Method for evaluating stability of wall of intracranial aneurysm

Abstract

The invention provides a new method for evaluating the stability of the wall of the intracranial aneurysm, the TOF-MRA and 4D Flow MRI data detected by HR-VWI are fused by using 3DSlicer software, and the relationship between the vulnerability of the wall of the aneurysm and the hemodynamic characteristics is explored by calculating the time average wall shear force (TAWSS) and the oscillation shear index (OSI) through accurate modeling and hemodynamic analysis and applying spatial autocorrelation analysis. According to the method, the stability of the wall of the intracranial aneurysm is comprehensively and accurately evaluated, a scientific basis is provided for clinical diagnosis and treatment, and the management effect of the intracranial aneurysm and the prognosis of a patient are improved.

Description

Technical Field

[0001] The present invention relates to the field of medical image processing, and in particular to a method and a process for evaluating the stability of an intracranial aneurysm wall. Background Art

[0002] Intracranial aneurysms are abnormal bulges of the intracranial arterial wall, and their rupture risk is closely related to the stability of the aneurysm wall. Currently, the methods for evaluating the stability of intracranial aneurysm walls mostly rely on traditional imaging assessments, but this method often cannot accurately reflect the relationship between the local hemodynamic characteristics of the aneurysm wall and the stability of the aneurysm wall. Therefore, it is of great significance to develop a method that can integrate three-dimensional imaging and hemodynamic characteristics to accurately evaluate the stability of intracranial aneurysm walls.

[0003] In order to solve the above problems, the applicant proposes a method for evaluating the stability of the aneurysm wall of an intracranial aneurysm. Summary of the invention

[0004] The purpose of the present invention is to provide a method for evaluating the stability of the aneurysm wall of an intracranial aneurysm, so as to solve the problems in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: a method for evaluating the stability of an intracranial aneurysm wall, comprising the following steps:

[0006] Step 1: Obtain 4D Flow MRI and HR-VWI imaging data of patients with unruptured intracranial aneurysms;

[0007] Step 2: Apply three-dimensional spatial coupling technology to spatially register 4D Flow MRI and HR-VWI image data, and perform three-dimensional quantitative analysis of the hemodynamic parameters and aneurysm wall vulnerability in the aneurysm cavity in a unified spatial coordinate system;

[0008] Step 3: Based on the registered image data, calculate the hemodynamic parameters of each spatial point in the tumor cavity, including flow velocity, flow rate, wall shear force, and oscillation shear index, and extract the high-resolution enhancement information and thickness of the tumor wall;

[0009] Step 4: In a unified spatial coordinate system, construct a spatial atlas of the hemodynamics and aneurysm wall vulnerability of the intracranial aneurysm, which includes the hemodynamic parameter vectors of each spatial point in the aneurysm cavity and the corresponding quantitative index of the aneurysm wall vulnerability;

[0010] Step 5: Use bivariate spatial autocorrelation analysis to explore the correlation between the three-dimensional spatial law of hemodynamics in the tumor cavity and the vulnerability of the tumor wall at the corresponding coordinate points, and analyze the spatial correspondence between the hemodynamic characteristic domain and the tumor wall damaged domain.

[0011] Optionally, the hemodynamic parameters include time-averaged wall shear force, wall shear stress gradient, wall shear stress divergence and pressure gradient.

[0012] Optionally, the quantitative indicators of aneurysm wall vulnerability include wall enhancement index, aneurysm-pituitary stalk enhancement contrast ratio and wall enhancement rate.

[0013] Optionally, the spatial registration in step 2 includes using 3D Slicer software to spatially register the TOF-MRA sequence detected by HR-VWI with the 4D Flow MRI sequence, and performing three-dimensional modeling of the aneurysm and the parent artery through cerebral vascular segmentation and reconstruction.

[0014] Optionally, the construction of the spatial map in step 4 includes combining the hemodynamic parameters and aneurysm wall vulnerability quantitative indicators corresponding to each coordinate point of the aneurysm wall into a high-dimensional vector to form a spatial coordinate-hemodynamic-aneurysm wall vulnerability spatial map.

[0015] Optionally, it also includes applying a three-dimensional convolutional deep neural network model to train the constructed hemodynamics and aneurysm wall vulnerability spatial map to predict the stability and growth of the aneurysm.

[0016] Optionally, the three-dimensional modeling model includes an input layer, a convolution layer, a pooling layer, an upsampling layer and an output layer, wherein the input layer receives three-dimensional tensor data of the spatial map, and the output layer performs a binary classification prediction task and a three-dimensional tensor output task.

[0017] Optionally, it also includes regular clinical and imaging follow-up of patients with unruptured intracranial aneurysms based on a prospective cohort study, and validating and optimizing the predictive efficacy of the three-dimensional modeling model according to the follow-up results.

[0018] Optionally, the prospective cohort study includes enrolling patients with unruptured intracranial aneurysms who meet specific inclusion and exclusion criteria, and performing regular 4D Flow MRI and HR-VWI examinations on them to obtain dynamic changes in hemodynamics and aneurysm wall vulnerability data.

[0019] Optionally, the inclusion criteria include patient age, aneurysm morphological characteristics, and completeness of clinical and imaging baseline data; the exclusion criteria include surgical history, concomitant other vascular diseases, and contraindications to magnetic resonance imaging.

[0020] Beneficial effects: The present invention can accurately evaluate the stability of the aneurysm wall of intracranial aneurysms through spatial autocorrelation analysis and three-dimensional positioning coupling technology, comprehensively considers the three-dimensional imaging and hemodynamic characteristics, and provides a more scientific and comprehensive basis for the diagnosis and treatment of intracranial aneurysms. DETAILED DESCRIPTION

[0021] The following reference description introduces the preferred embodiments of the present invention to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms of embodiments, and the protection scope of the present invention is not limited to the embodiments mentioned in the text.

[0022] Example 1

[0023] Intracranial aneurysm (IA) is one of the most common causes of hemorrhagic stroke, and the complexity of its treatment and management poses a huge challenge to the medical community. Currently, there is a lack of non-invasive imaging technology to accurately assess the stability of IA, which leads to a dilemma for doctors when deciding whether to perform surgical intervention. This study proposed an innovative method to map the spatial atlas of IA hemodynamics and aneurysm wall vulnerability for the first time by combining 4D Flow MRI (four-dimensional flow magnetic resonance imaging) and high-resolution magnetic resonance vascular wall imaging (HR-VWI) technology, and constructed a three-dimensional convolutional deep neural network model in order to more accurately assess the stability of IA and predict its rupture risk.

[0024] The rupture of IA is often accompanied by extremely high mortality and disability rates, and the existing management methods make it difficult to make the best decision between conservative observation and surgical treatment. Studies have shown that the formation, progression and rupture of IA are jointly affected by the hemodynamic state and inflammatory damage of the tumor wall. However, traditional research methods are limited to hemodynamic analysis under static boundary conditions, ignoring the spatial heterogeneity of the hemodynamic environment and the three-dimensional coupling relationship with the vulnerability of the tumor wall.

[0025] The purpose of the present invention is to provide a method based on 4D Flow MRI and HR-VWI technology, by drawing a three-dimensional spatial map of IA hemodynamics and tumor wall vulnerability, constructing and optimizing a three-dimensional convolutional deep neural network model, so as to achieve accurate assessment of IA stability and rupture risk.

[0026] Based on the multicenter prospective cohort study (ClinicalTrial: NCT06308952; ChiCTR: 2400082984) previously participated in, a retrospective nested cohort study method was adopted. The inclusion and exclusion criteria of the research subjects were as follows:

[0027] Age ≥18 years;

[0028] The maximum diameter of saccular aneurysm is ≥4 mm / the maximum transverse diameter of fusiform aneurysm is ≥4 mm;

[0029] Perform blood tests on the same day as the imaging examination;

[0030] The clinical and imaging baseline and follow-up data were complete.

[0031] Exclusion criteria:

[0032] history of surgical clipping or endovascular treatment;

[0033] Combined with arteriovenous malformation, dural arteriovenous fistula, moyamoya disease, dissecting aneurysm, etc.

[0034] Patients with a history of acute or chronic infectious inflammation, autoimmune disease, heart disease, blood disease, or cancer;

[0035] Patients were lost to follow-up;

[0036] Contraindications to MRI (eg, pregnancy, breastfeeding, or presence of MRI-incompatible implants) and / or poor MRI image quality.

[0037] 4D Flow MRI Data Acquisition

[0038] Using advanced 3.0T MRI equipment, 4D Flow MRI scans were performed on patients with unruptured IA to obtain real-time IA hemodynamic characteristics during the cardiac cycle, including parameters such as flow velocity, flow, volume flow, turbulence, and wall shear force.

[0039] HR-VWI data collection

[0040] In the same MRI scan, patients underwent HR-VWI scanning to obtain high-resolution aneurysm wall enhancement (AWE) and aneurysm wall thickness (AWT) information.

[0041] 3D space mapping

[0042] Image Registration and Modeling

[0043] 3D Slicer software was used to spatially register the TOF-MRA sequence with the 4D Flow MRI sequence to exclude confounding signal shadows, and the three-dimensional model of the aneurysm and the parent artery was reconstructed by cerebral vascular segmentation.

[0044] Calculation of hemodynamic parameters

[0045] Based on 4D Flow MRI data, the hemodynamic parameters of each spatial coordinate point of the tumor wall were calculated, including wall shear stress (WSS), time-averaged wall shear stress (TAWSS), oscillatory shear index (OSI), etc.

[0046] HR-VWI data were used to map the aneurysm wall enhancement signal onto the aneurysm 3D model using the minimum Euclidean distance method. The vulnerability quantitative index of each coordinate point of the aneurysm wall was obtained by calculating the ratio of the aneurysm wall signal intensity to the pituitary stalk signal intensity (CRstalk).

[0047] Bivariate spatial autocorrelation analysis

[0048] Global spatial autocorrelation analysis

[0049] The global bivariate Moran'I index was calculated to explore the correlation between hemodynamic parameters and tumor wall vulnerability in the global spatial distribution.

[0050] Local spatial autocorrelation analysis

[0051] The inverse Euclidean distance and Gaussian kernel function spatial weight matrices were constructed, and the local bivariate Moran'I index was calculated to analyze the spatial correlation between the local hemodynamic environment and the vulnerability of the tumor wall.

[0052] The hemodynamic-tumor wall vulnerability spatial atlas data were preprocessed into a unified standard three-dimensional tensor as input data.

[0053] The 3D U-Net CNN model was used to construct a binary classification prediction task (stable / unstable IA) and a 3D tensor output task (predicting the quantitative index of tumor wall vulnerability). The Adam optimization algorithm was used for model training.

[0054] The model performance was verified by F1 score and AUC-ROC of K-fold cross validation and compared with the hemodynamic analysis model based on traditional static boundary conditions.

[0055] Calculate the required sample size according to the statistical formula, clarify the inclusion and exclusion criteria, and ensure the homogeneity of the research subjects and the reliability of the data.

[0056] The model is validated based on prospective cohort data, and the model architecture is adjusted based on the prediction results to optimize model performance.

[0057] Specific implementation steps

[0058] Data collection and processing

[0059] Equipment preparation: Select advanced medical equipment with 3.0T MRI scanning function to ensure that the image resolution and scanning speed meet the research needs.

[0060] Patient preparation: The patient should be fully prepared before scanning, including body positioning, breathing training, etc., to reduce motion artifacts.

[0061] Scan parameter setting: According to the MRI equipment manual and research requirements, set appropriate scanning parameters, including layer thickness, TR, TE, FOV, etc.

[0062] Image reconstruction: Use image processing software such as 3D Slicer to reconstruct and preprocess the original image to obtain a clear three-dimensional model of the aneurysm and the parent artery.

[0063] Hemodynamics and quantification of aneurysm wall vulnerability

[0064] Calculation of hemodynamic parameters: Based on 4D Flow MRI data, the WSS, TAWSS, OSI and other parameters of each coordinate point on the tumor wall are calculated using a fluid dynamics algorithm.

[0065] Quantification of tumor wall vulnerability: Using HR-VWI data, CRstalk was calculated by the signal intensity ratio method to obtain the quantitative vulnerability index of each coordinate point of the tumor wall.

[0066] Spatial Mapping

[0067] Spatial registration: The hemodynamic model is spatially registered with the aneurysm 3D model to ensure the spatial consistency of the data.

[0068] Data fusion: Fusion of hemodynamic parameters with quantitative indicators of tumor wall vulnerability to construct a three-dimensional spatial map.

[0069] Bivariate spatial autocorrelation analysis

[0070] Global analysis: The global bivariate Moran'I index was calculated to evaluate the correlation between hemodynamic parameters and aneurysm wall vulnerability in the global space.

[0071] Local analysis: construct a spatial weight matrix, calculate the local bivariate Moran'I index, and analyze local spatial correlation.

[0072] Model building and training

[0073] Data preprocessing: Convert the spatial atlas data into a unified standard three-dimensional tensor as input data.

[0074] Model building: Use a deep learning framework such as TensorFlow or PyTorch to build a 3D U-Net CNN model.

[0075] Model training: Use the Adam optimization algorithm to train the model until convergence.

[0076] Model validation and optimization

[0077] Cross-validation: K-fold cross-validation is used to evaluate model performance and ensure the generalization ability of the model on different datasets.

[0078] Prospective validation: The model is validated based on prospective cohort data, and the model architecture is adjusted based on the prediction results to optimize model performance.

[0079] This paper proposes an innovative method based on 4D Flow MRI and HR-VWI technology. By drawing a three-dimensional spatial map of IA hemodynamics and aneurysm wall vulnerability and constructing a three-dimensional convolutional deep neural network model, it achieves an accurate assessment of IA stability and rupture risk. This method not only overcomes the limitations of traditional research methods, but also provides a scientific basis for the individualized treatment of IA. In the future, with the continuous development of technology and the continuous accumulation of data, this model is expected to be widely used in clinical practice, bringing better treatment effects and quality of life to patients.

[0080] Through the implementation of this invention, we look forward to publishing high-quality scientific papers in the future, sharing research results, and providing new ideas and methods for the medical community. At the same time, we will continue to devote ourselves to research in this field, constantly explore new technologies and methods, and contribute to the precise treatment of IA. Conclusion: The method of the present invention achieves accurate assessment of the stability of the aneurysm wall of intracranial aneurysms through the fusion of multimodal MRI data and advanced image processing technology. This method has the advantages of being non-invasive, accurate, and comprehensive, and is of great significance for improving the diagnosis and treatment of intracranial aneurysms.

[0081] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the attached claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any mark in the claims should not be regarded as limiting the claims involved.

[0082] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. A method for evaluating the stability of an intracranial aneurysm wall, characterized in that: The following steps are involved: Step 1: Obtain 4D Flow MRI and HR-VWI imaging data of patients with unruptured intracranial aneurysms; Step 2: Apply three-dimensional spatial coupling technology to spatially register 4D Flow MRI and HR-VWI image data, and perform three-dimensional quantitative analysis of the hemodynamic parameters and aneurysm wall vulnerability in the aneurysm cavity in a unified spatial coordinate system; Step 3: Based on the registered image data, the hemodynamic parameters of each spatial point in the tumor cavity are calculated, including flow velocity, flow rate, wall shear force, and oscillatory shear index, and the high-resolution enhancement information and thickness of the tumor wall are extracted; Step 4: In a unified spatial coordinate system, construct a spatial atlas of the hemodynamics and wall vulnerability of the intracranial aneurysm, which includes the hemodynamic parameter vectors of each spatial point in the aneurysm cavity and the corresponding quantitative index of the wall vulnerability; Step 5: Use bivariate spatial autocorrelation analysis to explore the correlation between the three-dimensional spatial law of hemodynamics in the tumor cavity and the vulnerability of the tumor wall at the corresponding coordinate points, and analyze the spatial correspondence between the hemodynamic characteristic domain and the tumor wall damaged domain.

2. The method according to claim 1, characterized in that The hemodynamic parameters include time-averaged wall shear force, wall shear stress gradient, wall shear stress divergence and pressure gradient.

3. The method according to claim 1, characterized in that The quantitative indicators of aneurysm wall vulnerability include wall enhancement index, aneurysm-pituitary stalk enhancement contrast ratio and wall enhancement rate.

4. The method according to claim 1, characterized in that: The spatial registration in step 2 includes using 3DSlicer software to spatially register the TOF-MRA sequence detected by HR-VWI with the 4D Flow MRI sequence, and performing three-dimensional modeling of the aneurysm and the parent artery through cerebral vascular segmentation and reconstruction.

5. The method according to claim 1, characterized in that The construction of the spatial map in step 4 includes combining the hemodynamic parameters and the aneurysm wall vulnerability quantitative indicators corresponding to each coordinate point of the aneurysm wall into a high-dimensional vector to form a spatial coordinate-hemodynamic-aneurysm wall vulnerability spatial map.

6. The method according to claim 1, characterized in that It also includes the application of a three-dimensional convolutional deep neural network model to train the constructed hemodynamics and aneurysm wall vulnerability spatial map to predict the stability and growth of aneurysms.

7. The method according to claim 6, characterized in that The three-dimensional modeling model includes an input layer, a convolution layer, a pooling layer, an upsampling layer and an output layer, wherein the input layer receives three-dimensional tensor data of the spatial atlas, and the output layer performs a binary classification prediction task and a three-dimensional tensor output task.

8. The method according to claim 1, characterized in that It also includes regular clinical and imaging follow-up of patients with unruptured intracranial aneurysms based on a prospective cohort study, and validating and optimizing the predictive efficacy of the three-dimensional modeling model based on the follow-up results.

9. The method according to claim 8, characterized in that The prospective cohort study included patients with unruptured intracranial aneurysms who met specific inclusion and exclusion criteria and underwent regular 4D Flow MRI and HR-VWI examinations to obtain data on dynamic changes in hemodynamics and aneurysm wall vulnerability.

10. The method according to claim 9, characterized in that The inclusion criteria included patient age, aneurysm morphological characteristics, and completeness of clinical and imaging baseline data; the exclusion criteria included surgical history, concomitant other vascular diseases, and contraindications to magnetic resonance imaging.