A method for calculating and analyzing attribute features of biological tissue materials

Abstract

The application provides a biological tissue material attribute feature calculation and analysis method, comprising the following steps: S1, diffusion tensor image preprocessing; S2, calculating the diffusion tensor of a region of interest; S3, converting the diffusion tensor into a permeation tensor, converting the diffusion characteristics and the diffusion tensor of each voxel, and obtaining the permeation characteristics and the permeation coefficient of the scanned tissue. The method provided by the application establishes a method for obtaining the second-order permeation tensor of each voxel and commonly used permeation characteristic parameters from the diffusion tensor image, realizes the material permeability quantification of the biological tissue, is suitable for diffusion image feature calculation of multiple parts, has strong universality, can expand the application field of diffusion tensor imaging, and gets rid of the function limitation of the existing software which is limited to nerve fiber tracking.

Description

Technical Field

[0001] This invention belongs to the field of materials permeability technology, specifically relating to a method for calculating and analyzing the material properties of biological tissues. Background Technology

[0002] Permeability is the ability of a material to allow fluid to pass through without damaging the structure of the medium; it is a fundamental property of materials. The permeability coefficient is an important parameter describing permeability, reflecting the ease with which a fluid passes through a medium in different directions. Water is the main component of human body fluids and the most abundant substance in the human body; therefore, osmosis is prevalent in many organs, such as brain white matter, myocardium, liver, and intestines. Diffusion tensor imaging (DTI) is a special magnetic resonance imaging (MRI) technique. Its basic principle is to track the Brownian motion of free water molecules in the body. By applying diffusion-sensitive gradients in multiple directions, the diffusion coefficient of each voxel is obtained, thereby measuring the degree and directionality of water molecule diffusion. The diffusion coefficient is an indicator characterizing the strength of the diffusion phenomenon (the diffusion coefficient is a second-order tensor, referred to as diffusion tensor D for ease of explanation in subsequent technical descriptions). DTI is currently the only non-invasive examination method that can effectively observe and track brain white matter fiber tracts, and it is widely used in brain development and cognitive function research, examination of brain pathological changes, and surgical evaluation.

[0003] Currently, the application of Dispersion Imaging (DTI) is concentrated in neurological diseases, and image post-processing is limited to fiber tracing technology, restricting the scope of DTI's application in the biomedical field. The diffuseability of water molecules corresponds to the permeability of biological tissues. The diffusion coefficient and the permeability coefficient in anisotropic media are both second-order tensors with similar properties. The permeability coefficient (a second-order tensor, hereinafter referred to as the permeability tensor) is a fundamental property of materials and can be widely used in numerical simulation of human physiological processes, the development of biomaterials such as biomembranes, medical materials, and drug therapy research. Converting diffusion characteristics into permeability characteristics can fully utilize the non-invasive detection feature of DTI and expand its application areas; however, no research has yet proposed a method for obtaining permeability characteristics from DTI images.

[0004] Therefore, to address the aforementioned issues, it is necessary to propose a method and analysis system for calculating the features of diffusion tensor images, thereby expanding the application areas of diffusion tensor imaging. Summary of the Invention

[0005] The main objective of this invention is to propose a specific technical solution for acquiring diffusion tensor image features and to establish an analysis system for converting diffusion tensor features into permeability features. This method is not limited by diffusion tensor image parameters or scanning locations, and can extract material permeability characteristics from diffusion tensor images, thus expanding the application field of diffusion tensor imaging.

[0006] A method for calculating and analyzing the property characteristics of biological tissue materials includes the following steps:

[0007] S1. Preprocessing of Diffusion Tensor Image

[0008] The raw data obtained by the magnetic resonance imaging (MRI) device is converted into NIFTI format, and the diffusion gradient and diffusion sensitivity coefficient in each direction are extracted. The diffusion gradient is rearranged and a direction number is added before storage.

[0009] S2. Calculate the diffusion tensor of the region of interest.

[0010] Draw the region of interest and create a mask;

[0011] The image signal intensity after applying diffusion gradient pulses in different directions is related to the signal intensity before applying the diffusion gradient pulses, the diffusion sensitivity coefficient, the diffusion gradient applied in each direction, and the diffusion tensor. Except for the diffusion tensor that needs to be calculated in this module, which is related to the properties of the scanned tissue, all other data are obtained from the image, and the diffusion tensor is calculated based on this relationship.

[0012] By solving the system of equations for dispersion rates in different directions, the second-order dispersion tensor matrix at that location is obtained; by traversing each voxel within the region of interest, the dispersion tensor of each voxel is obtained.

[0013] Furthermore, in S2, the image signal intensity S after applying diffusion gradient pulses in different directions i The image signal intensity S0 before the diffusion gradient pulse is applied, the diffusion sensitivity coefficient b, and the applied diffusion gradient field g. i =[x i y i The diffusion tensor is related to the diffusion tensor D; the diffusion tensor is a 3×3 symmetric positive definite matrix containing 6 unknown parameters, corresponding to the diffusion rates in 6 directions. The direction i of the diffusion gradient g of the scan must be greater than 6; the diffusion tensor image at a certain voxel diffusion signal S i Relationship with S0:

[0014]

[0015] T represents matrix g i Transpose of;

[0016] Solving this relation yields the diffusion tensor of the voxel location. Automatically traversing all voxels within the region of interest and calculating the diffusion tensor yields the diffusion tensor within the region of interest.

[0017] The diffusion tensor data in the same direction are organized and summarized, and after data mapping, a 3×3 image matrix is ​​obtained. Each image position and data corresponds to a different diffusion rate in a different direction. The difference in diffusion rate in each direction of the output image matrix is ​​calibrated by adding the diffusion gradient g direction of the scan. This method can reduce the impact of image noise on the calculation of the diffusion tensor.

[0018] S3, Conversion of diffusion tensor to penetration tensor

[0019] First, some characteristic parameters are calculated based on the diffusion tensor to replace the diffusion tensor in representing the diffusion characteristics. These include the eigenvalues ​​of the diffusion tensor, the eigenvectors corresponding to the principal eigenvalues, the radial diffusion rate, and the axial diffusion rate.

[0020] The aforementioned dispersion characteristics are transformed into permeability characteristics. Radial dispersion rate corresponds to the transverse permeability of the scanned tissue, and axial dispersion rate corresponds to the longitudinal permeability of the scanned tissue. The permeability coefficient, as a fundamental property of materials, is represented by the permeability tensor in anisotropic materials.

[0021] Based on the conversion relationship, the diffusion characteristics and diffusion tensors of each voxel are converted to obtain the permeation characteristics and permeation coefficient of the scanned tissue.

[0022] In S3, the obtained diffusion tensor is used to calculate the image diffusion features, including: the tensor eigenvalues ​​of the image in the three principal directions, the eigenvectors corresponding to the principal eigenvalues, the radial diffusion rate, and the axial diffusion rate. The eigenvalue of the image matrix is ​​λ. i Let i = 1, 2, 3, and assume λ1 > λ2 > λ3; the principal eigenvector corresponding to the principal eigenvalue λ1 is v1; the relationship for calculating the above parameters is as follows:

[0023] Radial dispersion:

[0024] Axial dispersion ratio: AD = λ1.

[0025] S3 also includes establishing an analysis system for obtaining the permeation tensor from the characteristic parameters of the diffusion tensor, specifically:

[0026] Establish a formula for calculating the penetration tensor from the characteristics of the diffusion tensor:

[0027] K =RD· I 3+AD·v1×v1

[0028] in I 3 represents a 3×3 identity matrix;

[0029] Based on the above calculation formula, a 3×3 second-order tensor corresponding to each voxel is obtained, namely the permeability tensor, to describe the permeability strength of each voxel.

[0030] The method proposed in this invention establishes a method for obtaining the second-order diffusion tensor values ​​of each voxel and commonly used diffusion feature parameters from diffusion tensor images, realizing the numerical quantification of the degree of water molecule diffusion in all directions. It is applicable to the calculation of diffusion tensor image features in multiple parts, has strong universality, can expand the application field of diffusion tensor imaging, and get rid of the functional limitations of existing software that are limited to fiber tracking.

[0031] This invention establishes an analysis system that utilizes the aforementioned image feature calculation method to transform the obtained diffusion features into permeability features. This analysis system acquires lateral permeability, longitudinal permeability, and permeability coefficient from diffusion tensor images, achieving non-invasive quantification of the permeability characteristics of scanned tissues. It can be used for numerical simulation of organs such as the brain, liver, and heart; simulation and prediction of drug treatment effects; and the development of medical biomembrane materials. It is of great significance to research fields such as biology and medicine where permeability coefficients are difficult to measure experimentally and where high data accuracy and patient specificity are required. Attached Figure Description

[0032] Figure 1 This is a flowchart of the present invention;

[0033] Figure 2 This is a detailed flowchart of the present invention. Detailed Implementation

[0034] The embodiments of the present invention are described in detail below. These embodiments are for illustrative purposes only, and the present invention can be implemented in various different ways and application fields as defined and covered by the claims.

[0035] A method for calculating and analyzing the properties of biological tissue materials, which extracts the diffusion tensor values ​​and image features of each voxel from a diffusion tensor image, includes the following steps:

[0036] Step 1: Use magnetic resonance imaging (MRI) to acquire diffusion tensor images of the human heart, taking cardiac diffusion tensor as an example;

[0037] Step 2: Preprocess the acquired cardiac diffusion tensor images, including data format conversion, acquisition of necessary scan parameters, and format preparation. The specific methods are as follows:

[0038] Data format conversion: Converting raw DICOM (Digital Imaging and Communications in Medicine) format images acquired by MRI scans into NIFTI (Neuroimaging Informatics Technology Initiative) format, which is commonly used in neuroimaging research, to facilitate data extraction, processing, and analysis.

[0039] Obtain scan parameters: Obtain the diffusion gradient matrix and diffusion sensitivity coefficient from the NIFTI format header file. The size of the diffusion gradient matrix and the diffusion sensitivity coefficient are related to the magnetic resonance imaging scan settings.

[0040] Data formatting: Add direction numbers to the diffusion gradient matrix to facilitate subsequent calculation of diffusion rates in different directions.

[0041] Step 3: For cardiac diffusion tensor imaging, the anisotropy of the myocardium is most significant. This embodiment uses the steps for obtaining the myocardial diffusion tensor as an example. The calculation methods for other organs are the same as in the embodiment. The specific steps are as follows:

[0042] The region containing the myocardium is selected as the region of interest (ROI), and a mask is created to obtain the image and signal data of the ROI. Calculating only the ROI reduces computational load and improves computational efficiency.

[0043] Image signal intensity S after applying diffuse gradient pulses in different directions i The image signal intensity S0 before the diffusion gradient pulse is applied, the diffusion sensitivity coefficient b, and the applied diffusion gradient field g. i =[x i y i The diffusion tensor (D) is related to the diffusion tensor. The diffusion tensor is a 3×3 symmetric positive definite matrix containing 6 unknown parameters, corresponding to the diffusion rates in 6 directions. Therefore, the direction i of the diffusion gradient g in the scan must be greater than 6. The diffusion tensor image at a certain voxel diffusion signal S... i Relationship with S0:

[0044]

[0045] Solving this relation yields the diffusion tensor of the voxel position. By automatically traversing all voxels within the region of interest and calculating their diffusion tensors, the diffusion tensors of each voxel within the region of interest can be obtained.

[0046] The diffusion tensor data in the same direction are organized and summarized, and after data mapping, a 3×3 image matrix is ​​obtained. Each image position and data corresponds to a different diffusion rate in a different direction. The difference in diffusion rate in each direction of the output image matrix is ​​calibrated by adding the diffusion gradient g direction of the scan. This method can reduce the impact of image noise on the calculation of the diffusion tensor.

[0047] Step 4: Calculate the diffusion features of the image using the diffusion tensor obtained by the above method. Analyze the extracted diffusion tensor image features, including calculating diffusion tensor feature parameters, obtaining the permeation tensor of the scanned tissue from the diffusion tensor features, and analyzing the permeation characteristics of the tissue material. This invention analyzes the material properties of the scanned tissue, extracts permeation features from the diffusion tensor image, and establishes a unique diffusion tensor image post-processing analysis system.

[0048] This includes: calculating the tensor eigenvalues ​​of the image in the three principal directions using the diffusion tensor, the eigenvectors corresponding to the principal eigenvalues, and the radial and axial diffusion rates. The eigenvalues ​​of the image matrix are λ. i Let i = 1, 2, 3 and assume λ1 > λ2 > λ3; let v1 be the principal eigenvector corresponding to the principal eigenvalue λ1; the relationship for calculating the above parameters is as follows:

[0049] Radial diffusivity (RD):

[0050] Axial diffusivity (AD): AD = λ1

[0051] Step 5: Convert the diffusion feature into a penetration feature. In the penetration feature, the penetration rate perpendicular to the direction of the scanned tissue layering is called the longitudinal penetration rate, which corresponds to the axial diffusion rate of the diffusion feature. The penetration rate along the direction of the scanned tissue layering is called the transverse penetration rate, which is usually greater than the longitudinal penetration rate and corresponds to the radial diffusion rate of the diffusion feature.

[0052] Step 6: Since the microscopic dispersion of water molecules and the macroscopic flow of fluids are subject to the same physiological constraints, the dispersion tensor and the osmotic tensor have the same eigenvalues ​​and eigenvectors. Utilizing the correlation of these tensors, an analytical system is established to obtain the osmotic tensor from the characteristic parameters of the dispersion tensor. The specific steps are as follows:

[0053] Establish a formula for calculating the penetration tensor from the characteristics of the diffusion tensor:

[0054] K =RD· I 3+AD·v1×v1

[0055] in I 3 is a 3×3 identity matrix.

[0056] Based on the above calculation formula, a 3×3 second-order tensor corresponding to each voxel can be obtained, namely the permeability tensor, to describe the permeability strength of each voxel.

[0057] At this point, the calculation and analysis of the diffusion tensor image features have been completed.

[0058] The present invention establishes an analysis system for obtaining the permeability tensor, i.e. the permeability coefficient of a material, from diffusion tensor imaging. This system can non-invasively acquire the permeability characteristics of various scanned tissues, providing specific quantitative data for organ numerical simulation, medical material research and development, drug permeation process treatment simulation, and other aspects with precise quantification.

Claims

1. A method of biological tissue material property feature calculation and analysis, characterized by, The method comprises the following steps: S1, pre-processing of the diffusion tensor image The raw data format obtained by the magnetic resonance device scanning is converted into NIFTI format, and the diffusion gradient and diffusion sensitivity coefficient of each direction are extracted. After the diffusion gradient is rearranged and the direction number is added, it is stored; S2, calculating the diffusion tensor of the region of interest Draw the region of interest and create a mask; Image signal intensity after applying diffusion gradient pulses in different directions Image signal intensity before applying diffusion gradient pulses , diffusion sensitivity coefficient b, applied diffusion gradient field and diffusion tensor D; the diffusion tensor is a 3x3 symmetric positive definite matrix, containing 6 unknown parameters, corresponding to 6 directional diffusivities, when the number of scanned diffusion gradient g directions i is greater than 6, the diffusion tensor image in a certain voxel and relationship: ， T denotes the transpose of the matrix g i The voxel position diffusion tensor is obtained by solving the equation, and the diffusion tensor in the region of interest is obtained by automatically traversing all voxels in the region of interest and calculating the diffusion tensor. After mapping the diffusion tensor of the same direction, a 3x3 image matrix is obtained. Each image position and data corresponds to a different direction of diffusion rate. The output image matrix is used to calculate the difference calibration of each direction of diffusion rate, and the diffusion gradient g direction of the scanning is increased; S3, converting the diffusion tensor into a permeation tensor First, calculate the partial characteristic parameters of the diffusion tensor to replace the diffusion tensor to represent the diffusion characteristics, including the eigenvalue of the diffusion tensor, the eigenvector corresponding to the principal eigenvalue, the radial diffusion rate and the axial diffusion rate; Convert the above diffusion characteristics into permeation characteristics. The radial diffusion rate corresponds to the transverse permeability of the scanned tissue, and the axial diffusion rate corresponds to the longitudinal permeability of the scanned tissue. The permeability coefficient is a basic property of the material, and the permeation tensor is used to represent the anisotropic material; According to the conversion relationship, the diffusion characteristics and the diffusion tensor of each voxel are converted to obtain the permeation characteristics and the permeability coefficient of the scanned tissue.

2. The method according to claim 1, wherein, In S3, the diffusion tensor calculation image is obtained to calculate the diffusion features, including: the tensor eigenvalue of the diffusion tensor calculation image in three principal directions, the eigenvector corresponding to the principal eigenvalue, the radial diffusion rate and the axial diffusion rate; the image matrix eigenvalue is and assuming ; the principal eigenvalue corresponding to the principal eigenvector is ; the parameter calculation relationship is as follows: Radial dispersion: , Axial dispersion: .

3. The method according to claim 2, wherein, S3 also includes establishing an analysis system between the diffusion tensor characteristic parameters and the permeation tensor, specifically: Establishing a formula for permeability tensor from diffusion tensor characteristics of the formula: = ， wherein is a 3 x 3 identity matrix; According to the above calculation formula, a 3x3 second-order tensor, i.e. a permeation tensor, is obtained for each voxel to describe the permeability characteristics of each voxel.

Images