Method, device, storage medium and processor for parameter determination of epicardial fat

Abstract

The application discloses a method and device for measuring epicardial fat parameters, a storage medium and a processor. In the scheme, the cross-sectional profile point tracing data of the coronary CT angiography image to be measured is obtained; the epicardial fat contour line of the cross-sectional image corresponding to the cross-sectional profile point tracing data is fitted by using the cross-sectional profile point tracing data; the pericardial contour curved surface passing through the plurality of epicardial fat contour lines is fitted and generated by using lofting technology according to the plurality of fitted epicardial fat contour lines; the pericardial cavity three-dimensional model is generated according to the pericardial contour curved surface; and the epicardial fat parameters corresponding to the coronary CT angiography image are measured by performing voxel analysis on the pericardial cavity three-dimensional model. The technical scheme of the application can realize the purpose of measuring the epicardial fat parameters with high accuracy in an automatic manner, and improves the auxiliary value of the measured epicardial fat parameters for the research on cardiovascular diseases.

Description

Technical Field

[0001] This application relates to the field of medical image processing technology, and in particular to a method, apparatus, storage medium and processor for measuring parameters of epicardial fat. Background Technology

[0002] Epicardial adipose tissue (EAT) is a layer of fatty tissue located between the myocardium and the visceral pericardium. It has attracted considerable attention due to its unique anatomical and physiological relationship with the heart. EAT has various physiological functions, such as mechanically protecting the myocardium, providing heat and energy to the myocardium, helping to protect coronary circulation, improving endothelial function, and reducing oxidative stress and inflammation, thereby achieving the goal of protecting the heart. Measuring the parameters of epicardial adipose tissue is of great significance for research on cardiovascular diseases.

[0003] Currently, the parameters of epicardial fat are typically determined by manually outlining the pericardial contour layer by layer in coronary CT angiography images, and then calculating the parameters of epicardial fat from the outlined EAT region. However, manual measurement of epicardial fat parameters is not only time-consuming, but also prone to errors due to unclear pericardial contours in some CT scans and a lack of continuity between layers, resulting in low accuracy in parameter measurement. Therefore, it cannot effectively support research on cardiovascular diseases. Summary of the Invention

[0004] To address the aforementioned issues, this application provides a method, apparatus, storage medium, and processor for measuring epicardial fat parameters, with the aim of achieving highly accurate measurement of epicardial fat parameters in an automated manner.

[0005] The embodiments of this application disclose the following technical solutions:

[0006] In a first aspect, this application provides a method for measuring parameters of epicardial fat, the method comprising:

[0007] Acquire cross-sectional contour point recording data of the coronary CT angiography image to be measured; the coronary CT angiography image is a three-dimensional image, which includes multi-layer two-dimensional cross-sectional images;

[0008] The epicardial fat contour line of the cross-sectional image corresponding to the cross-sectional contour point tracing data is obtained by fitting the cross-sectional contour point tracing data;

[0009] Based on the multiple epicardial fat contour lines obtained by fitting, a pericardial contour surface passing through the multiple epicardial fat contour lines is generated by fitting using the lofting technique.

[0010] A three-dimensional model of the pericardial cavity is generated based on the pericardial contour surface.

[0011] By performing voxel analysis on a three-dimensional model of the pericardial cavity, the epicardial fat parameters corresponding to the coronary CT angiography images were determined.

[0012] In an optional implementation, the cross-sectional contour point recording data includes the positional information of the cross-sectional contour points recorded on the corresponding cross-sectional image; the epicardial fat contour line of the cross-sectional image corresponding to the cross-sectional contour point recording data is obtained by fitting the cross-sectional contour point recording data, including:

[0013] Using the positional information of the cross-sectional contour points marked on the target cross-sectional image, an initial contour line of epicardial fat passing through all cross-sectional contour points on the target cross-sectional image is generated;

[0014] Based on the positional information of the cross-sectional contour points marked on the target cross-sectional image and the initial contour line of the epicardial fat, multiple curve segments of the initial contour line of the epicardial fat are determined.

[0015] Multiple curve control points are generated based on the positional information of the cross-sectional contour points marked on the target cross-sectional image and multiple curve segments.

[0016] Based on the location information of multiple control points and cross-sectional contour points, a closed curve passing through all cross-sectional contour points and all curve control points on the target cross-sectional image is fitted, which serves as the epicardial fat contour line of the target cross-sectional image.

[0017] In an optional implementation, based on the fitted multiple epicardial fat contour lines, a pericardial contour surface passing through the multiple epicardial fat contour lines is generated by fitting using a lofting technique, including:

[0018] Using multiple epicardial fat contour lines, the directional variables of the corresponding pericardial sections and the directional variables of the corresponding perpendicular lines of the pericardial sections are extracted respectively;

[0019] Based on the weights of the control points of each curve on multiple epicardial fat contour lines, the extracted directional variables of each pericardial section and the directional variables of the vertical line, a pericardial contour surface composed of a quadrilateral finite element network is generated by lofting technology.

[0020] In an optional implementation, the number of epicardial fat contour lines used to generate the pericardial contour surface is M, where M is less than the total number of slices in the cross-sectional images contained in the coronary CT angiography image; by performing voxel analysis on the three-dimensional model of the pericardial cavity, the epicardial fat parameters corresponding to the coronary CT angiography image are determined, including:

[0021] Based on the three-dimensional model of the pericardial cavity, the epicardial fat contour line of the cross-sectional image of the coronary CT angiography examination that did not obtain the epicardial fat contour line was calculated by interpolation.

[0022] Filter out voxels in the 3D model of the pericardial cavity that do not conform to the voxel CT value range of adipose tissue;

[0023] Based on the analysis of the retained voxels, the epicardial fat parameters corresponding to the coronary CT angiography images were determined.

[0024] In an optional implementation, analysis is performed based on the retained voxels to determine epicardial fat parameters corresponding to coronary CT angiography images, including:

[0025] Based on the analysis of the retained voxels, the volume of epicardial fat and CT value characteristic parameters are obtained; the CT value characteristic parameters include at least one of the following: minimum CT value, maximum CT value, average CT value, median CT value or standard deviation of CT value;

[0026] The CT value characteristic parameter is used as the density of epicardial fat.

[0027] In optional implementation methods, the method for measuring epicardial fat parameters further includes: using the volume and density of epicardial fat obtained from coronary CT angiography images as predictive factors, obtaining the first predictive efficacy improvement information of the first predictive model compared to the baseline model for heart failure with improved ejection fraction through multiple regression analysis, obtaining the second predictive efficacy improvement information of the second predictive model compared to the baseline model for heart failure with improved ejection fraction through multiple regression analysis, and obtaining the third predictive efficacy improvement information of the third predictive model compared to the baseline model for heart failure with improved ejection fraction through multiple regression analysis.

[0028] The first prediction model uses the volume of epicardial fat as one of the predictors; the second prediction model uses the density of epicardial fat as one of the predictors; the third prediction model uses both the volume and density of epicardial fat as two of the predictors; the baseline model does not use the volume and density of epicardial fat as predictors.

[0029] In a second aspect, this application provides a device for measuring parameters of epicardial fat, the device comprising:

[0030] The acquisition module is used to acquire the cross-sectional contour point recording data of the coronary CT angiography image to be measured; the coronary CT angiography image is a three-dimensional image, which includes multiple two-dimensional cross-sectional images.

[0031] The curve fitting module is used to fit the cross-sectional image corresponding to the cross-sectional image using the cross-sectional contour point tracing data to obtain the epicardial fat contour line.

[0032] The surface generation module is used to generate a pericardial contour surface that passes through multiple epicardial fat contour lines by using lofting technology, based on the multiple epicardial fat contour lines obtained through fitting.

[0033] The 3D model generation module is used to generate a 3D model of the pericardial cavity based on the pericardial contour surface.

[0034] The parameter measurement module is used to measure the epicardial fat parameters corresponding to the coronary CT angiography images by performing voxel analysis on the three-dimensional model of the pericardial cavity.

[0035] Optionally, the cross-sectional contour point recording data includes the positional information of the cross-sectional contour points recorded on the corresponding cross-sectional image; the curve fitting module includes:

[0036] The first generation unit is used to generate an initial epicardial fat contour line that passes through all cross-sectional contour points on the target cross-sectional image by utilizing the position information of the cross-sectional contour points marked on the target cross-sectional image.

[0037] The determination unit is used to determine multiple curve segments of the initial contour line of the epicardial fat based on the position information of the cross-sectional contour points marked on the target cross-sectional image and the initial contour line of the epicardial fat.

[0038] The second generation unit is used to generate multiple curve control points based on the position information of multiple curve segments and cross-sectional contour points marked on the target cross-sectional image;

[0039] The fitting unit is used to fit a closed curve passing through all cross-sectional contour points and all curve control points on the target cross-sectional image based on the position information of multiple control points and the position information of cross-sectional contour points, as the epicardial fat contour line of the target cross-sectional image.

[0040] In a third aspect, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for measuring parameters of epicardial fat.

[0041] In a fourth aspect, this application provides a processor for running a computer program, which, when running, executes the above-described method for measuring parameters of epicardial fat.

[0042] Compared with the prior art, this application has the following beneficial effects:

[0043] In this application's technical solution, the epicardial fat contour lines of the corresponding cross-sectional images are automatically fitted using the cross-sectional contour point recording data of the coronary CT angiography images to be measured. This eliminates the need for manual delineation of the pericardial contour, avoiding errors caused by unclear pericardial contours in some CT scans, which leads to low accuracy in epicardial fat parameter measurement. Then, based on the fitted multiple epicardial fat contour lines, a pericardial contour surface passing through these contour lines is generated using lofting techniques. A three-dimensional model of the pericardial cavity is automatically generated from this surface, and voxel analysis is performed on this model to determine the epicardial fat parameters corresponding to the coronary CT angiography images. This achieves highly accurate measurement of epicardial fat parameters in an automated manner, enhancing the auxiliary value of the measured epicardial fat parameters in cardiovascular disease research. Furthermore, it provides valuable reference for medical professionals (e.g., drug administration and treatment of relevant personnel). Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 A flowchart illustrating a method for measuring parameters of epicardial fat provided in an embodiment of this application;

[0046] Figure 2A A schematic diagram of the epicardial fat contour in a cross-sectional image provided in an embodiment of this application;

[0047] Figure 2B A schematic diagram of the epicardial fat contour line in another cross-sectional image provided in an embodiment of this application;

[0048] Figure 3 A flowchart illustrating the fitting process of an epicardial fat contour line provided in this application embodiment;

[0049] Figure 4 A schematic diagram of a pericardial contour surface provided for an embodiment of this application;

[0050] Figure 5 A schematic diagram of a three-dimensional model of the pericardial cavity provided in an embodiment of this application;

[0051] Figure 6A flowchart illustrating the generation process of a pericardial contour surface provided in this application embodiment;

[0052] Figure 7 A flowchart illustrating the process of measuring epicardial fat parameters is provided in this application embodiment;

[0053] Figure 8A A schematic diagram of a cross-section of the pericardium before voxel filtering in a three-dimensional model of a cavity that does not conform to the voxel CT value range of adipose tissue, provided for an embodiment of this application;

[0054] Figure 8B A schematic diagram of a cross-section of the pericardium after voxel filtering in a three-dimensional model of a pericardium that does not conform to the voxel CT value range of adipose tissue, provided for an embodiment of this application;

[0055] Figure 9 A schematic diagram illustrating comparative prediction performance improvement information provided in an embodiment of this application;

[0056] Figure 10 This is a schematic diagram of a device for measuring parameters of epicardial fat provided in an embodiment of this application. Detailed Implementation

[0057] As described earlier, current methods for measuring epicardial fat parameters typically involve manually outlining the pericardial contour layer by layer in coronary CT angiography images and calculating the parameters of the epicardial fat based on the outlined EAT region. However, this manual method is not only time-consuming but also prone to errors due to unclear pericardial contours in some CT scans and a lack of continuity between layers, resulting in low accuracy in epicardial fat parameter measurements. Therefore, it cannot effectively support research on cardiovascular diseases.

[0058] Through research, the inventors have developed a method to automatically fit the epicardial fat contour lines of the corresponding cross-sectional images from coronary CT angiography images. This eliminates the need for manual delineation of the pericardial contour, avoiding errors caused by unclear pericardial contours in some CT scans and resulting in low accuracy in epicardial fat parameter measurement. Then, based on the fitted multiple epicardial fat contour lines, a pericardial contour surface passing through these contour lines is generated using lofting techniques. A three-dimensional model of the pericardial cavity is automatically generated from this surface, and voxel analysis is performed on this model to determine the epicardial fat parameters corresponding to the coronary CT angiography images. This automated method achieves high accuracy in epicardial fat parameter measurement, enhancing its auxiliary value in cardiovascular disease research and providing valuable reference for medical professionals (e.g., drug administration and treatment).

[0059] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0060] Method Implementation Examples

[0061] This application provides an embodiment of a method for measuring parameters of epicardial fat. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system, such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, it is merely an example order provided, and in some cases, the steps shown or described may be performed in a different order than that presented here.

[0062] See Figure 1 The figure is a flowchart of a method for measuring parameters of epicardial fat according to an embodiment of this application. Figure 1 As shown, the method includes the following steps:

[0063] Step S101: Obtain the cross-sectional contour point recording data of the coronary artery CT angiography image to be measured.

[0064] In one optional embodiment, an epicardial fat parameter measurement system can be used as the execution subject of the epicardial fat parameter measurement method of this application embodiment. For ease of description, the epicardial fat parameter measurement system will be referred to as the system below.

[0065] In step S101, the coronary CT angiography images are three-dimensional images, including multi-layered two-dimensional cross-sectional images. The coronary CT angiography images are obtained using a dual-source CT system. After obtaining the coronary CT angiography images using the dual-source CT system, the system can automatically upload the images to a dedicated workstation for image post-processing, which includes, but is not limited to, automatic segmentation and image reconstruction.

[0066] Currently, methods for measuring EAT include echocardiography and cardiac magnetic resonance imaging (MRI). However, because echocardiography is a two-dimensional examination while EAT is a three-dimensional structure, uneven distribution may exist in a two-dimensional plane, making it impossible for echocardiography to accurately measure the volume and density of EAT. Cardiac MRI, on the other hand, has a long examination time, many contraindications, and is not suitable for all patients, exhibiting poor versatility.

[0067] To address the aforementioned issues, this application proposes to obtain three-dimensional coronary CT angiography images via CT. These images allow for accurate determination of EAT parameters and provide accurate data for subsequent epicardial fat parameter measurements.

[0068] In this embodiment, the cross-sectional contour point recording data consists of epicardial fat contour points manually recorded in multiple two-dimensional cross-sectional images that clearly display the pericardial contour. Specifically, it can be as follows: Figure 2A The outline points D0, D1, D2, D3, ..., D of the blue marker shown are... n n+1 data points (i.e., epicardial fat contour points) are roughly evenly set along the pericardial contour. i =(x i y i ), where i = 0, 1, 2, ..., n. In this application, only a portion of the two-dimensional cross-sectional images from the multi-layer two-dimensional cross-sectional images are manually marked with epicardial fat contour points.

[0069] Step S102: The epicardial fat contour line of the cross-sectional image corresponding to the cross-sectional contour point recording data is obtained by fitting the cross-sectional contour point recording data.

[0070] To improve the accuracy of epicardial fat parameter measurement and address the problem that existing technologies rely on manual delineation of the pericardial contour, which often results in errors due to unclear pericardial contours in some CT scans, leading to low accuracy in epicardial fat parameter measurement, the system of this application can automatically fit the epicardial fat contour line of the cross-sectional image corresponding to the obtained cross-sectional contour point recording data of the coronary artery CT angiography image to be measured.

[0071] Optionally, Figure 3 A flowchart illustrating the fitting process of the epicardial fat contour line provided in this application embodiment is shown below. Figure 3 As shown, the process includes the following steps, wherein the cross-sectional profile point recording data includes the position information of the cross-sectional profile points recorded on the corresponding cross-sectional image.

[0072] Step S1021: Using the position information of the cross-sectional contour points marked on the target cross-sectional image, generate the initial contour line of the epicardial fat that passes through all cross-sectional contour points on the target cross-sectional image.

[0073] In this embodiment, the system can utilize the position information D of n+1 data points (i.e., cross-sectional contour points) that are set approximately uniformly along the pericardial contour. i =(x i y i This process involves making the pericardial contour pass through all data points and form a closed curve, thus obtaining the initial contour of the epicardial fat. Figure 2A The blue curve shown is the initial outline of the epicardial fat.

[0074] Step S1022: Based on the position information of the cross-sectional contour points marked on the target cross-sectional image and the initial contour line of the epicardial fat, determine multiple curve segments of the initial contour line of the epicardial fat.

[0075] In this embodiment, the system can divide the initial contour line of the epicardial fat into multiple curve segments based on the position information of the cross-sectional contour points, which can be respectively as follows: Figure 2A The C1(t), C2(t), C3(t), ..., C shown are... n (t), for a certain curve segment C i (t), whose starting points are respectively D i-1 and D i .

[0076] Step S1023: Generate multiple curve control points based on the position information of multiple curve segments and cross-sectional contour points marked on the target cross-sectional image.

[0077] In this embodiment, the system can determine the location information of the cross-sectional contour points and multiple curve segments C. i(t) Fit a cubic NURBS curve, which has n+1 control points P. i , Figure 2A The red marker in the diagram is control point P. i Where i = 0, 1, 2, ..., n. Since this spline curve is a closed curve, to ensure that the spline maintains C at the beginning and end joints... 2 Continuity, two new control points P are added. n+1 and P n+2 and make P n =P0, P n+1 =P1 and P n+2 =P2. The specific formula for defining a cubic NURBS curve is shown in formula (1).

[0078]

[0079] Where C(i) is a cubic NURBS curve; P (i mod n ) N is the control point in the Cartesian coordinate system; i,3 (i) represents the basis functions of the cubic spline, where the non-decreasing node vector of the spline is: Ξ={ξ1,ξ2,ξ3,…,ξ n+p ,ξ n+p+1}. Where ξ is the direction variable of the pericardial cross-section, for [ξ1, ξ2]... n+p+1 For any node ξ in ], its basis function is defined according to the Cox de-boor recursive formula, as shown in formula (2) and formula (3).

[0080] When p = 0

[0081] When p = 1, 2, 3...

[0082]

[0083] The system can generate multiple curve control points based on the above formulas (1), (2), and (3) and the position information of the cross-sectional contour points. Figure 2A The red markers in the diagram are the control points for the generated curve.

[0084] Step S1024: Based on the position information of multiple control points and the position information of cross-sectional contour points, fit a closed curve that passes through all cross-sectional contour points and all curve control points on the target cross-sectional image, and use it as the epicardial fat contour line of the target cross-sectional image.

[0085] In this embodiment, the system can use the position information of multiple control points and the position information of cross-sectional contour points to make the fitted cubic NURBS curve correctly delineate the pericardial contour, obtaining a closed curve passing through all cross-sectional contour points and all curve control points on the target cross-sectional image, as... Figure 2B The target cross-sectional image shown shows the epicardial fat contour.

[0086] Step S103: Based on the multiple epicardial fat contour lines obtained by fitting, a pericardial contour surface passing through the multiple epicardial fat contour lines is generated by fitting using lofting technology.

[0087] Step S104: Generate a three-dimensional model of the pericardial cavity based on the pericardial contour surface.

[0088] To further improve the accuracy of epicardial fat parameter measurement, the system in this application can generate a model based on multiple fitted epicardial fat contour lines, using these contour lines as a framework and employing B-spline lofting technology. Figure 4 The pericardial contour surface shown is drawn through multiple epicardial fat contour lines, and is generated based on the pericardial contour surface. Figure 5 The three-dimensional model of the pericardial cavity is shown in the figure.

[0089] Optionally, Figure 6 A flowchart illustrating the generation process of a pericardial contour surface, as provided in this application embodiment, is shown below. Figure 6 As shown, the process includes the following steps:

[0090] Step S1031: Using multiple epicardial fat contour lines, extract the direction variables of the corresponding pericardial cross sections and the direction variables of the corresponding perpendicular lines of the pericardial cross sections.

[0091] In this embodiment, the system can utilize multiple epicardial fat contour lines to extract the direction variable ξ of the pericardial cross-section corresponding to each epicardial fat contour line, and extract the direction variable η of the perpendicular line of the pericardial cross-section corresponding to each epicardial fat contour line.

[0092] Step S1032: Based on the weights of the control points of each curve on multiple epicardial fat contour lines, the extracted directional variables of each pericardial section and the directional variables of the vertical line, a pericardial contour surface composed of a quadrilateral finite element network is generated by lofting technology.

[0093] Optionally, B-spline lofting is a modeling technique that can be used to create complex 3D models and can handle objects with varying cross-sectional shapes. In this embodiment, during the process of generating a pericardial contour surface composed of a quadrilateral finite element network using B-spline lofting, the system needs to ensure that any two NURBS curves of the lofting technique do not intersect. Therefore, the system can ensure that any two NURBS curves do not intersect using the following formulas (4) and (5) to generate a pericardial contour surface composed of a quadrilateral finite element network.

[0094]

[0095] Where C(ξ,η) is the NURBS curve, ξ is the direction variable of the pericardial cross-section, and η is the direction variable of the vertical line of the cross-section. i,j Let (X, Y, Z) be the control point in the Cartesian coordinate system; R is the shape formula, and the specific binary NURBS shape formula is as follows:

[0096]

[0097] Where N and M are the i-th p-th and j-th q-th basis functions, respectively, and w ij These are the weights corresponding to the control points.

[0098] It should be noted that this application allows for adjustment of the surface shape and curvature by adjusting control points, which has significant advantages for analyzing changes in EAT at different times in CT scans of the same patient.

[0099] Step S105: By performing voxel analysis on the three-dimensional model of the pericardial cavity, the epicardial fat parameters corresponding to the coronary CT angiography images are determined.

[0100] To further improve the accuracy of epicardial fat parameter measurement, the system in this application can perform voxel analysis on the three-dimensional model of the pericardial cavity, filter out voxels in the three-dimensional model that do not conform to the voxel CT value range of adipose tissue, and measure parameters such as volume and density of epicardial fat corresponding to the coronary CT angiography image based on the retained voxels.

[0101] In this embodiment, the number of epicardial fat contour lines used to generate the pericardial contour surface is M, where M is less than the total number of slices in the cross-sectional images contained in the coronary CT angiography image. Figure 7 A flowchart illustrating the process of measuring epicardial fat parameters provided in this application embodiment is shown below. Figure 7 As shown, the process includes the following steps:

[0102] Step S1051: Based on the three-dimensional model of the pericardial cavity, interpolate to calculate the epicardial fat contour of the cross-sectional image of the coronary CT angiography examination image where the epicardial fat contour was not obtained.

[0103] In this embodiment, the system only needs to perform interpolation calculations on the cross-sectional images of coronary CT angiography images that do not have the epicardial fat contour line, to obtain the epicardial fat contour line corresponding to the cross-sectional image without the epicardial fat contour line.

[0104] Step S1052: Filter out voxels in the pericardial cavity 3D model that do not conform to the voxel CT value range of adipose tissue.

[0105] In this embodiment, the pre-voxel-filtered pericardial cross-section in the three-dimensional model of the pericardial cavity that does not conform to the voxel CT value range of adipose tissue can be as follows: Figure 8A As shown, the voxel CT value range that conforms to epicardial fat is -190 to -30 hnsf'U. The system can filter voxels that do not conform to the voxel CT value range of adipose tissue to obtain the results shown below. Figure 8B The pericardial cross-section shown is composed of voxels that have been preserved.

[0106] Step S1053: Based on the retained voxels, analyze and determine the epicardial fat parameters corresponding to the coronary CT angiography images.

[0107] In this embodiment, the system can perform analysis based on the retained voxels to obtain parameters such as the volume and density of epicardial fat. The density of epicardial fat may include maximum density, minimum density, average density, median density, and density standard deviation.

[0108] It should be noted that by interpolating to calculate the epicardial fat contour in cross-sectional images of coronary CT angiography where the epicardial fat contour was not obtained, the problem of low accuracy in epicardial fat parameter measurement caused by the lack of continuity between the contours of each two-dimensional cross-sectional image drawn manually is overcome, and the accuracy of epicardial fat parameter measurement is further improved.

[0109] Optionally, based on the retained voxels, the system analyzes and determines the epicardial fat parameters corresponding to the coronary CT angiography images, including: the system can analyze the retained voxels to obtain the volume and CT value characteristic parameters of the epicardial fat; the CT value characteristic parameters include at least one of the following: minimum CT value, maximum CT value, mean CT value, median CT value, or standard deviation of CT value; then the system can use the CT value characteristic parameters as the density of the epicardial fat.

[0110] Based on the scheme described in steps S101 to S105 above, it can be seen that by utilizing the cross-sectional contour point recording data of the coronary CT angiography images to be measured, the epicardial fat contour lines of the corresponding cross-sectional images are automatically fitted, eliminating the need for manual delineation of the pericardial contour. This avoids the problem of low accuracy in epicardial fat parameter measurement due to unclear pericardial contours in some CT scans. Then, based on the fitted multiple epicardial fat contour lines, a pericardial contour surface passing through these contour lines is generated using lofting techniques. A three-dimensional model of the pericardial cavity is automatically generated based on this surface. By performing voxel analysis on the three-dimensional model, the epicardial fat parameters corresponding to the coronary CT angiography images are measured. This achieves the goal of high-accuracy measurement of epicardial fat parameters in an automated manner, enhancing the auxiliary value of the measured epicardial fat parameters for cardiovascular disease research. It can also provide valuable reference for medical professionals' work (e.g., drug administration and treatment of relevant personnel).

[0111] Left ventricular ejection fraction (LVEF) is the most important indicator for assessing left ventricular systolic function in heart failure patients and a key basis for heart failure classification. HfimpEF, or heart failure with reduced ejection fraction, is a unique type of heart failure characterized by ventricular remodeling and improved cardiac function. HfimpEF has specific pathophysiological mechanisms, clinical features, and prognostic outcomes. Compared to other heart failure types such as HFrEF (heart failure with reduced ejection fraction) or HFpEF (heart failure with preserved ejection fraction), HFimpEF has a better prognosis. Early identification of patients with HFimpEF or those at risk of developing HFimpEF helps explore the mechanisms of ventricular remodeling and myocardial rehabilitation, improving our understanding of structural and functional changes in heart failure. Therefore, identifying predictors of HFimpEF is crucial.

[0112] Current technologies for predicting a patient's potential to develop HFimpEF typically use predictive factors including age, sex, duration of heart failure, etiology, number of comorbidities, and metabolic factors such as blood glucose levels, insulin resistance, and blood glucose variability. However, because these predictive factors lack specificity, current technologies suffer from low predictive efficacy in assessing a patient's potential to develop HFimpEF.

[0113] To address the aforementioned issues and improve predictive efficacy, the system in this embodiment utilizes the volume and density of epicardial fat from acquired coronary CT angiography images as predictive factors. The system then uses multivariate regression analysis to obtain information on the first predictive efficacy improvement of the first predictive model compared to the baseline model for heart failure with improved ejection fraction. Furthermore, the system uses multivariate regression analysis to obtain information on the second predictive efficacy improvement of the second predictive model compared to the baseline model for heart failure with improved ejection fraction. Finally, the system uses multivariate regression analysis to obtain information on the third predictive efficacy improvement of the third predictive model compared to the baseline model for heart failure with improved ejection fraction.

[0114] The first prediction model uses the volume of epicardial fat as one of the predictors; the second prediction model uses the density of epicardial fat as one of the predictors; the third prediction model uses both the volume and density of epicardial fat as two of the predictors; the baseline model does not use the volume and density of epicardial fat as predictors.

[0115] In this embodiment, the baseline model is specifically shown in formula (6).

[0116]

[0117] The first prediction model obtained by the multiple regression analysis method is shown in formula (7).

[0118]

[0119] Among them, the polynomial function f(x) included in the EAT volume 年龄 ,x 性别 ,x 既往心梗史 …x EAT体积 The specific details are shown in formula (8).

[0120]

[0121] Among them, eGFR stands for estimated glomerular filtration rate; ARNI stands for angiotensin receptor neprilysin inhibitor; ARB stands for angiotensin II receptor blocker; and HbA1c stands for hemoglobin A1c.

[0122] The second prediction model obtained by the system through multiple regression analysis is shown in formula (9).

[0123]

[0124] Among them, the polynomial function incorporating density: f(x) 年龄 ,x 性别 ,x 既往心梗史 …x EAT密度 Specifically, as shown in formula (10).

[0125]

[0126] Optionally, the system obtains the formula for the third prediction model through multiple regression analysis, as shown in formula (11).

[0127]

[0128] The polynomial function incorporating EAT density and volume can be found in formulas (8) and (10) above. The independent variables for past myocardial infarction history and medication use are represented by 0 or 1 (0 for none, 1 for present).

[0129] In this embodiment, the system can use the volume and density of epicardial fat, as well as the volume and density of epicardial fat, as one of the predictive factors of the predictive model. By inputting relevant patient data (such as age, gender, duration of heart failure, etiology, number of comorbidities, blood glucose level, insulin resistance, blood glucose variability, volume and density of epicardial fat, etc.) into different predictive models (including the first predictive model, the second predictive model, and the third predictive model) for multivariate regression analysis, information on the improvement of predictive efficacy of different predictive models for heart failure with improved ejection fraction compared to the baseline model can be obtained.

[0130] For example, such as Figure 9 As shown, the conventional model (i.e., the baseline model) using parameters such as traditional risk factors, biochemical indicators, and drug treatment had a predictive power of 0.793 for HFimpEF [95% CI 0.731–0.856]. The first predictive model, which added EAT volume to the conventional model, showed a significantly increased predictive power of 0.846 [95% CI 0.792–0.899] compared to the conventional model (P = 0.018). The second predictive model, which added EAT density to the conventional model, further increased its predictive power to 0.833 [95% CI 0.777–0.889]. The third predictive model, which added both EAT volume and density to the conventional model, further increased its predictive power to 0.854 [95% CI 0.803–0.906] compared to the conventional model (P = 0.009).

[0131] It should be noted that the parameters of volume and density of epicardial fat in this application are imaging predictive indicators. Compared with the laboratory indicators in the prior art, the volume and density of epicardial fat have the advantages of being more intuitive, more specific, and more independent.

[0132] Optionally, the system can further evaluate the improvement in predicting HFimpEF by the new model incorporating EAT parameters compared to the conventional model (i.e., the baseline model) using the net reclassification improvement (NRI) and integrated discrimination improvement (IDI) index. Compared to the conventional model, adding the EAT volume significantly increases the predictive power for HFimpEF, with an NRI increase of 13.9% [95% CI 1.2%-26.7%] (P = 0.032) and an IDI increase of 10.7% [95% CI 6.2%-15.2%] (P < 0.001). Adding EAT volume and density to the conventional model further significantly improved its predictive power for HFimpEF, with NRI increasing by 21.3% [95% CI 9.0%-33.7%] (P = 0.001) and IDI increasing by 12.4% [95% CI 7.6%-17.2%] (P < 0.001).

[0133] It should be noted that this application may also use EAT parameters measured by echocardiography, cardiac magnetic resonance imaging, or CT scans using other software as a predictor of a patient’s potential to develop HFimpEF.

[0134] In an optional embodiment, parameters of epicardial fat in multiple patients with heart rate-pregnancy-induced cardiac arrest (HFrEF) can be measured. The mean age of these HFrEF patients was 58.4 ± 13.2 years, and 82.3% were male. After a mean follow-up of 8.6 (4.9–13.3) months, 51.2% of patients recovered to HFimpEF, while the remaining 48.8% remained with HFrEF. Compared to HFrEF patients, HFimpEF patients were predominantly female, younger, had higher diastolic blood pressure, and better renal function. Furthermore, compared to HFrEF patients, HFimpEF patients had fewer cases of ischemic cardiomyopathy, a history of revascularization, and antiplatelet therapy. There were no statistically significant differences between the two groups in other medical history, laboratory findings, or treatments.

[0135] Using the method for measuring epicardial fat parameters provided in this application, the mean epicardial fat volume of the aforementioned multiple HFrEF patients was determined to be 111.51 (101.33–184.36) mL, and the mean epicardial fat density (or attenuation) was -76.79 ± 6.83 HU. Spearman correlation analysis showed a negative correlation between epicardial fat density and epicardial fat volume (r = -0.33, P < 0.001). Compared with HFrEF patients, HFimpEF patients had significantly smaller epicardial fat volumes (115.36 [interquartile 87.08–154.78] mL vs 169.67 [interquartile 137.22–218.89] mL, P < 0.001) and significantly higher mean epicardial fat density (-74.92 ± 6.84 vs -78.76 ± 6.28 HU, P < 0.001). That is, there is a negative correlation between EAT volume and LVEF change (r = -0.35, P < 0.001), while there is a positive correlation between EAT density and LVEF change (r = 0.18, P = 0.009).

[0136] Optionally, after obtaining the parameters of epicardial fat, Spearman correlation analysis reveals that the parameters of epicardial fat (especially EAT volume) are closely related to left ventricular reverse remodeling. EAT volume is positively correlated with changes in left ventricular end-systolic volume index (r = 0.32, P < 0.001), left ventricular end-systolic diameter (r = 0.22, P = 0.002), and left ventricular end-diastolic diameter (r = 0.34, P < 0.001). EAT density, however, is negatively correlated with changes in left ventricular end-systolic volume index (r = -0.14, P = 0.045). That is, a larger EAT volume and a lower EAT density indicate more significant left ventricular shrinkage and more pronounced left ventricular reverse remodeling.

[0137] Optionally, after obtaining the parameters of epicardial fat, receiver operating characteristic (ROC) curve analysis showed that EAT volume had good predictive power for the occurrence of HFimpEF (area under the curve was 0.727 [95% CI 0.657–0.797]). Whether the EAT volume was less than 138 mL was the optimal diagnostic cutoff point for predicting whether a patient would progress to HFimpEF, with a predictive sensitivity of 65.4% and a specificity of 75.7%. Compared with patients with an EAT volume greater than 138 mL, heart failure patients with an EAT volume <138 mL had an approximately 6.1-fold increased probability of developing HFimpEF.

[0138] Meanwhile, EAT density also showed good predictive power for the occurrence of HFimpEF (area under the curve 0.655 [95% CI 0.580–0.729]). Whether the EAT density is greater than -78 HU is the optimal diagnostic cutoff point for predicting whether a patient will progress to HFimpEF, with a sensitivity of 69.2% and a specificity of 55.6%. Compared with patients with EAT density < -78 HU, heart failure patients with EAT density > -78 HU have an approximately 2.7-fold increased probability of developing HFimpEF.

[0139] The method for measuring epicardial fat parameters provided in this application achieves highly accurate and automated measurement, enhancing the auxiliary value of the measured epicardial fat parameters in cardiovascular disease research and providing valuable reference for medical professionals. Compared with existing echocardiography or magnetic resonance imaging methods for measuring epicardial fat (EAT), the method provided in this application is not only highly accurate but also less costly, making it suitable for widespread application. This method also allows epicardial fat parameters (EAT volume and density) to be used as one of the predictive factors for a patient's potential to develop HFimpEF, improving the predictive efficacy for this condition.

[0140] Device Examples

[0141] This application provides a device for measuring parameters of epicardial fat, wherein... Figure 10 This is a schematic diagram of a parameter measuring device for epicardial fat provided in an embodiment of this application, as shown below. Figure 10 As shown, the device includes: an acquisition module 11, a curve fitting module 12, a surface generation module 13, a 3D model generation module 14, and a parameter measurement module 15. From Figure 10 You can see the connections between several modules.

[0142] The acquisition module 11 is used to acquire the cross-sectional contour point recording data of the coronary CT angiography image to be measured; the coronary CT angiography image is a three-dimensional image, which includes multi-layer two-dimensional cross-sectional images.

[0143] The curve fitting module 12 is used to fit the cross-sectional image corresponding to the cross-sectional image using the cross-sectional contour point tracing data to obtain the epicardial fat contour line.

[0144] The surface generation module 13 is used to generate a pericardial contour surface passing through multiple epicardial fat contour lines by fitting multiple epicardial fat contour lines obtained by lofting technology.

[0145] The 3D model generation module 14 is used to generate a 3D model of the pericardial cavity based on the pericardial contour surface.

[0146] The parameter measurement module 15 is used to measure the epicardial fat parameters corresponding to the coronary CT angiography images by performing voxel analysis on the three-dimensional model of the pericardial cavity.

[0147] Optionally, the curve fitting module includes: a first generation unit, a determination unit, a second generation unit, and a fitting unit.

[0148] The cross-sectional contour point recording data includes the position information of the cross-sectional contour points recorded on the corresponding cross-sectional image. The first generation unit is used to generate an initial contour line of epicardial fat that passes through all cross-sectional contour points on the target cross-sectional image by using the position information of the cross-sectional contour points recorded on the target cross-sectional image.

[0149] The determination unit is used to determine multiple curve segments of the initial contour line of the epicardial fat based on the position information of the cross-sectional contour points marked on the target cross-sectional image and the initial contour line of the epicardial fat.

[0150] The second generation unit is used to generate multiple curve control points based on the position information of multiple curve segments and cross-sectional contour points marked on the target cross-sectional image;

[0151] The fitting unit is used to fit a closed curve passing through all cross-sectional contour points and all curve control points on the target cross-sectional image based on the position information of multiple control points and the position information of cross-sectional contour points, as the epicardial fat contour line of the target cross-sectional image.

[0152] Optionally, the surface generation module includes an extraction unit and a third generation unit.

[0153] The extraction unit is used to extract the directional variables of the corresponding pericardial cross section and the directional variables of the corresponding perpendicular line of the pericardial cross section using multiple epicardial fat contour lines.

[0154] The third generation unit is used to generate a pericardial contour surface composed of a quadrilateral finite element network by using lofting techniques, based on the weights of the control points of each curve on multiple epicardial fat contour lines, the extracted directional variables of each pericardial section, and the directional variables of the vertical lines.

[0155] Optionally, the parameter measurement module includes: a calculation unit, a filtering unit, and a parameter measurement unit.

[0156] Among them, the number of epicardial fat contour lines used to generate the pericardial contour surface is M, where M is less than the total number of layers of cross-sectional images contained in the coronary CT angiography image. The calculation unit is used to interpolate and calculate the epicardial fat contour lines of the cross-sectional images of the coronary CT angiography image where the epicardial fat contour lines were not obtained based on the three-dimensional model of the pericardial cavity.

[0157] The filtering unit is used to filter out voxels in the three-dimensional model of the pericardium cavity that do not conform to the voxel CT value range of adipose tissue.

[0158] The parameter measurement unit is used to analyze the retained voxels and measure the epicardial fat parameters corresponding to the coronary CT angiography images.

[0159] Optionally, the parameter measurement unit includes a parameter acquisition subunit and a parameter determination subunit.

[0160] The parameter acquisition subunit is used to perform analysis based on the retained voxels to obtain the volume of epicardial fat and CT value characteristic parameters. The CT value characteristic parameters include at least one of the following: minimum CT value, maximum CT value, average CT value, median CT value, or standard deviation of CT value.

[0161] The parameter determination subunit is used to use CT value characteristic parameters as the density of epicardial fat.

[0162] Optionally, the device for measuring the parameters of epicardial fat further includes: a predictive efficacy improvement information acquisition module, which uses the volume and density of epicardial fat in the acquired coronary CT angiography images as predictive factors, and obtains the first predictive efficacy improvement information of the first predictive model compared to the baseline model for heart failure with improved ejection fraction through a multiple regression analysis method, obtains the second predictive efficacy improvement information of the second predictive model compared to the baseline model for heart failure with improved ejection fraction through a multiple regression analysis method, and obtains the third predictive efficacy improvement information of the third predictive model compared to the baseline model for heart failure with improved ejection fraction through a multiple regression analysis method;

[0163] The first prediction model uses the volume of epicardial fat as one of the predictors; the second prediction model uses the density of epicardial fat as one of the predictors; the third prediction model uses both the volume and density of epicardial fat as two of the predictors; the baseline model does not use the volume and density of epicardial fat as predictors.

[0164] Storage Media Examples

[0165] This application provides a computer-readable storage medium storing a program that, when executed by a processor, implements some or all of the steps in the method for determining the parameters of epicardial fat described in the foregoing method embodiments of this application. The storage medium can be any medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0166] Processor Implementation

[0167] This application provides a processor for running a program, wherein, during program execution, some or all of the steps in the method for determining the parameters of epicardial fat described in the foregoing method embodiments are performed.

[0168] It should be noted that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for the device embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method embodiments. The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separate. The components indicated as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment solution according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0169] The above description is merely one specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method of measuring a parameter of epicardial fat, characterized by, The method comprises the following steps: Obtaining cross-section contour point tracing data of a coronary CT angiography image to be measured; the coronary CT angiography image is a three-dimensional image, and the coronary CT angiography image comprises a plurality of two-dimensional cross-section images; the cross-section contour point tracing data comprises position information of cross-section contour points traced on the corresponding cross-section images; Obtaining an epicardial fat contour line of a target cross-section image by using the position information of the cross-section contour points traced on the target cross-section image; Generating a pericardial contour curved surface passing through the plurality of epicardial fat contour lines by lofting technology according to the plurality of fitted epicardial fat contour lines; Generating a pericardial cavity three-dimensional model according to the pericardial contour curved surface; Determining an epicardial fat parameter corresponding to the coronary CT angiography image by voxel analysis on the pericardial cavity three-dimensional model; The cross-section contour point tracing data comprises position information of cross-section contour points traced on the corresponding cross-section images; The method for obtaining the epicardial fat contour line of the target cross-section image by using the cross-section contour point tracing data comprises the following steps: Generating an epicardial fat initial contour line passing through all the cross-section contour points on the target cross-section image by using the position information of the cross-section contour points traced on the target cross-section image; Determining a plurality of curve segments of the epicardial fat initial contour line according to the position information of the cross-section contour points traced on the target cross-section image and the epicardial fat initial contour line; Generating a plurality of curve control points according to the plurality of curve segments and the position information of the cross-section contour points traced on the target cross-section image; Fitting a closed curve passing through all the cross-section contour points and all the curve control points on the target cross-section image as the epicardial fat contour line of the target cross-section image according to the position information of the plurality of control points and the position information of the cross-section contour points.

2. The method of claim 1, wherein, The method for generating the pericardial contour curved surface passing through the plurality of epicardial fat contour lines by lofting technology according to the plurality of fitted epicardial fat contour lines comprises the following steps: Extracting a direction variable of a pericardial cross-section and a direction variable of a perpendicular line of the pericardial cross-section corresponding to each of the plurality of epicardial fat contour lines; Generating a pericardial contour curved surface composed of a quadrilateral finite element network by lofting technology according to the weight of each curve control point on the plurality of epicardial fat contour lines, the direction variable of each pericardial cross-section and the direction variable of the perpendicular line.

3. The method according to any one of claims 1-2, characterized in that, The number of epicardial fat contour lines used for generating the pericardial contour curved surface is M, and the M is less than the total number of cross-section images contained in the coronary CT angiography image; The method for determining the epicardial fat parameter corresponding to the coronary CT angiography image by voxel analysis on the pericardial cavity three-dimensional model comprises the following steps: According to the pericardial cavity three-dimensional model, interpolating and calculating an epicardial fat contour line of a cross-section image of the coronary CT angiography image for which no epicardial fat contour line is obtained. filtering out voxels in the pericardial cavity three-dimensional model that do not belong to the fat tissue voxel CT value interval; based on the remaining voxels, analyzing to determine the epicardial fat parameters corresponding to the coronary CT angiography image.

4. The method of claim 3, wherein, The analysis based on the remaining voxels to determine the epicardial fat parameters corresponding to the coronary CT angiography image comprises: based on the remaining voxels, analyzing to obtain the volume and CT value characteristic parameters of the epicardial fat; the CT value characteristic parameters include at least one of the minimum CT value, the maximum CT value, the average CT value, the median CT value, or the CT value standard deviation; The CT value characteristic parameters are used as the density of the epicardial fat.

5. The method of claim 4, wherein, Further comprising: using the volume and density of the epicardial fat obtained from the coronary CT angiography image as a predictor, obtaining first prediction model performance improvement information for ejection fraction improving heart failure compared to a baseline model by multivariate regression analysis method, obtaining second prediction model performance improvement information for ejection fraction improving heart failure compared to a baseline model by multivariate regression analysis method, and obtaining third prediction model performance improvement information for ejection fraction improving heart failure compared to a baseline model by multivariate regression analysis method; wherein the first prediction model uses the volume of the epicardial fat as one of the predictors of the model; the second prediction model uses the density of the epicardial fat as one of the predictors of the model; the third prediction model uses the volume and density of the epicardial fat as two of the predictors of the model; and the baseline model does not use the volume and density of the epicardial fat as predictors of the model.

6. An apparatus for measuring a parameter of epicardial fat, characterized by comprising: an acquisition module configured to acquire cross-sectional profile point tracing data of a coronary CT angiography image to be determined; the coronary CT angiography image is a three-dimensional image, and the coronary CT angiography image includes a plurality of two-dimensional cross-sectional images; the cross-sectional profile point tracing data includes position information of cross-sectional profile points traced on the corresponding cross-sectional images; a curve fitting module configured to fit the cross-sectional profile point tracing data to obtain epicardial fat contour lines of the cross-sectional images corresponding to the cross-sectional profile point tracing data; a surface generation module configured to generate a pericardial contour surface passing through the plurality of epicardial fat contour lines by lofting technology based on the fitted plurality of epicardial fat contour lines; a three-dimensional model generation module configured to generate a pericardial cavity three-dimensional model based on the pericardial contour surface; a parameter determination module configured to determine epicardial fat parameters corresponding to the coronary CT angiography image by voxel analysis of the pericardial cavity three-dimensional model; the cross-sectional profile point tracing data includes position information of cross-sectional profile points traced on the corresponding cross-sectional images; the curve fitting module comprises: The first generating unit is configured to generate an epicardial fat initial contour line passing through all the cross-section contour points on the target cross-section image according to the position information of the cross-section contour points drawn on the target cross-section image. The determining unit is configured to determine a plurality of curve segments of the epicardial fat initial contour line according to the position information of the cross-section contour points drawn on the target cross-section image and the epicardial fat initial contour line. The second generating unit is configured to generate a plurality of curve control points according to the plurality of curve segments and the position information of the cross-section contour points drawn on the target cross-section image. The fitting unit is configured to fit a closed curve passing through all the cross-section contour points and all the curve control points on the target cross-section image as an epicardial fat contour line of the target cross-section image according to the position information of the plurality of control points and the position information of the cross-section contour points.

7. A computer readable storage medium characterized in that, The computer readable storage medium stores a computer program, and when the computer program is run by a processor, a parameter measurement method of epicardial fat as claimed in any one of claims 1-5 is implemented.

8. A processor, comprising: A computer program for running, which when run performs a parameter measurement method of epicardial fat as claimed in any one of claims 1-5.

Images