Computer simulation system for analyzing blood flow state
The computer simulation system using MRI and CFD simulations addresses the complexity and invasiveness of existing methods by providing rapid and accurate blood flow analysis in cerebral and coronary arteries, facilitating non-invasive diagnosis and surgical planning.
Patent Information
- Application Number
- JP2023073088
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2023-04-27
- Publication Date
- 2024-11-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing methods for analyzing blood flow in cerebral arteries and coronary arteries are complex, time-consuming, and invasive, making it difficult for medical personnel to perform quick and accurate diagnoses.
A computer simulation system using MRI and CFD simulations to obtain blood flow velocity and pressure in blood vessels, employing phase contrast MRI for blood velocity data, MRA for vessel shape, and CFD simulations with optimized assimilation surfaces and boundary conditions to calculate blood flow parameters.
Enables rapid, non-invasive, and accurate diagnosis of blood flow conditions, allowing for quick visualization and prediction of surgical outcomes for aneurysms and vascular malformations.
Smart Images

Figure 2024158140000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to a computer simulation system for obtaining blood flow velocity and pressure in the cerebral arteries, coronary arteries, aorta, and the like of an individual subject. [Background technology]
[0002] Conventionally, technologies have been introduced that use devices that use US (Ultra Sonic) or MRI (Magnetic Resonance Imaging) to perform image analysis of blood vessels and determine the state of blood flow. Patent Document 1 introduces a technology that uses US to calculate the shear stress applied to the inner wall of a blood vessel. Patent Document 2 introduces a method of measuring the overall state of the heart using MRI. Patent Document 3 introduces an analytical method that uses MRI to calculate branched blood vessels. Patent Document 4 introduces an invention related to a computer system for determining the pressure inside a blood vessel. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Patent No. 5522719 [Patent Document 2] Patent No. 6301133 [Patent Document 3] Patent No. 6517031 [Patent Document 4] Patent No. 6959391 Summary of the Invention [Problem to be solved by the invention]
[0004] In order to estimate the state of blood flow in organs such as the cerebral arteries and the heart, it is necessary to know the state of blood flow in the blood vessels surrounding those organs. In other words, it is desirable to accurately grasp the speed of blood flow, the flow rate of blood in the blood vessels, and the pressure generated in the blood vessels by actual measurement. Furthermore, non-invasive measurement, including non-contrast measurement, is an essential condition. However, even with the known US and MRI (a technology that performs condition analysis by computer analysis from MRI and US images), for example, the pressure generated in the blood vessels is analyzed by computer, but it is not easy to perform the analysis in a short time of minutes or seconds while maintaining practical operability and appropriate accuracy.
[0005] That is, the previously introduced technologies require very complicated operations and calculations, and require a long time for analysis. Although theoretically possible, they do not realize a system that allows medical personnel to quickly obtain results for on-site diagnosis. Although the above patent documents pursue theories, they do not mention specific practical diagnostic methods. The only practical catheter diagnosis (a method of direct measurement) is an invasive test, which has the risk of causing a large burden and injury to the subject. [Means for solving the problem]
[0006] The following configurations are each a means for solving the above problems.
[0007] <Configuration 1> A computer simulation system that obtains blood flow velocity and pressure within blood vessels through the following process. (1) Set an observation region on the blood vessels. (2) Using pcMRI (phase contrast Magnetic Resonance Imaging), blood flow velocity data of a three-dimensional image of the observation area is obtained as a collection of voxels. Blood vessel shape data is obtained by simultaneously measuring with MRA (Magnetic Resonance Angiography). (3) Set multiple assimilation surfaces consisting of cross sections perpendicular to the centerline of the blood vessel shape between the inlet and outlet of the above observation area. (4) The flow velocity of the voxels on the set assimilation surface is measured to obtain blood flow velocity vector data. (5) In any assimilation surface, one or more of the above-mentioned observed blood flow velocity vector data is set as the observation data to be used for CFD (Computational Fluid Dynamics) simulation using the governing equations (Navier-Stokes equations or lattice Boltzmann equations). (6) In the CFD simulation, blood is an incompressible fluid, and the inlet boundary condition of the observation area is the average flow velocity calculated from the passing flow rate measured on the assimilation surface, the outlet boundary condition is zero pressure, and the boundary condition of the part in contact with the blood vessel wall is set to zero flow velocity. (7) The observational data for all of the above assimilation surfaces and the above boundary conditions are input, and the above CFD simulation is performed to obtain the blood flow velocity and pressure at any location in the observation region.
[0008] <Configuration 2> The computer simulation system according to configuration 1, characterized in that in the CFD simulation, a plurality of blood flow velocity vector data items having the largest absolute values, which are between 5% and 60% of the total, are selected and used from among the blood flow velocity vectors observed on each assimilation surface.
[0009] <Configuration 3> 3. The computer simulation system according to configuration 2, wherein the assimilation surface is disposed at a location 2 to 5 times the inner diameter of the blood vessel from the inlet and outlet of the observation region.
[0010] <Configuration 4> 4. The computer simulation system according to any one of configurations 1 to 3, characterized in that the pixels of the three-dimensional image data of the observation area are divided and isotropically subdivided into voxels. <Component 5> A computer simulation system according to any one of configurations 1 to 3, characterized in that an assimilation surface is automatically set within an observation area at a location where either the torsion or curvature of the blood vessel does not exceed a limit value, and blood flow velocity and pressure are automatically calculated using the blood vessel shape, torsion, and curvature as feature variables. <Component 6> A computer simulation for obtaining blood flow velocity and pressure in a blood vessel by carrying out the steps described in any one of configurations 1 to 3. Effect of the Invention
[0011] It will be possible to perform comparative diagnosis of the condition of subjects during surgery for aneurysms, vascular malformations, and intravascular placement, or during health checkups, and to visualize individual subjects' numerical values. Measurement results can be displayed quickly and easily from non-invasive tests. It will be possible to realize a digital twin that combines actual measurements and simulations to improve accuracy. [Brief description of the drawings]
[0012] [Figure 1] FIG. 13 is an explanatory diagram of the setting of a blood vessel observation region and an assimilation surface. [Diagram 2] FIG. 13 is an explanatory diagram of the boundary condition at the entrance of the observation region. [Diagram 3] FIG. 2 is a functional block diagram of the computer system of the present invention. [Figure 4] 2 is a flowchart showing the operation of the computer system of the present invention. [Diagram 5] FIG. 2 is an explanatory diagram of a governing equation of a fluid. [Figure 6] This is an explanation of the Navier-Stokes equations. [Figure 7] FIG. 1 is an explanatory diagram of the lattice Boltzmann equation. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0013] The present invention is a system that obtains the blood flow velocity and blood flow pressure of each subject by measuring the blood flow velocity of the subject using MRI (Magnetic Resonance Imaging) and performing fluid calculations using CFD (Computational Fluid Dynamics).
[0014] That is, a system is provided that utilizes non-invasive actual measured blood velocity values from MRI, supplements missing pressure values with CFD, and stores inaccurate flow rates, enabling necessary processing with simple operations in minutes or seconds. This makes it relatively easy to perform comparative diagnoses and predictions of the condition of individual subjects before and after surgery for aneurysms, vascular malformations, and intravascular placement. Below, the embodiments of the present invention will be described in detail for each example. EXAMPLES
[0015] In the present invention, the flow velocity and pressure of blood in blood vessels etc. are obtained by the following computer simulation. (1) Observation area setting The subject of the examination is blood vessels around the diagnosis target region in each part of the human body. For example, blood vessels containing diseased areas such as aneurysms. Accurately determining the internal blood flow velocity and pressure is useful for diagnosing abnormalities in blood vessels. It is not practical if it takes a long time to obtain the desired results by photographing the diseased area and processing the data using a computer. The present invention can shorten this processing time to a sufficiently practical level.
[0016] (2) Acquiring 3D image data This device measures the blood vessels of the subject along with the surrounding organs at the same time. MRI is used, which has little impact on the human body and allows for precise observation of blood flow. The shape of the blood vessels is measured using an MRI shooting sequence called MRA (a method of controlling the magnetic field, electric field and signal processing), and the blood flow velocity is measured using pcMRI (phase contrast Magnetic Resonance Imaging).
[0017] The diameter of blood vessels in the human body is approximately 2mm to 15mm. For example, the resolution of a general-purpose MRI device is approximately 1mm. These pixels are divided to obtain a 3D image with a resolution of isotropic voxels with a side length of 0.1mm. This is to improve the accuracy of calculations. If the measurement device has a high resolution and uses isotropic pixels, they can be used as measurement voxels as they are.
[0018] The shape and volume of blood vessels in the observation region are obtained from the voxels of the blood vessel shape of the MRA in the entire observation region. Then, blood flow velocity vectors for each voxel are obtained from the brightness values of pcMRI. The blood flow velocity vectors for each voxel can be directly used for the calculation of governing equations of fluids such as the Navier-Stokes equations and the lattice Boltzmann equations.
[0019] (3) Assimilation surface setting First, the center line of the blood vessel shape is automatically generated. Next, for example, the entire observation area is displayed on the operation screen, and the user is requested to specify the setting position of the assimilation surface. The assimilation surface is set near the entrance and exit of the observation area, and before and after the locations of blood vessel branches and aneurysms. The operator may specify the position on the screen, or an automatic display may be selected using the learning results.
[0020] Since the assimilation surface is used to obtain data that is the basis for calculating blood flow velocity, it is not suitable for locations where the blood vessel shape is curved, twisted, or significantly deformed. This is because the initial calculation error becomes large and additional time is required for calculation processing. Therefore, it is recommended to provide a function that automatically sets the assimilation surface by image analysis at a location before or after a location where the blood vessel branches or has a variance, and where either the twist or curvature of the blood vessel does not exceed a limit value. In addition, blood flow velocity and pressure can be automatically calculated using the blood vessel shape, torsion, and curvature as characteristic variables. For example, in locations where there is a large bend or twist, it is recommended to correct the effective blood flow velocity to be lower than the calculated blood flow velocity, and the effective pressure to be higher than the calculated pressure.
[0021] It is advisable to place the assimilation surface at a location about 2 to 5 times the inside diameter of the blood vessel from the entrance and exit of the observation area. Boundary conditions are set at the entrance and exit of the observation area as described below. After the computer simulation is performed, the conditions for conservation of mass and momentum throughout the entire observation area are satisfied, and the blood flow velocity and blood flow pressure at any location within the observation area can be obtained. However, near the entrance and exit of the observation area, the blood flow gradually converges to the measured flow velocity value at the assimilation surface as the actual measured values are matched with the artificially set boundary conditions. Therefore, the calculation results for this area cannot be used, resulting in a wasted area.
[0022] Since the time required for simulation calculation processing is proportional to the length of the observation area, calculation time ends up being wasted on this wasted area. Therefore, an assimilation surface was installed at a position where consistency is possible near the entrance and exit of the observation area. Since a portion of the actual measured values is used in the assimilation surface, the range that satisfies the condition of conservation of mass (flow rate) can be brought closer to the entrance and exit.
[0023] Experiments have shown that by placing the assimilation surface at a location that is five times or less than the inside diameter of the blood vessel from the entrance and exit of the observation area, it is possible to shorten the wasted distance and expand the effective length of the entire observation area. The condition of flow conservation must be satisfied, so the surface cannot be placed closer to the entrance or exit than twice the inside diameter of the blood vessel. This can effectively reduce the simulation calculation time.
[0024] (4) Voxel blood flow velocity vector data Since the measurement accuracy is low for the low velocity portion of the blood flow velocity vector data of the voxels on the assimilation surface, one or more blood flow velocity vector data with a large absolute value of 60% or less of the total is selected and used. Increasing the number beyond 60% does not have a significant effect on the accuracy of the calculation. The most effective method for processing with the simplest processing and without affecting the calculation speed is about 5% to 30%.
[0025] Since it is not possible to measure the blood flow velocity vector data of all voxels with high accuracy, at most half of the data with high accuracy and large absolute values are selected and used for the CFD simulation. This improves the accuracy of the flow velocity and pressure values obtained in the CFD simulation and strengthens the robustness of the CFD calculation.
[0026] (5) Observation data In any assimilation surface, at least one of the above-mentioned observed blood flow velocity vector data is set as the observed data to be used in the CFD simulation. Either one blood flow velocity vector data set or multiple blood flow velocity vector data sets may be used for one assimilation surface. It has been found that the accuracy is greatly improved when multiple blood flow velocity vector data sets are used.
[0027] When using only one blood velocity vector data, it is recommended to use the maximum value. In the CFD simulation, input the position information of each blood velocity vector, and the position information and area of each assimilation surface, etc., as necessary. The inlet boundary condition of the observation region is the average flow velocity calculated from the passing flow rate measured at the assimilation surface.
[0028] Whether the flow velocity vector of the inlet boundary condition is large or small, the calculation time until the appropriate calculation result is obtained becomes long, and the calculation becomes unstable and not robust. Therefore, using the blood flow velocity vector data obtained on the assimilation surface, the inlet boundary condition of the observation area is set to the average flow velocity calculated from the passing flow rate measured on the assimilation surface near the inlet. This reduces the calculation time with minimal matching and is effective in maintaining robustness.
[0029] (6) Boundary conditions for CFD simulation Blood is an incompressible fluid, and the inlet boundary condition of the observation area is the average flow velocity calculated from the flow rate measured on the assimilation surface, the outlet boundary condition is set to zero pressure, and the boundary condition of the part in contact with the blood vessel wall is set to zero flow velocity. This allows the Navier-Stokes equations and the lattice Boltzmann equations to be used in CFD simulations.
[0030] (7) CFD Simulation As described above, the observation data on all of the above assimilation surfaces and the above boundary conditions are input, the above CFD simulation is executed, and the blood flow velocity and pressure at any location in the observation area are obtained. By automating the above series of operations, robust LBM / CFD can be executed with simple operations. In one example, the simulation calculation results were obtained in a short time of less than a few minutes. By using the above-mentioned techniques, the calculation time was shortened to a practical range. EXAMPLES
[0031] In the computer simulation system of the present invention, as shown in the block diagram of Fig. 3, when observation data 42 acquired in the observation region shown in Fig. 1 and boundary conditions 44 are input to a CFD simulator 40, calculation processing is automatically performed using governing equations 46 of fluid to obtain calculated data 48 indicating the state of blood flow inside a blood vessel. The specific procedure will be described with reference to the flowchart of Fig. 4.
[0032] FIG. 1 shows a blood vessel 12 in the human skull. This is a shape image and blood flow velocity image data (three types of vector component images) captured by MRA and pcMRI. First, in step S11 of the flowchart, the shape of the blood vessel 12 is defined as an observation region 14. The locations of the blood flow inlet 16 and two outlets of the blood vessel 12, i.e., the left outlet 18 and the right outlet 20, are determined. For example, an operator may input instructions using a pointing device such as a mouse while viewing the image of FIG. 1 displayed on a display.
[0033] In the next step S12, the pixels of the captured images (both MRA and pcMRI) are converted into isotropic voxel data with sides of 0.1 mm. Then, in step S13, the center line 15 of the blood vessel 12 is drawn. A program module that executes such processing can be started. The drawing of the center line 15 can also be automatically processed by a known method. In this manner, the image shown in FIG. 1 is processed.
[0034] Next, in step S14, assimilation surface A, assimilation surface B, and assimilation surface C perpendicular to the center line 15 are placed near the entrance 16, left exit 18, and right exit 20 of the blood vessel 12, respectively. Furthermore, assimilation surface D, assimilation surface E, and assimilation surface F are placed on both sides of the branching portion. Assimilation surface D is placed very close to the aneurysm. If there is an aneurysm, the blood flow in the vicinity will be disturbed, for example, in a swirling manner. This is to ensure that this state is clearly reflected in the processing results. As already explained, all of these assimilation surfaces are placed in optimal positions to increase the processing speed of the analysis calculations and improve the accuracy of the analysis results. The operator can specify the placement locations using a pointing device. Automatic processing may also be performed as already explained.
[0035] FIG. 2 is an enlarged view of the vicinity of the blood vessel entrance 16 shown in FIG. 1. An assimilation surface A is placed in the immediate vicinity of this entrance 16. The front view of the assimilation surface A perpendicular to the center line 15 is shown within circle m. The illustrated mesh is an isotropic voxel 24. The entire observation region is represented by a single Cartesian coordinate system. The voxel 24 is a cube, and it is advisable to use the center point as the position coordinate of the voxel 24. In step S15, the flow velocity and direction of the voxel 24 (center point) are measured, and in step S16, the blood flow velocity vector 22 shown in circle n is obtained.
[0036] In step S17, for example, several blood flow velocity vectors 22 with large absolute values are selected as observed data, as shown in circle p in Fig. 2. In step S18, the same process is performed for all assimilation surfaces shown in Fig. 1 to obtain observed data 44 shown in Fig. 3.
[0037] In step S19, the flow rates through all assimilation surfaces shown in Fig. 1 are calculated. For each assimilation surface, the blood flow velocity vectors 22 of all voxels 24 can be measured, as shown in circle p in Fig. 2. By adding up those that are reasonably reliable, the flow rates through each individual assimilation surface can be obtained. From these, the average calculation taking into account branches and small branches not imaged, or the flow rate of the assimilation surface near the inlet can be set. This can be a rough calculation, so high accuracy is not required. This is set as the initial value (boundary condition) of the blood flow at the inlet 16.
[0038] In step S20, as shown in Fig. 1, boundary conditions are set based on the shapes of the inlet 16, the left outlet 18, the right outlet 20, and the blood vessel 12. Then, in step S21, the observation data 44 and the boundary conditions 42 in Fig. 2 are input to the CFD simulator 40 to execute simulation calculation processing. After that, automatic processing can be performed. In step S22, the blood flow velocity and pressure in the observation area 14 in Fig. 1 can be obtained as the processing results. For example, it is advisable to output a list of the states of each part of the blood vessel 12 that has been specified in advance.
[0039] The governing equations 46 of the fluid shown in Figure 3 are the Navier-Stokes equations or the lattice Boltzmann equations, as shown in Figure 5. The general operation when these equations are used in a CFD simulation will be explained using Figures 4 and 5.
[0040] Regardless of which equation is used, the external force F is defined as shown in the figure. The equivalent values calculated from the observed blood flow values are inserted into each location corresponding to the external force F, and a CFD simulation is performed to satisfy the conditions of conservation of momentum and conservation of mass.
[0041] When using the Navier-Stokes equations, the blood flow velocity at coordinates (x, y, z) within a blood vessel and the pressure at that location can be calculated by simultaneously solving equation (1) (the momentum conservation equation) and equation (2) (the continuity equation) in Figure 6.
[0042] On the other hand, when the lattice Boltzmann equation is used, equations (4), (5), and (6) are used, which are modified versions of equation (3) in FIG. 7. In this equation, fluid motion can be reproduced by evolving the particle velocity distribution function, which represents the proportion of particles moving in a limited direction, over time according to a simple algorithm. In other words, the blood flow velocity vector of the voxel can be directly used for f in equation (4). The pressure is derived from the state equation in a single calculation. Furthermore, since the particle velocity distribution function is independent, parallel calculation can be performed efficiently. For this reason, using this equation in the present invention has the greatest effect of shortening the calculation time. [Explanation of symbols]
[0043] 12 Blood vessels 14 Observation Area 15 Center line 16 Entrance 18 Left exit 20 Right exit 22 Blood flow velocity vector 24 Voxels 40 CFD Simulator 42 Observational Data 44 Boundary Conditions 46 Governing Equations of Fluids 48 Calculated Data A~F assimilation surface
Claims
1. A computer simulation system that obtains blood flow velocity and pressure in blood vessels, etc. through the following processing: (1) An observation region is set in the blood vessel. (2) Blood flow velocity data from pcMRI (Phase contrast Magnetic Resonance Imaging) and vascular shape data from MRA (Magnetic Resonance Angiography) are obtained as a collection of three-dimensional voxel data of the observation area. (3) A plurality of assimilation surfaces consisting of cross sections perpendicular to the center line of the blood vessel shape are set between the entrance and exit of the above-mentioned observation area. (4) The flow velocity of the voxels on the set assimilation surface is measured to obtain blood flow velocity vector data. (5) In any assimilation surface, at least one of the above-mentioned observed blood flow velocity vector data is set as the observation data to be used in a CFD (Computational Fluid Dynamics) simulation using the governing equation (the Navier-Stokes equation or the lattice Boltzmann equation). (6) In the CFD simulation, blood is an incompressible fluid, and the inlet boundary condition of the observation area is the average flow velocity calculated from the passing flow rate measured on the assimilation surface, the outlet boundary condition is zero pressure, and the boundary condition of the part in contact with the blood vessel wall is set to zero flow velocity. (7) The observation data on all of the above assimilation surfaces and the above boundary conditions are input, and the above CFD simulation is executed to obtain the blood flow velocity and pressure at any location in the observation region.
2. The computer simulation system according to claim 1, characterized in that in the CFD simulation, a plurality of blood flow velocity vector data items having the largest absolute values, 5% to 60%, of the blood flow velocity vectors observed on each assimilation surface are selected and used.
3. 3. The computer simulation system according to claim 2, wherein the assimilation surface is located at a distance of 2 to 5 times the inner diameter of the blood vessel from the inlet and outlet of the observation region.
4. 4. The computer simulation system according to claim 1, wherein pixels of the three-dimensional image data of the observation area are subdivided into isotropic voxels.
5. 4. The computer simulation system according to claim 1, wherein the assimilation surface is automatically set before or after a location where a blood vessel branches or has an aneurysm, and where either the torsion or curvature of the blood vessel does not exceed a limit value.
6. A computer simulation method for obtaining blood flow velocity and pressure in blood vessels, etc., by the following process. (1) An observation region is set in the blood vessel. (2) Three-dimensional image data of the observation area is obtained as a collection of voxels using pcMRI (Phase contrast Magnetic Resonance imaging) and MRA. (3) A plurality of assimilation planes consisting of cross sections perpendicular to the center line of the blood vessel are set between the entrance and exit of the above-mentioned observation area. (4) The flow velocity of the voxels on the set assimilation surface is measured to obtain blood flow velocity vector data. (5) In any assimilation surface, one or more of the above-mentioned observed blood flow velocity vector data is set as observation data to be used in a CFD (Computational Fluid Dynamics) simulation using the governing equation (the Navier-Stokes equation or the lattice Boltzmann equation). (6) In the CFD simulation, blood is an incompressible fluid, and the inlet boundary condition of the observation area is the average flow velocity calculated from the passing flow rate measured on the assimilation surface, the outlet boundary condition is zero pressure, and the boundary condition of the part in contact with the blood vessel wall is set to zero flow velocity. (7) The observation data on all of the above assimilation surfaces and the above boundary conditions are input, and the above CFD simulation is executed to obtain the blood flow velocity and pressure at any location in the observation region.
Citation Information
Patent Citations
Liquid developer for electrostatic photography
JP1980022719A
Multiplex transmission transmitter
JP1988001133A
Medical image processing device and magnetic resonance imaging device
JP6517031B2
Method and system for patient-specific blood flow modeling
JP6959391B2
