Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH): uncertainty quantification of geometric rupture risk parameters.

BACKGROUND: Geometric parameters have been proposed for prediction of cerebral aneurysm rupture risk. Predicting the rupture risk for incidentally detected unruptured aneurysms could help clinicians in their treatment decision. However, assessment of geometric parameters depends on several factors, including the spatial resolution of the imaging modality used and the chosen reconstruction procedure. The aim of this study was to investigate the uncertainty of a variety of previously proposed geometric parameters for rupture risk assessment, caused by variability of reconstruction procedures. MATERIALS: 26 research groups provided segmentations and surface reconstructions of five cerebral aneurysms as part of the Multiple Aneurysms AnaTomy CHallenge (MATCH) 2018. 40 dimensional and non-dimensional geometric parameters, describing aneurysm size, neck size, and irregularity of aneurysm shape, were computed. The medians as well as the absolute and relative uncertainties of the parameters were calculated. Additionally, linear regression analysis was performed on the absolute uncertainties and the median parameter values. RESULTS: A large variability of relative uncertainties in the range between 3.9 and 179.8% was found. Linear regression analysis indicates that some parameters capture similar geometric aspects. The lowest uncertainties < 6% were found for the non-dimensional parameters isoperimetric ratio, convexity ratio, and ellipticity index. Uncertainty of 2D and 3D size parameters was significantly higher than uncertainty of 1D parameters. The most extreme uncertainties > 80% were found for some curvature parameters. CONCLUSIONS: Uncertainty analysis is essential on the road to clinical translation and use of rupture risk prediction models. Uncertainty quantification of geometric rupture risk parameters provided by this study may help support development of future rupture risk prediction models.

A review on the reliability of hemodynamic modeling in intracranial aneurysms: why computational fluid dynamics alone cannot solve the equation.

Computational blood flow modeling in intracranial aneurysms (IAs) has enormous potential for the assessment of highly resolved hemodynamics and derived wall stresses. This results in an improved knowledge in important research fields, such as rupture risk assessment and treatment optimization. However, due to the requirement of assumptions and simplifications, its applicability in a clinical context remains limited.This review article focuses on the main aspects along the interdisciplinary modeling chain and highlights the circumstance that computational fluid dynamics (CFD) simulations are embedded in a multiprocess workflow. These aspects include imaging-related steps, the setup of realistic hemodynamic simulations, and the analysis of multidimensional computational results. To condense the broad knowledge, specific recommendations are provided at the end of each subsection.Overall, various individual substudies exist in the literature that have evaluated relevant technical aspects. In this regard, the importance of precise vessel segmentations for the simulation outcome is emphasized. Furthermore, the accuracy of the computational model strongly depends on the specific research question. Additionally, standardization in the context of flow analysis is required to enable an objective comparison of research findings and to avoid confusion within the medical community. Finally, uncertainty quantification and validation studies should always accompany numerical investigations.In conclusion, this review aims for an improved awareness among physicians regarding potential sources of error in hemodynamic modeling for IAs. Although CFD is a powerful methodology, it cannot provide reliable information, if pre- and postsimulation steps are inaccurately carried out. From this, future studies can be critically evaluated and real benefits can be differentiated from results that have been acquired based on technically inaccurate procedures.

A Review of Hemolysis Prediction Models for Computational Fluid Dynamics.

Flow-induced hemolysis is a crucial issue for many biomedical applications; in particular, it is an essential issue for the development of blood-transporting devices such as left ventricular assist devices, and other types of blood pumps. In order to estimate red blood cell (RBC) damage in blood flows, many models have been proposed in the past. Most models have been validated by their respective authors. However, the accuracy and the validity range of these models remains unclear. In this work, the most established hemolysis models compatible with computational fluid dynamics of full-scale devices are described and assessed by comparing two selected reference experiments: a simple rheometric flow and a more complex hemodialytic flow through a needle. The quantitative comparisons show very large deviations concerning hemolysis predictions, depending on the model and model parameter. In light of the current results, two simple power-law models deliver the best compromise between computational efficiency and obtained accuracy. Finally, hemolysis has been computed in an axial blood pump. The reconstructed geometry of a HeartMate II shows that hemolysis occurs mainly at the tip and leading edge of the rotor blades, as well as at the leading edge of the diffusor vanes.

Computational Fluid Dynamics-Based Design Optimization Method for Archimedes Screw Blood Pumps.

An optimization method suitable for improving the performance of Archimedes screw axial rotary blood pumps is described in the present article. In order to achieve a more robust design and to save computational resources, this method combines the advantages of the established pump design theory with modern computer-aided, computational fluid dynamics (CFD)-based design optimization (CFD-O) relying on evolutionary algorithms and computational fluid dynamics. The main purposes of this project are to: (i) integrate pump design theory within the already existing CFD-based optimization; (ii) demonstrate that the resulting procedure is suitable for optimizing an Archimedes screw blood pump in terms of efficiency. Results obtained in this study demonstrate that the developed tool is able to meet both objectives. Finally, the resulting level of hemolysis can be numerically assessed for the optimal design, as hemolysis is an issue of overwhelming importance for blood pumps.

The Computational Fluid Dynamics Rupture Challenge 2013--Phase II: Variability of Hemodynamic Simulations in Two Intracranial Aneurysms.

With the increased availability of computational resources, the past decade has seen a rise in the use of computational fluid dynamics (CFD) for medical applications. There has been an increase in the application of CFD to attempt to predict the rupture of intracranial aneurysms, however, while many hemodynamic parameters can be obtained from these computations, to date, no consistent methodology for the prediction of the rupture has been identified. One particular challenge to CFD is that many factors contribute to its accuracy; the mesh resolution and spatial/temporal discretization can alone contribute to a variation in accuracy. This failure to identify the importance of these factors and identify a methodology for the prediction of ruptures has limited the acceptance of CFD among physicians for rupture prediction. The International CFD Rupture Challenge 2013 seeks to comment on the sensitivity of these various CFD assumptions to predict the rupture by undertaking a comparison of the rupture and blood-flow predictions from a wide range of independent participants utilizing a range of CFD approaches. Twenty-six groups from 15 countries took part in the challenge. Participants were provided with surface models of two intracranial aneurysms and asked to carry out the corresponding hemodynamics simulations, free to choose their own mesh, solver, and temporal discretization. They were requested to submit velocity and pressure predictions along the centerline and on specified planes. The first phase of the challenge, described in a separate paper, was aimed at predicting which of the two aneurysms had previously ruptured and where the rupture site was located. The second phase, described in this paper, aims to assess the variability of the solutions and the sensitivity to the modeling assumptions. Participants were free to choose boundary conditions in the first phase, whereas they were prescribed in the second phase but all other CFD modeling parameters were not prescribed. In order to compare the computational results of one representative group with experimental results, steady-flow measurements using particle image velocimetry (PIV) were carried out in a silicone model of one of the provided aneurysms. Approximately 80% of the participating groups generated similar results. Both velocity and pressure computations were in good agreement with each other for cycle-averaged and peak-systolic predictions. Most apparent "outliers" (results that stand out of the collective) were observed to have underestimated velocity levels compared to the majority of solutions, but nevertheless identified comparable flow structures. In only two cases, the results deviate by over 35% from the mean solution of all the participants. Results of steady CFD simulations of the representative group and PIV experiments were in good agreement. The study demonstrated that while a range of numerical schemes, mesh resolution, and solvers was used, similar flow predictions were observed in the majority of cases. To further validate the computational results, it is suggested that time-dependent measurements should be conducted in the future. However, it is recognized that this study does not include the biological aspects of the aneurysm, which needs to be considered to be able to more precisely identify the specific rupture risk of an intracranial aneurysm.

Cerebral blood flow in a healthy Circle of Willis and two intracranial aneurysms: computational fluid dynamics versus four-dimensional phase-contrast magnetic resonance imaging.

Computational fluid dynamics (CFD) opens up multiple opportunities to investigate the hemodynamics of the human vascular system. However, due to numerous assumptions the acceptance of CFD among physicians is still limited in practice and validation through comparison is mandatory. Time-dependent quantitative phase-contrast magnetic resonance imaging PC-MRI measurements in a healthy volunteer and two intracranial aneurysms were carried out at 3 and 7 Tesla. Based on the acquired images, three-dimensional (3D) models of the aneurysms were reconstructed and used for the numerical simulations. Flow information from the MR measurements were applied as boundary conditions. The four-dimensional (4D) velocity fields obtained by CFD and MRI were qualitatively as well as quantitatively compared including cut planes and vector analyses. For all cases a high similarity of the velocity patterns was observed. Additionally, the quantitative analysis revealed a good agreement between CFD and MRI. Deviations were caused by minor differences between the reconstructed vessel models and the actual lumen. The comparisons between diastole and systole indicate that relative differences between MRI and CFD are intensified with increasing velocity. The findings of this study lead to the conclusion that CFD and MRI agree well in predicting intracranial velocities when realistic geometries and boundary conditions are provided. Due to the considerably higher temporal and spatial resolution of CFD compared to MRI, complex flow patterns can be further investigated in order to evaluate their role with respect to aneurysm formation or rupture. Nevertheless, special care is required regarding the vessel reconstruction since the geometry has a major impact on the subsequent numerical results.

On flow fluctuations in ruptured and unruptured intracranial aneurysms: resolved numerical study.

Flow fluctuations have emerged as a promising hemodynamic metric for understanding of hemodynamics in intracranial aneurysms. Several investigations have reported flow instabilities using numerical tools. In this study, the occurrence of flow fluctuations is investigated using either Newtonian or non-Newtonian fluid models in five patient-specific intracranial aneurysms using high-resolution lattice Boltzmann simulation methods. Flow instabilities are quantified by computing power spectral density, proper orthogonal decomposition, and fluctuating kinetic energy of velocity fluctuations. Our simulations reveal substantial flow instabilities in two of the ruptured aneurysms, where the pulsatile inflow through the neck leads to hydrodynamic instability, particularly around the rupture position, throughout the entire cardiac cycle. In other monitoring points, the flow instability is primarily observed during the deceleration phase; typically, the fluctuations begin just after peak systole, gradually decay, and the flow returns to its original, laminar pulsatile state during diastole. Additionally, we assess the rheological impact on flow dynamics. The disparity between Newtonian and non-Newtonian outcomes remains minimal in unruptured aneurysms, with less than a 5% difference in key metrics. However, in ruptured cases, adopting a non-Newtonian model yields a substantial increase in the fluctuations within the aneurysm sac, with up to a 30% higher fluctuating kinetic energy compared to the Newtonian model. The study highlights the importance of using appropriate high-resolution simulations and non-Newtonian models to capture flow fluctuation characteristics that may be critical for assessing aneurysm rupture risk.

Image-based hemodynamic simulations for intracranial aneurysms: the impact of complex vasculature.

PURPOSE: Hemodynamics play an important role in the assessment of intracranial aneurysm (IA) development and rupture risk. The purpose of this study was to examine the impact of complex vasculatures onto the intra-vessel and intra-aneurysmal blood flow. METHODS: Complex segmentation of a subject-specific, 60-outlet and 3-inlet circle of Willis model captured with 7T magnetic resonance imaging was performed. This model was trimmed to a 10-outlet model version. Two patient-specific IAs were added onto both models yielding two pathological versions, and image-based blood flow simulations of the four resulting cases were carried out. To capture the differences between complex and trimmed model, time-averaged and centerline velocities were compared. The assessment of intra-saccular blood flow within the IAs involved the evaluation of wall shear stresses (WSS) at the IA wall and neck inflow rates (NIR). RESULTS: Lower flow values are observed in the majority of the complex model. However, at specific locations (left middle cerebral artery 0.5 m/s, left posterior cerebral artery 0.25 m/s), higher flow rates were visible when compared to the trimmed counterpart. Furthermore, at the centerlines the total velocity values reveal differences up to 0.15 m/s. In the IAs, the reduction in the neck inflow rate and WSS in the complex model was observed for the first IA (IA-A δNIRmean = - 0.07ml/s, PCA.l δWSSmean = - 0.05 Pa). The second IA featured an increase in the neck inflow rate and WSS (IA-B δNIRmean = 0.04 ml/s, PCA.l δWSSmean = 0.07 Pa). CONCLUSION: Both the magnitude and shape of the flow distribution vary depending on the model's complexity. The magnitude is primarily influenced by the global vessel model, while the shape is determined by the local structure. Furthermore, intra-aneurysmal flow strongly depends on the location in the vessel tree, emphasizing the need for complex model geometries for realistic hemodynamic assessment and rupture risk analysis.

Variability of computational fluid dynamics solutions for pressure and flow in a giant aneurysm: the ASME 2012 Summer Bioengineering Conference CFD Challenge.

Stimulated by a recent controversy regarding pressure drops predicted in a giant aneurysm with a proximal stenosis, the present study sought to assess variability in the prediction of pressures and flow by a wide variety of research groups. In phase I, lumen geometry, flow rates, and fluid properties were specified, leaving each research group to choose their solver, discretization, and solution strategies. Variability was assessed by having each group interpolate their results onto a standardized mesh and centerline. For phase II, a physical model of the geometry was constructed, from which pressure and flow rates were measured. Groups repeated their simulations using a geometry reconstructed from a micro-computed tomography (CT) scan of the physical model with the measured flow rates and fluid properties. Phase I results from 25 groups demonstrated remarkable consistency in the pressure patterns, with the majority predicting peak systolic pressure drops within 8% of each other. Aneurysm sac flow patterns were more variable with only a few groups reporting peak systolic flow instabilities owing to their use of high temporal resolutions. Variability for phase II was comparable, and the median predicted pressure drops were within a few millimeters of mercury of the measured values but only after accounting for submillimeter errors in the reconstruction of the life-sized flow model from micro-CT. In summary, pressure can be predicted with consistency by CFD across a wide range of solvers and solution strategies, but this may not hold true for specific flow patterns or derived quantities. Future challenges are needed and should focus on hemodynamic quantities thought to be of clinical interest.

Is Accurate Lumen Segmentation More Important than Outlet Boundary Condition in Image-Based Blood Flow Simulations for Intracranial Aneurysms?

PURPOSE: Image-based blood flow simulations are increasingly used to investigate the hemodynamics in intracranial aneurysms (IAs). However, a strong variability in segmentation approaches as well as the absence of individualized boundary conditions (BCs) influence the quality of these simulation results leading to imprecision and decreased reliability. This study aims to analyze these influences on relevant hemodynamic parameters within IAs. METHODS: As a follow-up study of an international multiple aneurysms challenge, the segmentation results of five IAs differing in size and location were investigated. Specifically, five possible outlet BCs were considered in each of the IAs. These are comprised of the zero-pressure condition (BC1), a flow distribution based on Murray's law with the exponents n = 2 (BC2) and n = 3 (BC3) as well as two advanced flow-splitting models considering the real vessels by including circular cross sections (BC4) or anatomical cross sections (BC5), respectively. In total, 120 time-dependent blood flow simulations were analyzed qualitatively and quantitatively, focusing on five representative intra-aneurysmal flow and five shear parameters such as vorticity and wall shear stress. RESULTS: The outlet BC variation revealed substantial differences. Higher shear stresses (up to Δ9.69 Pa), intrasaccular velocities (up to Δ0.15 m/s) and vorticities (up to Δ629.22 1/s) were detected when advanced flow-splitting was applied compared to the widely used zero-pressure BC. The tendency of outlets BCs to over- or underestimate hemodynamic parameters is consistent across different segmentations of a single aneurysm model. Segmentation-induced variability reaches Δ19.58 Pa, Δ0.42 m/s and Δ957.27 1/s, respectively. Excluding low fidelity segmentations, however, (a) reduces the deviation drastically (>43%) and (b) leads to a lower impact of the outlet BC on hemodynamic predictions. CONCLUSION: With a more realistic lumen segmentation, the influence of the BC on the resulting hemodynamics is decreased. A realistic lumen segmentation can be ensured, e.g., by using high-resolved 2D images. Furthermore, the selection of an advanced outflow-splitting model is advised and the use of a zero-pressure BC and BC based on Murray's law with exponent n = 3 should be avoided.

In vitro and in silico assessment of flow modulation after deploying the Contour Neurovascular System in intracranial aneurysm models.

BACKGROUND: The novel Contour Neurovascular System (Contour) has been reported to be efficient and safe for the treatment of intracranial, wide-necked bifurcation aneurysms. Flow in the aneurysm and posterior cerebral arteries (PCAs) after Contour deployment has not been analyzed in detail yet. However, this information is crucial for predicting aneurysm treatment outcomes. METHODS: Time-resolved three-dimensional velocity maps in 14 combinations of patient-based basilar tip aneurysm models with and without Contour devices (sizes between 5 and 14 mm) were analyzed using four-dimensionsal (4D) flow MRI and numerical/image-based flow simulations. A complex virtual processing pipeline was developed to mimic the experimental shape and position of the Contour together with the simulations. RESULTS: On average, the Contour significantly reduced intra-aneurysmal flow velocity by 67% (mean w/ = 0.03m/s; mean w/o = 0.12m/s; p-value=0.002), and the time-averaged wall shear stress by more than 87% (mean w/ = 0.17Pa; mean w/o = 1.35Pa; p-value=0.002), as observed by numerical simulations. Furthermore, a significant reduction in flow (P<0.01) was confirmed by the neck inflow rate, kinetic energy, and inflow concentration index after Contour deployment. Notably, device size has a stronger effect on reducing flow than device positioning. However, positioning affected flow in the PCAs, while being robust in effectively reducing flow. CONCLUSIONS: This study showed the high efficacy of the Contour device in reducing flow within aneurysms regardless of the exact position. However, we observed an effect on the flow in PCAs, which needs to be investigated further.

Integrated biophysical matching of bacterial nanocellulose coronary artery bypass grafts towards bioinspired artery typical functions.

Revascularization via coronary artery bypass grafting (CABG) to treat cardiovascular disease is established as one of the most important lifesaving surgical techniques worldwide. But the shortage in functionally self-adaptive autologous arteries leads to circumstances where the clinical reality must deal with fighting pathologies coming from the mismatching biophysical functionality of more available venous grafts. Synthetic biomaterial-based CABG grafts did not make it to the market yet, what is mostly due to technical hurdles in matching biophysical properties to the complex demands of the CABG niche. But bacterial Nanocellulose (BNC) Hydrogels derived by growing biofilms hold a naturally integrative character in function-giving properties by its freedom in designing form and intrinsic fiber architecture. In this study we use this integral to combine impacts on the luminal fiber matrix, biomechanical properties and the reciprocal stimulation of microtopography and induced flow patterns, to investigate biomimetic and artificial designs on their bio-functional effects. Therefore, we produced tubular BNC-hydrogels at distinctive designs, characterized the structural and biomechanical properties and subjected them to in vitro endothelial colonization in bioreactor assisted perfusion cultivation. Results showed clearly improved functional properties and gave an indication of successfully realized stimulation by artery-typical helical flow patterns.

CFD simulations of inhalation through a subject-specific human larynx - Impact of the unilateral vocal fold immobility.

BACKGROUND: The larynx of the human respiratory tract plays a vital role in breathing and voice production. Both can be influenced by functional and/or morphological changes of the larynx, e.g., immobility of one or both vocal folds (VF). The immobile VF can become stationary in different positions such as the median, paramedian, intermediate or lateral position. The impact of unilateral vocal fold immobility (UVFI) on inhalation is the focus of this study. METHODS: Transient numerical simulations of the inhalation process in patient-specific airways are performed. Five configurations are considered: paramedian and intermediate VF positions on the left and right, and healthy. Large eddy simulations are used to describe the complex laryngeal turbulent flow. Airway resistance, power loss, and spectral entropy are calculated to quantify the work of inspiration and evaluate flow regimes. RESULTS: The laryngeal jet intensity and flow disturbance increase with the severity of immobility. In comparison to the healthy configuration, UVFI with right/left intermediate and right/left paramedian VF position increases the airway resistance over the oropharynx to the trachea by 69%/58% and 310%/285%, respectively. When the entire respiratory system is considered, an increase of up to 48% is estimated. Spectral entropy increases of up to 2.5 times indicate higher turbulence levels due to UVFI. CONCLUSIONS: Surgery of immobile VF aims to improve glottis closure. However, this can have a negative impact on breathing efficiency. To that end, this study provides initial insights into the conflicting objectives of open versus closed VFs.

Impact of stents and flow diverters on hemodynamics in idealized aneurysm models.

Cerebral aneurysms constitute a major medical challenge as treatment options are limited and often associated with high risks. Statistically, up to 3% of patients with a brain aneurysm may suffer from bleeding for each year of life. Eight percent of all strokes are caused by ruptured aneurysms. In order to prevent this rupture, endovascular stenting using so called flow diverters is increasingly being regarded as an alternative to the established coil occlusion method in minimally invasive treatment. Covering the neck of an aneurysm with a flow diverter has the potential to alter the hemodynamics in such a way as to induce thrombosis within the aneurysm sac, stopping its further growth, preventing its rupture and possibly leading to complete resorption. In the present study the influence of different flow diverters is quantified considering idealized patient configurations, with a spherical sidewall aneurysm placed on either a straight or a curved parent vessel. All important hemodynamic parameters (exchange flow rate, velocity, and wall shear stress) are determined in a quantitative and accurate manner using computational fluid dynamics when varying the key geometrical properties of the aneurysm. All simulations are carried out using an incompressible, Newtonian fluid with steady conditions. As a whole, 72 different cases have been considered in this systematic study. In this manner, it becomes possible to compare the efficiency of different stents and flow diverters as a function of wire density and thickness. The results show that the intra-aneurysmal flow velocity, wall shear stress, mean velocity, and vortex topology can be considerably modified thanks to insertion of a suitable implant. Intra-aneurysmal residence time is found to increase rapidly with decreasing stent porosity. Of the three different implants considered in this study, the one with the highest wire density shows the highest increase of intra-aneurysmal residence time for both the straight and the curved parent vessels. The best hemodynamic modifications are always obtained for a small aneurysm diameter.

Central moments multiple relaxation time LBM for hemodynamic simulations in intracranial aneurysms: An in-vitro validation study using PIV and PC-MRI.

The lattice Boltzmann method (LBM) has recently emerged as an efficient alternative to classical Navier-Stokes solvers. This is particularly true for hemodynamics in complex geometries. However, in its most basic formulation, i.e. with the so-called single relaxation time (SRT) collision operator, it has been observed to have a limited stability domain in the Courant/Fourier space, strongly constraining the minimum time-step and grid size. The development of improved collision models such as the multiple relaxation time (MRT) operator in central moments space has tremendously widened the stability domain, while allowing to overcome a number of other well-documented artifacts, therefore opening the door for simulations over a wider range of grid and time-step sizes. The present work focuses on implementing and validating a specific collision operator, the central Hermite moments multiple relaxation time model with the full expansion of the equilibrium distribution function, to simulate blood flows in intracranial aneurysms. The study further proceeds with a validation of the numerical model through different test-cases and against experimental measurements obtained via stereoscopic particle image velocimetry (PIV) and phase-contrast magnetic resonance imaging (PC-MRI). For a patient-specific aneurysm both PIV and PC-MRI agree fairly well with the simulation. Finally, low-resolution simulations were shown to be able to capture blood flow information with sufficient accuracy, as demonstrated through both qualitative and quantitative analysis of the flow field while leading to strongly reduced computation times. For instance in the case of the patient-specific configuration, increasing the grid-size by a factor of two led to a reduction of computation time by a factor of 14 with very good similarity indices still ranging from 0.83 to 0.88.

Hemodynamic Data Assimilation in a Subject-specific Circle of Willis Geometry.

PURPOSE: The anatomy of the circle of Willis (CoW), the brain's main arterial blood supply system, strongly differs between individuals, resulting in highly variable flow fields and intracranial vascularization patterns. To predict subject-specific hemodynamics with high certainty, we propose a data assimilation (DA) approach that merges fully 4D phase-contrast magnetic resonance imaging (PC-MRI) data with a numerical model in the form of computational fluid dynamics (CFD) simulations. METHODS: To the best of our knowledge, this study is the first to provide a transient state estimate for the three-dimensional velocity field in a subject-specific CoW geometry using DA. High-resolution velocity state estimates are obtained using the local ensemble transform Kalman filter (LETKF). RESULTS: Quantitative evaluation shows a considerable reduction (up to 90%) in the uncertainty of the velocity field state estimate after the data assimilation step. Velocity values in vessel areas that are below the resolution of the PC-MRI data (e.g., in posterior communicating arteries) are provided. Furthermore, the uncertainty of the analysis-based wall shear stress distribution is reduced by a factor of 2 for the data assimilation approach when compared to the CFD model alone. CONCLUSION: This study demonstrates the potential of data assimilation to provide detailed information on vascular flow, and to reduce the uncertainty in such estimates by combining various sources of data in a statistically appropriate fashion.

Super-resolution and denoising of 4D-Flow MRI using physics-Informed deep neural nets.

BACKGROUND AND OBJECTIVE: Time resolved three-dimensional phase contrast magnetic resonance imaging (4D-Flow MRI) has been used to non-invasively measure blood velocities in the human vascular system. However, issues such as low spatio-temporal resolution, acquisition noise, velocity aliasing, and phase-offset artifacts have hampered its clinical application. In this research, we developed a purely data-driven method for super-resolution and denoising of 4D-Flow MRI. METHODS: The flow velocities, pressure, and the MRI image magnitude are modeled as a patient-specific deep neural net (DNN). For training, 4D-Flow MRI images in the complex Cartesian space are used to impose data-fidelity. Physics of fluid flow is imposed through regularization. Creative loss function terms have been introduced to handle noise and super-resolution. The trained patient-specific DNN can be sampled to generate noise-free high-resolution flow images. The proposed method has been implemented using the TensorFlow DNN library and tested on numerical phantoms and validated in-vitro using high-resolution particle image velocitmetry (PIV) and 4D-Flow MRI experiments on transparent models subjected to pulsatile flow conditions. RESULTS: In case of numerical phantoms, we were able to increase spatial resolution by a factor of 100 and temporal resolution by a factor of 5 compared to the simulated 4D-Flow MRI. There is an order of magnitude reduction of velocity normalized root mean square error (vNRMSE). In case of the in-vitro validation tests with PIV as reference, there is similar improvement in spatio-temporal resolution. Although the vNRMSE is reduced by 50%, the method is unable to negate a systematic bias with respect to the reference PIV that is introduced by the 4D-Flow MRI measurement. CONCLUSIONS: This work has demonstrated the feasibility of using the readily available machinery of deep learning to enhance 4D-Flow MRI using a purely data-driven method. Unlike current state-of-the-art methods, the proposed method is agnostic to geometry and boundary conditions and therefore eliminates the need for tedious tasks such as accurate image segmentation for geometry, image registration, and estimation of boundary flow conditions. Arbitrary regions of interest can be selected for processing. This work will lead to user-friendly analysis tools that will enable quantitative hemodynamic analysis of vascular diseases in a clinical setting.

Quantitative assessment of 4D hemodynamics in cerebral aneurysms using proper orthogonal decomposition.

BACKGROUND AND PURPOSE: The comparison of different time-varying three-dimensional hemodynamic data (4D) is a formidable task. The purpose of this study is to investigate the potential of the proper orthogonal decomposition (POD) for a quantitative assessment. METHODS: The complex spatial-temporal flow information was analyzed using proper orthogonal decomposition to reduce the complexity of the system. PC-MRI blood flow measurements and computational fluid dynamic simulations of two subject-specific IAs were used to compare the different flow modalities. The concept of Modal Assurance Criterion (MAC) provided a further detailed objective characterization of the most energetic individual modes. RESULTS: The most energetic flow modes were qualitatively compared by visual inspection. The distribution of the kinetic energy on the modes was used to quantitatively compare pulsatile flow data, where the most energetic mode was associated to approximately 90% of the total kinetic energy. This distribution was incorporated in a single measure, termed spectral entropy, showing good agreement especially for Case 1. CONCLUSION: The proposed quantitative POD-based technique could be a valuable tool to reduce the complexity of the time-dependent hemodynamic data and to facilitate an easy comparison of 4D flows, e.g., for validation purposes.

Novel feature-based visualization of the unsteady blood flow in intracranial aneurysms with the help of proper orthogonal decomposition (POD).

The recognition and interpretation of pulsatile subject-specific blood flow is a challenging task. Animations of various quantities - such as blood flow velocity, pressure, or wall shear stress - can be depicted to visualize the complex time-varying flow features, normally in a region of interest. Traditional visualization methods however can hardly convey the dynamic information of the system. Proper orthogonal decomposition (POD), a mathematical tool, allows for the complex spatial-temporal information to be decomposed into individual spatial modes. In the present study, the most energetic blood flow features are extracted with the help of POD analysis. The first mode, representing the most energetic flow feature, characterizes the temporal mean of the flow velocity. It is considered as the primary flow. The second most energetic mode corresponds to the secondary flow features. Visualization techniques combining the primary and the secondary flows are suggested in the present paper in order to create a simplified visualization of the unsteady blood flow. The methods are presented for intracranial aneurysms for both measured as well as simulated data, illustrating the application for Phase-Contrast Magnetic Resonance Imaging (PC-MRI) and computational fluid dynamics (CFD) results.

Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH)-Phase Ib: Effect of morphology on hemodynamics.

BACKGROUND: Image-based blood flow simulations have been increasingly applied to investigate intracranial aneurysm (IA) hemodynamics. However, the acceptance among physicians remains limited due to the high variability in the underlying assumptions and quality of results. METHODS: To evaluate the vessel segmentation as one of the most important sources of error, the international Multiple Aneurysms AnaTomy CHallenge 2018 (MATCH) was announced. 26 research groups from 13 different countries segmented three datasets, which contained five IAs in total. Based on these segmentations, 73 time-dependent blood flow simulations under consistent conditions were carried out. Afterwards, relevant flow and shear parameters (e.g., neck inflow rate, parent vessel flow rate, spatial mean velocity, and wall shear stress) were analyzed both qualitatively and quantitatively. RESULTS: Regarding the entire vasculature, the variability of the segmented vessel radius is 0.13 mm, consistent and independent of the local vessel radius. However, the centerline velocity shows increased variability in more distal vessels. Focusing on the aneurysms, clear differences in morphological and hemodynamic parameters were observed. The quantification of the segmentation-induced variability showed approximately a 14% difference among the groups for the parent vessel flow rate. Regarding the mean aneurysmal velocity and the neck inflow rate, a variation of 30% and 46% was observed, respectively. Finally, time-averaged wall shear stresses varied between 28% and 51%, depending on the aneurysm in question. CONCLUSIONS: MATCH reveals the effect of state-of-the-art segmentation algorithms on subsequent hemodynamic simulations for IA research. The observed variations may lead to an inappropriate interpretation of the simulation results and thus, can lead to inappropriate conclusions by physicians. Therefore, accurate segmentation of the region of interest is necessary to obtain reliable and clinically helpful flow information.