CN-121981986-A - Method and system for predicting aneurysm wall enhancement region based on hemodynamic parameters

CN121981986ACN 121981986 ACN121981986 ACN 121981986ACN-121981986-A

Abstract

The invention discloses a method and a system for predicting an aneurysm wall enhancement region based on hemodynamic parameters, wherein the method comprises the steps of acquiring TOF-MRA images and carrying out segmentation optimization, processing triangular patch grids, extracting hemodynamic parameters, determining the aneurysm wall enhancement state matched with each spatial node on the surface of an aneurysm and the prediction characteristics affecting AWE distribution, constructing a mapping relation model of the hemodynamic parameters and the AWE, calculating a corresponding aneurysm wall enhancement state value M, analyzing the aneurysm wall enhancement state value M matched with each spatial node on the surface of the predicted aneurysm to obtain a point-level enhancement prediction distribution map, detecting the connected region and the calculation area of the point-level enhancement prediction distribution map, and determining the spatial distribution of the aneurysm wall enhancement state and the effective enhancement region and outputting the spatial distribution. According to the invention, by analyzing the association between the hemodynamic parameters and the aneurysm wall enhancement, the high-efficiency prediction of the aneurysm wall enhancement is realized, and the accuracy of clinical diagnosis is ensured.

Inventors

MEI YUQIAN
Pang Zhangyu
FENG XIN
HUANG CHI
MA JINGTAO
DUAN CHUANZHI
Yan Linhan

Assignees

川北医学院

Dates

Publication Date: 20260505
Application Date: 20260120

Claims (10)

1. A method of predicting an enhanced region of an aneurysm wall based on hemodynamic parameters, comprising: Acquiring a TOF-MRA image and performing segmentation optimization to obtain a triangular patch grid; processing the triangular patch grids, and extracting hemodynamic parameters of corresponding space nodes at the geometric centers of the triangular patches in the triangular patch grids; The method comprises the steps of determining a tumor wall enhancement state matched with each spatial node on the surface of an aneurysm and a prediction feature affecting AWE distribution, and completing the construction of a mapping relation model of hemodynamic parameters and AWE according to the tumor wall enhancement state matched with each spatial node on the surface of the aneurysm and the prediction feature; calculating a corresponding tumor wall enhancement state value M according to the hemodynamic parameters and an AWE mapping relation model; Analyzing the corresponding tumor wall enhancement state value M to obtain a tumor wall enhancement state and a point level enhancement prediction distribution map matched with each spatial node on the surface of the predicted aneurysm; Analyzing the area of the communication area to determine the enhanced state of the aneurysm wall and the spatial distribution of the effective enhanced area; the type of aneurysm wall and the spatial distribution of the effective enhancement region are output.
2. The method of predicting an enhanced region of an aneurysm wall based on hemodynamic parameters of claim 1, wherein the acquiring TOF-MRA images and performing segmentation optimization results in a triangular patch grid, comprising: Acquiring TOF-MRA sequences of a patient to obtain continuous two-dimensional tomographic images; dividing the two-dimensional tomographic image by using preset medical image dividing software according to a preset dividing algorithm, and reconstructing the two-dimensional tomographic image into a high-resolution three-dimensional model containing the aneurysm and the aneurysm-carrying artery thereof; And carrying out smoothing and grid optimization on the high-resolution three-dimensional model, and deriving a triangular patch grid with a preset format.
3. The method of predicting an enhanced region of an aneurysm wall based on hemodynamic parameters of claim 1, wherein processing the triangular patch grid to extract hemodynamic parameters of corresponding spatial nodes at a geometric center of each triangular patch in the triangular patch grid comprises: importing the triangular patch grid into preset CFD preprocessing software, reserving a Willis ring geometry structure, and creating a multilayer prism boundary layer; A preset non-Newtonian fluid viscosity model is selected to describe the blood viscosity characteristics and input typical parameters of human blood; numerical simulation is performed by combining the multi-layer prismatic boundary layer with a preset non-Newtonian fluid viscosity model, and one or more hemodynamic parameters are extracted from the numerical simulation.
4. The method of predicting an aneurysm wall enhancement region based on hemodynamic parameters of claim 1, wherein determining a matching aneurysm wall enhancement state for each spatial node of the surface of the aneurysm and affecting the predicted features of the AWE profile comprises: synchronously acquiring the enhanced T1-VWI sequences of all patients in the training set, and acquiring signal intensity values SI a of all voxels in the region; Accurately positioning a pituitary stalk region, and obtaining signal intensity values SI b of all voxels in the pituitary stalk region; The mean SI c of all voxel signal intensity values in the pituitary stalk region is calculated: ; Wherein n is the number of all voxels in the pituitary stalk region, i is the ith voxel in the pituitary stalk region; The contrast ratio CR of the signal intensity value SI a for each voxel in the region to the average value SI c of the signal intensity values for all voxels in the pituitary stalk region is calculated: ; when CR is larger than or equal to a preset contrast ratio, judging that the enhancement state of the tumor wall matched with each spatial node on the surface of the aneurysm in the voxel region is surface patch enhancement; when CR is smaller than a preset contrast ratio, judging that the enhancement state of the tumor wall matched with each spatial node on the surface of the aneurysm in the voxel region is a patch non-enhancement state; The method comprises the steps of collecting hemodynamic parameters of each patient in a training set, establishing a point-to-point space mapping relation between the hemodynamic parameters and the characteristics of the aneurysm wall enhancement signals in an MRI image based on the topological structure of triangular patch grids, calculating the consistency or the correlation degree of the spatial distribution of each hemodynamic parameter and the aneurysm wall enhancement signal distribution by adopting a preset statistical analysis method, identifying the hemodynamic parameters which meet the preset obvious statistical correlation with the enhancement signal distribution, and determining the hemodynamic parameters as the prediction characteristics of a prediction model.
5. The method for predicting an aneurysm wall enhancement region based on hemodynamic parameters of claim 1, wherein the constructing of the model of the mapping relationship between hemodynamic parameters and AWE is accomplished according to the aneurysm wall enhancement state and the prediction features matched by each spatial node of the surface of the aneurysm, comprising: taking the prediction characteristic as an independent variable X, taking the tumor wall enhancement state matched with each spatial node on the surface of the aneurysm as an independent variable, and calculating the average value of single independent variables And standard deviation : ; ; Wherein w is the total number of independent variables in a single class, and e is the e-th independent variable; z-score normalization of independent variable X to X bn : ; Constructing a mapping relation model of hemodynamic parameters and AWE, wherein the model is in the form of: z=β 0 +β 1 ·X b1 +β 2 ·X b2 +...+β n ·X bn ; Wherein z is a linear predictive value, β 0 is an intercept term, and β 1 to β n are regression coefficients of the standardized variables; After construction, model parameters including intercept and regression coefficients of each normalized variable are saved, including mean μ and standard deviation The parameters are normalized for each variable included.
6. The method of predicting an aneurysm wall enhancement region based on hemodynamic parameters of claim 5, wherein the calculating a corresponding tumor wall enhancement state value M from the hemodynamic parameters and a model of a mapping relationship of the hemodynamic parameters to AWE comprises: for each spatial node, performing Z-score normalization on each hemodynamic parameter according to each variable normalization parameter of the hemodynamic parameter and AWE mapping relation model to obtain a normalization parameter X an : ; Wherein X cn is the nth hemodynamic parameter; Substituting the standardized parameters into a mapping relation model of the hemodynamic parameters and AWE, and calculating to obtain a corresponding tumor wall enhancement state value M: M=β 0 +β 1 ·X a1 +β 2 ·X a2 +...+β n ·X an 。
7. The method according to claim 2, wherein analyzing the corresponding tumor wall enhancement state value M to obtain a predicted tumor wall enhancement state and point-level enhancement prediction profile for each spatial node match of the surface of the aneurysm comprises: Converting the corresponding tumor wall enhancement state value M into enhancement probability P through a Sigmoid function: P=1/(1+exp(-M)); Setting a probability threshold, wherein when P is larger than or equal to the probability threshold, the prediction result of the space node is enhancement, and the prediction enhancement state of the triangular patch corresponding to the space node is determined to be enhancement; And mapping the prediction result of each triangular patch back to the three-dimensional model to generate a point-level enhanced prediction distribution map of the surface of the aneurysm.
8. The method of predicting an aneurysm wall enhancement region based on hemodynamic parameters of claim 1, wherein detecting spatially connected regions of the point-level enhancement prediction profile and calculating connected region areas comprises: Based on the point-level enhancement prediction distribution diagram, establishing an adjacent relation between triangular patches, and if two triangular patches share one edge, confirming the two triangular patches as adjacent enhancement patches; detecting the connected areas of triangular patches with the predicted enhancement state being enhanced, and identifying all the connected areas consisting of adjacent enhancement patches by adopting a preset algorithm; and accumulating all triangular patch areas of each communication area to obtain the area of the communication area.
9. The method of predicting an aneurysm wall enhancement region based on hemodynamic parameters of claim 1, wherein analyzing the connected region area to determine a spatial distribution of an aneurysm wall enhancement state and an effective enhancement region comprises: Setting a communication area threshold, judging the communication area as an effective enhancement area when the area of the communication area is larger than or equal to the communication area threshold, and recording the space distribution of the effective enhancement area; if an effective enhancement area exists on the surface of the aneurysm, judging that the type of the wall of the aneurysm is a wall enhanced aneurysm; if there is no effective enhancement region on the surface of the aneurysm, the type of the wall of the aneurysm is determined to be a non-wall enhanced aneurysm.
10. A system for predicting an enhanced region of an aneurysm wall based on a hemodynamic parameter, the method for implementing the predicting an enhanced region of an aneurysm wall based on a hemodynamic parameter according to any one of claims 1-9, comprising: the image acquisition optimization module is used for acquiring TOF-MRA images and carrying out segmentation optimization to obtain triangular patch grids; the parameter extraction module is used for processing the triangular patch grids and extracting hemodynamic parameters of space nodes corresponding to the geometric centers of the triangular patches in the triangular patch grids; The model construction module is used for determining a tumor wall enhancement state matched with each spatial node on the surface of the aneurysm and a prediction feature affecting AWE distribution, and completing the construction of a mapping relation model of hemodynamic parameters and AWE according to the tumor wall enhancement state matched with each spatial node on the surface of the aneurysm and the prediction feature; the model prediction module is used for calculating a corresponding tumor wall enhancement state value M according to the hemodynamic parameters and a mapping relation model of the hemodynamic parameters and AWE; the connected region analysis module is used for analyzing the corresponding tumor wall enhancement state value M to obtain a tumor wall enhancement state and a point level enhancement prediction distribution map matched with each spatial node on the surface of the predicted aneurysm; The result judging module is used for analyzing the area of the communication area and determining the reinforcing state of the aneurysm wall and the spatial distribution of the effective reinforcing area; and the result output module outputs the type of the aneurysm wall and the spatial distribution of the effective enhancement area.

Description

Method and system for predicting aneurysm wall enhancement region based on hemodynamic parameters Technical Field The present application relates to the field of aneurysm wall enhancement prediction, and more particularly to a method and system for predicting an aneurysm wall enhancement region based on hemodynamic parameters. Background With the rapid development of medical imaging technology and hemodynamic analysis technology, the importance of aneurysm wall enhanced region prediction as a core component of intracranial aneurysm risk assessment is increasingly prominent. However, traditional AWE assessment methods mostly rely on enhanced MRI scanning and manual interpretation analysis, which makes it difficult to circumvent clinical risks of contrast agent use and individual application limitations. Especially when facing the situations of renal insufficiency patients, contrast media allergy crowd and the situation of needing prospective risk assessment, the traditional method often has difficulty in combining the safety and the accuracy of diagnosis. Therefore, the prior art has defects, and improvement is needed. Disclosure of Invention In view of the above problems, an object of the present invention is to provide a method and a system for predicting an aneurysm wall enhancement region based on hemodynamic parameters, which are capable of realizing accurate prediction of the aneurysm wall enhancement region by analyzing the association between hemodynamic parameters and aneurysm wall enhancement and constructing a quantitative mapping relation model, thereby ensuring the safety and the foresight of evaluation in cerebrovascular disease diagnosis. The present invention provides in a first aspect a method of predicting an enhanced region of an aneurysm wall based on hemodynamic parameters, comprising: Acquiring a TOF-MRA image and performing segmentation optimization to obtain a triangular patch grid; processing the triangular patch grids, and extracting hemodynamic parameters of corresponding space nodes at the geometric centers of the triangular patches in the triangular patch grids; The method comprises the steps of determining a tumor wall enhancement state matched with each spatial node on the surface of an aneurysm and a prediction feature affecting AWE distribution, and completing the construction of a mapping relation model of hemodynamic parameters and AWE according to the tumor wall enhancement state matched with each spatial node on the surface of the aneurysm and the prediction feature; calculating a corresponding tumor wall enhancement state value M according to the hemodynamic parameters and an AWE mapping relation model; Analyzing the corresponding tumor wall enhancement state value M to obtain a tumor wall enhancement state and a point level enhancement prediction distribution map matched with each spatial node on the surface of the predicted aneurysm; Analyzing the area of the communication area to determine the enhanced state of the aneurysm wall and the spatial distribution of the effective enhanced area; the type of aneurysm wall and the spatial distribution of the effective enhancement region are output. In this scheme, the acquiring TOF-MRA images and performing segmentation optimization to obtain triangular patch grids includes: Acquiring TOF-MRA sequences of a patient to obtain continuous two-dimensional tomographic images; dividing the two-dimensional tomographic image by using preset medical image dividing software according to a preset dividing algorithm, and reconstructing the two-dimensional tomographic image into a high-resolution three-dimensional model containing the aneurysm and the aneurysm-carrying artery thereof; And carrying out smoothing and grid optimization on the high-resolution three-dimensional model, and deriving a triangular patch grid with a preset format. In this scheme, the processing the triangular patch grid, extracting hemodynamic parameters of space nodes corresponding to geometric centers of triangular patches in the triangular patch grid, includes: importing the triangular patch grid into preset CFD preprocessing software, reserving a Willis ring geometry structure, and creating a multilayer prism boundary layer; A preset non-Newtonian fluid viscosity model is selected to describe the blood viscosity characteristics and input typical parameters of human blood; numerical simulation is performed by combining the multi-layer prismatic boundary layer with a preset non-Newtonian fluid viscosity model, and one or more hemodynamic parameters are extracted from the numerical simulation. In this solution, the determining the enhancement state of the tumor wall matched with each spatial node on the surface of the aneurysm and the prediction feature affecting AWE distribution includes: synchronously acquiring the enhanced T1-VWI sequences of all patients in the training set, and acquiring signal intensity values SI a of all voxels in the region; Accurately posi