CN-122025154-A - Personalized heart model reconstruction method, device and medium with infarcted area

CN122025154ACN 122025154 ACN122025154 ACN 122025154ACN-122025154-A

Abstract

The application relates to the technical field of medical images, and provides a personalized heart model reconstruction method, device and medium with an infarct area. The method comprises the steps of preprocessing a cardiac magnetic resonance delay gadolinium enhanced CMR-LGE image of a patient to obtain a first high-resolution label image containing a ventricular cavity and infarcted tissues, constructing a heart finite element grid model based on high-precision heart cavity data of the patient, adjusting the direction of the first high-resolution label image to obtain a second high-resolution label image, and mapping an infarcted area in the second high-resolution label image into the heart finite element grid model to obtain a personalized heart model with the infarcted area. Based on the scheme of the application, the accuracy of personalized heart model reconstruction with the infarcted area can be improved.

Inventors

WANG ZEFENG
DENG DONGDONG
DONG RUIQING
CHENG LITING
Liu Huanfu
XIA LING
WU YONGQUAN

Assignees

首都医科大学附属北京安贞医院

Dates

Publication Date: 20260512
Application Date: 20251212

Claims (10)

1. A method for reconstructing a personalized cardiac model with an infarcted area, comprising the steps of: preprocessing a cardiac magnetic resonance delay gadolinium enhanced CMR-LGE image of a patient to obtain a first high-resolution label image containing a ventricular cavity and infarcted tissue, and constructing a cardiac finite element grid model based on high-precision cardiac cavity data of the patient; Performing direction adjustment on the first high-resolution tag image to obtain a second high-resolution tag image; And mapping the infarcted region in the second high-resolution label image into the heart finite element grid model to obtain a personalized heart model with the infarcted region.
2. The method of claim 1, wherein preprocessing the CMR-LGE image of the patient to obtain a first high resolution label image comprising the ventricular cavity and infarcted tissue, comprises: dividing a CMR-LGE image of a patient to obtain a ventricular muscle image and an infarct tissue image, wherein the infarct tissue image comprises a full infarct tissue region and a half infarct tissue region; performing interpolation reconstruction based on the ventricular muscle image to obtain a ventricular muscle high-resolution label image; Performing interpolation reconstruction based on the infarcted tissue image to obtain an infarcted tissue high-resolution label image; combining the ventricular muscle high-resolution labeled image with the infarcted tissue high-resolution labeled image to generate a first high-resolution labeled image comprising a ventricular cavity and infarcted tissue.
3. The method of claim 1, wherein the constructing a cardiac finite element mesh model based on the patient's high-precision cardiac cavity data comprises: Determining a heart model contour based on the high accuracy heart cavity data; And carrying out surface mesh division, mesh quality optimization and volume mesh division on the outline of the heart model to generate a heart finite element mesh model, wherein the heart finite element mesh model is stored as a node coordinate file for recording node coordinates and a mesh topology file for recording vertex indexes of tetrahedron units.
4. The method of claim 1, wherein the performing the direction adjustment on the first high-resolution label image to obtain a second high-resolution label image comprises: Identifying that the apex of the heart in the first high-resolution label image points to the direction of the bottom of the heart, and carrying out direction adjustment on the first high-resolution label image to change the direction of the apex of the heart to the direction of the bottom of the heart into the positive direction of the z-axis; identifying the direction of the left ventricle in the first high-resolution label image to the right ventricle, and carrying out direction adjustment on the first high-resolution label image to change the direction of the left ventricle to the right ventricle to the negative direction of the x-axis; Identifying that a rear wall in the first high-resolution tag image points to the direction of a front wall, and carrying out direction adjustment on the first high-resolution tag image, wherein the direction of the rear wall points to the direction of the front wall is changed into the direction of the negative y axis; and generating a second high-resolution label image consistent with the coordinate system of the heart finite element grid model according to the first high-resolution label image after the direction is adjusted.
5. The method of claim 4, wherein the identifying that the apex in the first high resolution label image is pointing in the bottom of the heart direction, and the performing a direction adjustment on the first high resolution label image to change the apex to the bottom of the heart direction to the positive z-axis direction comprises: Determining each two-dimensional slice along the z-axis direction in the first high-resolution label image as a layer, wherein the layer with the largest z-coordinate value is the topmost layer, and the layer with the smallest z-coordinate value is the bottommost layer; respectively counting the number of myocardial pixels contained in the topmost layer and the bottommost layer, identifying a layer with a large number of myocardial pixels as a cardiac bottom layer, and identifying a layer with a small number of myocardial pixels as a cardiac apex layer; comparing the z coordinate values of the bottom and top layers to determine the direction of the apex toward the bottom; if the apex pointing direction indicates that the apex layer is located above the bottom layer, the first high-resolution label image is flipped along the z-axis so that the apex pointing direction is changed to point in the positive z-axis direction.
6. The method of claim 4, wherein the identifying that the left ventricle in the first high resolution label image is pointing in the right ventricle direction, and the performing a direction adjustment on the first high resolution label image to change the left ventricle to point in the right ventricle direction to point in the negative x-axis direction, comprises: Dividing a non-myocardial region in the first high-resolution label image into a left ventricle cavity, a right ventricle cavity and an image background according to a connectivity rule, wherein the circularity of the left ventricle cavity at a heart bottom layer is closer to 1 than that of the right ventricle cavity at the heart bottom layer; determining a vector R between the center of gravity of the left ventricular chamber and the center of gravity of the right ventricular chamber; comparing the angle between the vector R and the negative x-axis direction to determine the left ventricular direction to the right ventricular direction; and if the left ventricle points to the right ventricle direction and does not point to the negative x-axis direction, rotating and/or turning the first high-resolution label image so that the left ventricle points to the right ventricle direction instead of pointing to the negative x-axis direction.
7. The method of claim 4, wherein the identifying that the back wall in the first high resolution label image is pointing in the front wall direction, the first high resolution label image being directionally adjusted, the back wall pointing in the front wall direction instead pointing in the y-axis negative direction, comprises: Respectively acquiring the barycentric coordinates of the right ventricle cavity of the bottom layer, the 1/4 height layer at the bottom side of the heart and the 1/2 height layer at the bottom side of the heart in the first high-resolution label image; Analyzing the y coordinate change trend of the gravity center of the right ventricle cavity when the heart is changed from the bottom of the heart to the apex of the heart, and determining the direction of the rear wall to the front wall based on the y coordinate change trend; And if the direction of the rear wall pointing to the front wall is not pointing to the negative y-axis direction, the first high-resolution label image is turned upside down, so that the direction of the rear wall pointing to the front wall is changed to be pointing to the negative y-axis direction.
8. The method of claim 1, wherein said mapping the infarcted region in the second high resolution label image into the heart finite element mesh model results in a personalized heart model with the infarcted region, comprising: determining, for each tetrahedral mesh in the cardiac finite element mesh model, a mesh centroid of the tetrahedral mesh; Performing coordinate conversion and downward rounding on the grid gravity center according to the image resolution of the second high-resolution tag image to obtain a pixel coordinate corresponding to the grid gravity center in the second high-resolution tag image; Acquiring a corresponding first label value in the second high-resolution label image according to the pixel coordinates under the condition that the pixel coordinates belong to a myocardial tissue area, and endowing the first label value to the tetrahedral mesh; under the condition that the pixel coordinates belong to a non-myocardial area, sequentially searching non-zero pixels in a first-order neighborhood to a third-order neighborhood of the pixel coordinates, and endowing the tetrahedral mesh with a second label value of the first non-zero pixel; A personalized heart model with the infarcted region is generated based on all tetrahedral meshes assigned tag values in the heart finite element mesh model.
9. An electronic device comprising a processor and a memory storing a computer program, characterized in that the processor implements the steps of the personalized heart model reconstruction method with infarcted zone of any one of claims 1 to 8 when said computer program is executed.
10. A non-transitory computer readable storage medium, having stored thereon a computer program, characterized in that the computer program, when executed by a processor, implements the steps of the personalized heart model reconstruction method with infarcted zone of any of claims 1 to 8.

Description

Personalized heart model reconstruction method, device and medium with infarcted area Technical Field The invention relates to the technical field of medical images, in particular to a personalized heart model reconstruction method, device and medium with an infarct area. Background The personalized heart model with infarcted areas plays an important role in electrophysiological simulation of heart diseases, risk stratification and optimization of therapeutic regimens. LGE-CMR imaging techniques provide a data basis for the construction of this model. However, in practice, the geometric form and spatial position of the infarcted region of the heart model constructed based on clinical image data are often difficult to ensure complete consistency with the real anatomical structure, and a certain risk of precision loss exists. The deviation in the accuracy of the model can directly influence the reliability of the subsequent electrophysiological simulation result, and the effectiveness of the model in guiding clinical decisions and researches can be weakened. Therefore, how to improve the accuracy of personalized heart model reconstruction with an infarcted area becomes a technical problem to be solved urgently. Disclosure of Invention The embodiment of the application provides a personalized heart model reconstruction method with an infarcted area, equipment and a medium, which are used for solving the technical problem of how to improve the accuracy of the personalized heart model reconstruction with the infarcted area. In a first aspect, an embodiment of the present application provides a method for reconstructing a personalized cardiac model with an infarct zone, including: preprocessing a CMR-LGE image of a patient to obtain a first high-resolution label image containing a ventricular cavity and infarcted tissue, and constructing a heart finite element grid model based on high-precision heart cavity data of the patient; Performing direction adjustment on the first high-resolution label image to obtain a second high-resolution label image; And mapping the infarcted region in the second high-resolution label image into a heart finite element grid model to obtain a personalized heart model with the infarcted region. With reference to the first aspect, in some possible implementations, preprocessing the CMR-LGE image of the patient to obtain a first high resolution label image containing the ventricular cavity and infarcted tissue includes: Dividing a CMR-LGE image of a patient to obtain a ventricular muscle image and an infarct tissue image, wherein the infarct tissue image comprises a full infarct tissue region and a half infarct tissue region; interpolation reconstruction is carried out based on the ventricular muscle image, and a ventricular muscle high-resolution label image is obtained; performing interpolation reconstruction based on the infarcted tissue image to obtain an infarcted tissue high-resolution label image; combining the ventricular muscle high-resolution labeled image with the infarcted tissue high-resolution labeled image to generate a first high-resolution labeled image comprising the ventricular cavity and the infarcted tissue. With reference to the first aspect and the foregoing implementation manner, in some possible implementation manners, constructing a cardiac finite element mesh model based on high-precision cardiac cavity data of a patient includes: determining a heart model contour based on the high accuracy heart cavity data; And carrying out surface meshing, mesh quality optimization and volume meshing on the outline of the heart model to generate a heart finite element mesh model, wherein the heart finite element mesh model is stored as a node coordinate file for recording node coordinates and a mesh topology file for recording vertex indexes of tetrahedron units. With reference to the first aspect and the foregoing implementation manner, in some possible implementation manners, performing direction adjustment on the first high-resolution label image to obtain a second high-resolution label image, including: Identifying that the apex of the heart in the first high-resolution label image points to the direction of the bottom of the heart, and carrying out direction adjustment on the first high-resolution label image to change the direction of the apex of the heart to the direction of the bottom of the heart into the positive direction of the z-axis; Identifying the direction of the left ventricle in the first high-resolution label image to the right ventricle, and carrying out direction adjustment on the first high-resolution label image to change the direction of the left ventricle to the right ventricle into the direction of the negative x-axis; identifying that a rear wall in the first high-resolution label image points to the direction of a front wall, and carrying out direction adjustment on the first high-resolution label image, wherein the direction of the rear wall poin