Cerebral aneurysms 1

Introduction: Between 3% and 5% of the adult population may be diagnosed with an intracranial cerebral aneurysm anatomically featured by protruding sacs in specific cerebrovascular sites. The hemodynamics of cerebral aneurysms could be dominated by the complex geometries of aneurysms and blood vessels, and variable flow conditions. Many studies have examined the association between hemodynamic factors and the geometries [1-2] as well as boundary conditions [3-4]. However, how age-associated alterations in waveforms influence the intra-aneurysmal hemodynamics has rarely been explored.

Methods: In the present study, computational fluid dynamic (CFD) simulations were conducted for 45 cerebral aneurysms (13 ruptured vs 32 unruptured) located at the internal carotid artery (ICA) with typical young and older adult waveforms. Obtained results were statistically analyzed to compare hemodynamic characteristics in the ruptured and unruptured aneurysms, with emphasis placed on examining the sensitivity of hemodynamic factors to varying waveforms and associated frequency harmonics.

Results: Statistically, time-averaged wall shear stress (TAWSS), maximum WSS (MaxWSS), relative residence time (RRT), low WSS area (LSA), and pressure loss coefficient (PLC) of both rupture and unruptured aneurysms were found insensitive to the variation of young and older adult waveforms (P > 0.05). In comparison with young-adult waveform, a significant increase in oscillatory shear index (OSI) was observed under inflow condition of older-adult waveform (P < 0.05), with average differences of 54.98% and 51.46% for ruptured and unruptured aneurysms respectively.Further results shown that the WSS oscillations were dominated by the low-frequency spectrum, especially the 0th and 1st harmonics.

Discussion: In summary, our study demonstrated the strong effects of older and young waveforms on OSI magnitudes, but minimal influence on the time-averaged hemodynamic factors. This suggests that special care should be taken on the fluctuating parameters for the evaluation of rupture risk involved a large number of cerebral aneurysms.

Acknowledgements: FYL was supported by the National Natural Science Foundation of China (Grant no. 81370438).

References

1. Cebral JR, Mut F, Weir J, et al. Quantitative characterization of the hemodynamic environment in ruptured and unruptured brain aneurysms. American Journal of Neuroradiology 2011; 32: 145-51.
2. Xiang J, Natarajan SK, Tremmel M, et al. Hemodynamic–morphologic discriminants for intracranial aneurysm rupture. Stroke 2011; 42: 144-52.
3. Liang F, Liu X, Yamaguchi R, et al. Sensitivity of flow patterns in aneurysms on the anterior communicating artery to anatomic variations of the cerebral arterial network. Journal of biomechanics 2016; 49: 3731-40.
4. Sarrami-Foroushani A, Lassila T, Gooya A, et al. Uncertainty quantification of wall shear stress in intracranial aneurysms using a data-driven statistical model of systemic blood flow variability. Journal of Biomechanics 2016; 49: 3815-23.

Introduction

While many studies have associated aneurysm growth with rupture, the pathophysiology of aneurysms evolution is poorly understood. Hemodynamics is thought to be an important factor in the progressive degeneration and remodeling of the aneurysm wall. However, the changes in the aneurysm flow condition during its evolution are poorly understood [1]. Our objective was to study the changes in hemodynamic characteristics of cerebral aneurysms during their evolution and to investigate whether those changes are different between ruptured and unruptured aneurysms.

 

Methods

A total of 20 cerebral aneurysms at a single location (the anterior communicating artery) were studied with image-based computational fluid dynamics (CFD). For each aneurysm, a synthetic sequence approximating the aneurysm geometry at six earlier stages was generated by shrinking the aneurysm sac through Laplacian smoothing while keeping the neck fixed. A second sequence was generated for each aneurysm by selecting from our database aneurysms at the same location and with similar shape but smaller size and slightly smaller neck. Shape similarity was computed as the Euclidean distance in Zernike moment shape space [2]. CFD simulations were carried out for each aneurysm and each simulated evolution stage under similar flow boundary conditions. A total of 19 flow parameters were computed to characterize the aneurysmal hemodynamic environment and compare their values at different stages and between different aneurysms.

 

Results

When the aneurysm neck was kept fixed (first sequence), the mean velocity, viscous dissipation, wall shear stress (WSS), kinetic energy, vorticity, inflow rate and shear rate decreased as the aneurysms enlarged. However, when the aneurysm neck enlarged along with the sac (although at a slower pace), these parameters did not always decrease. Furthermore, WSS was observed to increase in ruptured aneurysms but to decrease or remain roughly the same in unruptured aneurysms (p = 0.039). Other variables that had different slopes in ruptured and unruptured aneurysm sequences are shown in Fig.1. These results indicate that different hemodynamic characteristics may be associated with aneurysm evolution and future rupture.

 

Discussion

Understanding cerebral aneurysm evolution is important for improved patient management since it enables a more precise rupture risk assessment. In this study we focused on a single location and observed differences in the evolution of flow characteristics of ruptured and unruptured aneurysms. These results suggest that different local flow conditions may predispose aneurysm for further degradation and ultimately rupture, or wall remodeling and stabilization. Future studies will focus on extending these results to other locations.

 

References

1. Sforza, Daniel M., Christopher M. Putman, and Juan Raul Cebral. "Hemodynamics of cerebral aneurysms." Annual review of fluid mechanics 41 (2009): 91-107.
2. Millán, Raul D., et al. "Morphological characterization of intracranial aneurysms using 3-D moment invariants." IEEE transactions on medical imaging 26.9 (2007): 1270-1282.

Introduction

    Predicting intracranial aneurysm evolution is one of the most significant biomechanics challenges because it depends critically on the interplay between hemodynamics stresses and vessel wall remodeling.  Current diagnostic imaging relies mostly on angiography, whether conventional digital subtraction angiography, or non-invasive (MRI, CT) techniques. These techniques do not provide vessel wall evaluation, limiting aneurysm characterization to aneurysm size and luminal appearance. A new combined approach is proposed here that incorporates patient-specific CFD modeling with a new imaging technique developed for direct visualization of the vessel wall and incorporation of vessel wall degradation information[1-2]: intracranial vessel-wall MRI (VWMRI). This study combines the strengths of both techniques to understand the hemodynamics role in vessel wall degradation, by independently correlating quantitative measures from CFD and VWMRI to qualitative pathology assessment of resected aneurysms.

Methods

    Ten consecutive patients with an unstable unruptured intracranial aneurysm (UIA), based on clinical and conventional luminal imaging assessment, scheduled for aneurysm clipping were recruited for the study. Patients underwent multi-contrast 3D VWMRI within 1 week prior to aneurysm clipping. 3D lumen reconstructions were created from contrast enhanced imaging using in-house software. Computational Fluid Dynamics simulations were performed using patient waveforms, with scaled and normalized to match 1.5 Pa average shear stress at the vessel inlets, and Resistive-Capacitance boundary conditions applied at vessel outlets. A combination of manual and iterative closest points registration was performed to align the VWMRI and CFD results. Each saccular histology slice was partitioned into four quadrants for qualitative assessments of inflammatory infiltration and wall thickness features. Histology slices registered to the VWMRI were averaged axially for the number of histology slices aligned within matching VWMRI slice.

Results

    Results show vessel wall thinning correlates strongly with high TAWSS regions and negatively with oscillatory flow. Figure 1 depicts a typical result for one of the ten patients: the comparison between the VWMRI against OSI, TAWSS and DWSS shows wall thinning at high TAWSS and neighboring high OSI and impingement (high instantaneous WSS) zones.

Discussion

    There is qualitative evidence that VWMRI can be augmented with CFD results to improve the understanding of the degraded state of the vessel wall. Validation of these correlation with histology data will make it possible to quantitatively study the role of hemodynamic stresses on aneurysmal wall remodeling in vivo, by following patients that are not treated longitudinally. The potential for combining hemodynamic information with clinical imaging can lead to a translational holy grail: quantitative assessment of rupture risk. Additional patient data and statistical analysis of the histology information for the 10-patient cohort will be presented to fully quantify this relationship.

Introduction

Intracranial side aneurysms (IA) are pathological blood-filled bulges in cerebral blood vessels. Though material transport is known to play a dominant role in aneurysm growth and risk for rupture [1], there is lack of experimental studies addressing the mass transport features of aneurysms. In this work, we focus on experimentally studying, in a systematic manner, the convective vs. diffusive features of aneurysms with different geometries and subjected to a well-defined flow.

Methods

Bolus washout experiments, of either non-diffusive fluorescent particles (1 micron) or small diffusive molecules (fluorescein), were performed in transparent silicone phantom models of side aneurysms. The models, were connected to a custom perfusion system which produced well defined flow profiles. Following bolus injection, Normalized Fluorescent Intensity (NFI) maps were acquired and NFI curves within a defined Region of Interest (ROI) were extracted to evaluate the mass distribution inside the cavity, see Figure 1a.  Additionally, computational simulations were performed to examine the flow field and mass transport under the studied conditions.

Results and Discussion

As expected, aneurysms with larger aspect ratios (AR) have slower washout kinetics (35% increased clearance time of an AR3.2 compared to AR1.6). Additionally, parent vessel geometry was shown to have a significant effect on the washout time of both particles and dye and are significantly larger for geometries where the cavity is located inside the artery bend compared geometries with outer position of the same aspect ratio, see Figure 1B. In larger AR aneurysm models (AR 3.2) low Peclet number regions have been identified in the apex and vortex center of the aneurysm, where particles were retained for a long period of time.

Figure 1: (a) Representative experimental results of time series fluorescence intensity maps following bolus injection of the dye solution. The maps show the filling and washout of dye in the aneurysm cavity (b) Normalized fluorescence intensity (NFI) curves at different aneurysm geometries following either dye or particle injections

Discussion

Altogether our results suggest that diffusive and non-diffusive materials interact differently with saccular side aneurysms and that there are regions within saccular aneurysms prone to non-diffusive mass accumulation which may correlate with zones of plaque and blood clot formation.

 

References

[1] Lieber, B. B., & Gounis, M. J. (2002). The physics of endoluminal stenting in the treatment of cerebrovascular aneurysms. Neurological Research, 24 Suppl 1(April), S33-42. https://doi.org/10.1179/016164102101200014

 

The majority of cerebral aneurysms are thought to occur at vessel bifurcations, however minimally-invasive treatment options for aneurysms in these locations remain limited.[1] The effective treatment of cerebral bifurcation aneurysms by flow-diversion is a promising technique, but it remains challenging and controversial in clinical practice.[2] In particular, achieving complete aneurysm isolation without jailing daughter vessel branches that supply blood to other brain regions is impossible with conventional device designs.

In an extension of previous analysis of the novel Sphere flow-diverter device[3], this study virtually compares the performance of the dedicated bifurcation aneurysm device to treatment with conventional flow-diverter devices. Both novel and conventional flow-diverter device designs are virtually deployed to treat 13 cerebral bifurcation aneurysms in 21 unique treatment scenarios. Transient CFD simulations are performed to compare a number of flow and wall shear stress (WSS) based metrics that have been correlated with successful aneurysm occlusion. Both designs of novel device and the conventional flow-diverter are found to substantially reduce aneurysm inflow and WSS following deployment, with typical reductions of around 55% and 75% for inflow and mean WSS respectively. The second generation Sphere device, Sphere II, is shown to offer equivalent performance in all treatment metrics considered, when compared to a conventional flow diverter device, with a significance level of P=0.05.

 

[1] Alfano, J. M. et al “Intracranial Aneurysms Occur More Frequently at Bifurcation Sites that Typically Experience Higher Hemodynamic Stresses” Neurosurgery 73(3):497-505,  2013

[2] Caroff, J. et al “Flow-Diverter Stents for the Treatment of Saccular Middle Cerebral Artery Bifurcation Aneurysms” Am J Neuroradiol. 37(2):279-284, 2016

[3] Peach, T. W. et al “The ‘Sphere’: A Dedicated Bifurcation Aneurysm Flow-Diverter Device,” Cardiovasc. Eng. Technol. 5(4):334–347, 2014

Introduction: We have been developing a microporous covered stent (MCS [1]) for the embolization treatment of intracranial aneurysms (IAs). IA embolization mechanism by the stenting is based on thrombus formation in IA, induced by flow stagnation in IA after shielding the neck of IA. Therefore, change in flow intensity (ex. shear rate) in IA before and after the stenting is closely related to IA embolization activity. In this study, we investigated the shear rate in an aneurysm model depending on geometry of an aneurysm or curvature of a parent vessel, and flow change in an aneurysm by MCS or flow diverter (FD), a commercially available device for IAs treatment.

Methods: IA flow was visualized by particle image velocimetry (PIV) method using in vitro flow simulator [2]. Aneurysm depth, dome size and neck width of 2-dimensional IA model (thickness: 5 mm) was varied on basic size of 9.1 mm, 10.0 mm and 7.1 mm, respectively. Curvature ratio of parent vessel (width: 5 mm) was varied 0 to 0.5. Stent models, imitated general expandable stent for supporting a stenotic vessel (bare stent, BS), FD or MCS, were placed between IA and parent vessel model.

Results: Area mean shear rate of IA flow tend to higher for IA with smaller volume and/or higher aspect ratio (aneurysm depth / neck width). In measurement with high curvature parent vessel, velocity and shear rate of IA flow was significantly high due to main flow from parent vessel directly entered (Fig. 1A). Following placing stent model, area mean shear rate was reduced to 35% at BS, 9% at FD and 3% at MCS in the measurement with basic size IA, straight parent vessel (Fig. 1B).

 

Fig.1: (A) Velocity (vector) and shear rate (color scale) in IA flow. (B) Reduction of area mean shear rate in IA flow by placing stent models. Reynolds number of parent vessel’s flow was about 600.

 

Discussion: These PIV results indicate that IA with smaller volume, wider-neck and higher curvature ratio parent vessel need high shielding device for IA embolization treatment. Conversely, giant and/or high aspect ratio IA is relative easy to treat by stent placement. It was considerable that MCS has flow reduction property about 3 times FD, and more reliable IA embolization can be expected in the treatment of IA with fast flow.

 

Acknowledgements

This study was funded in part by a Grant-in-Aid for Young Scientist (B) (15K19985) from the Japan Society for the Promotion of Science and grants from Ministry of Health, Labour and Welfare of Japan.

Reference

1. Nakayama, Y., et al., (2016). J Artif Organs, 19(2) p179

2. Tajikawa, T., et al., (2013). Trans Jap Soc Mech Eng B, 79(801) p992

The authors have been developing a honeycomb microporous covered stent as a newly device for cerebral aneurysm treatment. This stent has a porous film to control both embolization and rapid neointimal formation.  According to previous study, it is expected that the neointimal formation becomes earlier by using much larger pores.  But there is a possibility that the covered stent with large pares cannot embolize by increasing inflow from the parent vessel.  The purpose of this study is estimation of stent embolization and optimization of pore design to ensure compatibility between reliable embolization and controlling early neointimal formation, in-vitro experiment under the steady flow condition based on flow similarity law has been conducted by using a 2D parent vessel model with various saccular aneurysm models and various simplified micro-porous film models as shown in Fig. 1.  The flow visualization result shows that three flow patterns in the aneurysm cavity were observed namely, viscous shear induced swirl flow, friction pressure loss induced semi-swirl flow and those coexistence and transitional flow.  This result indicated that viscous shear force caused by WSS drives blood in the aneurysm cavity through the micro-pores on the film.  The authors had calculated moment of inertia of blood in the cavity and its driving force (accurately torque) due to WSS.  The estimating result showed that the ratio of the driving torque to the moment of inertia showed good agreement with the amount of shear rate in the aneurysm cavity flow.  As the result of those theoretical analysis and in-vitro experiment as shown in Fig. 2, the optimized pore design could introduced; the covered stent with 60% porosity and 200μm in pore diameter is one of the best balance embolization and neointimal formation.  Finally, to confirm the optimized pore design, follow-up animal experiments have been conducted.  The result of stenting, a newly designed stent can be embolized immediately just after stenting.

Acknowledgments

A part of this research was supported by JSPS. KAKENHI (15K19985) and the Health Labour Sciences Research Grant.

References

[1] Tajikawa T, et. al., Trans. JSME Series B，79(801), (2013)，pp.992-1004.

Introduction

 Detailed hemodynamic analysis of blood flow in pathological segments close to lesion has provided physicians with invaluable information about the local flow patterns leading to vascular disease. However, these diseases have both local and distal effects on the circulation of blood within the cerebral tree. This study quantifies the hemodynamic changes that occur throughout the entire arterial tree due to an aneurysm before and after treatment.

Method

 A 74‑year‑old female with a saccular aneurysm in the right posterior inferior cerebellar artery was treated with aneurysm clipping. Time-dependent blood flow data and anatomical information were obtained using time-of-flight angiography and magnetic resonance angiography, at pre‑intervention (PRI), and post‑intervention (PSI). The measured blood flow profiles were used as velocity boundary conditions in ANSYS 18.1. To enable large-scale 3D modeling, we used a hybrid meshing technique at the aneurysm vessel interface to enable a combination of highly dense unstructured meshes with economic reconstruction of parametric mesh generation¹ for the rest of the vascular tree. The reconstructed computational models for the large portion of arterial tree are shown in Figure 1A.

Results and Discussion

 Figure 1B illustrates the inflow jet stream into the aneurysm sac on different planes oriented perpendicularly and horizontally to the aneurysm ostium before treatment. High-risk prolonged RRT (relative residence time) sites were located on a bleb close to the separation of the inflow streamline in the aneurysm sac.

The waveform shapes are depicted from lesion site to downstream distal vessels in Figure 1C. In the post-aneurysm region, there was a phase lag and an augmentation of peak‑diastolic velocity in the pre-intervention model.

We also calculated the hemodynamic indexes for waveform analysis using pulsatility index (PI), resistance index (RI), systole to diastole ratio (S/D) of the velocity profile. Waveform indexes such as PI, RI and S/D ratio reduced by 16.3%, 11.52%, and 18.86% after aneurysm clipping, respectively.

Conclusion

Since in vivo blood flow acquisition is not simultaneous, such phase shift differences cannot be studied with current in vivo blood flow measurements. The lesion regions not only affect blood flow streamline of the proximal sites, but also generate pulse wave shift and disturbed flow in downstream vessels. Distal dynamic boundary condition provides wave propagation and maintains more realistic phase lag between blood flow and pressure and for a large portion of arterial tree². This requires the use of large-scale simulation to visualize both local and large-scale effects of pathological lesions.

Acknowledgments

This project was supported by NIH-1R21NS099896 and NSF-CBET-1301198.

References

Ghaffari, M. et al. Comput. Biol. Med. 91, 353–365 (2017).

Olufsen, M. S. et al, Am. J. Physiol. - Heart Circ. Physiol. 276, 257–268 (1999).

Computational fluid dynamics (CFD) based assessment of cerebral aneurysms has received much attention in the last decade. The usability of these methods depends on the quality of the computations, highlighted in recent discussions. The purpose of this study is to investigate the convergence of common hemodynamic indicators with respect to numerical resolution and to propose methods for improving the boundary conditions.

Methods

First, 38 middle cerebral artery bifurcation aneurysms were studied at two different resolutions (one comparable to most studies, and one finer) [1]. Relevant hemodynamic indicators were collected from two of the most cited studies, and were compared at the two refinements. Second, we present an algorithm for deriving the boundary conditions from 4D PC-MRI using a data assimilation algorithm and assess the robustness with respect to noise [2].

Results 

Most of the hemodynamic indicators were very well resolved at the coarser resolutions, correlating with the finest resolution with a correlation coefficient >0.95. The oscillatory shear index (OSI) had the lowest correlation coefficient of 0.83. However, there are noticeable variations in the proportion of the aneurysm under low shear, as well as in spatial and temporal gradients not captured by the correlation alone. The data assimilation algorithm was robust with respect to Gaussian noise, even up to signal-to-noise ratio of 1.

Conclusion

Statistically, hemodynamic indicators agree well across the different resolutions studied here. However, there are clear outliers visible in several of the hemodynamic indicators, which suggests that special care should be taken when considering individual assessment. The proposed data assimilation technique is robust with high levels of noise and may improve the boundary conditions.

References

[1] Evju, Øyvind, et al. "Robustness of common hemodynamic indicators with respect to numerical resolution in 38 middle cerebral artery aneurysms." PloS one 12.6 (2017): e0177566.

[2] Funke, S. W., et al. "Variational data assimilation for transient blood flow simulations." arXiv preprint arXiv:1607.03013(2016).

Introduction

Flow-diverter stent becomes a primary tool for the treatment of wide-necked complex aneurysms due to high occlusion rate and low post-treatment complications [1]. However in some cases the flow-diverter deployment procedure associated with various technical complications, resulting in unpredictable flow alterations in aneurysm or in adjacent vessels, which could potentially affect the clinical outcome. In this study we investigate the flow alterations in cerebral artery induced by incomplete expansion of the distal end of the flow-diverter stent and correlate these alterations with a known clinical outcome.

 

Methods

A hemodynamics in a patient-specific model of the middle cerebral artery before and after the flow-diverter placement was retrospectively studied by numerical simulations. Two deployment scenarios were analyzed: with complete and incomplete (50%) expansion of the distal end of the stent. A geometrical model of the flow-diverter used in clinics was created according to the manufacturer specification and deployed in the aneurysm model by fast virtual stenting technique [2]. A realistic inlet velocity curve was imposed at the inlet of the model, which corresponds to the physiological flow rate in the middle cerebral artery. The CFD simulations were performed using computational resources of “Lomonosov” supercomputer and high performance technology MPI.

 

Results

The parameter values obtained for the non-treated aneurysm were used as reference to evaluate the flow alterations induced by stenting. We found that the resultant flow field in the aneurysm sac was similar for both studied deployment scenarios. In both cases the average intra-aneurysmal flow was sufficiently reduced to initiate thrombosis. The velocity and wall shear stress (WSS) distributions as well as streamlines in the aneurysm sac were not significantly influenced by incomplete expansion of the distal end of the flow-diverter compared to the fully expanded one. However for the incompletely expanded stent we found a zone of non-physiologically high WSS in the middle cerebral artery near the distal end of the stent, which was not observed for the fully expanded stent scenario (Fig. 1). This high WSS values could affect the endothelial cells in this region and initiate the wall remodeling, which could lead to post-treatment complications after the flow-diverter deployment.

Discussion

The results correlate with available clinical outcome for the studied case. The endothelial hyperplasia was observed during the post-treatment period, possibly caused by abnormal WSS, which was detected by numerical simulations. The results of the study could be used by physicians for prediction of possible post-treatment complications after incomplete flow-diverter expansion and further planning of therapeutic procedures.

 

Acknowledgements

This work was supported by Russian Science Foundation (Project 16-15-10327).

 

References

1. Walcott. B., et al., (2016). JAMA Neurol, 73(8) p1002

2. Berg, P., et al., (2017). In: Computing and Visualization for Intravascular Imaging and Computer-Assisted Stenting, p371

Although rupture risk assessment for intracranial aneurysms based on hemodynamic simulations is addressed in numerous recent research studies, the acceptance among physicians remains limited. In the majority of these studies, surface parameters, e.g., wall shear stress or oscillatory shear, are considered exclusively [1]. However, the corresponding image-based blood flow simulations, which provide highly-resolved time-dependent information, are mostly discarded.

To account for the analysis of hemodynamic phenomena that cause certain loads on the luminal surface of the aneurysm, a new dynamic line filtering approach is applied. Therefore, unsteady blood flow simulations are carried out in three ruptured intracranial aneurysms. As a reference, three unruptured aneurysms were chosen to allow for a comparison between both groups [2]. Within the analysis, flow structures are coarsely filtered based on their distance to user selected local surface parameter extrema and visualized as pathlines. The filtered line bundles are color-coded according to the selected surface region, allowing for a visual assessment of the overall flow course. Each bundle can be further refined by mapping flow parameters such as velocity or vorticity to color, opacity or line thickness. Additional filtering based on parameter range selections is available using scatterplots and parallel coordinate views, respectively.

As indicated in Figure 1 for one ruptured and one unruptured aneurysm, individual flow structures can be analyzed based on interesting hemodynamic surface parameters. Hence, complex flow interactions that lead to effects such as increased oscillatory shear can be separated. Additionally, stable flow phenomena, which are observed in unruptured aneurysms, are confirmed allowing a better classification between the two groups.

The exploration of the blood flow within intracranial aneurysm using surface patch-based dynamic line filtering enables a profound analysis of relevant flow structures. Especially the consideration of conspicuous hemodynamic surface parameters allows for a better evaluation of the causing flow structures. Hence, our approach provides detailed information towards a better understanding of the rupture risk and therefore an improved therapy planning for physicians.

Figure 1: Representative illustration of one ruptured (R) and one unruptured (UR) intracranial aneurysm. Dynamic line filtering is used to identify flow structures causing interesting surface parameters.

Acknowledgements

The work is funded by the Federal Ministry of Education and Research in Germany within the Research Campus STIMULATE under Grant No. 13GW0095A and the European Regional Development Fund under the operation number ’ZS /2016/04/78123’ as part of the initiative "Sachsen-Anhalt WISSENSCHAFT Schwerpunkte".

 

References

[1] Chung B, Cebral JR (2015) CFD for evaluation and treatment planning of aneurysms: review of proposed clinical uses and their challenges. ANN BIOMED ENG, 43(1): 122–138. doi: 10.1007/s10439-014-1093-6

[2] Berg P, Beuing O (2017) Multiple intracranial aneurysms: A direct hemodynamic comparison between ruptured and unruptured vessel malformations, J COMPUT ASSIST RADIOL SURG, 10.1007/s11548-017-1643-0

Introduction
Cerebral aneurysms (CA) are often treated with flow-diverting stents (FDS) to reduce blood flow into the aneurysm sac, promoting the development of a stable thrombus. Successful treatment is highly dependent on the degree of flow reduction and the altered hemodynamics inside the aneurysm following treatment[1]. Establishing a causal connection between hemodynamic metrics of FDS-treated CAs and long-term clinical outcomes requires a rigorous parametric characterization of this flow environment.

 

Methods
In this study, we use 3D particle image velocimetry to measure the flow field inside of idealized silicone aneurysm models treated with commercially-available FDS. Four side-branch aneurysm models are created with a parent-vessel radius (R_PV) of 2mm and an aneurysm diameter of 7mm. Parent-vessel curvature (κ) at the aneurysm neck for each of the models is 0.0, 0.03, 0.07 and 0.11 mm^-1. Each model is treated with a 4x20mm Pipeline stent. The flow-field is measured under both steady and pulsatile flow conditions, with parent-vessel flow-rates ranging from 100-400 mL/min. Using heart rates of 0 (steady), 30, 60 and 90 BPM, we explore a wide range of the relevant non-dimensional parameters: Reynolds: 150-600, Womersley: 0-7, Dean = Re√κR_PV :0-280.

 

Results
The velocity field measured at the aneurysm mid-plane under steady-flow conditions is shown in figure 1A (flow: right-left), with parent-vessel curvature increasing across columns and parent-vessel flow-rate increasing down rows. At low values of the parent-vessel De (upper left), the flow enters the aneurysm through the proximal neck region and exits distally (circulating clockwise). However, as De increases, the flow begins to separate at the leading edge and recirculate (circulating counterclockwise), until, at a De > 180, a single counter-rotating vortex exists inside the sac. The Re number at the aneurysm neck, based on the area-averaged neck velocity neck and neck diameter, is plotted against parent-vessel De number in figure 1B, where the magnitude of aneurysm velocity is shown to be highly dependent on De number. Under pulsatile flow, the inertia of the flow inside the aneurysm is strengthened, and thus the parent-vessel De required to induce counter-rotation of flow inside the aneurysm is higher than it is under steady flow conditions.

Discussion
In this study, we show that the structure and magnitude of flow entering aneurysms treated with FDS is dominated by parent-vessel De and Wo. The presence of a strong coherent vortex inside the aneurysmal sac has been shown to negatively effect stable thrombus formation, and thus act against the successful treatment of CAs. These results improve our understanding of FDS hemodynamics, providing a first principles approach to characterizing aneurysms for treatment selection, and enabling more predictive modeling of treatment progression.


Acknowledgments
NIH 1R01NS088072

References
[1]Oured,R J.NeuroIntventSurg0:1-6(2016)

Intracranial aneurysm (IA) is a ‘ballooning’ of the vascular lumen of an intracranial artery that occurs most frequently at the branching points of the major arteries. Aneurysm rupture is believed to occur when the wall stress exceeds the strength of the wall tissue. Therefore, knowledge of the stress distribution in an intact IA wall could be valuable in assessing its risk of rupture. Finite element method is useful tool for noninvasive estimation the in vivo wall stress distribution for patient-specific IAs. However, microstructure of aneurysmal wall has been shown to vary among patients. While all IA walls are characterized by lack of elastic lamina, some of the IA walls had plenty of disorganized mural cells, whereas some of the walls had lost most of them [1]. For that reason, in this study we investigate possible physiological ranges of wall stress distribution of human intracranial aneurysm depending wall microstructure, as well as the role of active stress from smooth muscle cells (SMCs).

Three-dimensional patient-specific intracranial aneurysm and surrounding healthy arteries were reconstructed from digital subtraction angiography (DSA) imaging. The geometry was imported into finite element analysis program (FEAP), and discretized with hexahedral finite elements. A nonlinear mixture material model from [2], consisting of isotropic elastin, four collagen fiber families and smooth muscle cells, was used to describe mechanical behavior of soft tissue. It was assumed that healthy arteries are comprised of all three structurally important constituents, whereas in aneurysmal part either only collagen remained or collagen and some SMCs [1]. Stresses were computed under the mean blood loading. Because of the constant turnover of collagen it is likely that its pre-stretch in the stable IAs is at homeostatic value, whereas in the enlarging aneurysms it would be increased. Orientations of collagen in IAs were studied in several studies [3, 4]. Stiffness was estimated such that displacements in loaded configuration remain zero, i.e. simulated geometry correlates to the imaging.

Loss of mural cells is believed to shift the balance between “repair and maintenance” and wall degeneration leading to rupture, yet in this study we aim to investigate the role of SMCs on wall stress distribution, and thus growth of aneurysm.

Acknowledgements

This work has been fully supported by Croatian Science Foundation under the project IP-2014-09-7382.

References

1. J. Frösen, R. Tulamo, A. Paetau, E. Laaksamo, M. Korja, A. Laakso, M. Niemela, J. Hernesniemi, Acta Neuropathol. 123 (2012).
2. J. D. Humphrey, K.R. Rajagopal, Math Method Appl Sci 12(2002).
3. D.J. MacDonald, H.M. Finlay, P.B. Canham, Ann. Biomed. Eng. 28 (2000).
4. B.K. Wicker, H.P. Hutchens, Q. Wu, A.T. Yeh, J.D. Humphrey, Comput. Methods Biomech. Biomed. Engin. 11(2008).

The symmetric anatomic configuration of the Circle of Willis (CoW)  is encountered in 22,99% of healthy population. There are many anatomical variations of the CoW such as the missing of one or both posterior communication arteries, case encountered in 15,20% from the analyzed population. In the present study we will approach the transitive behavior of the hemodynamics in the CoW. Due to the fact that the action of the heart is pulsating, the flow through the CoW arteries will have variations during the cardiac cycle. In the human physiology, the arterial walls are deformed by the blood flow. In the case of small arteries this deformation can be neglected.

Due to the interaction of the fluid with the artery’s wall, pressure waves appear in the sanguine flow. The arterial wall has a complex structure and the aim of the present study is to model the CoW from the point of view of the unsteady flow in the presence of flexible arterial walls. For this reason in the present paper will be studied the interaction problems of fluid-structure based on the numerical analysis. The study is structured in two parts: 1) the fluid-structure approach based on the blood and arteries properties and on the cerebral-vascular characteristics and 2) the flow and simulation based on the geometric and computational model, numerical approach and the flow’s characteristics.

Introduction: Histological studies have shown that intracranial aneurysms can have very different wall characteristics [1]. Some aneurysm walls exhibit hyperplastic remodeling characterized by cell proliferation and wall thickening, while others have undergone decellularization and wall thinning. These characteristics are often observed during surgery as changes in the appearance (color) in regions of thin or thickened walls [2]. Flow conditions have been associated with histological changes of the aneurysm wall and inflammation [3]. Our study focused on the association of local flow conditions and focal remodeling of the aneurysm wall.

Methods: This study included 65 aneurysms that were treated surgically, and for which intra-operative videos and pre-surgical 3D images were available. Computational fluid dynamics models with patient-specific geometries were constructed from the 3D images. The following regions of the aneurysm wall were identified in the intra-operative videos according to their visual appearance, and virtually labeled on the CFD model: a) atherosclerotic regions (yellow), b) hyperplastic regions (white), and c) thin regions (red). The remainder of the aneurysm wall was labeled as “normal” looking. Local flow conditions in each region were characterized by calculating several variables on the aneurysm wall: wall shear stress (WSS), oscillatory shear index (OSI), relative residence time (RRT), WSS gradient and divergence, and gradient oscillatory number (GON). Regions were statistically compared using the Mann-Whitney test.

Results: Hyperplastic regions had lower mean WSS (p=0.006) and pressure (p=0.009) than normal looking regions. Atherosclerotic and hyperplastic regions look quite similar, only minimum pressure was higher in atherosclerotic regions (p=0.03). However, thick regions had higher RRT (p<0.001), OSI (p=0.03), and GON (p=0.006) than thin regions. Additionally, thin regions had lower RRT (p<0.001), OSI (p=0.02), and GON (p<0.001) than normal looking regions, and had higher WSS (p=0.006) and pressure (p=0.009) than hyperplastic regions. Qualitatively, thin regions tend to be aligned with the flow stream in the aneurysm, while thick (atherosclerotic and hyperplastic) regions tend to be aligned with the ends of the intrasaccular vortices (normal to the flow stream). A representative example is presented in Fig.1.

Conclusions: The flow conditions in thickened aneurysm wall regions (atherosclerotic and hyperplastic), are characterized by low and oscillatory WSS produced by slow and recirculating flows located near vortex ends towards the sides of the intra-aneurysmal flow stream. In contrast, thin regions are aligned with the flow stream and are subjected to faster flows with higher and less oscillatory WSS and WSS gradients. Local flow conditions could in principle be used to predict regions of the wall prone for adaptive remodeling (thickening) or degeneration (thinning).

References
1. Kataoka, Taneda, et al. Stroke, 30, 1396-1401, 1999.
2. Kadasi, Dent, Malek, JNIS, 5, 201-6, 2013.
3. Cebral, Ollikainen, Chung, et al. AJNR, 38(1), 119-126, 2017.

Funding: NIH R01NS097457

Computational fluid dynamic simulation (CFD) of cerebral artery and aneurysm began to be reported in early 2000s. In 2002, Oshima et al succeeded in the segmentation of tortuous internal carotid artery from medical images and visualized complex secondary flows using CFD. In the next year, Steinmann et al published the first report of image-based CFD of a real cerebral aneurysm. These epoch making studies may be the sources of the today's CFD.

 

CFD started its role as a research tool. At first, the complex flow patterns in cerebral aneurysms were depicted qualitatively in writing. Then, some researchers combined the research technique in medicine with that in engineering. That is, hemodynamic parameters, such as wall shear stress, were calculated using CFD from multiple cases and the statistics was used to draw a generalized conclusion. This approach drew attentions from many medial and engineering researchers and led to the prosperity of CFD.

 

As a clinical tool, CFD was expected to predict the rupture risk of cerebral aneurysms in early 2000s. There were some misunderstandings about CFD in physicians, who felt some atmosphere of future prediction in the word of "simulation". As all engineers know, CFD simulates only the spatial distributions of velocity and pressure in complex geometry and CFD are not calculating the future. But the engineers didn't correct the misunderstanding of physicians at that time. After some discourage, physicians began to demand accuracy and validity from CFD. It is difficult to meet this demand since there is no standard methods to measure the blood velocity with sufficient accuracy. In such timing, an innovation came from the CFD of coronary arteries. Heart flow FFR-ct team overcame the problem of accuracy and validity in a smart manner. They measured pressure instead of velocity. Pressure is much easier to measure with higher accuracy. In addition, to validate the CFD they conducted clinical trials. Physician does not care how to measure each item of blood examination if the results are reproducible and useful in the diagnosis. Heart flow FFR-ct caused physicians to become aware that high technology such as fluid structure interaction and data assimilation are not necessary for CFD. As a clinical tool, CFD does not have to be validated with real measurements. Clinical usefulness based on clinical trials are more preferred by physicians.

 

There may be two future directions for CFD. As a research tool, it will be more complex and more sophisticated, and as a clinical tool, it will be simpler and  more robust. There are many researchers on the former side, but I would like more researchers to be on the latter side.

Introduction

Direct assessment of hemodynamic parameters as potential indicators for unruptured intracranial aneurysm (UIA) rupture risk is difficult to realize. Alternatively, investigating their relationship with surrogate markers for rupture-risk may be able to reveal their potential clinical efficacy. In this study we related selective hemodynamic parameters to aneurysm wall enhancement (AWE) [1] and to the PHASES score, a coded sum of pertinent clinical risk factors [2]. 

 

Methods

From 257 consecutive UIA patients presenting in 2016 at the Changhai Hospital affiliated to the Second Military Medical University, 65 sacular UIA with AWE and three-dimensional subtraction angiography data available were selected. PHASES score was used to classify aneurysms into three groups according to rupture risk less than 1 %, 1 % to 3 % and larger than 3 % in 5 years. Investigated morphological parameters included size and size-related quantities (figure 1). AWE images were acquired as post-contrast T1-weighted images (in-plane resolution 0.4 m; slice thickness 1.5 mm) immediately after intravenous injection of Gd-DTPA (0.1mmol/kg). AWE images were differentiated as no (NAWE), partial (PAWE) and circular (CAWE) enhancement. From transient CFD simulations, hemodynamic parameters were derived including normalized wall shear stress (NWSS), percent area of low WSS (LSA), oscillatory shear index (OSI) and relative residence time (RRT) (figure1). Statistical significance (p<0.05) between groups was assessed by the Kruskal-Wallis H test.

 

Results

With increasing PHASES score, size-related parameters increased and NWSS decreased significantly (figure 1). LSA, OSI and RRT increased significantly. Number of NAWE aneurysms decreased and PAWE and CAWE increased significantly.

Discussion

In agreement with previous studies, which demonstrated presence of inflammatory cell infiltrations associated with aneurysm wall weakening and rupture [3], our results demonstrate increase of AWE with rupture risk as expressed by the PHASES score. WSS, OSI and RRT varied significantly with rupture risk, thereby supporting their use for identifying high-risk aneurysms.  As AWE imaging might not be available in standard clinical care of aneurysm patients or may not be feasible due to contraindications, CFD simulations serve as an alternative approach to identify high-risk aneurysms in need of immediate treatment. Our results were derived from a retrospective review and a prospective study is warranted to further explore the clinical significance of our results.

 

References

 

1. Edjlali M, et al.  Stroke. 2014;45:3704-3706
2. Greving JP, et al. Lancet Neurol. 2014;13:59-66
3. Frosen J, et al. Stroke. 2004;35:2287-2293

Introduction: Current methods of medical imaging provide key information for the clinical assessment of intracranial aneurysms. To date, broadly accepted quantitative criteria to compare aneurysm morphology are yet lacking. In this study, we relate established shape descriptors to expert assessments of aneurysm irregularity in view of establishing clinically meaningful descriptors.

Method: To address morphology in an isolated manner, from 3D angiographies, we produced replicas of aneurysms and adjacent segments of the parent vascular tree, extracted computed geometry indices and measures based on curvature, surface writhe number and geometry moments. Independently, experts evaluated the morphology of these aneurysms in a 3D viewer by inspecting their shapes and rating the degree of irregularity on a 9-point rating scale. Besides these 3D views, no further information was given to the raters. We then examined the univariate and multivariate correlations between the aggregated ratings and the quantitative descriptions of the 3D models.

Results: Preliminary results are based on 134 aneurysm models and 15 raters, who all assessed the perceived irregularity of the aneurysm shape. Univariate analysis shows that curvature metrics reproduce the rater assessment of shape irregularity the best (Spearman correlation ρ=0.86, p<0.001). Furthermore, a writhe-based index (entropy of the writhe number distribution, ρ=0.75, p<0.001) and shape indices (non-sphericity, ρ=0.71, p<0.001) predict the ratings on average well. Consistent with observational data, our rater assessment of irregularity degree increased with increasing aneurysm size (ρ=0.78, p<0.001).

Discussion: We identified and validated quantitative measures mirroring well the qualitative raters’ assessments of morphological irregularity characteristics. In the future, such quantitative measures could be used in rating schemes for aneurysm assessment. Furthermore, these metrics could be employed in the context of a databank, allowing for retrieving and grouping cases of similar morphological properties.

Introduction
The pathophysiology of intracranial saccular aneurysms is associated with hemodynamics, but the role of hemodynamics is still poorly understood [1]. This study aims at clarifying the mechanism of wall degeneration and rupture by combining the hemodynamics with aneurysmal wall pathology.

Methods
An inter-aneurysmal lesion heterogeneity, or translucent, hypertrophic, atherosclerotic, was compared with hemodynamics by CFD (unruptured 23 cases). Harvested tissues were analyzed in wall thickness by Micro-CT, endothelial cells by SEM and TEM, smooth muscle cells, inflammatory cells, and extracellular matrixes by histological and immunochemical analyses (HE, EM, α-SMA and CD68 staining) (unruptred 26 and ruptured 5 cases).   

Results
As for unruptured, translucent and atherosclerotic lesions were correlated with flow impingement and stasis, respectively, with a detection rate of approximately 80%. Wall-thinning lesions (<100 μm) showed (1) endothelial cells were mostly absent or seriously injured (P<0.05), (2) smooth muscle cells were sparsely distributed due to cell death (P<0.05), (3) no macrophage was identified at all, and (4) platelets and leucocytes adhered to luminal surfaces sometimes with fibrin networks. The occupancy rate of wall-thinning lesions was greater in ruptured cases (P<0.01) (Fig.1 (A)). The rupture location corresponded to wall-thinning lesions (5/5). Those lesions showed sparse smooth muscle cells and no macrophage, being coincident with those of unruptured. Macrophages were only observed in thick walls  (>100 μm) or thrombus (Fig.1 (B)). The ruptured wall-thinning lesions showed an infiltration of blood cells (4/5) and heterogeneous luminal thrombus of organized, organizing and fresh constitutes (5/5).   

Discussion
Flow impingements were found to prevent wall regeneration, which was characterized with a proliferation of smooth muscle cells, by hindering thrombus formation. As aneurysms enlarged over time by blood pressure, persistent flow impingements led wall-thinning lesions to be progressively extended, finally with a vulnerability due to over-extension. The vulnerability behaved like a microscopic dissection with blood-cell infiltrations, leading to form a spot thrombus. This event may be associated with a minor leak [2] and proceeds before rupture by considering a period of time required for thrombus organization. The growing spot thrombus was unfortunatelly aligned with flow impingements, and thereby became eroded fluid-dynamically. This flow-induced thrombus destabilization led underlying wall-thinning lesions to be disrupted and eventually ruptured.       

Conclusion
Flow impingement plays a primary role of wall thinning, wall vulnerability, thrombus destabilization, and an eventual rupture of intracranial saccular aneurysms. This type of wall degeneration was biomechanically induced without any biochemical proteolysis and autolysis due to inflammatory cells and thrombus formation.

Acknowledgements
This research was supported by a Health Science Research Grant from the Ministry of Health, Labour, and Welfare, Japan (H23-IRYOUKIKI-IPAN-006)(23090401).

References
[1] Frösen J et al., Acta Neuropathol (2012) 123: 773-786
[2] Frösen J et al., Stroke (2004) 35: 2287-2293.

INTRODUCTION: Is the morphology of a cerebral aneurysm a factor in any clinical outcome? Be that rupture of an aneurysm under observation or during flow diverter deployment or recurrence of an aneurysm that is coil embolized? How should one go about finding out? First, we need to find ways to quantify morphology with metrics and then test the prognostic capability of the metrics in adequately powered patient populations. If the first step - morphology quantification - is ineffective, it will adversely impact the fidelity of the second step - test prognostic capability of morphology. The first step is easier said than done. What exactly constitutes morphology? How many morphological features are there? How do we mathematically capture them with metrics? We submit that morphology quantification would be effective when we identify a minimum number of continuous metrics that are all physically meaningful, intuitive, non-redundant, determinable through various imaging modalities (uni-plane, bi-plane and volumetric) and describe aneurysm morphology adequately. Our group has spent about 15 years studying these questions and we propose here, an effective way to quantify morphology.

METHODS: We conducted a study of about 300 unruptured cerebral aneurysms from three independent populations. We catalogued all the aneurysm morphology metrics reported in the literature and evaluated each from a perspective of robustness and intuitiveness in definition, user sensitivity and through factor analysis. We also developed new metrics to describe aneurysm shape that avoid using boundary based variables in favor of space filling variables which we demonstrate to be especially less sensitive to image resolution.

RESULTS: We propose that cerebral aneurysms be described with five metrics that describe sac size, sac ellipticity, sac irregularity, sac size relative to vessel size and sac size relative to neck size in that order of importance as demonstrated by factor analysis which showed these indices describing 48%, 16%, 13%, 9% and 8% of the variances respectively such that they collectively describe 93% of the variance in the patient population. We present precise mathematical definitions for metrics that capture these features for both algorithmic and manual approaches such that they may be estimated highly precisely using volumetric image data or approximately using uni- or biplane angiograms and/or by visual observations.

CONCLUSION: We show that the story of morphology in the patient population can be told almost exhaustively by these five essential features and their corresponding metrics. Future studies seeking to study the prognostic utility of aneurysm morphology for any given clinical outcome may benefit by testing for the predictability of these five metrics.

Introduction

A previous high-resolution CFD study, using peak-systolic steady-inflow conditions, had found incidental correlation between flow instability in 5/12 intracranial aneurysms (IAs) and rupture status [1]. On the other hand, Varble et al., using mean steady-inflow conditions, reported lower prevalence of instabilities (8/56 IAs) and no correlation with rupture status [2]. Thus, the prevalence, nature, and correlation of these flow instabilities with rupture is still unclear. The primary goal of this study is to establish the prevalence and nature of flow instabilities in IAs using direct numerical simulations (DNS) of pulsatile flows, with the secondary goal of correlating hemodynamics with rupture status.

Methods

We simulated all of the bifurcation IAs from the Aneurisk database, totaling 55 IAs. For the purpose of this abstract, we have focused on 20/55 cases. We generated 3 million (3M) tetrahedral element meshes, run with second-order velocity shape functions, which are at least equivalent to ~24M linear tetrahedra (range:11-55M), resulting in nominal sac resolution of 0.06 mm. Simulations were run for 3 cycles of 20,000 time-steps each. The inflow rates were scaled to the ICA inlet area using u=0.27m/s, which resulted in a mean inflow rate of 3.8+0.9mL/s, consistent with in-vivo measured value of 4.08+1.0mL/s [3]. We post-processed TAWSS, OSI, and spectral-power index (SPI), our recently-proposed metric for quantifying high-frequency content in WSS and/or velocity field [4].

As shown in panel A, for unruptured IAs, only cases 78 and 71 show notable instabilities, while for the ruptured IAs, all cases, except case 94, show flow instabilities.  The SPI maps shown in panel B correlate well with the flow instabilities in panel A, and broadly indicate the presence of higher SPI in the ruptured cases. However, looking at the box-plots inset in panel B, no significant trends are evident.

Discussion

Approximately 45% of the N=20 cases showed flow instabilities, in agreement with Valen-Sendstad et al. [1], but inconsistent with Varble et al. [2], which could be due to their use of mean inflow rates. Secondly, we did not find significant correlation between flow instabilities and rupture status, at least when using 25Hz to quantify instabilities in the SPI maps. However, since SPI lumps together all frequencies >25Hz, it is possible that other frequency bands, for example those associated with aneurysm bruits (>100Hz), maybe correlated with rupture. Thus, we will further investigate discrete frequency bands to compute SPI. Lastly, to understand the inconsistency between Varble et al.’s vs. our findings, we will simulate a subset of IAs at mean inflow rates.

References

1. Valen-Sendstad et al. J. Biomech. 46 (2013).
2. Varble et al. J. Biomech. Eng. 138 (2016).
3. Hoi et al. Physiol Meas. 31 (2010).
4. Khan et al. J. Biomech. 52 (2017).