A method of calculating vessel wall compliance at both ends of an intracranial artery stenosis

Abstract

The application provides a method for calculating the compliance of the blood vessel wall at both ends of an intracranial artery stenosis blood vessel. The method comprises the following steps: S1: obtaining an intracranial blood vessel image, performing specific blood vessel segment image segmentation processing by using a combination of a blood vessel enhancement algorithm, a fast marching algorithm and a geodesic algorithm, and obtaining a blood vessel model; S2: meshing the blood vessel model and calculating a blood vessel outer wall model; S3: according to the blood vessel outer wall model, taking in-vitro blood pressure data as an iteration standard, designing a boundary condition model, and performing fluid-structure coupling simulation operation; and S4: according to the simulation result, calculating the blood vessel compliance and the wall expansion ratio. The method for calculating the compliance of the blood vessel wall at both ends of an intracranial artery stenosis blood vessel has the advantages that the expansion process of the blood vessel of a patient in a cardiac cycle can be simulated only by using ordinary blood vessel imaging images of the patient, and corresponding fluid mechanics and biomechanics parameters can be obtained.

Description

Technical Field

[0001] This invention relates to the field of medical imaging technology, and in particular to a method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery. Background Technology

[0002] Treatment options for early-stage intracranial atherosclerosis remain controversial. Several important clinical studies have shown that percutaneous transluminal angioplasty and stent placement offer limited improvement in efficacy compared to conservative treatment due to perioperative complications in the treatment of intracranial atherosclerosis. However, for patients prone to recurrence with conservative treatment, interventional therapy may be a better option. Therefore, patient risk stratification is crucial for optimizing treatment options.

[0003] Currently, the assessment of intracranial arteriosclerosis mainly relies on the degree of vascular stenosis in medical imaging. However, there is a discrepancy between the degree of vascular stenosis and the severity of vascular function, so hemodynamic parameters are needed to assist in diagnosis.

[0004] Fractional flow reserve (FFR) is the gold standard for assessing coronary hemodynamics; however, its relationship with functional impairments caused by intracranial vascular stenosis lacks supporting clinical research data. In coronary artery applications, algorithms for non-invasive FFR acquisition using hydrodynamic simulations have been validated in multicenter randomized controlled trials. In intracranial artery applications, the use of TOF-MRA or DSA imaging combined with hydrodynamic simulations to obtain similar parameters has also been reported. However, the results of intracranial vascular FFR have not shown a clear association with the occurrence of any particular symptom or prognosis.

[0005] Furthermore, the tangential force (wall shear stress) on the vessel wall, along with the circumferential and axial stresses, are all related to plaque growth and rupture within the vessel wall. Studies on wall shear stress have shown that endothelial cells are sensitive to changes in wall shear stress, and variations in the amplitude and pattern of wall shear stress are associated with endothelial cell dysfunction, inflammation, and symptom development. Non-invasive acquisition of intracranial vessel wall shear stress using medical imaging and hemodynamics has revealed that the ratio of wall shear stress at both ends of the stenosis is associated with functional prognosis. Simultaneously, the stress within the vessel wall is approximately ten thousand times greater than shear stress, and vascular remodeling caused by vessel wall stress compensates for changes in shear stress caused by stenosis. Research has shown that flow-induced vasodilation (FMD) is a characteristic of endothelial cell function; in asymptomatic atherosclerotic stenosis, impaired FMD is a manifestation of endothelial cell dysfunction. In carotid artery ultrasound studies, vascular compliance (stiffness) has also been used as a new research indicator for functional prognosis after acute ischemic stroke. The effects of blood flow on vasodilation, vascular stiffness, and vascular compliance may also be related to the brain's autoregulatory mechanisms of blood flow. This may be because, upon sensing a decrease in flow or pressure due to stenosis, the brain adjusts vascular elasticity to compensate for the reduced blood reserve, resulting in vascular dilation or changes in compliance (stiffness). Therefore, non-invasively obtaining reference parameters for the compliance of the responsible intracranial vascular segment can play a positive auxiliary role in the selection of treatment methods and prognosis prediction for arteriosclerotic vascular stenosis.

[0006] Therefore, it is necessary to provide a new method for calculating the compliance of the vessel walls at both ends of stenotic intracranial arteries to solve the above-mentioned technical problems. Summary of the Invention

[0007] The technical problem solved by this invention is to provide a method for calculating the compliance of the vessel walls at both ends of an intracranial arterial stenosis by calculating the difference in vascular compliance before and after a stenotic segment of the intracranial vessel and the coefficient of expansion of the vessel wall, in order to assess the impact of changes in the hemodynamics and mechanical environment of the stenotic segment of the vessel wall.

[0008] To solve the above-mentioned technical problems, the method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery provided by the present invention includes the following steps: S1: Acquire intracranial vascular images, and use a combination of vascular enhancement algorithm, fast traversal algorithm and geodesic algorithm to perform image segmentation processing of specific vascular segments to obtain vascular models; S2: Mesh the blood vessel model and calculate and generate the outer wall of the blood vessel to complete the complete blood vessel model with a solid blood vessel wall; S3: Perform fluid dynamics simulation calculations based on the complete vascular model and with in vitro blood pressure measurement data as boundary conditions; S4: Calculate the ratio of vascular compliance to wall expansion based on the simulation results.

[0009] Preferably, step S1 specifically includes the following steps: S11: Enhance the acquired intracranial vascular images: Linear interpolation resampling is performed on intracranial vascular images to convert the resolution of each direction to be consistent; multi-scale smoothing is performed on intracranial vascular images; the Hessian matrix and its eigenvalues ​​of each pixel are calculated in the smoothed images at each scale, and the eigenvalues ​​are used to enhance the tubular parts in the intracranial vascular images. S12: Using the fast traversal algorithm to select and segment tubular vessel segments in enhanced intracranial vascular images: The S11 calculation results are displayed in an interactive window. The starting point of the required blood vessel segment is manually marked. Within this range, the rapid traversal method is used to obtain the range of the blood vessel segment. S13: Using the geodesic algorithm, smooth the outer wall of the blood vessel segment obtained after step S12 to obtain a smooth outer wall of the blood vessel segment.

[0010] Preferably, the linear interpolation resampling calculates the coordinates of the resampling point based on the intracranial vascular image resolution and the target resolution, and calculates the intensity value within the new coordinates based on the intensity difference of the original image. During the multi-scale Gaussian smoothing process, the scale selection is based on the scale of the target blood vessel within the specified range, and the smoothing process follows formula (1):

[0011] Where * denotes convolution calculation, and G is a Gaussian function. Let I be the variance of the Gaussian function, I be the pixel intensity of the original image, and x represent the pixel position. The eigenvalues ​​of the matrix are obtained after calculating the Hessian matrix for each point. , , The three are arranged in ascending order of absolute value, and the enhancement value v of each pixel is calculated according to formula (2) to replace the original pixel intensity. Method (2) is as follows:

[0012] The enhanced head image, after thresholding, reveals the outline of intracranial blood vessels. The viewer rotates the 3D interactive window implemented in the VTK framework and selects the starting point of the desired blood vessel segment using mouse operation. After obtaining the starting point, the colliding fronts algorithm is used in the enhanced image to obtain the desired blood vessel segment. The obtained blood vessel segment is used as the input original image, and the inverse of the gradient of the enhanced image is used as the image feature. The geodesic algorithm is then used to smooth and optimize the blood vessel surface.

[0013] Preferably, step S2 specifically includes the following steps: S21: Use the Marching Cubes algorithm to complete the initial meshing of the outer wall of the smooth vessel segment and obtain the initial mesh data; S22: Use the Laplace algorithm to smooth the initial grid data to obtain intermediate grid data; S23: Manually trim the entrances and exits of the intermediate grid data in the interactive window to obtain the vascular inner wall grid model; S24: The Delaunay triangulation rule is used to re-divide the obtained vascular inner wall mesh model to obtain a new vascular inner wall mesh model. S25: Calculate the centerline of the new blood vessel inner wall mesh model: S26: Based on the centerline of the grid and the new blood vessel inner wall grid model generated in step S24, expand the blood vessel inner wall grid outward, and perform unequal displacement outward according to the centerline information to generate a smooth blood vessel outer wall. Then fill the position between the two blood vessel walls to form a complete blood vessel grid, calculate the volume grid, and finally generate the blood vessel outer wall model.

[0014] Preferably, when creating the outer wall model of a blood vessel, the center line of the inner wall of the blood vessel is first created, and the radius of the maximum inscribed circle is calculated. Then, the inner wall surface is moved outward according to the corresponding blood vessel radius at a certain ratio, and a mesh connection is established between the inner and outer walls to generate a volume mesh. Finally, the mesh of the inlet and outlet end faces of the blood vessel is calculated and generated to complete the complete blood vessel model including the outer wall required for fluid dynamics simulation.

[0015] Preferably, step S3 specifically includes the following steps: S31: Establishment of the boundary condition model: Fluid dynamics calculations were performed on the blood vessel model using flow inlet and pressure outlet, and a boundary condition algorithm model was designed. The outlet pressure uses a flow-resistance model, where pressure is the product of flow rate and resistance. Resistance is inversely proportional to the outlet area, and empirical values ​​are used for calculation. The inlet flow rate is obtained by using an iterative method to make the inlet flow rate in the simulation results equal to the blood pressure measured by the patient in vitro. S32: Perform fluid dynamics simulation calculations using the above boundary conditions on the complete blood vessel model: Based on the SIMPLE method for iteratively solving the partial differential equations in the fluid part of the vascular model, an inlet flow iterative loop is added to the outside. The inlet flow is adjusted to the point where the pressure result reaches the measured blood pressure value. The fluid-structure interaction part adopts implicit update bidirectional coupling. The outer loop is performed in the normal fluid-structure interaction mode. First, the simulation calculation of diastolic pressure is performed. After stable convergence, the inlet pressure is adjusted to systolic pressure, and the calculation is performed again until convergence, thus completing the simulation calculation of the corresponding states of diastolic and systolic pressure.

[0016] Preferably, in step S32, mesh re-division is not performed during fluid dynamics simulation calculations to ensure that the mesh nodes correspond between the two calculations.

[0017] Preferably, for fluid dynamics simulation, based on the conventional SIMPLE process for solving the Navier-Stokes equations and the fluid-structure interaction process, an iterative loop for fitting the inlet blood pressure is added to the outer layer. The pressure fitting iteration is achieved by adjusting the inlet flow rate, specifically by calculating the inlet flow rate according to formula (3), where For entry speed, The relaxation coefficient is... It is the fitting target for the inlet pressure. , The inlet and outlet pressures are calculated using formula (3) as follows:

[0018] Preferably, the diastolic and systolic blood pressure states are calculated once according to formula (3). After the patient's images and blood pressure data are calculated through the above process, the pressure, flow velocity, and wall shear force field distribution in the entire calculation domain are obtained. The vascular compliance is calculated using formula (4) based on the pressure field p and velocity v of the two states obtained. The subscripts s and d represent the systolic and diastolic phases, respectively. Formula (4) is as follows:

[0019] Preferably, step S4 specifically includes the following steps: After the simulation calculations for diastolic and systolic blood pressure are completed, vascular compliance is the pressure change caused by each unit flow change. Based on the simulation results, it is simplified to the pressure difference between systolic and diastolic phases divided by the flow difference. The ratio of the node difference between the two simulation results to the flow difference of the corresponding vascular segment is used as the result of vascular compliance. In addition, the changes in the area of ​​the vascular wall unit are summarized according to the closest center point position, as shown in formula (5). In the formula, the subscript n represents the center line point numbered n, and the subscript m represents the vascular wall triangle number corresponding to the nth center line point. Finally, the vascular surface area expansion ratio is obtained by dividing the larger side value by the smaller side value according to the relationship between the center line point and the stenosis position, as shown in formula (6). The surface area expansion sum in formula (6) is the abbreviation of the surface area expansion sum corresponding to the center line point in formula (5). Formulas (5) and (6) are as follows:

[0020]

[0021] in This is the sum of the changes in the surface area of ​​the blood vessel wall within the vertical section corresponding to the centerline point with index number n. Let be the area of ​​the corresponding m-th patch, with superscripts s and d representing the contraction and relaxation periods, respectively. SDR is the surface area expansion ratio, and in Formula 6, p and d represent the proximal and distal ends, respectively.

[0022] Compared with related technologies, the method for calculating the compliance of the vessel walls at both ends of intracranial arterial stenosis provided by the present invention has the following beneficial effects: This invention provides a method for calculating the wall compliance of stenotic intracranial arteries at both ends. Utilizing routine imaging information from intracranial vascular examinations and peripheral blood pressure data, and employing medical imaging algorithms and fluid dynamics simulation algorithms, iterative fitting is used to calculate hemodynamic parameters, vascular compliance, and the wall area expansion ratio of the stenotic vessel. Retrospective studies have demonstrated the effectiveness of these properties. Vascular compliance and the wall area expansion ratio, as predictive parameters for disease progression or prognosis, provide new insights for the research and treatment of atherosclerotic intracranial vascular stenosis. Attached Figure Description

[0023] Figure 1 A flowchart of the interactive vessel segmentation and meshing algorithm for the method of calculating the compliance of vessel walls at both ends of intracranial arterial stenosis provided by the present invention; Figure 2 The intracranial vessel segmentation algorithm flow is provided by the present invention for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery.

[0024] Figure 3 The flowchart of the vascular meshing process for the method of calculating the compliance of the vessel walls at both ends of stenotic intracranial arteries provided by the present invention is shown. Figure 4 A flow chart of the fluid dynamics simulation of the method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery provided by the present invention; Figure 5 A schematic diagram of the vascular compliance calculation results for the method of calculating the vascular wall compliance at both ends of an intracranial arterial stenosis provided by the present invention; Figure 6 A schematic diagram showing the calculation results of the wall area ratio of the method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery provided by the present invention. Figure 7 This is a schematic diagram of the statistical results of a retrospective study on the method for calculating the compliance of the vessel walls at both ends of stenotic intracranial arteries provided by the present invention. Detailed Implementation

[0025] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0026] Please refer to the following: Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 and Figure 7 ,in, Figure 1 A flowchart of the interactive vessel segmentation and meshing algorithm for the method of calculating the compliance of vessel walls at both ends of intracranial arterial stenosis provided by the present invention; Figure 2 This invention demonstrates the blood vessel segmentation algorithm process for the method of calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery provided by the present invention. Figure 3 The flowchart of the vascular meshing process for the method of calculating the compliance of the vessel walls at both ends of stenotic intracranial arteries provided by the present invention is shown. Figure 4 A flow chart of the fluid dynamics simulation of the method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery provided by the present invention; Figure 5 This is a schematic diagram illustrating the vascular compliance calculation results of the method for calculating the vascular wall compliance at both ends of a stenotic intracranial artery provided by the present invention. Figure 5 (a) Patients, (b) Normal controls; Figure 6 This is a schematic diagram showing the calculation results of the wall area ratio of the method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery provided by the present invention. Figure 6 In the middle, (a) patients with restenosis after interventional treatment and (b) patients who have not experienced restenosis one year after interventional treatment; Figure 7 This is a schematic diagram of the statistical results of a retrospective study on the method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery, as provided in this invention. The method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery is based on intracranial vascular imaging of the patient, including CTA or TOF-MRA images, and peripheral diastolic and systolic blood pressure measured by conventional methods. For patients with the above data, the calculation can be performed according to the method described in this patent, specifically including the following steps:

[0027] S1: Acquire intracranial vascular images, and use a combination of vascular enhancement algorithms, fast traversal algorithms, and geodesic algorithms to perform image segmentation of specific vascular segments to obtain vascular models: S11: Enhance the acquired intracranial vascular images: Linear interpolation resampling is performed on intracranial vascular images to convert the resolution of each direction to be consistent; multi-scale smoothing is performed on intracranial vascular images; the Hessian matrix and its eigenvalues ​​of each pixel are calculated in the smoothed images at each scale, and the eigenvalues ​​are used to enhance the tubular parts in the intracranial vascular images. S12: Using the fast traversal algorithm to select and segment tubular vessel segments in enhanced intracranial vascular images: The S11 calculation results are displayed in an interactive window. The starting point of the required blood vessel segment is manually marked. Within this range, the rapid traversal method is used to obtain the range of the blood vessel segment. S13: Using the geodesic algorithm, smooth the outer wall of the blood vessel segment obtained after step S12 to obtain a smooth outer wall of the blood vessel segment. S2: Mesh the blood vessel model and calculate and generate the outer wall model of the blood vessel: S21: Use the Marching Cubes algorithm to complete the initial meshing of the outer wall of the smooth vessel segment and obtain the initial mesh data; S22: Use the Laplace algorithm to smooth the initial grid data to obtain intermediate grid data; S23: Manually trim the entrances and exits of the intermediate grid data in the interactive window to obtain the vascular inner wall grid model; S24: The Delaunay triangulation rule is used to re-divide the obtained vascular inner wall mesh model to obtain a new vascular inner wall mesh model. S25: Calculate the centerline of the new blood vessel inner wall mesh model: S26: Based on the center line of the grid and the new blood vessel inner wall grid model generated in step S24, expand the blood vessel inner wall grid outward, and according to the center line information, perform unequal displacement outward to generate a smooth blood vessel outer wall. Then fill the position between the two blood vessel walls to form a complete blood vessel grid, calculate the volume grid, and finally generate the blood vessel outer wall model. S3: Perform fluid dynamics simulation calculations based on the blood vessel wall model and using in vitro blood pressure measurement data as boundary conditions: S31: Establishment of the boundary condition model: Fluid dynamics calculations were performed on the blood vessel model using flow inlet and pressure outlet, and boundary condition models were designed. The outlet pressure uses a flow-resistance model, where pressure is the product of flow rate and resistance. Resistance is inversely proportional to the outlet area, and empirical values ​​are used for calculation. An iterative method was used to make the simulated inlet flow rate equal to the patient's blood pressure measured externally. S32: Perform fluid dynamics simulation calculations on the boundary condition model: In the fluid part of the boundary condition model, based on the iterative solution of the partial differential equations using the SIMPLE method, an inlet flow rate iterative loop is added to the outside to adjust the inlet flow rate to the measured value. The fluid-structure interaction part adopts implicit update bidirectional coupling. In the outer loop, it is performed in the normal fluid-structure interaction mode. First, the simulation calculation at diastolic pressure is performed. After stable convergence, the inlet pressure is adjusted to systolic pressure, and the calculation is performed again until convergence is completed, thus completing the simulation calculation of the corresponding states of diastolic and systolic pressure. In the fluid dynamics simulation calculation, the mesh is not re-generated to ensure that the mesh nodes correspond in the two calculations. S4: Based on the simulation results, calculate the ratio of blood vessel compliance to wall expansion: After the simulation calculations for diastolic and systolic blood pressure are completed, vascular compliance is defined as the pressure change caused by each unit change in flow rate. Based on the simulation results, this is simplified to the pressure difference between systole and diastole divided by the flow rate difference. The ratio of the node difference between corresponding nodes in two simulations to the flow rate difference in the corresponding vessel segment is used as the result of vascular compliance. Furthermore, the changes in the area of ​​vessel wall units are summarized according to the nearest center point, and then summarized according to the relationship between the corresponding center point and the stenosis location, obtaining the ratio of wall area expansion within a certain range before and after stenosis as the expansion area ratio. The algorithm flow is as follows: Figure 1 The specific implementation methods for each part are as follows: In the image segmentation part, linear interpolation resampling calculates the coordinates of the resampling points based on the original image resolution and the target resolution, and calculates the intensity value within the new coordinates based on the intensity difference between the original image and the target resolution. During multi-scale Gaussian smoothing, the scale selection is based on the scale of the target blood vessel within the specified range; the closer the scale is to the target blood vessel size, the better the subsequent enhancement results. The smoothing process follows formula (1):

[0028] Where * denotes convolution calculation, and G is the Gaussian function. Let I be the variance of the Gaussian function, I be the pixel intensity of the original image, and x represent the pixel position. The eigenvalues ​​of the matrix are obtained after calculating the Hessian matrix for each point. , , The three are arranged in ascending order of absolute value, and the enhancement value v of each pixel is calculated according to formula (2) to replace the original pixel intensity. Method (2) is as follows:

[0029] Enhanced head images can now display the outlines of intracranial blood vessels after simple thresholding. Viewers can rotate the 3D interactive window and select the starting point of the desired vessel segment using mouse manipulation. After obtaining the starting point, the colliding fronts algorithm is used in the enhanced image to extract the desired vessel portion. The extracted vessel segment is used as the input original image, and the inverse of the gradient of the enhanced image is used as image features. A geodesic algorithm is then applied to smooth and optimize the vessel surface. The visualization process is illustrated below. Figure 2 , Figure 2 In the image, (a) the original image, (b) the enhanced intracranial blood vessels, (c) the result of the colliding front algorithm, and (d) the result of the geodesic algorithm.

[0030] The segmented blood vessel model is converted into a mesh using a standard Marching cubes algorithm for subsequent operations. To create the outer wall model of the blood vessel, the centerline of the inner wall needs to be created first, and the radius of the maximum inscribed circle needs to be calculated. Then, the inner wall surface is moved outward proportionally according to the corresponding blood vessel radius. The inlet and outlet cross-sections of the blood vessel are obtained by connecting the inner and outer walls using boundary lines and center points. After obtaining the two-dimensional meshes of all the required faces of the blood vessel model, a volume mesh is generated, completing the input model required for fluid dynamics simulation.

[0031] For the fluid dynamics simulation part, this invention, based on the conventional SIMPLE process for solving the Navier-Stokes equations and the fluid-structure interaction process, adds an iterative loop for fitting the inlet blood pressure in the outer layer, such as... Figure 4 Pressure fitting iteration is achieved by adjusting the inlet flow rate, specifically by calculating the inlet flow rate according to formula (3), where... For entry speed, The relaxation coefficient is... It is the fitting target for the inlet pressure. , The inlet and outlet pressures are calculated using formula (3) as follows:

[0032]

[0033] The algorithm requires separate calculations of diastolic and systolic blood pressure throughout the process, followed by further processing based on the results of the two calculations. Specifically, this can be achieved by adding an outer iterative loop within the OpenFOAM open-source package, or by implementing pressure fitting iterations through an additional outer loop in a commercial solver that supports user-defined functions.

[0034] After the patient's imaging and blood pressure data are processed through the above steps, the pressure, flow velocity, and wall shear force field distributions across the entire computational domain can be obtained. In this invention, vascular compliance is calculated using formula (4) based on the obtained pressure field and flow rate in two states, where the subscripts s and d represent the systolic and diastolic phases, respectively. The results are shown below. Figure 5 .

[0035]

[0036] To verify the utility of this parameter, a retrospective analysis was conducted on images of 36 patients who had undergone middle cerebral artery interventional surgery for intracranial vascular stenosis and 36 healthy controls. The results showed that the compliance ratio of the inlet and outlet vessels of the stenotic vessel was an independent predictor of the need for interventional surgery. Figure 6Furthermore, the wall area changes are summarized and compared with the corresponding centerline position according to formula (5). In the formula, the area of ​​each model unit is the area of ​​all model units closest to the center point n. After summing, the ratio of the summation of the centerline relative to the narrow point is obtained according to formula (6), resulting in the wall expansion area ratio. Figure 7 .

[0037]

[0038]

[0039] in This is the sum of the changes in the surface area of ​​the blood vessel wall within the vertical section corresponding to the centerline point with index number n. Let be the area of ​​the corresponding m-th patch, with superscripts s and d representing the contraction and relaxation periods, respectively. SDR is the surface area expansion ratio, and in Formula 6, p and d represent the proximal and distal ends, respectively.

[0040] To achieve fluid dynamics simulation based on patient images, a workflow is required, including vessel segmentation, meshing, fluid dynamics simulation, and post-processing. Each step involves several algorithms or combinations of algorithms. Commonly used algorithms for vessel segmentation include vessel enhancement algorithms, level set algorithms, and statistical plus posterior probability algorithms. Due to the needs of intracranial arterial fluid dynamics applications, vessel segmentation problems often only require specific vessel segments. Therefore, the workflow frequently uses the initial overall segmentation results for rendering and 3D interaction, selecting the starting point of the desired vessel segment before proceeding to the next step of fine segmentation. The image segmentation result is still a 3D voxel structure, while fluid dynamics calculations typically use mesh-based models. Therefore, the traveling block algorithm is first used to convert the result into a mesh structure. To improve mesh quality, the mesh is then smoothed using Labras smoothing. For high-resolution or multimodal images, the vessel wall can be directly obtained through image segmentation. For images where the vessel wall structure cannot be directly obtained from the image, a vessel wall model with a thickness proportional to the vessel lumen radius can be obtained using centerline and mesh expansion algorithms.

[0041] In fluid dynamics simulation algorithms, laminar flow models are generally considered suitable for arterial blood flow. For simulation boundary conditions, under ideal conditions, all necessary boundary condition measurements can be obtained, such as using phase-contrast NMR to obtain the flow velocities at all vessel inlets and outlets, in which case the measurements can be used directly. However, in most cases, only some or no measurements of the patient's own condition are available. For measurements of the patient's inlet flow rate, such as the vessel inlet velocity obtained from carotid ultrasound or transcranial Doppler ultrasound, the outlet pressure can be obtained as a boundary condition using a resistance model or bellows model. When no other measurements can be obtained as boundary conditions, some studies have proposed comparing and iterating the simulation results with in vitro measured diastolic and systolic blood pressure to obtain a simulation that reflects the individual patient's condition.

[0042] After simulation, existing post-processing parameters mainly include wall shear force, time-averaged shear force (TAWSS), shear oscillation coefficient (OS I), relative residence time (RRT), and lateral shear force (TransWSS). This patent proposes to calculate the blood pressure and flow distribution obtained by iterative calculation of blood flow state at diastolic and systolic pressure, and calculate the global vascular compliance result according to the definition of vascular compliance in the bellows model. The difference and the ratio of diastolic to systolic vascular area expansion can be used as important parameters for auxiliary diagnosis.

[0043] Compared with related technologies, the method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery provided by this invention has the following advantages: This invention provides a method for calculating the wall compliance of stenotic intracranial arteries at both ends. Utilizing only routine intracranial vascular imaging and peripheral blood pressure data, the method employs medical imaging algorithms and hydrodynamic simulation algorithms, through iterative fitting, to calculate hemodynamic parameters, vascular compliance, and the wall area expansion ratio of the stenotic vessel. Retrospective studies have demonstrated the effectiveness of these properties. Vascular compliance and the wall area expansion ratio, as predictive parameters for disease progression or prognosis, provide new insights for the research and treatment of atherosclerotic intracranial vascular stenosis.

[0044] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery, characterized in that, Includes the following steps: S1: Acquire intracranial vascular images, and use a combination of vascular enhancement algorithm, fast traversal algorithm and geodesic algorithm to perform image segmentation processing on specific vascular segments to obtain vascular models; S2: Mesh the blood vessel model and calculate and generate the outer wall of the blood vessel to complete the complete blood vessel model with a solid blood vessel wall; S3: Based on the complete vascular model and using in vitro blood pressure measurement data as boundary conditions, perform fluid dynamics simulation calculations; S3 specifically includes the following steps: S31: Establishment of the boundary condition model: Fluid dynamics calculations were performed on the blood vessel model using flow inlet and pressure outlet, and a boundary condition algorithm model was designed. The outlet pressure uses a flow-resistance model, where pressure is the product of flow rate and resistance. Resistance is inversely proportional to the outlet area, and empirical values ​​are used for calculation. An iterative method was used to adjust the inlet flow rate so that the blood vessel inlet pressure in the simulation results was equal to the blood pressure measured in vitro by the patient. S32: Perform fluid dynamics simulation calculations using the above boundary conditions on the complete blood vessel model: Based on the SIMPLE method for iteratively solving the partial differential equations in the fluid part of the vascular model, an inlet flow iterative loop is added to the outside to adjust the inlet flow until the pressure result reaches the measured blood pressure value. The fluid-structure interaction part adopts implicit update bidirectional coupling. The outer loop is performed in the normal fluid-structure interaction mode. First, the simulation calculation of diastolic pressure is performed. After stable convergence, the inlet pressure is adjusted to systolic pressure and calculated again until convergence is completed, thus completing the simulation calculation of the corresponding states of diastolic and systolic pressure. In fluid dynamics simulation, based on the conventional SIMPLE process for solving the Navier-Stokes equations and the fluid-structure interaction process, an iterative loop for fitting the inlet blood pressure is added to the outer layer. The pressure fitting iteration is achieved by adjusting the inlet flow rate, specifically by calculating the inlet flow rate according to formula (3). For entry speed, The relaxation coefficient is... It is the target for fitting the inlet pressure. , The inlet and outlet pressures are calculated using formula (3) as follows: ； S4: Based on the simulation results, calculate the ratio of blood vessel compliance to wall expansion; S4 specifically includes the following steps: After the simulation calculations for diastolic and systolic blood pressure are completed, vascular compliance is the pressure change caused by each unit flow change. Based on the simulation results, it is simplified to the pressure difference between systolic and diastolic phases divided by the flow difference. The ratio of the pressure difference at the corresponding node between the two simulation results to the flow difference in the corresponding vascular segment is used as the result of vascular compliance. Furthermore, the changes in the area of ​​the vascular wall unit are summarized according to the closest centerline point, as shown in formula (5), where the subscript n in the formula represents the centerline point numbered n, and the subscript m represents the vascular wall triangular facet number corresponding to the nth centerline point. Finally, the changes are summarized according to the relationship between the centerline point and the stenosis location, and the vascular surface area expansion ratio is obtained by dividing the larger side by the smaller side. ;in This is the sum of the changes in the surface area of ​​the blood vessel wall within the vertical section corresponding to the centerline point with index number n. Let be the area of ​​the corresponding m-th patch, with superscripts s and d representing the contraction and relaxation phases, respectively.

2. The method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery according to claim 1, characterized in that, S1 specifically includes the following steps: S11: Enhance the acquired intracranial vascular images: Linear interpolation resampling is performed on intracranial vascular images to convert the resolution of each direction to be consistent; multi-scale smoothing is performed on intracranial vascular images; the Hessian matrix and its eigenvalues ​​of each pixel are calculated in the smoothed images at each scale, and the eigenvalues ​​are used to enhance the tubular parts in the intracranial vascular images. S12: Using the fast traversal algorithm to select and segment tubular vessel segments in enhanced intracranial vascular images: The S11 calculation results are displayed in an interactive window. The starting point of the required blood vessel segment is manually marked. Within this range, the rapid traversal method is used to obtain the range of the blood vessel segment. S13: Using the geodesic algorithm, smooth the outer wall of the blood vessel segment obtained after step S12 to obtain a smooth outer wall of the blood vessel segment.

3. The method for calculating the compliance of the vessel walls at both ends of a stenotic intracranial artery according to claim 1, characterized in that, In step S32, mesh re-division is not performed during fluid dynamics simulation calculations to ensure that the mesh nodes correspond between the two calculations.

Images