Deep Learning for 3D Vascular Segmentation in Phase Contrast Tomography.

Automated blood vessel segmentation is critical for biomedical image analysis, as vessel morphology changes are associated with numerous pathologies. Still, precise segmentation is difficult due to the complexity of vascular structures, anatomical variations across patients, the scarcity of annotated public datasets, and the quality of images. Our goal is to provide a foundation on the topic and identify a robust baseline model for application to vascular segmentation using a new imaging modality, Hierarchical Phase-Contrast Tomography (HiP-CT). We begin with an extensive review of current machine learning approaches for vascular segmentation across various organs. Our work introduces a meticulously curated training dataset, verified by double annotators, consisting of vascular data from three kidneys imaged using Hierarchical Phase-Contrast Tomography (HiP-CT) as part of the Human Organ Atlas Project. HiP-CT, pioneered at the European Synchrotron Radiation Facility in 2020, revolutionizes 3D organ imaging by offering resolution around 20µm/voxel, and enabling highly detailed localized zooms up to 1µm/voxel without physical sectioning. We leverage the nnU-Net framework to evaluate model performance on this high-resolution dataset, using both known and novel samples, and implementing metrics tailored for vascular structures. Our comprehensive review and empirical analysis on HiP-CT data sets a new standard for evaluating machine learning models in high-resolution organ imaging. Our three experiments yielded Dice scores of 0.9523 and 0.9410, and 0.8585, respectively. Nevertheless, DSC primarily assesses voxel-to-voxel concordance, overlooking several crucial characteristics of the vessels and should not be the sole metric for deciding the performance of vascular segmentation. Our results show that while segmentations yielded reasonably high scores-such as centerline Dice values ranging from 0.82 to 0.88, certain errors persisted. Specifically, large vessels that collapsed due to the lack of hydro-static pressure (HiP-CT is an ex vivo technique) were segmented poorly. Moreover, decreased connectivity in finer vessels and higher segmentation errors at vessel boundaries were observed. Such errors, particularly in significant vessels, obstruct the understanding of the structures by interrupting vascular tree connectivity. Through our review and outputs, we aim to set a benchmark for subsequent model evaluations using various modalities, especially with the HiP-CT imaging database.

Micro to macro scale analysis of the intact human renal arterial tree with Synchrotron Tomography.

Background: The kidney vasculature is exquisitely structured to orchestrate renal function. Structural profiling of the vasculature in intact rodent kidneys, has provided insights into renal haemodynamics and oxygenation, but has never been extended to the human kidney beyond a few vascular generations. We hypothesised that synchrotron-based imaging of a human kidney would enable assessment of vasculature across the whole organ. Methods: An intact kidney from a 63-year-old male was scanned using hierarchical phase-contrast tomography (HiP-CT), followed by semi-automated vessel segmentation and quantitative analysis. These data were compared to published micro-CT data of whole rat kidney. Results: The intact human kidney vascular network was imaged with HiP-CT at 25 µm voxels, representing a 20-fold increase in resolution compared to clinical CT scanners. Our comparative quantitative analysis revealed the number of vessel generations, vascular asymmetry and a structural organisation optimised for minimal resistance to flow, are conserved between species, whereas the normalised radii are not. We further demonstrate regional heterogeneity in vessel geometry between renal cortex, medulla, and hilum, showing how the distance between vessels provides a structural basis for renal oxygenation and hypoxia. Conclusions: Through the application of HiP-CT, we have provided the first quantification of the human renal arterial network, with a resolution comparable to that of light microscopy yet at a scale several orders of magnitude larger than that of a renal punch biopsy. Our findings bridge anatomical scales, profiling blood vessels across the intact human kidney, with implications for renal physiology, biophysical modelling, and tissue engineering.

Three-dimensional modeling of Marfan syndrome with elastic and hyperelastic materials assumptions using fluid-structure interaction.

BACKGROUND: Marfan syndrome (MFS) is a genetic disorder of the connective tissue. It most prominently influences the skeletal, cardiovascular, and ocular systems, but all fibrous connective tissue throughout the body can be affected as well. OBJECTIVE: This study aims to investigate a realistic three-dimensional model of an aorta of a specific patient suffering from MFS by considering elastic and hyperelastic materials for the tissue using fluid-structure interaction (FSI). METHODS: Isotropic linear elastic and Mooney-Rivlin hyperelastic assumptions are implemented. Linear and nonlinear mechanical properties of the aneurysmal MFS aortic tissue are derived from an uniaxial experimental test. RESULTS: Vortex generation in the vicinity of the aneurysm region in both elastic and hyperelastic models and the maximum blood velocity at peak flow time is calculated as 0.517 and 0.533 m/s for the two materials, respectively. The blood pressure is not significantly different between the two models (±8 Pa) and the blood pressure difference between the points in the horizontal plane of the aneurysm region is obtained as ±10 Pa for both models. The maximum von Mises stress for the hyperelastic model (2.19 MPa) is 27% more than the elastic one (1.72 MPa) and takes place at the inner curvature and upper part of the aorta and somehow far from the aneurysm region. The wall shear stress (WSS) is also considered for the elastic and hyperelastic assumptions, which is 36.7 Pa for both elastic and hyperelastic models. CONCLUSION: The aneurysm region in the MFS affects the blood flow and causes the vortex to be generated which consequently affects the blood flow in the downstream by adding some perturbations to the blood flow. The WSS is obtained to be lower in the aneurysm region compared to other regions which indicated vascular remodeling.

Investigating the performance of four specific types of material grafts and their effects on hemodynamic patterns as well as on von Mises stresses in a grafted three-layer aortic model using fluid-structure interaction analysis.

One of the important parts of the cardiac system is aorta which is the fundamental channel and supply of oxygenated blood in the body. Diseases of the aorta represent critical cardiovascular bleakness and mortality around the world. This study aims at investigation of hemodynamic parameters in a two-dimensional axisymmetric model of three-layer grafted aorta using fluid-structure interaction (FSI). It assumes that a damaged part of aorta, which may happen as a result of some diseases like aneurysm, dissection and post-stenotic dilatation, is replaced with a biomaterial graft. Four types of grafts materials so-called Polyurethane, Silicone rubber, Polytetrafluoroethylene (PTFE) and Dacron are considered in the present study. The assumption of linear elastic and isotropic material is set for the both aorta's wall and aforementioned grafts. Blood is considered as an incompressible and Newtonian fluid. The results indicate higher displacement in Polyurethane and silicone rubber in comparison with other two. Furthermore, results reveal that blood flow velocity has slightly higher values in PTFE and Dacron grafted models compared to Polyurethane and Silicone rubber ones. Even though there are some differences in hemodynamic patterns in these grafted models, they are not considerable as much as von Mises stresses across the graft-aorta intersections are. This study shows that the types of material grafts play an important role in the amount of stresses particularly at intersections of aorta and graft.

A computational fluid-structure interaction model of the blood flow in the healthy and varicose saphenous vein.

OBJECTIVE: Varicose vein has become enlarged and twisted and, consequently, has lost its mechanical strength. As a result of the varicose saphenous vein (SV) mechanical alterations, the hemodynamic parameters of the blood flow, such as blood velocity as well as vein wall stress and strain, would change accordingly. However, little is known about stress and strain and there consequences under experimental conditions on blood flow and velocity within normal and varicose veins. In this study, a three-dimensional (3D) computational fluid-structure interaction (FSI) model of a human healthy and varicose SVs was established to determine the hemodynamic characterization of the blood flow as a function of vein wall mechanical properties, i.e. elastic and hyperelastic. METHODS: The mechanical properties of the human healthy and varicose SVs were experimentally measured and implemented into the computational model. The fully coupled fluid and structure models were solved using the explicit dynamics finite element code LS-DYNA. RESULTS: The results revealed that, regardless of healthy and varicose, the elastic walls reach to the ultimate strength of the vein wall, whereas the hyperelastic wall can tolerate more stress. The highest von Mises stress compared to the healthy ones was seen in the elastic and hyperelastic varicose SVs with 1.412 and 1.535 MPa, respectively. In addition, analysis of the resultant displacement in the vein wall indicated that the varicose SVs experienced a higher displacement compared to the healthy ones irrespective of elastic and hyperelastic material models. The highest blood velocity was also observed for the healthy hyperelastic SV wall. CONCLUSION: The findings of this study may have implications not only for determining the role of the vein wall mechanical properties in the hemodynamic alterations of the blood, but also for employing as a null information in balloon-angioplasty and bypass surgeries.

Case Report: Combination Therapy with Mesenchymal Stem Cells and Granulocyte-Colony Stimulating Factor in a Case of Spinal Cord Injury.

INTRODUCTION: Various neuroregenerative procedures have been recently employed along with neurorehabilitation programs to promote neurological function after Spinal Cord Injury (SCI), and recently most of them have focused on the acute stage of spinal cord injury. In this report, we present a case of acute SCI treated with neuroprotective treatments in conjunction with conventional rehabilitation program. METHODS: A case of acute penetrative SCI (gunshot wound), 40 years old, was treated with intrathecal bone marrow derived stem cells and parenteral Granulocyte-Colony Stimulating Factor (G-CSF) along with rehabilitation program. The neurological outcomes as well as safety issues have been reported. RESULTS: Assessment with American Spinal Injury Association (ASIA), showed neurological improvement, meanwhile he reported neuropathic pain, which was amenable to oral medication. DISCUSSION: In the acute setting, combination therapy of G-CSF and intrathecal Mesenchymal Stem Cells (MSCs) was safe in our case as an adjunct to conventional rehabilitation programs. Further controlled studies are needed to find possible side effects, and establish net efficacy.

Wall stress in media layer of stented three-layered aortic aneurysm at different intraluminal thrombus locations with pulsatile heart cycle.

At the point when the aorta ruptures suddenly, as opposed to as the after-effect of injury, it is for the most part in aortic aneurysm. Aortic aneurysm rupture happens when the wall stress surpasses the strength of the vascular tissue. Intraluminal thrombus (ILT) may have advantages as it can absorb tension and decrease aortic aneurysm wall stress. This study aims to investigate the presence and growth effects of ILT on the wall stress in a stented aneurysm in one heart cycle. A virtual stented aneurysm model with ILT was made to study the flow and wall dynamics using fluid-structure interaction (FSI) analysis. Wall stresses at the center line of media layer of aorta thickness were calculated by two-dimensional axisymmetric finite element analysis. Calculations were executed as thrombus elastic modulus increased from 0.1 to 2 MPa and calculations were repeated as thrombus depth was increased in 10% increment until thrombus filled the whole aneurysm cavity. The von Mises stresses were compared in three sections, namely proximal, aneurysm and distal sections in the abdominal aorta. The wall stress showed its maximum value during a peak flow and pressure and gradually decreased as the pressure and velocity of blood reduced in all three aforementioned sections. As the intraluminal thrombus depth increased from 10% to 100%, the wall stress in distal, proximal and centre of aneurysm during one heart cycle was decreased. Furthermore, increasing the elastic modulus of thrombus from 10% to 100% triggered a reduction in wall stress in proximal, centre of intraluminal thrombus and distal regions during one heart cycle. The achievements of this study may have implications not only for understanding the wall stress in ILT, but also for providing more detailed information about aortic aneurysm with intraluminal thrombus and can help surgeons to do their best.

Hemodynamic simulation of blood flow in a new type of cardiac assist device named AVICENA.

The purpose of this study is to investigate the hemodynamic parameters of blood flow in a balloon as a part of a new type of cardiac assist device named AVICENA, which is implanted into the descending aorta to improve the strength of pumping blood flow in a poor-performing left ventricle. Balloon is inflated and deflated during diastole and systole, respectively. The longitudinal velocity of blood flow during balloon inflation and deflation has been considered. Through these investigations, the result reveals that the balloon inflation causes the blood flow to accelerate through the balloon and compensates the blood flow velocity required for the normal circulation system. When the balloon deflates, a reverse flow is generated and improves the perfusion of coronary arteries. Furthermore, the inlet pressure and acting force on the aortic valve for the healthy, unhealthy, and assisted heart have been compared. Result indicates that the force acting on the aortic valve has been considerably reduced for the assisted heart compare to the unhealthy or unassisted heart.