R. Löhner, O. Soto, H. Marcos, and P.L. Choyke. New methods for computational fluid dynamics of carotid artery from magnetic resonance angiography. in SPIE Medical Imaging. 2001,Vol.4321, 177-187, San Diego, California.

38.  Cebral, J.R., R. Löhner, O. Soto, P.L. Choyke, and P.J. Yim, Patient-specific simulation of carotid artery stenting using computational fluid dynamics. Lecture Notes in Computer Sciences, 2001, 2208: 153-160.

39.  Cebral, J.R., R. Löhner, O. Soto, P.L. Choyke, and P.J. Yim. Image-based finite element modeling of hemodynamics in stenosed carotid artery. in SPIE Medical Imaging. 2002,Vol.4683, 297-304, San Diego, California.

40.  Cebral, J.R., C.M. Putman, and J.E. Burgess. Patient-specific modeling of atherosclerotic carotid arteries from multi-modality image data. in IMECE. 2003 Washington, DC,Nov. 16-24.

41.  Yim, P.J., J.R. Cebral, A. Weaver, R.J. Lutz, O. Soto, G. Boudewijn, B. Vasbinder, V.H. Ho, and P.L. Choyke, Estimation of the differential pressure at renal artery stenoses. Magnetic Resonance in Medicine, 2004, 51: 969-977.

42.  Yim, P.J., J.R. Cebral, R. Löhner, V.H. Ho, and P.L. Choyke. Estimation of mechanical stress on the carotid artery. in Biomedical Engineering Society (BMES). 2001.

43.  Yim, P.J., J.R. Cebral, Y. Zhang, R.J. Lutz, and P.L. Choyke, Evaluation of magnetic resonance angiography for measuring arterial wall shear stress. Annals of Biomedical Engineering, 2000, 28: S-59.

44.  Cebral, J.R. and R. Summers, Tracheal and central bronchial aerodynamics using virtual bronchoscopy and computational fluid dynamics. IEEE Transactions in Medical Imaging, 2004, 28(8): 1021-1033.

45.  He, X., P. Venugopal, J.R. Cebral, H. Schmitt, and D.J. Valentino. Reproducibility of brain hemodynamic simulations: an inter-solver comparison. in SPIE Medical Imaging: Physiology, Function, and Structure from Medical Images. 2006,Vol.6143, 154-165, San Diego, California.

46.  Yim, P.J., B. Vasbinder, V.H. Ho, and P.L. Choyke. A Deformable isosurface and vascular applications. in SPIE Medical Imaging. 2002,Vol.4684, 1390-1397, San Diego, California.

47.  Cebral, J.R., P.J. Yim, R. Löhner, O. Soto, and P.L. Choyke, Blood flow modeling in carotid arteries using computational fluid dynamics and magnetic resonance imaging. Academic Radiology, 2002, 9: 1286-1299.

48.  Yim, P.J., J.R. Cebral, R. Mullick, and P.L. Choyke, Vessel surface reconstruction with a tubular deformable model. IEEE Transactions in Medical Imaging, 2001, 20(12): 1411-1421.

49.  Cebral, J.R. Realistic modeling of arterial hemodynamics from anatomic and physiologic image data. in 14th U.S. National Congress on Theoretical and Applied Mechanics. 2002 Blacksburg, Virginia,June 23-28.

50.  Satoh, T., K. Onoda, and S. Tsuchimoto, Visualization of intraaneurysmal flow patterns with transluminal flow images of 3D MR angiograms in conjunction with aneurysmal configurations. AJNR American Journal of Neuroradiology, 2003, 24: 1436-1445.

51.  Cebral, J.R., R. Pergolizzi, and C.M. Putman, Computational fluid dynamcis modeling of intracranial aneurysms: qualitatively comparison with cerebral angiography. AJNR American Journal of Neuroradiology, 2006, submitted.

52.  Cebral, J.R., M.A. Castro, O. Soto, R. Löhner, and N. Alperin, Blood flow models of the circle of Willis from magnetic resonance data. Journal of Engineering Mathematics, 2003, 47(3-4): 369-386.

53.  Cebral, J.R., R. Löhner, and J.E. Burgess. Computer simulation of cerebral artery clipping: relevance to aneurysm neuro-surgery planning. in ECCOMAS. 2000 Barcelona, Spain,Sept. 11-14.

54.  Cebral, J.R., R. Löhner, P.L. Choyke, and P.J. Yim, Merging of intersecting triangulations for finite element modeling. Journal of Biomechanics, 2001, 34: 815-819.

55.  Castro, M.A., C.M. Putman, and J.R. Cebral, Computational fluid dynamics modeling of intracranial aneurysms: effects of parent artery segmentation on intra-aneurysmal hemodynamcis. AJNR American Journal of Neuroradiology, 2006, 27: 1703-1709.

56.  Dempere-Marco, L., E. Oubel, M.A. Castro, C.M. Putman, A.F. Frangi, and J.R. Cebral, Estimation of wall motion in intracranial aneurysms and its effects on hemodynamic patterns. Lecture Notes in Computer Science, 2006, 4191: 438-445.

57.  Dempere-Marco, L., E. Oubel, A.F. Frangi, C.M. Putman, M.A. Castro, and J.R. Cebral. Wall motion and hemodynamics of intracranial aneurysms. in World Congress on Biomechanics. 2006 Munich, Germany,July 29-August 4.

58.  Rueckert, D., L.I. Sonoda, C. Hayes, D.L.G. Hill, M.O. Leach, and D.J. Hawkes, Non-rigid registration using free-form deformations: application to breast MR images. IEEE Trans on Medical Imaging, 1999, 18(8): 712-721.

59.  Rueckert, D., M.J. Clarkson, D.L.G. Hill, and D.J. Hawkes, Non-rigid registration using higher-order mutual information. SPIE Medical Imaging: Image Processing, San Diego, CA, 2000.

60.  Studholme, C., D.L.G. Hill, and D.J. Hawkes, An overlap invariant entropy measure of 3D medical image alignment. Pattern Recognition, 1999, 32(1): 71-86.

61.  Frangi, A.F., D. Rueckert, J.A. Schnabel, and W.J. Niessen, Automatic construction of multiple-object three-dimensional statistical shape models: Application to cardiac modeling. IEEE Trans on Medical Imaging, 2002, 21(9): 1151-1166.

62.  Frangi, A.F., M. Laclaustra, and P. Lamata, A registration-based approach to quantify flow-mediated dilation (FMD) of the brachial artery in ultrasound image sequences. IEEE Trans on Medical Imaging, 2003, 22: 1458-1469.

63.  Ishida, F., H. Ogawa, T. Simizu, T. Kojima, and W. Taki, Visualizing the dynamics of cerebral aneurysms with four-dimensional computed tomography angiography. Neurosurgery, 2005, 57: 460-471.

64.  Steinman, D.A., J.S. Milner, C.J. Norley, S.P. Lownie, and D.W. Holdworth, Image-based computational simulation of flow dynamics in a giant intracranial aneurysm. AJNR American Journal of Neuroradiology, 2003, 24(4): 559-566.

65.  Castro, M.A., C.P. Putman, and J.R. Cebral, Computational Modeling of Cerebral Aneurysms in Arterial Networks Reconstructed from Multiple 3D Rotational Angiography Images. SPIE Medical Imaging Conference, San Diego, CA, 2005, 5746: 233-244.