Method and device for comprehensively evaluating coronary artery plaque risk based on medical image

Abstract

The invention provides a method and a device for comprehensively evaluating coronary artery plaque dangerousness based on a medical image. The method for comprehensively evaluating the risk of the coronary plaque based on the medical image comprises the following steps: acquiring coronary strain parameters according to a coronary tree segmentation image; acquiring plaque form parameter information according to the plaque segmentation image; acquiring hemodynamic parameter information according to the resting state time sequence image; and acquiring a plaque risk index according to the coronary strain parameter, the plaque morphological parameter information and the hemodynamic parameter information. The method comprises the following steps: firstly, segmenting a heart structure, and calculating coronary artery strain parameters according to the shape of the coronary artery and the strain difference between different time phases; and calculating morphological parameters of the plaques according to the positions, forms and texture structures of the plaques. And then calculating hemodynamic parameters of the coronary artery by adopting a computational fluid mechanics method. And finally, designing a comprehensive index to jointly judge the risk of the coronary plaque.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly relates to a method for comprehensively evaluating the risk of coronary artery plaques based on medical images and a device for comprehensively evaluating the risk of coronary artery plaques based on medical images. Background Art

[0002] Cardiac ischemia caused by coronary artery plaque shedding is an important cause of myocardial infarction. Therefore, judging the vulnerability of coronary artery plaques and predicting the occurrence probability and risk of coronary artery plaque shedding are of great significance for the prevention of myocardial infarction. The present invention proposes a method for comprehensively judging the risk of coronary artery plaques based on coronary strain parameters, plaque morphological parameters, and hemodynamic parameters in different phases of the stress state and resting state.

[0003] There are a large number of literatures on predicting the risk of plaques in the prior art, but they mainly focus on single-type indicators, such as based on hemodynamics or based on plaque morphology. However, since the human body is a complex system and the risk of plaques is affected by multiple complex factors, a single indicator is difficult to represent the risk of plaques and is far from the actual situation. There are also some literatures that use multiple methods to jointly evaluate the risk of plaques. The research results are as follows:

[0004] There is also a prior art that discloses a method for predicting plaque vulnerability based on plaque morphological parameters and hemodynamic parameters. This method is mainly based on CT and OCT images and uses a convolutional neural network for prediction. This method directly predicts the results using a neural network, and its black box effect makes it difficult for doctors to judge the specific influence degree of various indicators and make targeted judgments. Moreover, this method needs to use OCT, which reduces the scope of adaptation. This method also does not utilize the strain parameters of the coronary arteries between different phases.

[0005] There is also a prior art that discloses a study based on the radial wall strain (RWS) of the coronary arteries in angiography, which mainly focuses on using RWS to evaluate the plaque stability of coronary heart disease patients and predict the risk of future acute myocardial infarction (AMI) events. However, this method only evaluates the strain of the coronary arteries and does not consider the influence of factors such as plaques and hemodynamics.

[0006] Or combine plaque morphological parameters with hemodynamic parameters to analyze and evaluate the cardiovascular function and plaque characteristics of patients with acute coronary syndrome, including quantitative analysis of plaques and hemodynamics for each lesion and each blood vessel, which can identify culprit lesions and help identify patients with an increased risk of adverse cardiac events. This method has been verified in multi-country and multi-center studies, but this method does not consider the differences in the morphology of the human coronary arteries itself, and the same plaque may show different performances at different positions in different coronary arteries.

[0007] In summary, the plaque risk, the condition of the coronary artery, the condition of the plaque, and the condition of the blood flow are all highly relevant. However, existing technologies often only focus on one aspect and cannot comprehensively evaluate the plaque risk. Summary of the Invention

[0008] The present invention provides a method for comprehensively evaluating the risk of coronary artery plaques based on medical images. The method for comprehensively evaluating the risk of coronary artery plaques based on medical images includes:

[0009] Obtaining coronary artery strain parameters from the coronary artery tree segmentation image;

[0010] Obtaining plaque morphological parameter information from the plaque segmentation image;

[0011] Obtaining hemodynamic parameter information from the resting state time series image;

[0012] Obtaining a plaque risk index based on the coronary artery strain parameters, the plaque morphological parameter information, and the hemodynamic parameter information.

[0013] Optionally, the obtaining of the coronary artery strain parameters from the coronary artery tree segmentation image includes:

[0014] Obtaining a stress state time series image and a resting state time series image;

[0015] Respectively obtaining segmented coronary artery tree images at different time phases based on the stress state time series image and the resting state time series image;

[0016] Performing the following operations on the segmented coronary artery tree images at different time phases:

[0017] Extracting the center line of the coronary artery in the coronary artery tree image and straightening and unfolding the coronary artery to obtain a straightened coronary artery image;

[0018] Detecting calcified plaques and non-calcified plaques on the coronary artery based on the obtained coronary artery image. When a calcified plaque or a non-calcified plaque is detected, performing image segmentation on the detected calcified plaque or the detected non-calcified plaque to obtain a calcified plaque image or a non-calcified plaque image. The calcified plaque image or the non-calcified plaque image constitutes the plaque segmentation image. Among them, a calcified plaque has multiple calcified plaque images at different time phases, and a non-calcified plaque has multiple non-calcified plaque images at different time phases;

[0019] Obtaining the coronary artery strain parameter corresponding to each calcified plaque based on the calcified plaque images at different time phases of each calcified plaque;

[0020] Obtaining the coronary artery strain parameter corresponding to each non-calcified plaque based on the non-calcified plaque images at different time phases of each non-calcified plaque.

[0021] Optionally, obtaining the coronary artery strain parameter corresponding to each calcified plaque from the calcified plaque images at different phases of each calcified plaque includes:

[0022] Performing the following operations on each of the calcified plaque images at different phases belonging to the same calcified plaque:

[0023] Calculating the minimum diameter of the coronary artery lumen of each calcified plaque image respectively to obtain a sequence of the minimum coronary artery diameters, and taking the maximum value D max and the minimum value D min , and obtaining the coronary artery strain parameter of each calcified plaque through the following formula:

[0024] e = (D max - D min ) / D max ;

[0025] where e is the coronary artery strain parameter.

[0026] Optionally, obtaining the coronary artery strain parameter corresponding to each non-calcified plaque from the non-calcified plaque images at different phases of each non-calcified plaque includes:

[0027] Performing the following operations on each of the non-calcified plaque images at different phases belonging to the same non-calcified plaque:

[0028] Calculating the minimum diameter of the coronary artery lumen of each non-calcified plaque image respectively to obtain a sequence of the minimum coronary artery diameters, and taking the maximum value D max and the minimum value D min , and obtaining the coronary artery strain parameter of each non-calcified plaque through the following formula:

[0029] e = (D max - D min ) / D max ;

[0030] where e is the coronary artery strain parameter.

[0031] Optionally, the plaque morphological parameter information includes plaque composition information, plaque morphological information, plaque location information, and plaque density information.

[0032] Optionally, obtaining the hemodynamic parameter information from the resting state time series images includes:

[0033] Comparing the segmented coronary artery tree images at each phase with each other, so as to obtain the resting state time series image at the phase with the smallest coronary artery lumen diameter at each phase, and performing hemodynamic simulation on the coronary arteries in this group of resting state time series images, so as to obtain hemodynamic parameters.

[0034] Optionally, the plaque risk index obtained based on the coronary artery strain parameters, plaque morphology parameter information, and hemodynamic parameter information is obtained through the following formula:

[0035] F(X) = a1Q + a2(ΔP) + a3R + a4τ + a5(ΔR) + a6(ΔQ) + a7L + … + b1x1 + … + b n x n ;

[0036] where a and b are parameters to be fitted; x is other coronary artery strain parameters, plaque morphology parameter information, and hemodynamic parameters; Q is the flow rate; R is the coronary artery lumen radius at the narrowest point; ΔP is the pressure difference before and after stenosis; L is the length of the stenosis segment, μ is the blood viscosity; τ is the wall shear stress; ΔR is the difference in the lumen between systole and diastole, that is, strain, and ΔQ is the difference between the flow rate assuming the lumen R remains unchanged during systole and the flow rate assuming the lumen R remains unchanged during diastole.

[0037] Optionally, the difference ΔR in the lumen between systole and diastole is obtained through the following formula:

[0038]

[0039] where Q is the flow rate, R is the coronary artery lumen radius at the narrowest point, ΔP is the pressure difference before and after stenosis, L is the length of the stenosis segment, μ is the blood viscosity, and τ is the wall shear stress.

[0040] Optionally, the flow rate Q is obtained through the following formula:

[0041]

[0042] where Q is the flow rate, R is the coronary artery lumen radius at the narrowest point, ΔP is the pressure difference before and after stenosis, L is the length of the stenosis segment, μ is the blood viscosity, and τ is the wall shear stress;

[0043] The wall shear stress τ is obtained through the following formula:

[0044]

[0045] where Q is the flow rate, R is the coronary artery lumen radius at the narrowest point, ΔP is the pressure difference before and after stenosis, L is the length of the stenosis segment, μ is the blood viscosity, and τ is the wall shear stress.

[0046] This application also provides a device for comprehensively evaluating the risk of coronary artery plaques based on medical images, characterized in that the device for comprehensively evaluating the risk of coronary artery plaques based on medical images includes:

[0047] Coronary artery strain parameter acquisition module, which is used to acquire coronary artery strain parameters according to the coronary artery tree segmentation image;

[0048] Plaque morphological parameter information acquisition module, which is used to acquire plaque morphological parameter information according to the plaque segmentation image;

[0049] Hemodynamic parameter information acquisition module, which is used to acquire hemodynamic parameter information according to the resting state time series image;

[0050] Plaque risk index acquisition module, which is used to acquire the plaque risk index according to the coronary artery strain parameters, plaque morphological parameter information and hemodynamic parameter information.

[0051] The method for comprehensively evaluating the risk of coronary artery plaques based on medical images in this application first segments the cardiac structure, segmenting the coronary arteries, left ventricle, and plaques on the coronary arteries. According to the morphology of the coronary arteries and the strain differences between different time phases, the coronary artery strain parameters are calculated. According to the position, morphology and texture structure of the plaques, the plaque morphological parameters are calculated. Then, the computational fluid dynamics (CFD) method is used to calculate the hemodynamic parameters of the coronary arteries, including pressure, flow rate, flow velocity, wall shear stress, FFR, etc. Finally, a comprehensive index is designed to combine the coronary artery strain parameters, plaque morphological parameters and hemodynamic parameters to jointly judge the risk of coronary artery plaques. Brief Description of the Drawings

[0052] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0053] Figure 1 It is a schematic flow chart of the method for comprehensively evaluating the risk of coronary artery plaques based on medical images according to an embodiment of the present invention;

[0054] Figure 2 It is a schematic diagram of the coronary artery tree segmentation image according to an embodiment of the present invention;

[0055] Figure 3 It is a schematic diagram of the straightened coronary artery according to an embodiment of the present invention;

[0056] Figure 4 It is a schematic diagram of the coronary artery diameter according to an embodiment of the present invention;

[0057] Figure 5 It is a schematic diagram of the CFD mesh division according to an embodiment of the present invention. Detailed Description of the Embodiments

[0058] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only for the purpose of illustrating and explaining the present invention, and are not intended to limit the present invention.

[0059] As Figure 1 shown, the method for comprehensively evaluating the risk of coronary artery plaques based on medical images includes:

[0060] Obtaining coronary artery strain parameters from the coronary artery tree segmentation image;

[0061] Obtaining plaque morphological parameter information from the plaque segmentation image;

[0062] Obtaining hemodynamic parameter information from the resting state time series image;

[0063] Obtaining a plaque risk index based on the coronary artery strain parameters, plaque morphological parameter information, and hemodynamic parameter information.

[0064] In this embodiment, the obtaining of coronary artery strain parameters from the coronary artery tree segmentation image includes:

[0065] Obtaining a stress state time series image and a resting state time series image;

[0066] Respectively obtaining segmented coronary artery tree images at different time phases based on the stress state time series image and the resting state time series image;

[0067] Performing the following operations on the segmented coronary artery tree images at different time phases:

[0068] Extracting the centerline of the coronary artery in the coronary artery tree image and straightening and unfolding the coronary artery to obtain a straightened coronary artery image;

[0069] Detecting calcified plaques and non-calcified plaques on the coronary artery based on the obtained coronary artery image. When a calcified plaque or a non-calcified plaque is detected, performing image segmentation on the detected calcified plaque or the detected non-calcified plaque to obtain a calcified plaque image or a non-calcified plaque image. The calcified plaque image or the non-calcified plaque image constitutes the plaque segmentation image. Among them, one calcified plaque has multiple calcified plaque images at different time phases, and one non-calcified plaque has multiple non-calcified plaque images at different time phases;

[0070] Obtaining the coronary artery strain parameter corresponding to each calcified plaque based on the calcified plaque images at different time phases of each calcified plaque;

[0071] Obtaining the coronary artery strain parameter corresponding to each non-calcified plaque based on the non-calcified plaque images at different time phases of each non-calcified plaque.

[0072] In this embodiment, obtaining the coronary artery strain parameter corresponding to each calcified plaque from the calcified plaque images at different phases of each calcified plaque includes:

[0073] Performing the following operations on each of the calcified plaque images at different phases belonging to the same calcified plaque:

[0074] Calculating the minimum diameter of the coronary lumen of each calcified plaque image respectively to obtain a sequence of the minimum coronary diameters, and taking the maximum value D in the sequence of the minimum coronary diameters max and the minimum value D min , and obtaining the coronary artery strain parameter of each calcified plaque through the following formula:

[0075] e = (D max - D min ) / D max ;

[0076] where e is the coronary artery strain parameter.

[0077] In this embodiment, obtaining the coronary artery strain parameter corresponding to each non-calcified plaque from the non-calcified plaque images at different phases of each non-calcified plaque includes:

[0078] Performing the following operations on each of the non-calcified plaque images at different phases belonging to the same non-calcified plaque:

[0079] Calculating the minimum diameter of the coronary lumen of each non-calcified plaque image respectively to obtain a sequence of the minimum coronary diameters, and taking the maximum value D in the sequence of the minimum coronary diameters max and the minimum value D min , and obtaining the coronary artery strain parameter of each non-calcified plaque through the following formula:

[0080] e = (D max - D min ) / D max ;

[0081] where e is the coronary artery strain parameter.

[0082] In this embodiment, the plaque morphological parameter information includes plaque composition information, plaque morphological information, plaque location information, and plaque density information.

[0083] In this embodiment, obtaining the hemodynamic parameter information from the resting state time series images includes:

[0084] Comparing the segmented coronary artery tree images at each phase with each other, so as to obtain the resting state time series image at the phase with the smallest coronary lumen diameter at each phase, and performing hemodynamic simulation on the coronary arteries in this group of resting state time series images, so as to obtain hemodynamic parameters.

[0085] In this embodiment, the plaque risk index obtained according to the coronary artery strain parameters, plaque morphology parameter information, and hemodynamic parameter information is obtained through the following formula:

[0086] F(X) = a1Q + a2(ΔP) + a3R + a4τ + a5(ΔR) + a6(ΔQ) + a7L + … + b1x1 + … + b n x n ;

[0087] where a and b are parameters to be fitted; x are other coronary artery strain parameters, plaque morphology parameter information, and hemodynamic parameters, such as plaque volume, maximum plaque diameter, etc.; Q is the flow rate; R is the radius of the coronary artery lumen at the narrowest point; ΔP is the pressure difference before and after stenosis; L is the length of the stenotic segment, μ is the blood viscosity; τ is the wall shear stress; ΔR is the difference in the lumen between systole and diastole, that is, strain, and ΔQ is the difference between the flow rate assuming the lumen R remains unchanged during systole and the flow rate assuming the lumen R remains unchanged during diastole.

[0088] In this embodiment, the difference ΔR in the lumen between systole and diastole is obtained through the following formula:

[0089]

[0090] where Q is the flow rate, R is the radius of the coronary artery lumen at the narrowest point, ΔP is the pressure difference before and after stenosis, L is the length of the stenotic segment, μ is the blood viscosity, and τ is the wall shear stress.

[0091] In this embodiment, the flow rate Q is obtained through the following formula:

[0092]

[0093] where Q is the flow rate, R is the radius of the coronary artery lumen at the narrowest point, ΔP is the pressure difference before and after stenosis, L is the length of the stenotic segment, μ is the blood viscosity, and τ is the wall shear stress;

[0094] The wall shear stress τ is obtained through the following formula:

[0095]

[0096] where Q is the flow rate, R is the radius of the coronary artery lumen at the narrowest point, ΔP is the pressure difference before and after stenosis, L is the length of the stenotic segment, μ is the blood viscosity, and τ is the wall shear stress.

[0097] The present application will be further described in detail below with reference to the accompanying drawings. It can be understood that this example does not constitute any limitation to the present application.

[0098] See Figure 1, obtain cardiac images at different time phases, segment the coronary arteries, and obtain the segmented coronary artery tree. The segmentation of the coronary artery tree can adopt a deep learning method based on nn-Unet and topological analysis. First, use nn-Unet to segment the coronary arteries, and then perform topological analysis on the coronary arteries to ensure the continuity of the coronary arteries and remove non-coronary impurities.

[0099] See Figure 2 , then, extract the centerline of the coronary artery, straighten and unfold the coronary artery (specifically, based on the skeletonization method, skeletonize the coronary artery segmentation result into the centerline, and then according to the normal direction of the centerline, intercept the cross-section at each point on the centerline, and stack the two-dimensional cross-sections into three-dimensional volume data to obtain the three-dimensional volume after straightening the coronary artery, and the coronary artery is located in the center of this three-dimensional volume), and obtain the straightened coronary artery.

[0100] See Figure 3 , adopt a CNN neural network to detect calcified and non-calcified plaques on the coronary artery along the centerline. At the detected plaques, use nn-Unet to segment the plaques. Then, remove the plaques from the coronary artery tree to obtain the coronary artery tree after removing the plaques, and the coronary arteries of this coronary artery tree represent the coronary artery lumen through which blood flows. On the cross-section of the coronary artery lumen at the plaque, the maximum diameter and minimum diameter of the lumen can be calculated, where the minimum diameter represents the coronary artery diameter at the stenosis.

[0101] At different time phases, calculate the minimum diameter of the coronary artery lumen at the same plaque respectively to obtain a sequence of coronary artery minimum diameters, Dt1, Dt2......Dtn, and take the maximum value D max and the minimum value D min . Then the coronary artery strain can be defined as e = (D max - D min ) / D max ; where e is the coronary artery strain parameter.

[0102] In this embodiment, the plaque morphology parameters are calculated by the following method:

[0103] In the time series images, detect calcified plaques and non-calcified plaques on the coronary artery according to the obtained coronary artery images, select the sequence images with the clearest display in the images corresponding to each plaque, and calculate the morphology parameters for the plaques. Including but not limited to the following parameters:

[0104] Plaque composition:

[0105] Lipid core size: A larger lipid core is related to plaque vulnerability.

[0106] Fibrous cap thickness: A thin fibrous cap (usually <65 microns) is a characteristic of vulnerable plaques.

[0107] Degree of calcification: Calcified plaques are relatively stable, but mixed calcification may increase vulnerability.

[0108] Plaque morphology:

[0109] Plaque volume: Larger plaques are more likely to rupture.

[0110] Plaque area: The area occupied by the plaque on the vascular cross-section.

[0111] Positive remodeling: The plaque expands outwards, which may increase vulnerability.

[0112] Negative remodeling: The plaque contracts inwards, which may reduce vulnerability.

[0113] Plaque location:

[0114] Proximal or bifurcational plaque: Plaques at these locations are more likely to rupture.

[0115] Plaque eccentricity: Eccentric plaques are more vulnerable than concentric plaques.

[0116] Plaque density:

[0117] CT value (HU unit): Low-density plaques (<30 HU) are associated with lipid cores, and high-density plaques (>130 HU) are associated with calcification.

[0118] In this embodiment, the following technical solution can be adopted to calculate hemodynamic parameters:

[0119] In the time-series images, based on the coronary lumen calculated in Step 1, select the sequence with the smallest minimum lumen diameter value at the plaque. On the basis of the segmented coronary tree, perform hemodynamic simulation on the coronary artery. The steps are as follows:

[0120] 1 Mesh generation.

[0121] The delauney mesh generation algorithm can be used for mesh generation. This algorithm can ensure that the divided triangular meshes are all acute triangles. When generating the mesh, attention should also be paid to distinguishing the mesh size according to the radian. At larger volumes such as the aorta, larger meshes can be used, and the number of meshes is relatively sparse. While in the coronary artery, especially in the fine branches of the coronary artery, smaller meshes should be used, and the number of meshes is dense, so as to well represent the characteristics of the fine parts of the blood vessel. The coronary artery with the generated mesh is as Figure 4 shown.

[0122] 2 Perform simulation using the CFD method

[0123] CFD is an abbreviation for Computational Fluid Dynamics. After mesh generation, the present invention uses Computational Fluid Dynamics to calculate blood flow parameters in coronary arteries. First, it is necessary to calculate blood flow related boundary condition parameters, which can be obtained through calculation from images or may be measured separately outside the images. The blood flow related parameters that the present invention needs to extract include but are not limited to the patient's blood pressure, cardiac output, coronary artery inlet and outlet areas, etc. After extracting the blood flow related parameters, together with the mesh as boundary conditions, they are input into the CFD calculation solver, and the blood flow results at each grid point, such as blood pressure, flow rate, wall shear stress, flow velocity, and blood flow direction, can be calculated.

[0124] 3 Calculating Hemodynamic Parameters

[0125] After CFD simulation, hemodynamic parameters can be directly obtained. In another embodiment, the results of CFD can also be used as the gold standard to train an AI neural network, and then the AI neural network can be used to directly calculate hemodynamic parameters. The advantage of the AI method compared to using CFD simulation is that it is faster.

[0126] In this embodiment, the hemodynamic equation is introduced into the regression fitting process. While performing linear regression, the correlation between various parameters is ensured. In this way, the fitted riskiness can, on the one hand, be used to judge the patient's prognosis according to the risk index, and on the other hand, due to the sufficient correlation between various parameters, it is convenient for users to understand the influence of various parameters on the patient's prognosis based on the weights of the parameters and make a comprehensive judgment.

[0127] First, a hemodynamic model of each parameter reflecting plaque riskiness is established. The difference between this model and the conventional hemodynamic model is that it takes into account the strain of the coronary artery.

[0128] According to the dynamic equation of Poiseuille flow in a circular tube, the relationship equation between blood flow parameters at the stenosis and coronary artery morphological parameters can be obtained:

[0129]

[0130] Among them, Q is the flow rate, R is the radius of the coronary artery lumen at the narrowest part, ΔP is the pressure difference before and after stenosis, L is the length of the stenosis segment, μ is the blood viscosity, and τ is the wall shear stress.

[0131] We introduce the coronary artery strain parameter into the model. The coronary artery strain can be defined as the difference in the lumen between systole and diastole at the narrowest part, that is, ΔR. Then the relationship between blood flow and strain is:

[0132] Among them, Q is the flow rate, R is the radius of the coronary artery lumen at the narrowest part, ΔP is the pressure difference before and after stenosis, L is the length of the stenosis segment, μ is the blood viscosity, and τ is the wall shear stress.

[0133] Then, we establish a constraint-based linear regression model and solve the linear regression model based on the above constraints:

[0134] F(X) = a1Q + a2(ΔP) + a3R + a4τ + a5(ΔR) + a6(ΔQ) + a7L + … + b1x1 + … + b n x n ;

[0135] where a and b are parameters to be fitted; x is other coronary artery strain parameters, plaque morphology parameter information, and hemodynamic parameters; Q is the flow rate; R is the radius of the coronary artery lumen at the narrowest point; ΔP is the pressure difference before and after stenosis; L is the length of the stenotic segment, μ is the blood viscosity; τ is the wall shear stress; ΔR is the difference in the lumen between systole and diastole, that is, the strain, and ΔQ is the difference between the flow rate assuming the lumen R remains unchanged in systole and the flow rate assuming the lumen R remains unchanged in diastole.

[0136] To solve this constrained optimization equation, the constrained least squares method is usually used. Since the above constraint equation is a non-linear equation and the solution is relatively complex, it can be calculated by introducing an iterative numerical calculation method of the Lagrangian function.

[0137] This application proposes a method for comprehensively evaluating the risk of coronary artery plaques based on medical images, which has the following advantages:

[0138] 1. By combining the stress-state time series images with the resting-state time series images, extracting the coronary artery, and calculating the strain of the coronary artery at different time phases, more accurate strain values can be obtained.

[0139] 2. Considering coronary artery strain, plaque morphology parameters, and hemodynamic parameters simultaneously for evaluating the risk of plaques.

[0140] This application also provides a device for comprehensively evaluating the risk of coronary artery plaques based on medical images. The device for comprehensively evaluating the risk of coronary artery plaques based on medical images includes a coronary artery strain parameter acquisition module, a plaque morphology parameter information acquisition module, a hemodynamic parameter information acquisition module, and a plaque risk index acquisition module, where

[0141] The coronary artery strain parameter acquisition module is used to obtain coronary artery strain parameters according to the coronary artery tree segmentation image;

[0142] The plaque morphology parameter information acquisition module is used to obtain plaque morphology parameter information according to the plaque segmentation image;

[0143] The hemodynamic parameter information acquisition module is used to obtain hemodynamic parameter information according to the resting-state time series image;

[0144] The plaque risk index acquisition module is used to acquire a plaque risk index according to the coronary artery strain parameters, plaque morphology parameter information, and hemodynamic parameter information.

[0145] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. A method for comprehensively evaluating the risk of coronary artery plaques based on medical images, characterized in that, The method for comprehensively evaluating the risk of coronary artery plaques based on medical images includes: Obtaining coronary artery strain parameters according to the coronary artery tree segmentation image; Obtaining plaque morphological parameter information according to the plaque segmentation image; Obtaining hemodynamic parameter information according to the resting-state time series image; Obtaining a plaque risk index according to the coronary artery strain parameters, plaque morphological parameter information, and hemodynamic parameter information.

2. The method for comprehensively evaluating the risk of coronary artery plaques based on medical images according to claim 1, wherein The obtaining of coronary artery strain parameters according to the coronary artery tree segmentation image includes: Obtaining load-state time series images at different time phases and resting-state time series images at different time phases; Obtaining segmented coronary artery tree images at different time phases according to the load-state time series image and the resting-state time series image respectively; Performing the following operations on the segmented coronary artery tree images at different time phases: Extracting the center line of the coronary artery in the coronary artery tree image and straightening and unfolding the coronary artery to obtain a straightened coronary artery image; Detecting calcified plaques and non-calcified plaques on the coronary artery according to the obtained coronary artery image. When a calcified plaque or a non-calcified plaque is detected, image segmentation is performed on the detected calcified plaque or the detected non-calcified plaque to obtain a calcified plaque image or a non-calcified plaque image. The calcified plaque image or the non-calcified plaque image constitutes the plaque segmentation image; wherein, one calcified plaque has multiple calcified plaque images at different time phases, and one non-calcified plaque has multiple non-calcified plaque images at different time phases; Obtaining the coronary artery strain parameter corresponding to each calcified plaque according to the calcified plaque images at different time phases of each calcified plaque; Obtaining the coronary artery strain parameter corresponding to each non-calcified plaque according to the non-calcified plaque images at different time phases of each non-calcified plaque.

3. The method for comprehensively evaluating the risk of coronary artery plaques based on medical images according to claim 2, wherein, The obtaining of the coronary artery strain parameter corresponding to each calcified plaque according to the calcified plaque images at different time phases of each calcified plaque includes: Performing the following operations on each of the calcified plaque images at different time phases belonging to the same calcified plaque: Calculate the minimum coronary lumen diameter of each calcified plaque image respectively to obtain a sequence of the minimum coronary diameters, and take the maximum value D in the sequence of the minimum coronary diameters max and the minimum value D min , and obtain the coronary strain parameters of each calcified plaque through the following formula: e = (D max - D min ) / D max ; where e is the coronary artery strain parameter.

4. The method for comprehensively evaluating the risk of coronary artery plaques based on medical images according to claim 3, characterized in that, The obtaining of the coronary artery strain parameter corresponding to each non-calcified plaque according to the non-calcified plaque images at different time phases of each non-calcified plaque includes: Performing the following operations on each of the non-calcified plaque images at different time phases belonging to the same non-calcified plaque: Calculate the minimum coronary lumen diameter of each non-calcified plaque image respectively to obtain a sequence of the minimum coronary diameters, and take the maximum value D in the sequence of the minimum coronary diameters max and the minimum value D min , and obtain the coronary strain parameters of each non-calcified plaque through the following formula: e = (D max - D min ) / D max ; where e is the coronary artery strain parameter.

5. The method for comprehensively evaluating the risk of coronary artery plaques based on medical images according to claim 4, characterized in that, The plaque morphological parameter information includes plaque composition information, plaque morphological information, plaque location information, and plaque density information.

6. The method for comprehensively evaluating the risk of coronary artery plaques based on medical images according to claim 5, wherein, The obtaining of hemodynamic parameter information according to the resting-state time series image includes: Comparing the segmented coronary artery tree images at each time phase with each other to obtain a resting-state time series image at one time phase with the smallest coronary artery lumen diameter at each time phase, and performing hemodynamic simulation on the coronary artery in this group of resting-state time series images to obtain hemodynamic parameters.

7. The method for comprehensively evaluating the risk of coronary artery plaque based on medical images according to claim 6, wherein The obtaining of the plaque risk index according to the coronary artery strain parameters, plaque morphological parameter information, and hemodynamic parameter information is obtained through the following formula: F(X) = a1Q + a2(ΔP) + a3R + a4τ + a5(ΔR) + a6(ΔQ) + a7L + … + b1x1 + … + b n x n ; Among them, a and b are parameters to be fitted; x is other coronary artery strain parameters, plaque morphology parameter information, and hemodynamic parameters; Q is the flow rate; R is the coronary artery lumen radius at the narrowest point; ΔP is the pressure difference before and after stenosis; L is the length of the stenosis segment, μ is the blood viscosity; τ is the wall shear stress; ΔR is the difference in the lumen between systole and diastole, that is, the strain, and ΔQ is the difference between the flow rate assuming the lumen R remains unchanged during systole and the flow rate assuming the lumen R remains unchanged at the diastolic size.

8. The method for comprehensively evaluating the risk of coronary artery plaques based on medical images according to claim 7, characterized in that, The difference in the lumen between systole and diastole, ΔR, is obtained by the following formula: Among them, Q is the flow rate, R is the coronary artery lumen radius at the narrowest point, ΔP is the pressure difference before and after stenosis, L is the length of the stenosis segment, μ is the blood viscosity, and τ is the wall shear stress.

9. The method for comprehensively evaluating the risk of coronary artery plaques based on medical images according to claim 8, wherein The flow rate Q is obtained by the following formula: Among them, Q is the flow rate, R is the coronary artery lumen radius at the narrowest point, ΔP is the pressure difference before and after stenosis, L is the length of the stenosis segment, μ is the blood viscosity, and τ is the wall shear stress; The wall shear stress τ is obtained by the following formula: Among them, Q is the flow rate, R is the coronary artery lumen radius at the narrowest point, ΔP is the pressure difference before and after stenosis, L is the length of the stenosis segment, μ is the blood viscosity, and τ is the wall shear stress.

10. An apparatus for comprehensively evaluating the risk of coronary artery plaques based on medical images, characterized in that, The device for comprehensively evaluating the risk of coronary artery plaques based on medical images includes: A coronary artery strain parameter acquisition module, which is used to acquire coronary artery strain parameters according to the coronary artery tree segmentation image; A plaque morphology parameter information acquisition module, which is used to acquire plaque morphology parameter information according to the plaque segmentation image; A hemodynamic parameter information acquisition module, which is used to acquire hemodynamic parameter information according to the resting state time series image; A plaque risk index acquisition module, which is used to acquire a plaque risk index according to the coronary artery strain parameters, plaque morphology parameter information, and hemodynamic parameter information.

Images