Method for obtaining coronary flow reserve score based on cta images and readable storage medium

Abstract

The application relates to a method for obtaining a coronary flow reserve score based on CTA images and a readable storage medium, the method comprising the following steps: acquiring coronary CTA three-dimensional image data, obtaining centerline points and radius information along the centerline points according to the three-dimensional image data; traversing the centerline points to generate a blood vessel pipeline in combination with the radius information, merging to obtain a reconstructed coronary artery vessel model, extracting a centerline from the model and obtaining a blood vessel cross-sectional area along the centerline; performing flow distribution, including performing flow distribution by using the proportion of the cross-sectional area of the end of a blood vessel branch to the cumulative sum of the cross-sectional areas of all blood vessel branch ends or the proportion of the volume of the blood vessel branch to the cumulative sum of the volumes of all blood vessel branches; according to a bifurcation point of the coronary artery vessel model, dividing the coronary artery vessel model into a bifurcation section and a long straight section by using the distance between the centerline point on the extended branch of the bifurcation point and the bifurcation point; obtaining the pressure drop of the long straight section, obtaining the pressure drop of the bifurcation section, and then obtaining the coronary flow reserve score of each centerline point.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to a method and readable storage medium for obtaining coronary flow reserve based on CTA images. Background Technology

[0002] Fractional flow reserve (FFR) is defined as the ratio of the maximum blood flow that a coronary artery can provide to the myocardium in the presence of stenosis to the maximum blood flow that the myocardium could theoretically receive without stenosis. FFR reflects the relationship between stenotic coronary arteries and myocardial perfusion, and can accurately determine whether stenotic coronary arteries cause hemodynamic disturbances, thus aiding in treatment decisions.

[0003] With the strengthening of the concept of "precision treatment," the clinical application rate of examinations such as FFR and IVUS (intravascular ultrasound) is increasing. FFR is the gold standard for coronary function examination. It is a detection technique that uses a pressure guidewire or pressure microcatheter to directly measure aortic pressure (Pa) and distal pressure (Pd) of the lesion by intervening in the blood vessel. However, FFR examination requires the use of a pressure guidewire; the procedure is invasive, expensive, and some patients may have drug intolerance.

[0004] In recent years, with the development of computational fluid dynamics (CFD), non-invasive CT-FFR (coronary CT flow reserve fraction) has emerged and gradually gained popularity in clinical practice. Computational CT-FFR does not require the use of vasodilators during measurement, and its advantages of being non-invasive, simple, and safe have made it very popular.

[0005] However, computational CT-FFR results depend heavily on the quality of the reconstructed coronary artery model. In practical research applications, manual or semi-automatic methods are often used to segment the vessels, followed by model preprocessing. This process is cumbersome and highly susceptible to the subjectivity of the software operator. After obtaining a qualified model, volumetric mesh generation is required, and mesh refinement is necessary at stenotic areas to ensure computational accuracy. The entire process is complex and time-consuming, failing to meet clinical needs. Summary of the Invention

[0006] Therefore, it is necessary to provide a method for obtaining coronary flow reserve based on CTA images to address the aforementioned technical problems.

[0007] Methods for obtaining coronary flow reserve based on CTA images include:

[0008] Acquire three-dimensional image data of coronary CTA, and obtain the centerline point and the radius information along the centerline point based on the three-dimensional image data;

[0009] The centerline points are traversed, and the radius information is combined to generate vascular channels. After merging, a reconstructed coronary artery vascular model is obtained, and the centerline is extracted from it to obtain the vascular cross-sectional area along the centerline.

[0010] Flow distribution can be carried out by using the proportion of the terminal cross-sectional area of ​​a vascular branch to the sum of the terminal cross-sections of all vascular branches, or the proportion of the volume of a vascular branch to the sum of the volumes of all vascular branches.

[0011] Based on the bifurcation point of the coronary artery model, the coronary artery model is divided into bifurcation segments and long straight segments by using the distance between the centerline point on the extended branch of the bifurcation point and the bifurcation point.

[0012] The pressure drop of the long straight segment is obtained, the pressure drop of the bifurcation segment is obtained, and then the coronary blood flow reserve fraction at each centerline point is obtained.

[0013] Optionally, obtaining the pressure drop of the bifurcation segment specifically includes:

[0014] The pressure drop of the bifurcation segment is obtained by using the ratio of the cross-sectional area of ​​the parent branch to the cross-sectional area of ​​the child branch, the ratio of the flow rate of the parent branch to the flow rate of the child branch, and the angle between the tangent vectors of the extension directions of the parent branch and the child branch.

[0015] Optionally, the pressure drop of the bifurcation segment is obtained by the following formula:

[0016]

[0017] In the formula, ΔP bif The pressure drop at the bifurcation point is ρ, where ρ is the blood density, and u is the blood density. f Let be the blood flow velocity, C be the ratio of the cross-sectional area of ​​the parent branch to the cross-sectional area of ​​the child branch, D be the ratio of the flow rate of the parent branch to the flow rate of the child branch, and θ be the angle between the tangent vectors of the parent branch extension direction and the child branch extension direction.

[0018] Optionally, the traffic allocation is performed using the following formula:

[0019] Qn = Qtotf(A,L)

[0020] In the formula, Q n Let Q be the flow rate of branch n. tot Let A be the total flow rate, A be the array of blood vessel cross-sectional areas, and L be the array of coordinates for the centerline length.

[0021] If flow distribution is based on the proportion of the terminal cross-sectional area of ​​each vascular branch to the sum of the terminal cross-sectional areas of all vascular branches, then...

[0022]

[0023] In the formula, It is the cross-sectional area at the end of branch n.

[0024] Optionally, the traffic allocation is performed using the following formula:

[0025] Qn = Qtotf(A,L)

[0026] In the formula, Q n Let Q be the flow rate of branch n. tot Let A be the total flow rate, A be the array of blood vessel cross-sectional areas, and L be the array of coordinates for the centerline length.

[0027] If flow distribution is based on the proportion of the volume of each vascular branch to the sum of the volumes of all vascular branches, then...

[0028]

[0029] In the formula V n Let n be the volume of branch n, An,i be the cross-sectional area of ​​branch n at position i on the center line, and Ln,i be the length coordinate of branch n at position i on the center line.

[0030] Optionally, the method also includes:

[0031] The reference diameter of the long straight segment is obtained by fitting, which is used to determine whether there is a narrow section in the long straight segment;

[0032] If there is no narrow section, the pressure drop of the long straight section includes friction loss;

[0033] If a narrow section exists, the pressure drop of the long straight section includes friction loss, the widening loss of the narrow section, and the contraction loss of the narrow section.

[0034] Optionally, the fitting process obtains the reference diameter for each centerline point within the long straight segment using any one of the following formulas:

[0035]

[0036]

[0037] In the formula, d prox It is the diameter of the healthy blood vessel proximal to the narrowed point, d dist It is the diameter of the healthy blood vessel distal to the narrowed segment, L ste is the length between the proximal and distal ends of the stenosis, and x is the different center points.

[0038] Optionally, the gradual widening loss of the narrow segment is obtained by the following formula:

[0039]

[0040] In the formula, ΔP exp Let ρ be the gradual loss at each centerline point within the narrow segment, γ be a constant, δ be a constant, L be the length of the narrow segment in the long straight segment, Q be the flow rate in the long straight segment, d0 be the reference diameter at each centerline point, and A0 be the reference area at each centerline point in the narrow segment. m A is the minimum diameter of the narrow segment. m It represents the minimum area of ​​the narrow segment.

[0041] Optionally, obtaining the centerline point and the radius information along the centerline point based on the three-dimensional image data specifically includes:

[0042] Extract the initial center path point and the radius information along the initial center path point from the three-dimensional image data;

[0043] Based on the three-dimensional image data, two-dimensional cross-sections are sequentially obtained along the initial center path point, and the two-dimensional cross-sections include the blood vessel cross-section and its surrounding image.

[0044] For any two-dimensional plane, the search range is set according to the initial center path point in the plane, and the blood vessel wall is located according to the vascular CT threshold.

[0045] If the pixel with the largest CT value within the search range is less than the vascular CT threshold or the CT value of that pixel is greater than the calcification CT threshold, then the corresponding initial center path point is corrected to the vascular wall and mapped to the three-dimensional image data to obtain the corrected centerline point and the radius information along the centerline point.

[0046] This application also provides a readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method for obtaining coronary flow reserve based on CTA images as described in this application.

[0047] The method for obtaining coronary flow reserve based on CTA images in this application has at least the following effects:

[0048] This application utilizes a reconstructed coronary artery model to extract the centerline and the cross-sectional area along the line. Flow allocation is performed using the centerline length and the cross-sectional area along the line, simplifying the flow allocation process. The allocated flow is used to obtain the pressure drop loss, thereby obtaining the FFR value. Therefore, simplifying the flow allocation process also simplifies the FFR acquisition process, improving the efficiency of obtaining the coronary fractional flow reserve. This application obtains the pressure drop separately for the long straight segment and the bifurcation segment, ensuring the accuracy of the obtained coronary fractional flow reserve at each centerline point. Attached Figure Description

[0049] Figure 1This is a flowchart illustrating a method for obtaining coronary flow reserve based on CTA images in one embodiment of this application.

[0050] Figure 2 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0052] One embodiment of this application provides a method for obtaining coronary flow reserve based on CTA images, including steps S10 to S50. Wherein:

[0053] Step S10: Obtain three-dimensional image data of coronary CTA, and obtain the centerline point and the radius information along the centerline point based on the three-dimensional image data;

[0054] Step S10 specifically includes steps S100 to S300.

[0055] Step S100: Acquire three-dimensional image data of coronary CTA (coronary CT angiography).

[0056] Step S200: Extract the first center path point (initial center path point) and the radius information along the first center path point from the 3D image data. The extraction method can be, for example, using a trained convolutional neural network.

[0057] Step S300: Correct the first center path point to obtain the corrected second center path point and the radius information along the second center path point. The second center path point obtained in step S300 is the centerline point.

[0058] Step S300 specifically includes steps S310 to S330. Wherein:

[0059] Step S310: Based on the three-dimensional image data, obtain the two-dimensional cross-section along the first central path point in sequence. The two-dimensional cross-section includes the blood vessel cross-section and its surrounding image.

[0060] Step S310 specifically includes steps S311 to S330. Wherein:

[0061] Step S311: For any first central path point, obtain the tangent vector, normal vector, and binormal vector of the first central path point on the curve in which it is located.

[0062] Before proceeding to step S310, the first center path points can be connected sequentially to form a curve and smoothed multiple times. The smoothed first center path points are then used to execute step S310.

[0063] Step S312: Based on the tangent vector, normal vector, and binormal vector, as well as the three-dimensional coordinates of the first center path point, obtain the cutting matrix of the first center path point, and then obtain the two-dimensional tangent of the first center path point.

[0064] That is, the cutting matrix at each first center path point is calculated based on the tangent vector, normal vector, binormal vector, and three-dimensional coordinates, and the two-dimensional cross-section of the image is cut out at that point using the cutting matrix.

[0065] Step S320: For any two-dimensional plane, set the search range according to the first center path point in the plane and locate the blood vessel wall according to the blood vessel CT threshold.

[0066] Furthermore, the search range is set based on the first center path point within the plane, specifically including: using the first center path point as the center and setting the search range with a preset radius. The preset radius can be, for example, 1.5mm.

[0067] Step S330: If the pixel with the largest CT value within the search range is less than the vascular CT threshold, or the pixel with the smallest CT value within the search range is greater than the calcification CT threshold, then the corresponding first central path point is corrected to the vessel wall and mapped to the three-dimensional image data to obtain the corrected second central path point (i.e., the centerline point) and the radius information along the centerline point. The radius information along the centerline point can be the radius information of the first central path point before correction, and after correction, the corresponding radius information becomes the radius information of the second central path point, which is the centerline point.

[0068] In step S330, the vascular CT threshold is obtained by multiplying the average CT value of the aorta by a first preset coefficient, and the calcification CT threshold is obtained by multiplying the average CT value of the aorta by a second preset coefficient.

[0069] The first preset coefficient can be, for example, between 1.0 and 1.1. The second preset coefficient can be, for example, 1.2. The system searches for the pixel with the largest HU value within the search range. If the HU value of this pixel is less than a set threshold (vascular CT threshold), the two-dimensional coordinates of this pixel on the two-dimensional section are corrected to the position within the blood vessel. Similarly, the system searches for the pixel with the smallest HU value within the search range. If the HU value of this pixel is greater than a set threshold (calcification CT threshold), the two-dimensional coordinates of this pixel on the two-dimensional section are corrected to the position within the blood vessel.

[0070] After correction, the result is transformed into the corresponding 3D coordinate system of the original 3D image, and the coordinates of the original center path points are updated accordingly (updating the coordinates of the first center path points). If the above conditions are not triggered, the coordinates of the original first center path points do not need to be corrected. This step iterates through all first center path points, completes the correction, and obtains the corrected second center path points.

[0071] Step S20: Traverse the centerline points, combine the radius information to generate blood vessel channels, merge them to obtain the reconstructed coronary artery blood vessel model, and extract the centerline and obtain the blood vessel cross-sectional area along the centerline.

[0072] In step S20, the centerline points are traversed and the radius information is combined to generate the blood vessel channel, which is completed through steps S400 to S500.

[0073] Step S400: Identify the left coronary system (step S420) and the right coronary system (step S410), and remove interfering branches based on the identification results (step S430). This step utilizes the centerline point (second central path point) information and combines it with anatomical features to segment the blood vessels.

[0074] Step S500: Generate curve data based on centerline points, traverse centerline points and combine radius information to generate vascular channels (step S510), and merge them to obtain a reconstructed coronary artery vascular model (step S520).

[0075] Step S400 further includes: dividing all second central pathway points into two parts, and determining whether each part belongs to the left or right coronary system based on the coordinates of the second central pathway point at the coronary artery ostium. After the determination is completed, steps S410 to S430 are executed. It can be understood that before executing S410 to S420, this method can only distinguish between the left and right coronary systems, but cannot identify the sub-parts included in the left or right coronary system, that is, the right coronary trunk, left main branch, left anterior descending branch, and left circumflex branch are not identified and named for the time being.

[0076] Step S410, identifying the right crown system, includes executing steps S411 to S413 based on the second center path point information.

[0077] Step S411: Establish a recognition coordinate system based on the three-dimensional image data. The recognition coordinate system includes: the X-axis pointing from the right side to the left side, the Y-axis pointing from the front side to the back side, and the Z-axis pointing from the bottom side to the top side.

[0078] Step S412: Select the coronary artery with the longest span along the Y-axis as the candidate right coronary trunk;

[0079] Step S413: If the Y-axis coordinate of the end of the candidate right coronary trunk is greater than the opening of the right coronary trunk, then the candidate right coronary trunk is confirmed as the right coronary trunk; otherwise, the coronary artery with the longest span along the X-axis is taken as the right coronary trunk.

[0080] Specifically, using the sagittal plane as the X-axis, the coronal plane as the Y-axis, and the transverse plane as the Z-axis, the range of each branch of the right coronary artery along the Y-axis and Z-axis is calculated, as well as the projection of each branch onto the plane perpendicular to the X-axis. The branch with the largest span along the Y-axis is identified as the candidate right coronary artery trunk. The Y-axis coordinates of the end of the candidate right coronary artery trunk are compared with the Y-axis coordinates of the opening point of the right coronary artery. If the coordinates of the end of the candidate trunk are greater than the coordinates of the opening point of the right coronary artery, it is determined as the right coronary artery trunk. Otherwise, the branch with the largest span in the X-direction is searched as the right coronary artery trunk.

[0081] Step S420: Identify the left coronary system, including identifying the left main branch (step S421), and identifying the left anterior descending branch and the left circumflex branch (step S422).

[0082] Step S421, identifying the left main branch, specifically includes: obtaining the remaining coronary arteries excluding the already identified right coronary system, the remaining coronary arteries including the left main branch, left anterior descending branch, left circumflex branch, and lateral branches, and identifying the left main branch based on the overlap of the left anterior descending branch and left circumflex branch before identification.

[0083] Step S422, identifying the left anterior descending artery and the left circumflex artery, specifically including: (1) obtaining the first part and the second part of the left main branch to be removed, the first part including the left anterior descending artery and its collateral branches, and the second part including the left circumflex artery and its collateral branches; (2) selecting the coronary artery with the largest span or the longest span on the Y-axis from the first part as the left anterior descending artery; (3) selecting the coronary artery with the largest span or the longest span on the Z-axis from the second part as the left circumflex artery.

[0084] Specifically, the left main branch of the left coronary artery system is removed, and the remaining branches are divided into two parts. The first part contains the left anterior descending branch (LAD) and its lateral branches (part A), and the second part contains the left circumflex branch (LCX) and its lateral branches (part B). The left anterior descending branch is determined from the first part based on the length of each branch (longest length) or the span on the Y-axis (largest span). From the second part, the branch with the largest span on the Z-axis is selected as the left circumflex branch.

[0085] Step S430, removing interfering branches, specifically includes: based on the identified right coronary trunk, left main branch, left anterior descending branch, and left circumflex branch, obtaining the remaining lateral branches and their respective trunks (the trunks of the lateral branches refer to the right coronary trunk, left main branch, left anterior descending branch, and left circumflex branch). For one of the remaining lateral branches: if the remaining lateral branch is parallel to its respective trunk, then the remaining lateral branch is removed; if the angle between the extension direction of the remaining lateral branch and the extension direction of its respective trunk is greater than a preset angle, then the remaining lateral branch is removed.

[0086] The preset angle can be, for example, 120 degrees. The process of removing interfering branches also includes reordering the order in which each lateral branch (remaining lateral branch) appears relative to the main trunk. In this step, the angle between each lateral branch and its corresponding main trunk is determined. If the lateral branch direction vector and the main trunk direction vector are almost parallel, the lateral branch is deleted. If the angle between the lateral branch direction vector and the main trunk direction vector is greater than a preset value, the lateral branch is deleted. This completes step S430, i.e., removing venous branch interference.

[0087] In step S500, step S510 specifically includes: generating curve data based on the second center path point, sweeping along each curve (generating one by one by traversing the center path points), and generating a pipe with a set radius. The set radius is the radius information of the center line point.

[0088] Step S520: After merging, a reconstructed coronary artery model is obtained, specifically including:

[0089] Step S521: After merging, an initial vascular model is obtained;

[0090] Step S522: Convert the initial blood vessel model into image data along the second center path point after removing interfering branches, and use it as input to the first level set;

[0091] Step S523: Based on the initial blood vessel model, the three-dimensional image data is denoised and gradient calculated sequentially to obtain image edge features, which are used as input to the second level set.

[0092] Step S524: Combining the first level set and the second level set, the segmentation edge is found using an active contour model based on partial differential equations to obtain an image sequence along the second center path point; in each image sequence, the image inside the target contour is positive, the target contour is the zero boundary, and the image outside the target contour is negative.

[0093] Step S525: Perform surface reconstruction on the image sequence to obtain a reconstructed coronary artery model.

[0094] Specifically, multiple pipelines generated from multiple curves are merged to form the initial pipeline model (initial vessel model) of the entire coronary tree. The pipeline model data is converted into image data and used as input to the first initial level set. Based on the existing initial pipeline contours, potential image edge features are obtained by denoising and gradient calculation of the original 3D image data, serving as another input to the second level set. Inward / outward extension is performed to find segmentation edges, ultimately obtaining the image sequence corresponding to the vessel path along the centerline. The traveling cubes algorithm is used to reconstruct the surface of this image sequence, obtaining the coronary artery vessel model composed of vessels of interest. The reconstructed coronary artery vessel model has a smooth surface and removes interfering venous branches, replacing the originally complex manual preprocessing operations and greatly improving the reproducibility of the results.

[0095] The above steps S100 to S500 extract and correct the centerline and radius along the coronary artery, i.e. the second central path point and radius information, which can automatically reconstruct the coronary artery model. The obtained coronary artery central path point can accurately describe the entire coronary tree and can automatically identify the left and right coronary systems of the coronary arteries. The branch information is complete and can remove venous interference branches, which is convenient for the calculation of FFR value in the following text.

[0096] In step 20, the centerline is extracted from the reconstructed coronary artery model obtained after merging, which can be achieved by either of the following two methods.

[0097] Method 1: The model is skeletonized. Implicit Laplacian smoothing with global position constraints is used to iteratively shrink the mesh geometry into a zero-volume skeleton shape. The shrunken mesh is then converted into a 1D surface to remove all collapsed faces while preserving the shape of the shrunken mesh and the original topology. The final centerline is obtained by mapping the skeleton mesh.

[0098] Method 2: A Voronoi diagram of the blood vessel can be generated using the triangular mesh produced by the Delaunay triangulation algorithm and the normal vector of each point. The center of the largest inscribed sphere of the blood vessel is then generated based on the Voronoi diagram. Finally, using the set start and end points, the shortest path is found among these center points based on the radius of the largest inscribed sphere. This path represents the final centerline of the blood vessel.

[0099] In step 20, the cross-sectional area of ​​the blood vessel along the centerline is obtained, for example, by the following method: Calculate the tangent vector, normal vector, and binormal vector corresponding to each centerline point. Use the tangent vector, normal vector, binormal vector, and the three-dimensional coordinates of each centerline point to calculate the cutting matrix at that point. Then, use the cutting matrix to perform a shearing operation on the blood vessel model at that point to obtain the blood vessel contour point information corresponding to that point, and further calculate the area information.

[0100] Step S30, perform flow allocation, including using the ratio of the cross-sectional area of ​​the terminal branch of the blood vessel to the sum of the cross-sectional areas of all terminal branches of the blood vessel, or the ratio of the volume of the blood vessel branch to the sum of the volumes of all blood vessel branches to perform flow allocation.

[0101] Traffic allocation is performed using the following formula:

[0102] Qn = Qtotf(A,L)

[0103] In the formula, Q n Let Q be the flow rate of branch n. tot For total flow, A This is an array of blood vessel cross-sectional areas. L This is a coordinate array representing the centerline length. This formula allocates flow based on the vessel cross-sectional area array and the centerline length coordinate array. f(A,L) can have different forms; the following two are provided in various embodiments of this application.

[0104] If flow distribution is based on the proportion of the terminal cross-sectional area of ​​each vascular branch to the sum of the terminal cross-sectional areas of all vascular branches, then... In the formula, It is the cross-sectional area at the end of branch n.

[0105] If flow distribution is based on the volume of each vascular branch as a percentage of the total volume of all vascular branches, then...

[0106]

[0107] In the formula V n Let n be the volume of branch n, An,i be the cross-sectional area of ​​branch n at position i on the center line, and Ln,i be the length coordinate of branch n at position i on the center line.

[0108] Furthermore, Q tot The total flow rate at the coronary artery inlet under maximum congestion can be obtained from physiological parameters measured or from convolutional neural networks. Obtaining the flow rate from convolutional neural networks includes: using a convolutional neural network to obtain the myocardial contour of the left ventricle, performing surface reconstruction on the myocardial contour to obtain the myocardial volume, and using the obtained myocardial volume to obtain the total flow rate at the coronary artery inlet under maximum congestion based on cardiac physiological parameters (such as myocardial tissue density, blood density, etc.).

[0109] Step S40: Based on the bifurcation point of the coronary artery model, the coronary artery model is divided into bifurcation segments and long straight segments by using the distance between the centerline point on the extended branch of the bifurcation point and the bifurcation point.

[0110] In step S40, the distance between each point on the center line of each branch and the known bifurcation point is determined. If the distance is greater than the set threshold, it is determined to be a long straight segment. If the distance is less than or equal to the set distance threshold, it is determined to be a bifurcation segment. The set distance threshold can be, for example, 5mm.

[0111] Step S50: Obtain the pressure drop of the long straight segment (Step S52), obtain the pressure drop of the bifurcation segment (Step S51), and then obtain the coronary blood flow reserve fraction at each centerline point (Step S53).

[0112] Step S51, obtaining the pressure drop of the bifurcation segment specifically includes:

[0113] The pressure drop of the bifurcation segment can be obtained by using the ratio of the cross-sectional area of ​​the parent branch to that of the child branch, the ratio of the flow rate of the parent branch to that of the child branch, and the angle between the tangent vectors of the extension directions of the parent branch and the child branch.

[0114] The pressure drop at the bifurcation point is obtained using the following formula:

[0115]

[0116] In the formula, ΔP bif The pressure drop at the bifurcation point is ρ, where ρ is the blood density, and u is the blood density. f Let be the blood flow velocity, C be the ratio of the cross-sectional area of ​​the parent branch to the cross-sectional area of ​​the child branch, D be the ratio of the flow rate of the parent branch to the flow rate of the child branch, and θ be the angle between the tangent vectors of the parent branch extension direction and the child branch extension direction.

[0117] Step S52 involves obtaining the pressure drop of the long straight segment, which is divided into two cases. If there is no stenosis, the pressure drop of the long straight segment includes the flow loss; if there is a stenosis, the pressure drop of the long straight segment includes the flow loss, the dilation loss of the stenosis, and the constriction loss of the stenosis. That is, if there is no stenosis in the long straight segment of the vessel (i.e., there is no stenosis), then only the pressure drop due to flow loss is present. If there is a stenosis in this segment of the vessel, then the pressure drop is the sum of the flow loss, the dilation loss, and the constriction loss.

[0118] In this embodiment, the reference diameter obtained by fitting is used to determine whether there is a narrow section in the long straight segment. The reference diameter of each centerline point within the long straight segment is obtained by fitting using any one of the following formulas:

[0119]

[0120]

[0121] In the formula, dprox is the diameter of the healthy vessel proximal to the stenosis, ddist is the diameter of the healthy vessel distal to the stenosis, Lste is the length between the proximal and distal ends of the stenosis (i.e., the length of the stenosis segment), and x represents different center points. Of the two formulas above used to obtain the long straight segment, the former is a linear fit, and the latter is an exponential function fit.

[0122] In one embodiment, the friction loss at each centerline point is obtained by the following formula:

[0123]

[0124] In the formula, ΔP V For the flow loss along the flow path, β is a constant (e.g., 110.08), μ is the blood viscosity, L is the stenotic length of the vessel segment, Q is the flow rate of the segment, and d is the vessel diameter at the centerline point.

[0125] In one embodiment, the gradual loss at each centerline point within the narrow segment is obtained by the following formula:

[0126]

[0127] In the formula, ΔP exp Let ρ be the gradual loss at each centerline point within the narrow segment, γ be a constant (e.g., 1.21), δ be a constant (e.g., 0.08), L be the length of the narrow segment in the long straight segment, Q be the flow rate of the long straight segment, d0 be the reference diameter at each centerline point, and A0 be the reference area at each centerline point of the narrow segment. m A is the minimum diameter of the narrow segment. m It represents the minimum area of ​​the narrow segment.

[0128] In one embodiment, the contraction loss at each centerline point of the narrow segment is obtained by the following formula:

[0129]

[0130] In the formula, ΔP ent The constriction loss at each centerline point within the narrow segment is represented by ε, which is a constant (e.g., 57.6), μ is the blood viscosity, d0 is the reference diameter at the centerline point, and d m Let Q be the minimum diameter of the narrow section, and let Q be the flow rate of that section.

[0131] Step S53 then yields the coronary flow reserve fraction at each centerline point, specifically including:

[0132] Obtain the pressure drop at each center point (ΔP k (This is the pressure drop at the k-th center point), thus obtaining the FFR value for each centerline point. P is the pressure at the inlet, and ΔP is the pressure drop.

[0133] It should be understood that, Figure 1 The flowchart may include at least some steps that may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but may be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0134] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 2 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with external terminals via a network connection. When executed by the processor, the computer program implements a method for acquiring coronary flow reserve based on CTA images. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device casing, or an external keyboard, touchpad, or mouse.

[0135] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0136] Step S10: Obtain three-dimensional image data of coronary CTA, and obtain the centerline point and the radius information along the centerline point based on the three-dimensional image data;

[0137] Step S20: Traverse the centerline points, combine the radius information to generate blood vessel channels, merge them to obtain the reconstructed coronary artery blood vessel model, and extract the centerline and obtain the blood vessel cross-sectional area along the centerline.

[0138] Step S30, perform flow allocation, including using the ratio of the cross-sectional area of ​​the terminal branch of the blood vessel to the sum of the cross-sectional areas of all terminal branches of the blood vessel, or the ratio of the volume of the blood vessel branch to the sum of the volumes of all blood vessel branches to perform flow allocation.

[0139] Step S40: Based on the bifurcation point of the coronary artery model, the coronary artery model is divided into bifurcation segments and long straight segments by using the distance between the centerline point on the extended branch of the bifurcation point and the bifurcation point.

[0140] Step S50: Obtain the pressure drop of the long straight segment, obtain the pressure drop of the bifurcation segment, and then obtain the coronary blood flow reserve fraction at each centerline point.

[0141] In one embodiment, a readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0142] Step S10: Obtain three-dimensional image data of coronary CTA, and obtain the centerline point and the radius information along the centerline point based on the three-dimensional image data;

[0143] Step S20: Traverse the centerline points, combine the radius information to generate blood vessel channels, merge them to obtain the reconstructed coronary artery blood vessel model, and extract the centerline and obtain the blood vessel cross-sectional area along the centerline.

[0144] Step S30, perform flow allocation, including using the ratio of the cross-sectional area of ​​the terminal branch of the blood vessel to the sum of the cross-sectional areas of all terminal branches of the blood vessel, or the ratio of the volume of the blood vessel branch to the sum of the volumes of all blood vessel branches to perform flow allocation.

[0145] Step S40: Based on the bifurcation point of the coronary artery model, the coronary artery model is divided into bifurcation segments and long straight segments by using the distance between the centerline point on the extended branch of the bifurcation point and the bifurcation point.

[0146] Step S50: Obtain the pressure drop of the long straight segment, obtain the pressure drop of the bifurcation segment, and then obtain the coronary blood flow reserve fraction at each centerline point.

[0147] The methods for obtaining coronary flow reserve (FFR) based on CTA images provided in the embodiments of this application can automatically calculate the FFR without manual intervention, ensuring the repeatability of the calculation results. Compared with direct hemodynamic calculations using three-dimensional CFD simulation, the embodiments calculate the FFR value using a dimensionality-reduced model (calculated using the centerline and corresponding cross-sectional area), significantly improving computational efficiency and meeting clinical needs. Compared with the one-dimensional or reduced-order models used in other existing technologies, the bifurcation pressure drop calculation method used in this method yields more accurate results.

[0148] In this embodiment, the computer program product includes program code portions for performing the steps of the method for acquiring coronary flow reserve based on CTA images in the various embodiments of this application, when the computer program product is executed by one or more computing devices. The computer program product may be stored on a computer-readable recording medium. It may also be provided for download via a data network (e.g., via RAN, via the Internet, and / or via RBS). Alternatively or additionally, the method may be encoded in a field-programmable gate array (FPGA) and / or an application-specific integrated circuit (ASIC), or its functionality may be provided for download by means of a hardware description language.

[0149] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0150] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered to be within the scope of this specification. When technical features of different embodiments are embodied in the same drawing, it can be regarded as the drawing also disclosing examples of combinations of the various embodiments involved.

[0151] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A method for obtaining coronary flow reserve based on CTA images, characterized in that, include: Acquire three-dimensional image data of coronary CTA, and obtain the centerline point and the radius information along the centerline point based on the three-dimensional image data; The centerline points are traversed, and the radius information is combined to generate vascular channels. After merging, a reconstructed coronary artery vascular model is obtained, and the centerline is extracted from it to obtain the vascular cross-sectional area along the centerline. Flow distribution can be carried out by using the proportion of the terminal cross-sectional area of ​​a vascular branch to the sum of the terminal cross-sections of all vascular branches, or the proportion of the volume of a vascular branch to the sum of the volumes of all vascular branches. Based on the bifurcation point of the coronary artery model, the coronary artery model is divided into bifurcation segments and long straight segments by using the distance between the centerline point on the extended branch of the bifurcation point and the bifurcation point. The pressure drop of the long straight segment and the pressure drop of the bifurcation segment are obtained, and then the coronary flow reserve fraction at each centerline point is obtained; the reference diameter of the long straight segment is obtained by fitting, which is used to determine whether there is a stenotic segment in the long straight segment; if there is no stenotic segment, the pressure drop of the long straight segment includes the flow loss along the path; if there is a stenotic segment, the pressure drop of the long straight segment includes the flow loss along the path, the dilation loss of the stenotic segment, and the contraction loss of the stenotic segment; The fitting process obtains the reference diameter for each centerline point within the long straight segment, using any one of the following formulas: In the formula, It is the diameter of the healthy blood vessel proximal to the narrowed area. It is the diameter of the healthy blood vessel distal to the narrowed segment. is the length between the proximal and distal ends of the stenosis, and x represents the different center points; The gradual widening loss of the narrow segment is obtained by the following formula: In the formula, The gradual loss at each centerline point within the narrow segment, Blood density, It is a constant. Let L be a constant, L be the length of the narrow section in the long straight segment, and Q be the flow rate of the long straight segment. The reference diameter for each centerline point, This serves as the reference area for each centerline point of the narrow segment. The minimum diameter of the narrow segment. It represents the minimum area of ​​the narrow segment.

2. The method for obtaining coronary flow reserve based on CTA images according to claim 1, characterized in that, Obtaining the pressure drop of the bifurcation segment specifically includes: The pressure drop of the bifurcation segment is obtained by using the ratio of the cross-sectional area of ​​the parent branch to the cross-sectional area of ​​the child branch, the ratio of the flow rate of the parent branch to the flow rate of the child branch, and the angle between the tangent vectors of the extension directions of the parent branch and the child branch.

3. The method for obtaining coronary flow reserve based on CTA images according to claim 2, characterized in that, The pressure drop of the bifurcation segment is obtained by the following formula: In the formula, For the pressure drop at the bifurcation point, Blood density, For blood flow velocity, Let be the ratio of the cross-sectional area of ​​the parent branch to the cross-sectional area of ​​the child branch, and D be the ratio of the flow rate of the parent branch to the flow rate of the child branch. The angle between the tangent vectors of the parent branch extension direction and the child branch extension direction.

4. The method for obtaining coronary flow reserve based on CTA images according to claim 1, characterized in that, The traffic allocation is specifically performed using the following formula: In the formula, Let n be the flow rate of branch n. For total flow, This is an array of blood vessel cross-sectional areas. This is an array of coordinates representing the centerline length. If flow distribution is based on the proportion of the terminal cross-sectional area of ​​each vascular branch to the sum of the terminal cross-sectional areas of all vascular branches, then... In the formula, It is the cross-sectional area at the end of branch n.

5. The method for obtaining coronary flow reserve based on CTA images according to claim 1, characterized in that, The traffic allocation is specifically performed using the following formula: In the formula, Let n be the flow rate of branch n. For total flow, This is an array of blood vessel cross-sectional areas. This is an array of coordinates representing the centerline length. If flow distribution is based on the volume of each vascular branch as a percentage of the total volume of all vascular branches, then... In the formula It is the volume of branch n. It is the cross-sectional area of ​​branch n at position i on the centerline. It is the length coordinate of branch n at position i on the center line.

6. The method for obtaining coronary flow reserve based on CTA images according to claim 1, characterized in that, Obtaining the centerline point and the radius information along the centerline point based on the three-dimensional image data specifically includes: Extract the initial center path point and the radius information along the initial center path point from the three-dimensional image data; Based on the three-dimensional image data, two-dimensional cross-sections are sequentially obtained along the initial center path point, and the two-dimensional cross-sections include the blood vessel cross-section and its surrounding image. For any two-dimensional plane, the search range is set according to the initial center path point in the plane, and the blood vessel wall is located according to the vascular CT threshold. If the pixel with the largest CT value within the search range is less than the vascular CT threshold or the CT value of that pixel is greater than the calcification CT threshold, then the corresponding initial center path point is corrected to the vascular wall and mapped to the three-dimensional image data to obtain the corrected centerline point and the radius information along the centerline point.

7. A readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method for obtaining coronary flow reserve based on CTA images as described in any one of claims 1 to 6.

Images