Meniscus tear mapping and typing method, system and equipment

Abstract

The invention discloses a method, a system and equipment for mapping and typing meniscus tearing, and relates to the technical field of meniscus. The method comprises the following steps: acquiring biaxial two-dimensional knee joint MRI images of a meniscus tearing patient; according to different signal intensities of different tissues in the knee joint, respectively generating an integral mask and a background tissue mask of each axial two-dimensional knee joint MRI image; according to the integral mask and the background tissue mask of each axial two-dimensional knee joint MRI image, obtaining a meniscus two-dimensional mask of each axial direction by adopting a subtraction method; performing three-dimensional reconstruction of the meniscus according to the meniscus two-dimensional mask in each axial direction to obtain a meniscus three-dimensional model in each axial direction; fusing the biaxial meniscus three-dimensional model to obtain a meniscus three-dimensional digital model; according to the meniscus three-dimensional digital model, mapping and quantitative parting are conducted on meniscus tearing. The accuracy and the sensitivity of meniscus tearing typing can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of meniscus, and particularly relates to a method, system and device for meniscus tear mapping and typing. BACKGROUND

[0002] Meniscus is a crescent-shaped fibrocartilage tissue within the joint capsule between femoral condyle and tibial condyle. In sports, especially in high-impact sports, it absorbs compression and shear forces, coordinates cartilage lubrication and maintains joint mobility. Meniscus injury is a common orthopedic diagnosis, affecting about 0.6% of the population, with an incidence of more than 50% in patients with combined anterior cruciate ligament tear and up to 75% in patients with osteoarthritis. Existing surgical treatments include suturing, meniscectomy and allograft transplantation, etc. These options depend largely on the geometric characteristics (such as location, depth, edge width) of the tear and the quality of the meniscus.

[0003] At present, magnetic resonance imaging (MRI) provides cross-sectional details of meniscus tears in a non-surgical manner, but traditional two-dimensional (2D) MRI often underestimates complex tear types (such as radial, horizontal or bucket handle tears) due to its limited ability to analyze multi-planar spatial relationships. More than 65% of lateral meniscus posterior horn tears are missed in ordinary MRI. Secondly, different doctors may have different evaluations of the tear due to experience, cognitive fatigue and time pressure. SUMMARY

[0004] The purpose of the present application is to provide a method, system and device for meniscus tear mapping and typing, which can improve the accuracy and sensitivity of meniscus tear typing.

[0005] To achieve the above-mentioned purpose, the present application provides the following solutions: In a first aspect, the present application provides a method for meniscus tear mapping and typing, comprising: acquiring two-dimensional knee joint MRI images of a patient with meniscus tear in double axial directions; generating a whole mask and a background tissue mask of each axial two-dimensional knee joint MRI image according to different signal intensities of different tissues in the knee joint; the tissues in the knee joint other than the meniscus are background tissues; obtaining a meniscus two-dimensional mask of each axial direction by using subtraction method according to the whole mask and the background tissue mask of each axial two-dimensional knee joint MRI image; performing three-dimensional reconstruction of the meniscus according to the meniscus two-dimensional mask of each axial direction to obtain a meniscus three-dimensional model of each axial direction; fusing the meniscus three-dimensional models of the double axial directions to obtain a meniscus three-dimensional digital model; and mapping and quantitatively typing the meniscus tear according to the meniscus three-dimensional digital model.

[0006] In a second aspect, the application provides a system for meniscus tear mapping and typing, comprising: an MRI image acquisition module, a mask generation module, a subtraction module, a three-dimensional reconstruction module, a fusion module, and a tear mapping module.

[0007] The MRI image acquisition module is configured to acquire two-dimensional knee MRI images of a patient with meniscus tear in a double axial direction. The mask generation module is configured to generate a whole mask and a background tissue mask of the two-dimensional knee MRI image in each axial direction according to different signal intensities of different tissues in the knee joint. The subtraction module is configured to obtain a meniscus two-dimensional mask in each axial direction by using a subtraction method according to the whole mask and the background tissue mask of the two-dimensional knee MRI image in each axial direction. The three-dimensional reconstruction module is configured to perform three-dimensional reconstruction of the meniscus according to the meniscus two-dimensional mask in each axial direction to obtain a meniscus three-dimensional model in each axial direction. The fusion module is configured to fuse the meniscus three-dimensional models in the double axial directions to obtain a meniscus three-dimensional digital model. The tear mapping module is configured to map and quantitatively type the meniscus tear according to the meniscus three-dimensional digital model.

[0008] In a third aspect, the application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor executes the computer program to implement the method for meniscus tear mapping and typing.

[0009] According to the specific embodiments provided by the application, the application has the following technical effects: The application provides a method, system, and device for meniscus tear mapping and typing, which obtains a meniscus two-dimensional mask in each axial direction by using a subtraction method, fuses meniscus three-dimensional models in a double axial direction after three-dimensional reconstruction of the meniscus, and obtains a meniscus three-dimensional digital model to provide a three-dimensional perspective of the tear and directly observe the meniscus tear for objective and quantitative analysis of the position and size of the tear. Compared with the traditional two-dimensional MRI for meniscus tear typing and the subjective evaluation of the tear by a physician, the application realizes high-precision and high-sensitivity typing of the meniscus tear. BRIEF DESCRIPTION OF DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the application or the related art, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0011] Figure 1 A flowchart illustrating a method for mapping and classifying meniscus tears provided in this application embodiment; Figure 2 This is a schematic diagram illustrating the definition of the meniscus tear location provided in an embodiment of this application; Figure 3 A schematic diagram illustrating the process of establishing a 3D digital model of the meniscus using subtraction and biaxial surface model fusion as provided in this embodiment of the application; Figure 4 This is a schematic diagram of the tear quantitative typing results provided in Example 1 of this application; Figure 5 This is a schematic diagram of the quantitative tear typing results provided in Example 2 of this application; Figure 6 This is a schematic diagram of the quantitative tear typing results provided in Example 3 of this application; Figure 7 A schematic diagram comparing the quantitative typing results of conventional two-dimensional MRI technology and the method of this application in an embodiment of this application; Figure 8 A schematic diagram of the functional modules of a system for mapping and classifying meniscus tears provided in this application embodiment; Figure 9 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0014] In one exemplary embodiment, such as Figure 1 As shown, a method for mapping and classifying meniscus tears is provided, including steps 101 to 106. Wherein: Step 101: Acquire two-dimensional MRI images of the knee joint in a patient with a meniscus tear in both axes. Exemplarily, the two axes include the coronal axis and the sagittal axis.

[0015] Step 102: generating the whole mask and background tissue mask of each axial two-dimensional knee MRI image according to different signal intensity of different tissues in knee joint; the tissue other than meniscus in knee joint is background tissue.

[0016] Step 103: obtaining the meniscus two-dimensional mask of each axial according to the whole mask and background tissue mask of each axial two-dimensional knee MRI image by using subtraction method.

[0017] Step 104: three-dimensional reconstruction of meniscus according to the meniscus two-dimensional mask of each axial to obtain the meniscus three-dimensional model of each axial.

[0018] Step 105: fusing the meniscus three-dimensional model of two axial to obtain the meniscus three-dimensional digital model.

[0019] Step 106: mapping and quantitative typing of meniscus tear according to the meniscus three-dimensional digital model.

[0020] Implementing the above steps 101 to 106, using two axial subtraction method to construct meniscus digital model, thereby realizing direct observation and objective quantitative typing of meniscus tear, and improving the accuracy and sensitivity of typing.

[0021] The whole mask and background tissue mask of each axial two-dimensional knee MRI image generated in step 102, the meniscus two-dimensional mask of each axial obtained in step 103, and the meniscus three-dimensional digital model obtained in step 105 can be visualized.

[0022] In another exemplary embodiment of the present application, the clinical 2D-MRI contains signal encoding number set of tissues in frequency space (k-space), and the signal encoding number set of spatial domain image information of different tissues in knee joint is converted into image for displaying meniscus, cartilage, bone, ligament and other tissues. Then the above step 101 can be replaced by the following steps 201~step 202: Step 201: obtaining the signal encoding number set of meniscus tear patient in two axial; Step 202: obtaining the two-dimensional knee MRI image of each axial according to the signal encoding number set of each axial by using formula ; in the formula, represents image intensity of coordinate position, represents inverse Fourier transform, represents proton density, represents repetition time, represents echo time, represents signal recovery speed, represents signal decay speed, The constant representing hydrogen gas, , , The coordinates of the mark within the valid time period are: Gradient intensity of position-dependent phases, Indicates coordinates as Location integration into the residual noise and small artifacts in the real scan, Indicates the organizations involved. Indicates the effective gradient application duration in the corresponding k-space. =0.02s~0.1s, The value depends on the scan produced by the MRI scanner itself. (Subscript) The variable representing the organization type in the above formula is the summation symbol. Each voxel may contain multiple tissues (such as fat, muscle, cartilage, etc.), and each tissue has an undetermined proton density pt and longitudinal relaxation time T. 1，t and lateral relaxation time T 2，t Therefore, summation is actually superimposing the signal contributions of various tissues at the same voxel location.

[0023] This refers to the brightness seen in the image. The inverse Fourier transform is the mathematical step that converts raw measurements into a spatial image. ∑tissue sum means that multiple tissues within a voxel can contribute. Proton density represents how much water / fat a tissue contains. (The last phrase, "with tissue," appears to be unrelated and likely refers to another concept.) and Together, the brightness and darkness of the structure were set.

[0024] In another exemplary embodiment of this application, different tissues within the knee joint have different signal intensities, and the meniscus can be extracted from adjacent structures by using a threshold. A global mask covers the entire image, and the background tissue mask is a high-signal / density mask.

[0025] The overall mask for the two-dimensional knee MRI image of each axis is: ; In the formula, Indicates coordinates as binary mask of voxels The coordinates of the location image intensity dataset in a two-dimensional knee joint MRI image are: MRI intensity at voxels and These represent the lower and upper thresholds used to determine the voxel inclusion or exclusion intensity; The background tissue mask for each axis of the two-dimensional knee MRI image is: ; wherein, is a binary mask of background tissue voxels with coordinates , is the MRI intensity at background tissue voxels with coordinates , is the minimum threshold value.

[0026] In another exemplary embodiment of the present application, the high signal surrounding tissue is removed from the whole body mask according to the subtraction principle to obtain the meniscus mask. Then, the step 103 can be replaced by the following steps 301-302: Step 301: according to the whole body mask and the background tissue mask of each axial two-dimensional knee MRI image, the meniscus two-dimensional mask of each axial is obtained by using the formula , wherein, is a binary mask of meniscus voxels with coordinates .

[0027] Step 302: remove the noise and tissue fragments in the meniscus two-dimensional mask of each axial to obtain the final meniscus two-dimensional mask of each axial.

[0028] The discrete signal is regarded as an island region, and the fragments with a size exceeding 10 mm and being connected with each other are reserved to obtain a more accurate meniscus mask : ; is the connected component label of voxel , and 1{.} is an indicator function, wherein true is 1, and the element of its set is displayed. . is the label set to be reserved (the index of the component meeting the size rule). is the element belonging to or.

[0029] In another exemplary embodiment of the present application, the boundary surface is set to convert the two-dimensional MRI mask data of into 3D data for three-dimensional reconstruction: MRI is a cumulative of a series of 2D slices, and the following formula is suitable for reconstructing the MRI image into a 3D model. is 0, the boundary of the meniscus model is determined by the above . represents the equivalent intensity level (i.e., threshold value) for defining the surface position, and a global threshold value can be used, and the default value in the binary mask is 0.5. is the volume intensity after Gaussian smoothing filtering with an iteration number of 2 and a smoothing factor of 0.3: ; denotes the level set function, where 0 is the reconstructed meniscus surface, denotes the filtered MRI volume, is the iso-intensity used as the surface threshold.

[0030] In the actual reconstruction process, first the region of interest is defined, and then the boundary surface is extracted based on , and subsequently the position and signal intensity information of the boundary and internal voxels are assigned in 3D space, thereby obtaining a 3D meniscus model that has both geometric morphology and contains tissue properties.

[0031] The above step 104 can be replaced by the following steps 401-404: Step 401: Determine the boundary of the three-dimensional model of the meniscus according to the two-dimensional mask of the meniscus in each axial direction.

[0032] Step 402: Convert the two-dimensional mask of the meniscus in each axial direction to three-dimensional data based on the boundary of the three-dimensional model of the meniscus.

[0033] Step 403: Perform three-dimensional reconstruction using the three-dimensional data to obtain an initial three-dimensional model of the meniscus in each axial direction.

[0034] Step 404: Perform wrapping and smoothing processing on the initial three-dimensional model of the meniscus in each axial direction to obtain a three-dimensional model of the meniscus in each axial direction.

[0035] Perform wrapping and smoothing processing on the reconstructed model to reduce surface irregularities and artifacts: In a 3D model, smoothing the surface of the 3D model to create a more rounded or seamless appearance usually requires changing the position of the vertices; wrapping involves projecting a 2D image or texture onto the surface of a 3D model, that is, wrapping it around the object.

[0036] Wrapping: ; M denotes a binary three-dimensional mask, where 1 represents the interior, is a binary boundary operator (all surface points at the inner-outer interface), Close0.5 is the gap closing distance, and Open0.1 is the minimum detail.

[0037] Smoothing: ; denotes the iteration of the 3D position of the vertex vector, is the updated vertex with an iteration number of 100. This represents the smoothing factor, which is 1 in this embodiment. It is the number of vertices (the number of adjacent vertices directly connected in the grid). (Quantity). Represents the total number of all adjacent vertices of a vertex. and. A vector representing the position of a vertex from its current position to the positions of its neighboring vertices. . It is an iteration of 100, where = 0,1,....,99.

[0038] In another exemplary embodiment of this application, step 105 described above can be replaced by steps 501 to 502: Step 501: Align the biaxial 3D model of the meniscus.

[0039] The alignment formula is: ; This represents the original vertex position, i.e., the 3D coordinates of the vertex in the mesh generated from the first reconstruction. It is a rotation matrix that rotates the original coordinates to the target direction. It is the translation vector that moves the rotated model to the target position.

[0040] Step 502: Merge the aligned biaxial meniscus 3D models to obtain a 3D digital model of the meniscus.

[0041] The merging formula is: ; This represents the vertex coordinates of the sagittal reconstruction. And the vertex coordinates represent the coronal plane reconstruction.

[0042] In another exemplary embodiment of this application, step 106 described above can be replaced by steps 601 to 609: Step 601: Draw the arc between the front and rear foot endpoints in the three-dimensional digital model of the meniscus.

[0043] Step 602: Divide the arc equally from the front foot endpoint to the rear foot endpoint, and divide the three-dimensional digital model of the meniscus into the front segment, body segment and rear segment according to the equally divided arc.

[0044] like Figure 2 As shown, the front foot endpoint A H With the rear foot endpoint P H The arc between ( Figure 2The meniscus is divided into 8 equal parts by the red lines, with the center of the circle as C, and labeled as A1-A3 (anterior segment), M1-M3 (body segment), and P1-P3 (posterior segment). X represents the transverse axis, and Y represents the longitudinal axis along the direction of the meniscus. Therefore, X distinguishes the left and right sides of the body, and Y distinguishes the anterior and posterior directions of the body. Anterior refers to the direction of the arrow pointing to the front of the body.

[0045] Step 603: Define the meniscus region inside the circle as the inner meniscus or white zone, the meniscus region outside the circle as the outer meniscus or red zone, and the meniscus side as the peripheral region.

[0046] Step 604: Determine the location of the meniscus tear based on the segmentation of the meniscus three-dimensional digital model and the definition of the meniscus region.

[0047] Step 605: Define the shape of the meniscus tear along the direction of the circle as longitudinal tear, the shape perpendicular to the direction of the circle as radial tear, and the shape in the horizontal direction of the periphery as horizontal tear.

[0048] Step 606: Determine the shape of the meniscus tear in the meniscus three-dimensional digital model based on the defined shape of the meniscus tear.

[0049] Step 607: Define the extent of the meniscus tear; the extent of the meniscus tear includes the length defined along the direction of the circle, the width defined perpendicular to the direction of the circle, and the depth defined perpendicular to the direction of the meniscus cross-section.

[0050] When quantitatively described, such as a peripheral horizontal tear of 10mm x 1mm x 2mm; a lesion that forms >1 / 3 size in any dimension is defined as a severe tear.

[0051] Step 608: Determine the extent of the meniscus tear in the meniscus three-dimensional digital model based on the defined extent of the meniscus tear.

[0052] Step 609: Determine whether it constitutes a through tear based on whether the tear causes damage to at least two geometric faces of the meniscus.

[0053] The present application is the first to quantitatively classify tears based on a meniscus digital model, through technical innovations such as subtraction, and double-axis superposition of coronal and sagittal planes, to construct a digital model and map the tear, thereby quantitatively classifying the tear and improving accuracy.

[0054] The above formula is applied to practice using Materialise Interactive Medica Control System (Mimics) 21.0 and 3-matic Research 13.0 software to demonstrate meniscus digital model reconstruction and tear classification.

[0055] Step 1, Obtain patient clinical MRI images DICOM files, including routine knee examination sequences and preferred scan sequences: The routine knee examination sequences include: Turbo Spin Echo (TSE) sequence for obtaining Proton Density (PD) fat-saturated images in coronal, sagittal and axial planes. Next, use T1 -weighted sequence to image again in sagittal plane. MRI includes T1 -weighted images with short repetition time (TR) and echo time (TE), and proton density (PD) weighted images with long TR and short TE sequence.

[0056] Coronal plane: repetition time 3050 ms, echo time 42 ms, flip angle 150°, pixel bandwidth 170 Hz, matrix 320 x 320, 30 layers, layer thickness 3 mm, no gap between layers, voxel size 0.5 x 0.5 x 0.5 cubic mm.

[0057] Sagittal plane: repetition time 3140 ms, echo time 36 ms, flip angle 150°, pixel bandwidth 200 Hz, matrix 320 x 320, 30 layers, layer thickness 3 mm, no gap between layers, voxel size 0.5 x 0.5 x 0.5 cubic mm.

[0058] Step 2, processing of 2D MRI images, preparation for 3D model reconstruction. Each plane will be imported and processed individually according to the same procedure, including: Step 2.1, import routine sequence and preferred sequence scan data, establish MRI project for each sequence name; Step 2.2, generate and calculate . .

[0059] Step 3, establish 3D model of sagittal and coronal MRI images, each plane is processed individually according to the same procedure before merging, including: Step 3.1, convert processed 2D MRI meniscus to 3D model, filter effect can be customized in the conversion settings for creating 3D model, iteration is 2, smoothing coefficient is 0.3 to obtain the best quality: .

[0060] Step 3.2, wrapping processing of model surface, in Mimics, for 3D tools, wrapping can be set to minimum detail 0.1 (as shown in the formula Open 0.1 ) and gap closure distance of 0.5 (as shown in the formula Close 0.5 ): .

[0061] Step 3.3, Model surface smoothing, in Mimics, set smoothing to 100 iterations with a smoothing factor of 1, where 100 iterations is smoothing 100 times with a factor of 1, t represents each model smoothing. For each iteration, it changes the vertices of the model v ), v i (t+1) represents the addition of iterations for each smoothing t +1): .

[0062] Step 4, Global registration calculation. The original DCM format file contains the anatomical position of the meniscus in the knee joint, so here the automatic alignment is performed, merging the sagittal and coronal models.

[0063] Figure 3 The process of using subtraction method to fuse with biaxial surface model to establish 3D digital model of meniscus is shown.

[0064] Step 5, Quantitative classification of the tear is performed on the reconstructed digital model.

[0065] Example 1: As shown in Figure 4 , the right knee discoid meniscus anterior segment 2.42 mm x 0.15 mm longitudinal tear and petal formation. Figure 4 , the short horizontal line represents the ruler, indicating a length of 5 mm. A represents the forward direction of the body, P represents the back direction of the body, and R represents the right side of the body. The red star represents the position of the structural lesion that can be observed by the naked eye.

[0066] Example 2: As shown in Figure 5 , the left knee lateral discoid meniscus anterior segment to posterior segment through serious lamellar tear, medial meniscus peripheral 14 mm x 2.0 mm serious lamellar tear. Figure 5 , the black triangle represents the position of the structural lesion that can be observed by the naked eye on the model, and L represents the left side of the body.

[0067] Example 3: As shown in Figure 6 , the right knee lateral meniscus body red area 13.2 mm x 2.1 mm through longitudinal tear, peripheral posterior segment 13.4 mm x 1.59 mm lamellar tear, and anterior segment red area 16.2 x 2.3 mm longitudinal tear and petal formation.

[0068] The prior art classifies meniscus tears based on traditional two-dimensional MRI, and the present application is the first to classify and calculate tears based on a meniscus digital model. The technical solution mainly includes two innovations: first, the establishment of a subtraction method and the integration of the sagittal and coronal planes to realize the construction of a meniscus digital model; second, high-accuracy and quantitative classification of tears based on a meniscus digital model. For example Figure 7 Part (a) is a subjective qualitative judgment by traditional two-dimensional MRI technology, Figure 7 Part (b) is a quantitative classification based on an inner-outer meniscus digital model.

[0069] The technical solution of the present application mainly uses the technical points of target signal and background signal mask recognition and subtraction in the transformed spatial domain image in the third part, the merging of coronal and sagittal 3D models, and reference voxel coordinates and dimensions for tear classification and calculation to realize the classification and measurement of tears based on a meniscus 3D model. Compared with the traditional two-dimensional image method, the biggest advantage is to provide a three-dimensional perspective of the tear, directly show the meniscus tear, and objectively and quantitatively analyze the location and size of the tear, realize high-precision and high-sensitivity classification of meniscus tears, and is the first in the world.

[0070] Based on the same inventive concept, the present application also provides a meniscus tear mapping and classification system for implementing the above-mentioned meniscus tear mapping and classification method. The problem-solving implementation scheme provided by the system is similar to the implementation scheme described in the above method, so the specific limitations in one or more meniscus tear mapping and classification system embodiments provided below can refer to the limitations of the meniscus tear mapping and classification method described above, and will not be repeated here.

[0071] In one exemplary embodiment, as shown in Figure 8 a meniscus tear mapping and classification system is provided, which includes an MRI image acquisition module, a mask generation module, a subtraction module, a three-dimensional reconstruction module, a fusion module, and a tear mapping module.

[0072] The system includes the following modules: an MRI image acquisition module for acquiring biaxial two-dimensional knee MRI images of patients with meniscus tears; a mask generation module for generating an overall mask and a background tissue mask for each axis of the two-dimensional knee MRI image based on the different signal intensities of different tissues within the knee joint; the tissues outside the meniscus in the knee joint are considered background tissue; a subtraction module for obtaining a two-dimensional meniscus mask for each axis using a subtraction method based on the overall mask and background tissue mask of the two-dimensional knee MRI image for each axis; a three-dimensional reconstruction module for performing three-dimensional reconstruction of the meniscus based on the two-dimensional meniscus masks for each axis, obtaining a three-dimensional model of the meniscus for each axis; a fusion module for fusing the biaxial three-dimensional meniscus models to obtain a three-dimensional digital model of the meniscus; and a tear mapping module for mapping and quantitatively classifying the meniscus tear based on the three-dimensional digital model of the meniscus.

[0073] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 9 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores the results of meniscus tear mapping and quantitative classification. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for meniscus tear mapping and classification.

[0074] Those skilled in the art will understand that Figure 9 The structures shown are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0075] Any technical features in the above embodiments can be combined, and for the sake of brevity, not all possible combinations are described above, however, it should be understood that the application encompasses all possible combinations of the technical features unless such a combination is not technically possible.

[0076] The principles and implementations of the present application have been described in specific examples, and the above descriptions of the embodiments are only used to help understand the method and its core idea of the present application; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation and application range can be changed. In summary, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for mapping and classifying meniscus tears, characterized in that, include: Acquire two-dimensional MRI images of the knee joint in patients with meniscus tears in both axes; Based on the different signal intensities of different tissues within the knee joint, a global mask and a background tissue mask for a two-dimensional knee joint MRI image of each axis are generated; the tissues outside the meniscus in the knee joint are the background tissues. Based on the overall mask and background tissue mask of the two-dimensional knee MRI image for each axis, a two-dimensional mask of the meniscus for each axis is obtained by subtraction. Based on the two-dimensional mask of the meniscus in each axis, the three-dimensional reconstruction of the meniscus is performed to obtain the three-dimensional model of the meniscus in each axis. By integrating the biaxial 3D models of the meniscus, a 3D digital model of the meniscus is obtained. Based on the three-dimensional digital model of the meniscus, the meniscus tear was mapped and quantitatively classified.

2. The method for mapping and classifying meniscus tears according to claim 1, characterized in that, Acquire biaxial two-dimensional MRI images of the knee joint of patients with meniscus tears, specifically including: Acquire biaxial frequency-space signal encoding datasets from patients with meniscus tears; Based on the frequency spatial signal encoding set for each axis, the formula is used. Two-dimensional MRI images of the knee joint are obtained for each axis; where, Indicates coordinates as Image intensity at location, This represents the inverse Fourier transform. Represents proton density, Indicates the repetition time. Indicates echo time. Indicates the signal recovery speed. Indicates the signal fading rate. The constant representing hydrogen gas, , , The coordinates of the mark within the valid time period are: Gradient intensity of position-dependent phases, Indicates coordinates as Location integration into the residual noise and small artifacts in the real scan, Indicates the organizations involved. The subscript indicates the duration of the effective gradient application in the corresponding k-space. Indicates the organization type.

3. The method for mapping and classifying meniscus tears according to claim 1, characterized in that, The overall mask for the two-dimensional knee MRI image of each axis is: ； In the formula, Indicates coordinates as binary mask of voxels The coordinates of the location image intensity dataset in a two-dimensional knee joint MRI image are: MRI intensity at voxels and These represent the lower and upper thresholds used to determine the voxel inclusion or exclusion intensity; The background tissue mask for each axis of the two-dimensional knee MRI image is: ； In the formula, Indicates coordinates as The binary mask of the background tissue voxels, Indicates coordinates as MRI intensity at the background tissue voxel This represents the minimum threshold.

4. The method for mapping and classifying meniscus tears according to claim 1, characterized in that, Based on the overall mask and background tissue mask of the two-dimensional knee MRI image for each axis, a two-dimensional mask of the meniscus for each axis is obtained by subtraction, specifically including: Based on the global mask and background tissue mask of the two-dimensional knee MRI image for each axis, the formula is used... This yields a two-dimensional mask of the meniscus for each axis; where, Indicates coordinates as The binary mask of the meniscus voxels; Noise and tissue debris are removed from the 2D mask of the meniscus in each axis to obtain the final 2D mask of the meniscus in each axis.

5. The method for mapping and classifying meniscus tears according to claim 1, characterized in that, Based on the two-dimensional mask of the meniscus in each axis, a three-dimensional reconstruction of the meniscus is performed to obtain a three-dimensional model of the meniscus in each axis, specifically including: Determine the boundary of the three-dimensional model of the meniscus based on the two-dimensional mask of the meniscus in each axis; Based on the boundary of the three-dimensional model of the meniscus, the two-dimensional mask of the meniscus in each axis is converted into three-dimensional data; Using the aforementioned three-dimensional data, a three-dimensional reconstruction is performed to obtain an initial three-dimensional model of the upper half of each axial meniscus; The initial 3D model of the upper meniscus in each axis is wrapped and smoothed to obtain the 3D model of the meniscus in each axis.

6. The method for mapping and classifying meniscus tears according to claim 1, characterized in that, By fusing biaxial 3D models of the meniscus, a 3D digital model of the meniscus is obtained, specifically including: A 3D model of a meniscus aligned along two axes; The aligned biaxial 3D models of the meniscus are merged to obtain a 3D digital model of the meniscus.

7. The method for mapping and classifying meniscus tears according to claim 1, characterized in that, Based on the aforementioned three-dimensional digital model of the meniscus, the meniscus tear is mapped and quantitatively classified, specifically including: Draw the arc between the anterior and posterior ends of the meniscus in the three-dimensional digital model of the meniscus; Divide the arc equally from the front foot endpoint to the rear foot endpoint, and then divide the three-dimensional digital model of the meniscus into the front segment, body segment, and rear segment according to the equally divided arc. The meniscus region inside the arc is defined as the interior of the meniscus or the white area, the meniscus region outside the arc is defined as the exterior of the meniscus or the red area, and the side of the meniscus is defined as the peripheral area. Based on the segmentation of the three-dimensional digital model of the meniscus and the definition of the meniscus region, the location of the meniscus tear is determined. The morphology of a meniscus tear along the arc direction is defined as a longitudinal tear, the morphology horizontally perpendicular to the arc direction is defined as a radial tear, and the morphology in the outer peripheral horizontal direction is defined as a horizontal tear. Based on the defined meniscus tear morphology, the meniscus tear morphology in the three-dimensional digital model of the meniscus is determined; Define the degree of meniscal tear; the degree of meniscal tear includes: the length defined along the arc direction, the width defined perpendicular to the arc direction, and the depth defined perpendicular to the cross-section of the meniscus; The degree of meniscus tear in the three-dimensional digital model of the meniscus is determined based on the defined degree of meniscus tear. Whether a tear constitutes a penetrating tear is determined based on whether the tear simultaneously damages at least two geometric surfaces of the meniscus.

8. The method for mapping and classifying meniscus tears according to claim 1, characterized in that, The dual axes include the coronal axis and the sagittal axis.

9. A system for mapping and classifying meniscus tears, characterized in that, include: The MRI image acquisition module is used to acquire two-dimensional MRI images of the knee joint in a biaxial direction for patients with meniscus tears. The mask generation module is used to generate an overall mask and a background tissue mask for a two-dimensional knee MRI image along each axis, based on the different signal intensities of different tissues within the knee joint; the tissues outside the meniscus in the knee joint are the background tissues. The subtraction module is used to obtain a two-dimensional mask of the meniscus for each axis by subtraction based on the overall mask and background tissue mask of the two-dimensional knee MRI image for each axis. The 3D reconstruction module is used to perform 3D reconstruction of the meniscus based on the 2D mask of the meniscus in each axis, and obtain the 3D model of the meniscus in each axis. The fusion module is used to fuse biaxial meniscus 3D models to obtain a 3D digital model of the meniscus. The tear mapping module is used to map and quantitatively classify meniscus tears based on the three-dimensional digital model of the meniscus.

10. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that the processor executes the computer program to implement the method for mapping and classifying meniscal tears according to any one of claims 1-7.

Images