3D Vessel Model Research Articles

Measurement of transcranial Doppler insonation angles from three-dimensional reconstructions of CT angiography scans.

Blood velocities measured by Transcranial Doppler (TCD) are dependent on the angle between the incident ultrasound beam and the direction of blood flow (known as the Doppler angle). However, when TCD examinations are performed without imaging the Doppler angle for each vessel segment is not known. We have measured Doppler angles in the basal cerebral arteries examined with TCD using three-dimensional (3D) vessel models generated from computed tomography angiography (CTA) scans. This approach produces angle statistics that are not accessible during non-imaging TCD studies. We created 3D models of the basal cerebral arteries for 24 vasospasm patients. Standard acoustic windows were mapped to the specific anatomy of each patient. Virtual ultrasound transmit beams were generated that originated from the acoustic window and intersected the centerline of each arterial segment. Doppler angle measurements were calculated and compiled for each vessel segment. Doppler angles were smallest for the middle cerebral artery M1 segment (median 24.6°) and ophthalmic artery (median 25.0°), and largest for the anterior cerebral artery A2 segment (median 76.4°) and posterior cerebral artery P2 segment (median 75.8°). The ophthalmic artery had the highest proportion of Doppler angles that were less than 60° (99%) while the anterior cerebral artery A2 segment had the lowest proportion of Doppler angles that were less than 60° (10%). These angle measurements indicate the expected deviation between measured and true velocities in the cerebral arteries, highlighting specific segments that may be prone to underestimation of velocity.

Patient-specific cerebral 3D vessel model reconstruction using deep learning

Three-dimensional vessel model reconstruction from patient-specific magnetic resonance angiography (MRA) images often requires some manual maneuvers. This study aimed to establish the deep learning (DL)-based method for vessel model reconstruction. Time of flight MRA of 40 patients with internal carotid artery aneurysms was prepared, and three-dimensional vessel models were constructed using the threshold and region-growing method. Using those datasets, supervised deep learning using 2D U-net was performed to reconstruct 3D vessel models. The accuracy of the DL-based vessel segmentations was assessed using 20 MRA images outside the training dataset. The dice coefficient was used as the indicator of the model accuracy, and the blood flow simulation was performed using the DL-based vessel model. The created DL model could successfully reconstruct a three-dimensional model in all 60 cases. The dice coefficient in the test dataset was 0.859. Of note, the DL-generated model proved its efficacy even for large aneurysms (> 10 mm in their diameter). The reconstructed model was feasible in performing blood flow simulation to assist clinical decision-making. Our DL-based method could successfully reconstruct a three-dimensional vessel model with moderate accuracy. Future studies are warranted to exhibit that DL-based technology can promote medical image processing.Graphical

The new era of cardiovascular research: revolutionizing cardiovascular research with 3D models in a dish.

Cardiovascular research has heavily relied on studies using patient samples and animal models. However, patient studies often miss the data from the crucial early stage of cardiovascular diseases, as obtaining primary tissues at this stage is impracticable. Transgenic animal models can offer some insights into disease mechanisms, although they usually do not fully recapitulate the phenotype of cardiovascular diseases and their progression. In recent years, a promising breakthrough has emerged in the form of invitro three-dimensional (3D) cardiovascular models utilizing human pluripotent stem cells. These innovative models recreate the intricate 3D structure of the human heart and vessels within a controlled environment. This advancement is pivotal as it addresses the existing gaps in cardiovascular research, allowing scientists to study different stages of cardiovascular diseases and specific drug responses using human-origin models. In this review, we first outline various approaches employed to generate these models. We then comprehensively discuss their applications in studying cardiovascular diseases by providing insights into molecular and cellular changes associated with cardiovascular conditions. Moreover, we highlight the potential of these 3D models serving as a platform for drug testing to assess drug efficacy and safety. Despite their immense potential, challenges persist, particularly in maintaining the complex structure of 3D heart and vessel models and ensuring their function is comparable to real organs. However, overcoming these challenges could revolutionize cardiovascular research. It has the potential to offer comprehensive mechanistic insights into human-specific disease processes, ultimately expediting the development of personalized therapies.

Assessment of intracranial aneurysm rupture risk using a point cloud-based deep learning model.

Background and Purpose: Precisely assessing the likelihood of an intracranial aneurysm rupturing is critical for guiding clinical decision-making. The objective of this study is to construct and validate a deep learning framework utilizing point clouds to forecast the likelihood of aneurysm rupturing. Methods: The dataset included in this study consisted of a total of 623 aneurysms, with 211 of them classified as ruptured and 412 as unruptured, which were obtained from two separate projects within the AneuX morphology database. The HUG project, which included 124 ruptured aneurysms and 340 unruptured aneurysms, was used to train and internally validate the model. For external validation, another project named @neurIST was used, which included 87 ruptured and 72 unruptured aneurysms. A standardized method was employed to isolate aneurysms and a segment of their parent vessels from the original 3D vessel models. These models were then converted into a point cloud format using open3d package to facilitate training of the deep learning network. The PointNet++ architecture was utilized to process the models and generate risk scores through a softmax layer. Finally, two models, the dome and cut1 model, were established and then subjected to a comprehensive comparison of statistical indices with the LASSO regression model built by the dataset authors. Results: The cut1 model outperformed the dome model in the 5-fold cross-validation, with the mean AUC values of 0.85 and 0.81, respectively. Furthermore, the cut1 model beat the morphology-based LASSO regression model with an AUC of 0.82. However, as the original dataset authors stated, we observed potential generalizability concerns when applying trained models to datasets with different selection biases. Nevertheless, our method outperformed the LASSO regression model in terms of generalizability, with an AUC of 0.71 versus 0.67. Conclusion: The point cloud, as a 3D visualization technique for intracranial aneurysms, can effectively capture the spatial contour and morphological aspects of aneurysms. More structural features between the aneurysm and its parent vessels can be exposed by keeping a portion of the parent vessels, enhancing the model's performance. The point cloud-based deep learning model exhibited good performance in predicting rupture risk while also facing challenges in generalizability.

Aneurysmal Inflow Rate Coefficient Predicts Ultra-early Rebleeding in Ruptured Intracranial Aneurysms: Preliminary Report of a Computational Fluid Dynamics Study.

Rebleeding from a ruptured intracranial aneurysm has poor outcomes. Although numerous factors are associated with rebleeding, studies on computational fluid dynamics (CFD) on hemodynamic parameters associated with early rebleeding are scarce. In particular, no report of rebleeding in ultra-early phase exists. We aimed to elucidate the specific hemodynamic parameters associated with ultra-early rebleeding using CFD. In this study, the rebleeding group included patients with aneurysmal subarachnoid hemorrhage (aSAH) that rebled within 6 h from the onset. The control group included patients without rebleeding, observed for >10 h following the initial rupture. Clinical images after initial rupture and before rebleeding were used to build 3D vessel models for hemodynamic analysis focusing on the following parameters: time-averaged wall shear stress (WSS), normalized WSS, low shear area, oscillatory shear index, relative residence time, pressure loss coefficient, and aneurysmal inflow rate coefficient (AIRC). Five and 15 patients in the rebleeding and control groups, respectively, met the inclusion criteria. The World Federation of Neurosurgical Surgeons grade was significantly higher in the rebleeding group (p = 0.0088). Hemodynamic analysis showed significantly higher AIRC in the rebleeding group (p = 0.042). The other parameters were not significantly different between groups. There were no significant differences or correlations between SAH severity and AIRC. AIRC was identified as a hemodynamic parameter associated with ultra-early rebleeding of ruptured intracranial aneurysms. Thus, AIRC calculation may enable the prediction of ultra-early rebleeding.

Modelling Lymph Propulsion in a 3D Model of Murine Collecting Vessel with Three Lymphangions in Series

The lymphatic system maintains tissue homeostasis by transporting the excess fluid from the interstitium and ultimately returning it to the venous circulation against an adverse pressure gradient and gravitational force. The spontaneous contractions of lymphangions, the building blocks of collecting vessels, and the secondary lymphatic valves play key roles in lymph propulsion. The aim of this study was to investigate lymph propulsion in a series of three contracting lymphangions in a 3D reconstructed model segmented from micro-CT scans of the collecting lymphatics in the hind limb of mice. Computational Fluid Dynamics and Fluid–Structure Interaction were used to study the behavior of flow within the collecting vessel, as well as the behavior and deformations of the vessel wall and the poroelastic interstitium. The secondary valves were modelled as porous membranes with closed or open states depending on their permeability. A sensitivity study revealed that the parameters having the most impact on the total volume of lymph propelled by active contraction of the lymphangions were the elastic modulus of the interstitium and the permeability of the secondary valves during the open states.

High-throughput 3D microvessel-on-a-chip model to study defective angiogenesis in systemic sclerosis

In early systemic sclerosis (Scleroderma, SSc), the vasculature is impaired. Although the exact etiology of endothelial cell damage in SSc remains unclear, it is hypothesized that endothelial to mesenchymal transition (EndoMT) plays a key role. To perform physiologically relevant angiogenic studies, we set out to develop an angiogenesis-on-a-chip platform that is suitable for assessing disease parameters that are relevant to SSc and other vasculopathies. In the model, we substituted Fetal Bovine Serum (FBS) with Human Serum without impairing the stability of the culture. We showed that 3D microvessels and angiogenic factor-induced sprouts exposed to key pro-inflammatory and pro-fibrotic cytokines (TNFα and TGFβ) undergo structural alterations consisting of destructive vasculopathy (loss of small vessels). We also showed that these detrimental effects can be prevented by compound-mediated inhibition of TGFβ-ALK5 signaling or addition of a TNFα neutralizing antibody to the 3D cultures. This demonstrates that our in vitro model is suitable for compound testing and identification of new drugs that can protect from microvascular destabilization or regression in disease-mimicking conditions. To support this, we demonstrated that sera obtained from SSc patients can exert an anti-angiogenic effect on the 3D vessel model, opening the doors to screening for potential SSc drugs, enabling direct patient translatability and personalization of drug treatment.

BETHSY 9.1b test calculation with TRACE using 3D vessel component

Recently, several advanced multidimensional computational tools for simulating reactor system behavior during real and hypothetical transient scenarios were developed. One of such advanced, best-estimate reactor systems codes is TRAC/RELAP Advanced Computational Engine (TRACE), developed by the U.S. Nuclear Regulatory Commission. The advanced TRACE comes with a graphical user interface called SNAP (Symbolic Nuclear Analysis Package). It is intended for pre- and post-processing, running codes, RELAP5 to TRACE input deck conversion, input deck database generation etc. The TRACE code is still not fully development and it will have all the capabilities of RELAP5. The purpose of the present study was therefore to assess the 3D capability of the TRACE on BETHSY 9.1b test. The TRACE input deck was semi-converted (using SNAP and manual corrections) from the RELAP5 input deck. The 3D fluid dynamics within reactor vessel was modeled and compared to 1D fluid dynamics. The TRACE 3D calculation was compared both to TRACE 1D calculation and RELAP5 calculation. Namely, the geometry used in TRACE is basically the same, what gives very good basis for the comparison of the codes. The only exception is 3D reactor vessel model in case of TRACE 3D calculation. The TRACE V5.0 Patch 1 and RELAP5/MOD3.3 Patch 4 were used for calculations. The BETHSY 9.1b test (International Standard Problem no. 27 or ISP-27) was 5.08 cm equivalent diameter cold leg break without high pressure safety injection and with delayed ultimate procedure. BETHSY facility was a 3-loop replica of a 900 MWe FRAMATOME pressurized water reactor. In general, all presented code calculations were in good agreement with the BETHSY 9.1b test. The TRACE 1D calculation results are comparable to RELAP5 calculated results. For some parameters they are better, this is mostly due to better tuning of the break flow, what influences timing of the transient. When comparing TRACE 1D and TRACE 3D calculation, the latter is slightly better. One reason for comparable results is already good agreement of 1D calculations and there was not much space to further improve the results. The other reason may be that in the facility the phenomena were mostly one dimensional (for example, external downcomer was used for reactor vessel modeling). However, when 3D behavior of the heater rod temperatures was investigated, the advantage of three dimensional treatment was clearly demonstrated.

Design of multi-layer grid in a giant electrostatic collection vessel of ultra-high sensitivity radon monitor

Developing an ultra-high sensitivity electrostatic collection radon monitor benefits the scientific experiments of China Jinping Underground Laboratory. Here, a one cubic meter electrostatic collection vessel with a multi-layer hemispherical metal grid was designed to increase the collection efficiency of positively charged Po-218 ions. The 3D model of the giant electrostatic collection vessel was constructed using the COMSOL Multiphysics simulation software, and the potential and electric field distributions in the vessel were simulated. Numerical simulation results were obtained according to the different radii and voltages applied to the grid. The electric field between the vessel wall and grid, between two grids, and between the grid and surface of the PIPS detector must be set uniformly to reduce the collection time of the positively charged Po-218 ions. Simulation results showed that setting a charged metal grid in the vessel can optimize the electric field distribution, and setting a two-layer charged metal grid in the giant vessel can further increase the cost performance. The average collection times of the electrostatic collection vessel with the two-layer grid along the vertical and oblique lines approximately 15% and 13% of that without the grid. The rates of positively charged Po-218 ions that could pass through the one and two-layer metal grids were 86.78% and 50%. Optimizing the electric field can greatly increase the sensitivity of radon monitors and reduce the humidity restrictions.

Aneurysmal wall enhancement and hemodynamics: pixel-level correlation between spatial distribution.

Inflammation and hemodynamics are interrelated risk factors for intracranial aneurysm rupture. This study aimed to identify the relationship between these risk factors from an individual-patient perspective using biomarkers of aneurysm wall enhancement (AWE) derived from high-resolution magnetic resonance imaging (HR-MRI) and hemodynamic parameters by four-dimensional flow MRI (4D-flow MRI). A total of 29 patients with 29 unruptured intracranial aneurysms larger than 4 mm were included in this prospective cross-sectional study. A total of 24 aneurysms had AWE and 5 did not have AWE. A three-dimensional (3D) vessel model of each individual aneurysm was generated with 3D time-of-flight magnetic resonance angiography (3D TOF-MRA). Quantification of AWE was sampled with HR-MRI. Time-averaged wall shear stress (WSS) and oscillatory shear index (OSI) were calculated from the 4D-flow MRI. The correlation between spatial distribution of AWE and hemodynamic parameters measured at pixel-level was evaluated for each aneurysm. In aneurysms with AWE, the spatial distribution of WSS was negatively correlated with AWE in 100% (24/24) of aneurysms, though 2 had an absolute value of the correlation coefficient <0.1. The OSI was positively correlated with AWE in 91.7% (22/24) of aneurysms; the other 2 aneurysms showed a negative correlation with AWE. In aneurysms with no AWE, there was no correlation between WSS (100%, 5/5), OSI (80%, 4/5), and wall inflammation. The spatial distribution of WSS was negatively correlated with AWE in aneurysms with AWE, and OSI was positively correlated with AWE in most aneurysms with AWE. While aneurysms that did not contain AWE showed no correlation between hemodynamics and wall inflammation.

Offline Programming Guidance for Swarm Steering of Micro-/Nano Magnetic Particles in a Dynamic Multichannel Vascular Model

Magnetically targeted drug delivery (MTD) systems are used in the treatment of various diseases. However, few studies on the targeting of micro-/nano-sized magnetic particles (MPs) inside multi-bifurcations vessel with fluid flow have appeared. Here, we present a user-interface offline programming guidance (OLPG) scheme that controls MPs within a multi-channel dynamic vascular model. The OLPG scheme can simplify the guidance complexity for MTD and overcome the difficulties in real-time sensing of magnetic nanoparticles (MPs). Calibration between real and virtual environments minimizes OLPG errors due to the aggregation properties of the MPs. A Swarm of Aggregated MPs (SAMPs) can be defined experimentally as the equivalent diameter of a single MP. The joystick position is linearly related to the MP magnetic forces of a real electromagnetic actuator. SAMPs were controlled inside the MTD simulator using the joystick and their control commands can be downloaded to the real controller of the <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in vitro</i> multi-channel vessel model. We performed both simulations and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in vitro</i> studies in the multi-channel vascular model. A user guided MPs to the desired locations in ∼50% of simulations and ∼49.5% of in vitro studies, in the absence of visual feedback. Also, a realistic 3D blood vessel model was simulated to evaluate the feasibility of the OLPG scheme. Our system has a potential to guide <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">in vivo</i> drug delivery.

Quantification of aneurysm wall enhancement in intracranial fusiform aneurysms and related predictors based on high-resolution magnetic resonance imaging: a validation study.

Background:Aneurysm wall enhancement (AWE) in high-resolution magnetic resonance imaging (HR-MRI) has emerged as a new imaging biomarker of intracranial aneurysm instability.Objective:To determine a standard method of AWE quantification for predicting fusiform intracranial aneurysms (FIAs) stability by comparing the sensitivity of each parameter in identifying symptomatic FIAs. The predictors of AWE and FIA types were also identified.Methods:We retrospectively analyzed consecutive fusiform aneurysm patients who underwent HR-MRI from two centers. The aneurysm-to-pituitary stalk contrast ratio (CRstalk), aneurysm enhancement ratio, and aneurysm enhancement index were extracted, and their sensitivities in discriminating aneurysm symptoms were compared using the receiver-operating characteristic curve. Morphological parameters of fusiform aneurysm were extracted based on 3D vessel model. Uni- and multivariate analyses of related predictors for AWE, CRstalk, and FIA types were performed, respectively.Results:Overall, 117 patients (mean age, 53.3 ± 11.7 years; male, 75.2%) with 117 FIAs underwent HR-MRI were included. CRstalk with the maximum signal intensity (CRstalk-max) had the highest sensitivity in identifying symptomatic FIAs with an area under the curve value (0.697) and a cut-off value of 0.90. The independent predictors of AWE were aneurysm symptoms [(odds ratio) OR = 3.754, p = 0.003], aspirin use (OR = 0.248, p = 0.037), and the maximum diameter of the cross-section (OR = 1.171, p = 0.043). The independent predictors of CRstalk-max were aneurysm symptoms (OR = 1.289, p = 0.003) and posterior circulation aneurysm (OR = 1.314, p = 0.001). Transitional-type showed higher rates of hypertension and mural thrombus over both dolichoectatic- and fusiform-type FIAs.Conclusion:CRstalk-max may be the most reliable parameter to quantify AWE to distinguish symptomatic FIAs. It also has the potential to identify unstable FIAs. Several factors contribute to the complex pathophysiology of FIAs and need further validation in a larger cohort.

The thermal performance of five different viscosity models in the kidney blood vessel with multi-phase mixture of non-Newtonian fluid models using computational fluid dynamics

Computational hemodynamic (CHD) is an engineering tool and a good approach that helped many physicians to obtain much information about the situation of the patient in a lot of diseases like cardiovascular disease, even surgery, etc. The dispersion of blood cells in the plasma is heterogeneous. Therefore, the blood fluid is a multi-phase mixture of non-Newtonian fluid. Numerical calculation of non-Newtonian viscosity models of blood flow parameter includes Reynolds number; different wall heat fluxes in three situations of the body (sleeping, standing and running), etc. are investigated. To construct a 3D model of the kidney blood vessel, an open-source software program using Digital Imaging and Communications in Medicine (DICOM) and Magnetic Resonance Image (MRI) is used. Additionally, the vessel wall is considered solid. The finite volume approach and SIMPLE scheme are used. The non-Newtonian blood flow is considered as a laminar flow. All of these heat fluxes generated by the body in different situations have their impact on the reported parameters in this paper. The reported parameters included dimensionless numbers like pressure drop, average wall shear stress, heat transfer coefficient, and temperature. The power-law non-Newtonian viscosity model makes the velocity gradient more than other non-Newtonian viscosity models. Also, the power-law model represents a higher heat transfer.

A deep learning oriented method for automated 3D reconstruction of carotid arterial trees from MR imaging.

The scope of this paper is to present a new carotid vessel segmentation algorithm implementing the U-net based convolutional neural network architecture. With carotid atherosclerosis being the major cause of stroke in Europe, new methods that can provide more accurate image segmentation of the carotid arterial tree and plaque tissue can help improve early diagnosis, prevention and treatment of carotid disease. Herein, we present a novel methodology combining the U-net model and morphological active contours in an iterative framework that accurately segments the carotid lumen and outer wall. The method automatically produces a 3D meshed model of the carotid bifurcation and smaller branches, using multispectral MR image series obtained from two clinical centres of the TAXINOMISIS study. As indicated by a validation study, the algorithm succeeds high accuracy (99.1% for lumen area and 92.6% for the perimeter) for lumen segmentation. The proposed algorithm will be used in the TAXINOMISIS study to obtain more accurate 3D vessel models for improved computational fluid dynamics simulations and the development of models of atherosclerotic plaque progression.

Effect of 2D/3D registration of 3D vessel model on the accuracy of 3D position estimation of endovascular device

Effect of 2D/3D registration of 3D vessel model on the accuracy of 3D position estimation of endovascular device

Abstract 385: Ultracentrifugation and Depletion Methods of High-density Lipoprotein Isolation Yield Particles That Are Functionally Distinct in Some of Their Vasoprotective Functions

The evaluation of the functional capabilities of high-density lipoproteins (HDL) has been shown to be a better predictor of several cardiometabolic diseases than the traditional marker of plasma HDL cholesterol. An important, yet often unconsidered, factor in clinically evaluating HDL function is the method used to purify HDL from plasma. It has been shown that the method of HDL isolation can impact the composition of the HDL produced yet the effect of these compositional changes on HDL function has been explored very little. The evaluation of isolation method is especially important when studying novel HDL functions, such as the vasoprotective functions relevant to Alzheimer’s disease (AD) recently discovered on brain vascular cells and in 3-dimensional brain vessel models. We compared the high-throughput HDL isolation method of polyethylene glycol (PEG) precipitation to the more time-consuming but purer method of sequential density gradient ultracentrifugation (UC) in several of these novel, brain-relevant vasoprotective functions. HDL isolated from young, healthy human plasma by PEG precipitation and UC were equal in their capacity for effluxing cholesterol from macrophages, suppressing TNFα-induced inflammation in brain endothelial cells, and preventing the pathological accumulation of the AD protein amyloid beta (Aβ) in 3D vessel models. However, only HDL isolated by UC, and not PEG precipitation, could induce nitric oxide production and inhibit Aβ-induced inflammation in brain endothelial cells. These findings highlight the importance of HDL isolation method in the composition and function of the HDL produced. Furthermore, that HDL isolation method affected some but not all HDL functions shows that HDL may act on brain vascular cells to protect against AD through more than one pathway.

Fabrication of 3D Capillary Vessel Models with Circulatory Connection Ports.

Bionic microscopic vessel models can contribute to the development of vascular treatment skills and techniques for clinical training. Most microscopic vessel models are limited to two dimensions, but three-dimensional (3D) models are important for surgery, such as on retina microscopic vessels, for the observation of colon microvessels, for measuring the deformability of red blood cell (RBC), and so on. Therefore, bionic 3D blood vessel models are increasingly in demand. For this reason, it is necessary to establish 3D fabrication techniques for microchannels. In this study, we established two fabrication methods for 3D microfluidic devices for the development of microscopic vessel models. First, we employed an exposure method using photolithographic technology. Second, we employed a 3D method using femtosecond laser and mask hybrid exposure (FMEx). Both methods made it possible to fabricate a millimeter-scale 3D structure with a submicrometer resolution and achieve an easy injection of solution. This is because it was possible to fabricate typical microfluidic channels used for model inlet and outlet ports. Furthermore, in the FMEx method, we employed an acid-diffusion effect using a chemically amplified resist to form a circular channel cross-section. The acid-diffusion effect made it realizable to fabricate a smooth surface independent of the laser scanning line width. Thus, we succeeded in establishing two methods for the fabrication of bionic 3D microfluidic devices with microfluidic channels having diameters of 15–16 µm for mimicking capillary vessels.

Detecting Regions of Stenosis and Aneurysm in a 3D Blood Vessel Model

Detecting Regions of Stenosis and Aneurysm in a 3D Blood Vessel Model

AN INTERACTIVE VIRTUAL INTRACORONARY STENTING SYSTEM BASED ON INTRAVASCULAR ULTRASOUND

Percutaneous transluminal coronary angioplasty (PTCA) is a minimally invasive surgery in the clinical treatment of coronary artery diseases. The successful operation highly depends on the size and accurate placement of the stent. In this study, a virtual stenting (VS) system was designed and implemented to facilitate the planning of PTCA. A three-dimensional (3D) vessel model was reconstructed through fusing intravascular ultrasound (IVUS) sequential images and simultaneously recorded X-ray angiograms during cardiac intervention. The user is allowed to intuitively explore the vessel in an endoscopic manner by creating/updating a fly-through trajectory in the lumen. A virtual stent library including a better variety of commercially available bare metal heart stents was built. The user is allowed to select a proper stent according to the morphology of the vessel and lesion and to move it to the lesion. Also, the user can visually observe the stent expansion/apposition and flexibly adjust its position. The system is used to assist visual diagnosis of the vascular diseases, evaluation of interventional treatment and training of the medical personnel.

Accurate three-dimensional modeling of blood vessels using computer tomography, intravascular ultrasound, and biplane angiogram images

In this paper, we propose a novel method for three-dimensional (3D) blood vessel modeling. Given the incomplete nature of medical images, a 3D blood vessel model that includes the intima and adventitia cannot be reconstructed with a single medical image. Intravascular ultrasound (IVUS) images are used to obtain detailed intimal and adventitial information. Then, X-ray angiogram images are used to calculate the position and orientation of IVUS images. Combining these medical images can generate a 3D blood vessel model. To visualize a blood vessel via IVUS, a catheter that is attached to a miniaturized ultrasound device is inserted into the blood vessel. The insertion of the catheter deforms the blood vessel. This deformation is reflected in the 3D blood vessel model that is developed from combining IVUS and X-ray angiogram images. To address this problem, we propose a novel method for 3D blood vessel modeling using undeformed intimal and adventitial information from IVUS images. Our proposed method has potential applications in hemodynamic analysis.