Accurate diagnosis of Chronic Mesenteric Ischemia (CMI) requires a comprehensive understanding of both anatomical and hemodynamic factors affecting mesenteric blood flow. This study employed Computational Fluid Dynamics (CFD) simulations to assess blood flow in the superior mesenteric artery (SMA) under varying conditions. Simulations on idealized geometries demonstrated the significant impact of hyperemia on flow dynamics in stenosed vessels, particularly in moderate stenosis. The presence of a collateral vessel from a patent celiac artery (CA) was shown to enhance distal SMA perfusion and reduce pressure gradients. A patient-specific computational workflow was also developed, incorporating CT-based segmentation, boundary condition setup, simulation, and post-processing. Results underscore that anatomical stenosis severity alone is insufficient to determine the functional significance of an SMA lesion. Instead, patient-specific hemodynamics, especially the presence and configuration of collateral pathways, must be considered. While promising, these findings require further validation with larger patient datasets.
link to dataset:
https://doi.org/10.4121/845085c4-9999-41bd-8a4f-02cbdea61164

Background: Cerebrovascular diseases, which often involve a disruption in blood flow in the Circle of Willis (CoW) and its branching arteries, pose a major global health risk. Computational fluid dynamics (CFD) analyses present an opportunity to study their pathophysiology but require high-quality vessel segmentations. Methods: To generate pseudo-labeled training data, computed tomography angiography (CTA) images were preprocessed and inference was run using two pretrained models: an nnU-Net for multi-class CoW segmentation and a DTUNet for binary cerebral vessel segmentation. These outputs were combined using a region-growing approach; the resulting pseudolabels were used to train an nnU-Net V2 with a topologyaware loss function. CFD analyses were performed on both a model-generated segmentation and a ground truth segmentation derived from a CTA scan which had been expert-labeled by a neuroradiologist. The resulting velocity, pressure and wall shear stress (WSS) profiles for both segmentations were compared across 10 cross-sections of the middle cerebral artery (MCA). Results: After filtering out inaccurate labels, 1,709 of 2,201 pseudo-labeled images were retained for training and testing. Common errors included over-segmentation of small vessels, under-segmentation of large vessels and poor separation of the anterior cerebral arteries when compared to expert-annotated ground truth segmentations. The proposed segmentation model was evaluated on the test set, which used pseudo-labels as a reference standard, and achieved a mean Dice score of 62%, clDice of 40%, IoU of 51%, a HD of 16.9 voxels and an ASD of 3.9 voxels. In terms of centerline-based metrics, the model achieved a mean overlap (OV) of 72% and an average ASCD of 4.26 voxels. In CFD simulations, the predicted segmentation yielded absolute errors of 48.96 �} 30.69 mm/s, 7.47 �} 6.07 Pa and 1.22 �} 0.81 Pa for blood flow velocity, pressure and WSS, respectively, compared to the expert-annotated reference (p < 0.05 for all). Conclusions: This study demonstrates that a deep learning model, trained using pseudo-labels, can successfully generate anatomically plausible multi-class segmentations of the CoW suitable for downstream CFD analysis. However, discrepancies in key hemodynamic metrics compared to expert-annotated data highlight the need for improved pseudo-label accuracy, especially in regions of complex vascular geometry.

Vascular calcification, the deposition of calcium in the vessel wall, is associated with several vascular diseases, including atherosclerosis, diabetes mellitus, and hypertension.
Fluid-structure interaction (FSI) is recommended to simulate blood flow incorporating vascular calcification. However, FSI applied to a three-dimensional (3D) model takes several days to simulate.
To reduce the computational complexity, 1D reduced order models (ROMs) are often used instead.

Reduced order modeling decreases the computational complexity of a model by removing dimensions of the coordinate system within a model. The cylindrical coordinate system is used in hemodynamics, especially in ROMs. The 1D ROM for hemodynamics is obtained by removing the azimuthal dimension (accomplished by assuming axial symmetry for all properties within arteries) and the radial dimension (accomplished by applying a predefined velocity profile to blood flow) from the 3D model. However, incorporating vascular calcification can make the geometry of arteries and flow within arteries asymmetric. A 2D ROM can increase the accuracy of the 1D ROM by including one of the two removed dimensions. Research regarding 2D blood flow mainly focuses on including the radial dimension, which cannot implement asymmetric calcification since axisymmetry is assumed.

This study obtains a 2D ROM for blood flow by removing the dimension corresponding to the radial distance from the three-dimensional model and by assuming that axial velocity is continuous in the neighborhood near the artery's origin. The 2D ROM obtains axisymmetric velocity by only allowing a single velocity profile. However, enabling a family of velocity profiles can make flow within arteries asymmetric. Hence, this study contributes to hemodynamics by studying blood flow that allows a family of velocity profiles.

A non-physiological steady-state solution has been obtained analytically, in which the volumetric flow rate vanishes, and numerical methods are developed to simulate the 2D ROM, which incorporates dimensional (Godunov) splitting, linear approximate solvers, and high-resolution methods. Jump-discontinuities within the mechanical properties of the vascular walls are smoothened for the 2D simulations. Numerical methods for the 2D ROM yield significant errors within the smoothening region for simulations with coarse grids.
The numerical method obtains the non-physiological steady-state solutions for arteries without calcification and has a relative error of O(Δx^1.500) for arteries with axisymmetric calcification. The 2D ROM cannot numerically obtain the non-physiological steady-state solution for arteries with asymmetric calcification due to the numerical errors within the smoothening range.

 3D and 2D numerical simulations with pulsatile blood flow are compared. The 3D simulation without calcification has a significantly higher diastolic pressure, larger inner wall radii, and larger volumetric flow rates than the 2D simulation. The differences in blood flow observed between pulsatile blood flow without calcification and with calcification match decently between the 3D simulations and the 2D simulations, except for locations within the smoothening region.

Cerebral small vessel disease (CSVD) is a leading cause of stroke and dementia, making the study of small vessel hemodynamics vital for advancing diagnostic and therapeutic strategies. This research focuses on fabricating microfluidic devices that replicate the lateral lenticulostriate arteries (LSA) to validate computational flow models of small cerebral vessels. A key challenge in studying CSVD is the lack of experimental validation for computational fluid dynamics (CFD) models, which are widely used to simulate hemodynamics. To address this, additive masked Stereolithography (mSLA) was employed to fabricate a microfluidic model of the LSAs. The study explored the impact of orientation, exposure time, and layer height on the roundness and error of the intended area of printed micro-pores, to optimize the manufacturing of a microfluidic device. The smallest printed pore measured 270 μm in diameter. Pores printed at a larger angle as assessed from the build plate were more likely to remain open, but exhibited a larger decrease in area compared to smaller angles. A lower exposure time exhibited a larger pore area, whereas a larger layer height showed a decrease in area from intended. The layer height and angle did not influence the roundness, whereas an increase in exposure time decreased the roundness of the pores. Additionally, flow experiments were conducted using a 3D printed microfluidic device to compare empirical data with CFD and analytical simulations. The encountered resistance was larger for the experimental results (3.19 ⋅10^11 Pa⋅s/m^3) compared to the analytical result (1.96 ⋅10^11 Pa⋅s/m^3) and the computational result (1.90 ⋅10^11 Pa⋅s/m^3), likely due to deviation from the intended size. Finally, arterial microfluidic devices were printed and flow was induced to showcase their functionality. Achieving precise channel dimensions remains the primary challenge in mSLA printing due to cumulative dosage effects.
This research bridges the gap between computational modeling and experimental validation, providing a platform for studying cerebral microcirculation. The findings demonstrate the feasibility of using commercially available 3D-printed microfluidic devices to replicate small cerebral vessels. The outcomes of this study contribute to the advancement of vascular biomodeling, with implications for future clinical applications in stroke and neurovascular research.

Background
When the heart cannot fully contract, less oxygen-rich blood is delivered to organs and tissues in the body, which can lead to organ failure and eventually death. A left ventricular assist device (LVAD) supports the heart by increasing blood flow from the left ventricle to the aorta. A novel LVAD is being developed that consists of an inflatable balloon, which is placed in the left ventricle via a small insertion in the groin. The balloon is actuated at high frequencies by making use of a diaphragm pump, which is placed next to the patient’s bed and is connected to the balloon with a catheter filled with helium. The diaphragm pump can inflate the balloon by compressing the helium and deflate the balloon with expansion of the helium. Actuating the diaphragm at high frequencies generates heat in the pressure chambers of the pump due to friction. However, the maximum temperature of parts of the system that can come into contact with a patient or medical personal, cannot exceed 41◦C according to medical regulation. Therefore, the aim of this project is to create a concept design for cooling the helium gas flow used to inflate and deflate an intraventricular balloon at high frequencies for a novel left ventricle assist device.

Method
First, an overview of existing cooling techniques is created. Eight of the 24 cooling techniques explored met the list of requirements. These eight techniques are assessed against the list of weighted criteria using the Harris profile method. A concept design is created for the three best scoring techniques: thermoelectric cooling, vortex cooling, and forced convection using air for which the forced convection concept design appears to be most feasible. The final cooling concept features two air fans blowing air through a heat sink incorporated in the aluminium base structure around the pressure chambers. To verify the concept, a (simplified) simulation model is created in Ansys Fluent.

Results
The simulation model is used to simulate air flow through the channels and heat conduction in the base structure around the pressure chambers. The optimal fan velocity is determined by analysing the pressure drop and the temperature behaviour of the system. Furthermore, the model is 3D printed in plastic to validate the simulation model by performing velocity and pressure tests. Since the test results match the simulation results, the simulation model is considered validated.

Conclusion
In conclusion, after assessing different cooling techniques, a concept based on forced convection is developed in detail. The design is verified with a simulation model, and the simulation results are validated with a 3D printed plastic model. Although the concept design allows for room for optimization, it has proven to be a feasible solution for cooling the pressure chambers, and thus the helium gas, in the diaphragm pump of the novel LVAD.

Cerebrovascular diseases, including stroke and dementia, rank among the top 10 causes of death globally. Where the Circle of Willis is an anatomical structure at the base of the brain that redistributes blood flow. Vrselja et al. challenged the traditional view that the primary function of the Circle of Willis is to provide collateral circulation, suggesting instead that it dissipates high pressures in the cerebrovasculature.

One-dimensional (1D) modeling, based on the simplified Navier-Stokes equations, offers a way to accurately predict blood flow with only a fraction of the computational expense of a three-dimensional (3D) model. SimVascular, a computational modeling software tailored for representing blood flow, recently added a 1D modeling feature. This study assessess the suitability of the 1D model in SimVascular for representing the cerebrovasculature, by testing it on the Circle of Willis. Additionally, preliminary research investigates whether the Circle of Willis functions as a pressure dissipative system, particularly through its communicating arteries.

The modeling pipeline included centerline extraction, mesh generation, boundary conditions, material model and solver parameters. Challenges were identified during centerline extraction, as SimVascular is inherently unable to handle loops, and inlet boundary conditions, where it could not manage multiple inlets. Considering these limitations, a series of 1D linear elastic models were created from the 3D geometry of the Circle of Willis to test whether it serves as a pressure dissipative system.

While SimVascular was unable to handle closed-loop geometries and multiple inlets, results indicated that the absence of a communicating artery induced relatively minor pressure oscillations. This suggests first, that SimVascular, in its current form, is not well-suited for 1D modeling of the cerebrovascular system's complexities, including loops and multiple inlets. Second, that the hypothesis of the Circle of Willis functioning primarily as a pressure dissipative system holds promise but requires further validation.

The study of haemodynamics, or the mechanics of blood flow, has been a topic of significant interest since ancient times, as it provides critical insights into both the normal functioning of the circulatory system and the mechanisms underlying various diseases. A robust scientific understanding of these dynamics facilitates the development of innovative prevention and intervention strategies, such as advanced, less invasive surgical techniques and cost-effective medical devices, which can dramatically improve global health outcomes. Over the centuries, experimental, analytical, and, more recently, computational methods have enhanced our knowledge of cardiovascular and cerebrovascular conditions, leading to improved diagnostic and therapeutic approaches that enhance life quality and expectancy.

Despite these advancements, substantial challenges remain, particularly in diagnosing and understanding haemodynamic complications such as stroke, stenosis, and regurgitation. Modern imaging techniques like MRI and sonography are limited by spatial and temporal resolution, requiring significant hardware improvements to overcome these constraints. Moreover, while analytical models have evolved to address the complex nature of blood flow and vessel behavior, they often struggle to accurately represent in vivo scenarios, especially in areas like the cerebral vasculature that are difficult to image. The presence of anomalies such as medial calcifications - which are responsible behind numerous life altering conditions and diseases such as, diabetes, kidney disease and hypertension - further complicates the accurate modeling of blood flow dynamics by introducing departure from axisymmetry, changes in material properties of the tissue and impacting the physics at localized scale.

To address the impact of medial calcifications, this thesis presents a detailed approach starting with the development of a three-dimensional (3D) model that represents the physiological fluid-structure interaction (FSI) phenomenon of pulsatile blood flow through an Internal Carotid Artery (ICA) with medial calcification. ICA is a critical vessel within the Circle of Willis which itself is one of the most crucial cerebro-vascular networks. Following this, a physics based one-dimensional (1D) Reduced Order Model (ROM) is developed, integrated with which is a data-driven constitutive modelling framework trained on the 3D simulation data, to account for changes in material proeprties introduced by medial calcifications. This ROM aims to provide a computationally efficient yet accurate representation of the physiological changes caused by medial calcifications. The thesis includes a comprehensive theoretical framework, detailed validation and verification processes, and comparisons between the 1D and 3D models to ensure accuracy and clinical relevance. The overarching goal is to create a well-validated ROM that maintains essential physical details while reducing computational costs, with an emphasis on evaluating its performance in clinical scenarios and exploring its potential applications and future research directions.

Trans-catheter aortic valve replacement (TAVR) has been widely used for treating Aortic stenosis (AoS). Numerical simulation of TAVR process is valuable in terms of reducing surgical risk for patients and cost for valve design. Finite element simulations are performed in this thesis to investigate the risk of coronary obstruction for patients who went through Aortic Valve Neo-cuspitization (AVNeo or Ozaki) procedure. The Expected Leaflet-to-ostium Distance was used as a criterion for assessing coronary obstruction risk. This value in 3 different deployment configurations is compared to determine the optimal deployment strategy. It is found that there is a large risk of coronary obstruction after TAVR for such patients under distal and midplane configuration but no risk under proximal configuration.

Ascending thoracic aortic aneurysm is a dangerous condition which is hard to locate in patients. CFD can be used to assess someone’s risk of suffering from such an aneurysm and also to assess any possible relation of an aneurysm with relevant parameters, like the shape of the aorta or how blood flows into the aorta.

Existing research on ascending aortic CFD flow generally focuses on assessing relatively high wall shear stress, seen as the cause for aortic wall damage and aneurysms, in patients and not much on any possible relationships between aortic shape and inflow with ascending aortic aneurysm. Work that does exist often uses 4D flow MRI, shown to produce results regarding peak wall shear stresses with less accuracy compared to CFD.

This thesis aims to supply that information using an aortic geometry from a large dataset of synthetically generated aortas together with synthetically generated inlet velocity profiles, showing a possibility to work with-out the need for patient measurements. The main work in this thesis focuses on finding any possible relationship between aortic inflow angle and relatively high wall shear stress, such that the results may better explain how aortic inflow can influence the appearance of aneurysms.

For this, a workflow has been established that allows working with large aortic geometry and inflow profile datasets with relative ease. This workflow process uses a mix of OpenFoam and Ansys Fluent usage.

Results have shown that the flow jet angle in the core of the flow has a significant negative correlation with peak WSS within the ascending aorta for the chosen aortic geometry. This is in contrast with other work showing positive correlations with WSS. This may suggest aortic geometry dependence together with the need to look at other flow variables, like jet impingement angle, to explain how aortic inflow can influence high WSS within the ascending aorta.

Recommendations for future work include: The same study but with a focus on aortas of healthy young people to see if aortas at risk can be found, as existing work mainly focuses on old people with ascending thoracic aortic aneurysm, a similar study with a focus on the inclusion of various aortic geometries to find how aortic geometry may influence the effect inflow variables can have on high WSS, the creation of a tool to easily calculate the impingement angle from aortic flow data and research into if aortic jet flow disruption could reduce peak WSS in the ascending aorta for use in aortic reconstruction.

Cardiovascular diseases (CVDs) are a group of disorders of the heart and blood vessels.
CVDs are the leading cause of death worldwide. To diagnose and treat CVDs, clinicians and cardiologists use multiple noninvasive imaging techniques. These scans are used to segment certain structures of the heart. Deep learning-based cardiac segmentation on short-axis cardiac magnetic resonance images (CMRI) has gained popularity over the past few years because of its generalisability and accuracy. This has exponentially reduced contouring times for clinicians. The development of such deep learning techniques has seen a common trend. In order to accommodate learning for larger cardiac datasets, the depth and size of segmentation networks have been increased. Unfortunately, the environmental impact of exploding such networks is not taken into account. One solution to mitigate having computationally expensive networks is to incorporate anatomical knowledge in the form of shape priors. The Gridnet and UNet with a shape prior are computationally efficient networks that are used to evaluate segmentation performance on a large and varied cardiac dataset (Combination of the Automated Cardiac Diagnosis Challenge - ACDC and Multi-Centre, Multi-Vendor and Multi-Disease Cardiac Segmentation challenge - M&M datasets). On average, these networks segment CMRIs with an average dice score of 0.87 and a Hausdorff distance of 11.7mm. In parallel, one of the major issues in cardiac technology is the under-representation of women in cardiac datasets. Purposefully curated cardiac datasets such as ACDC and M&M try and maintain equal representation. In real-world scenarios, this might not always be the case. Clinical trials to collect such data often report female representation as low as 25%. Evaluation of segmentation performance between a balanced and skewed dataset is conducted. This is to address if bias in such cardiac training datasets affects the performance of segmentation networks between male and female test patients.