Method and device for registering pathological sections based on biomechanical elastic deformation with MRI images

Abstract

The application provides a pathological section and MRI image registration method and device based on biomechanical elastic deformation, which comprises the following steps: obtaining a three-dimensional digital twin of a tumor region, obtaining a digital resection region in the three-dimensional digital twin based on a preset surgical margin; performing ex vivo dynamics simulation on the digital resection region to obtain a tissue deformation result; obtaining an optimal deformation section with maximum mutual information in the tissue deformation result and a pathological section; obtaining an inverse transformation deformation field between the optimal deformation section and the three-dimensional digital twin, and projecting a tumor region edge of the pathological section to the three-dimensional digital twin based on the inverse transformation deformation field. The scheme realizes fine local nonlinear adaptation of the section by constructing an ex vivo dynamics finite element simulation technology of a volume shrinkage function and performing local deformation with the pathological section edge information as a constraint, quantitatively restoring the global non-uniform shrinkage of the ex vivo tissue and the differential local deformation of the tumor and normal tissue.

Description

Technical Field

[0001] This application relates to the biomedical field, and in particular to a method and apparatus for registering pathological sections and MRI images based on biomechanical elastic deformation. Background Technology

[0002] In the field of clinical precision diagnosis and surgical treatment of solid tumors, preoperative magnetic resonance imaging (MRI) can provide macroscopic three-dimensional anatomical information of lesions and surrounding tissues in patients with its advantages of being non-invasive and having high soft tissue resolution. Meanwhile, postoperative digital pathological sections (WSI) of ex vivo tissues, as the gold standard for tumor diagnosis, can show the distribution range and invasion boundary of malignant cells at the micron scale, and are the decisive basis for evaluating surgical outcomes and formulating subsequent treatment plans.

[0003] Currently, medical image registration technology has formed a relatively mature technical system. Among them, rigid registration technology can achieve rapid registration between medical images of the same modality with small deformation through linear transformations such as translation, rotation, and scaling, and is widely used in the preliminary alignment of preoperative and postoperative images. Traditional elastic registration technology, by introducing a nonlinear transformation model, can adapt to local deformation between images and has been applied to some extent in the registration of multimodal macroscopic images such as CT and MRI.

[0004] However, existing registration techniques face significant technical bottlenecks in clinical applications. Firstly, the irreversible drastic morphological changes of excised tissues after surgery lead to insufficient registration accuracy. After surgical removal from the body, excised tissues undergo significant non-uniform shrinkage, distortion, and volume collapse due to the loss of physiological stress support, fixation fluid immersion, and paraffin embedding. The overall shrinkage rate typically reaches 10%-30%, and tumor tissue exhibits differentiated local deformations due to differences in physical properties compared to surrounding normal soft tissues. Existing rigid registration techniques can only achieve global linear transformations and cannot accommodate such complex non-linear deformations. Traditional elastic registration techniques, lacking consideration of the actual mechanical properties of biological tissues, rely solely on image features for unconstrained deformation fitting, easily producing spurious deformations such as tissue tearing and mesh overlap—deformations unlikely in pathology—resulting in significant deviations between the registration results and the actual tissue morphology, failing to meet the demands of precise clinical matching.

[0005] Secondly, current technologies cannot solve the problem of precise cross-dimensional matching between three-dimensional macroscopic images and two-dimensional microscopic pathological sections. Preoperative MRI images are three-dimensional volumetric data containing complete organ and lesion information, while pathological sections are two-dimensional ultrathin surface data with a thickness of only 5-10 μm. There is an order of magnitude difference between the two in terms of spatial dimension and resolution scale. Current technologies cannot accurately locate the spatial section that completely corresponds to the pathological section in the three-dimensional MRI volumetric data, nor can they establish a continuous mapping relationship between the two-dimensional pathological pixel coordinates and the three-dimensional MRI spatial coordinates, resulting in the inability to achieve precise spatial localization in preoperative three-dimensional images.

[0006] In summary, current technologies cannot achieve the in-situ retrospective of pathological gold standard lesion information to preoperative MRI images, nor can they quantify the actual physical distance between the tumor lesion edge and the surgical resection margin. They can only rely on clinical experience to judge whether the surgical resection is thorough, which easily leads to postoperative recurrence due to positive resection margins or damage to normal tissue due to excessive resection. Summary of the Invention

[0007] This application provides a method and apparatus for registering pathological sections and MRI images based on biomechanical elastic deformation. By constructing an in vitro dynamic finite element simulation technology based on a volume contraction function, and using the edge information of the pathological section as a constraint for local deformation, the method quantifies and restores the global non-uniform contraction of the in vitro tissue and the differentiated local deformation of tumor and normal tissue. Based on the global biomechanical deformation simulation, it achieves refined local nonlinear adaptation of the section, thereby improving the pixel-level registration accuracy between the pathological section and the image section.

[0008] In a first aspect, embodiments of this application provide a method for registering pathological sections and MRI images based on biomechanical elastic deformation, the method comprising:

[0009] Obtain 3D MRI images of the tumor region, construct a 3D digital twin corresponding to the 3D MRI images based on image segmentation algorithms, and obtain the digital resection area in the 3D digital twin based on the preset surgical margins;

[0010] A volume shrinkage function was constructed based on the fixative concentration and tissue ex vivo time. Based on the volume shrinkage function, the tissue deformation results were obtained by performing ex vivo dynamic simulation of the digitally excised area.

[0011] Pathological sections were obtained from the ex vivo tissue after surgery. The two-dimensional section with the largest mutual information with the pathological section in the tissue deformation results was selected as the optimal section. The edge information of the pathological section was used as a constraint to perform local deformation on the optimal section until the mutual information between the optimal section and the pathological section was maximized, thus obtaining the optimal deformation section.

[0012] The inverse transformation deformation field is obtained based on the mapping relationship between the optimal deformation section and the three-dimensional digital twin. The edge of the tumor region of the pathological slice is projected into the three-dimensional digital twin based on the inverse transformation deformation field.

[0013] Secondly, embodiments of this application provide a device for registering pathological sections and MRI images based on biomechanical elastic deformation, comprising:

[0014] The acquisition module is used to acquire MRI three-dimensional images of the tumor region, construct a three-dimensional digital twin corresponding to the MRI three-dimensional images based on image segmentation algorithms, and acquire the digital resection area in the three-dimensional digital twin based on the preset surgical margin.

[0015] The in vitro deformation module constructs a volume shrinkage function based on the concentration of the fixative and the time of tissue ex vivo, and performs in vitro dynamics simulation on the digitally excised area based on the volume shrinkage function to obtain the tissue deformation results.

[0016] The local deformation module is used to obtain pathological sections from ex vivo tissues after surgery. The optimal section is the two-dimensional section with the largest mutual information with the pathological section in the tissue deformation result. The optimal section is locally deformed with the edge information of the pathological section as a constraint until the mutual information between the optimal section and the pathological section is maximized.

[0017] The registration module obtains the inverse transformation deformation field based on the mapping relationship between the optimal deformation section and the three-dimensional digital twin, and projects the edge of the tumor region of the pathological slice into the three-dimensional digital twin based on the inverse transformation deformation field.

[0018] Thirdly, embodiments of this application provide an electronic device including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform a method for registering pathological sections and MRI images based on biomechanical elastic deformation.

[0019] Fourthly, embodiments of this application provide a readable storage medium storing a computer program, which, when executed by a processor, implements a method for registering pathological sections and MRI images based on biomechanical elastic deformation.

[0020] The main contributions and innovations of this invention are as follows:

[0021] This application's embodiments utilize a 3D digital twin construction technology based on deep learning image segmentation and tetrahedral finite element mesh modeling. This technology accurately segments organs and tumor regions in MRI images, restoring the 3D anatomical structure and true biomechanical properties of tissues in vivo. This provides a high-precision digital base that conforms to physiological characteristics for subsequent deformation simulation and registration. The scheme employs an in vitro dynamics finite element simulation technology based on the concentration of the fixative and the tissue's time outside the body to construct a volume contraction function. This quantifies and restores the global non-uniform contraction of the in vitro tissue and the differentiated local deformation of tumors and normal tissues, avoiding the false tissue tearing and mesh overlap problems caused by unconstrained deformation in traditional elastic registration, thus ensuring the pathological realism of the deformation simulation. Furthermore, this scheme uses pathological slide edge information as a constraint and a gradient-based optimization technique for local fine-tuning of the free deformation mesh. Based on global biomechanical deformation simulation, it achieves refined local nonlinear adaptation of the cross-section, ensuring the biomechanical rationality of the deformation and further improving the pixel-level registration accuracy between pathological slides and image cross-sections.

[0022] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0023] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0024] Figure 1 This is a schematic diagram of a biomechanical elastic deformation-based method for registering pathological sections and MRI images according to an embodiment of this application.

[0025] Figure 2 This is a schematic diagram of the contraction of ex vivo tissue after surgery according to an embodiment of this application;

[0026] Figure 3 This is a schematic diagram illustrating the adjustment of nodes within a free-deformation mesh based on gradient information, according to an embodiment of this application.

[0027] Figure 4 This is a structural block diagram of a biomechanical elastic deformation-based registration device for pathological slides and MRI images, according to an embodiment of this application.

[0028] Figure 5 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0029] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with one or more embodiments of this specification. Rather, they are merely examples of apparatuses and methods consistent with some aspects of one or more embodiments of this specification as detailed in the appended claims.

[0030] It should be noted that the steps of the corresponding methods are not necessarily performed in the order shown and described in this specification in other embodiments. In some other embodiments, the methods may include more or fewer steps than described in this specification. Furthermore, a single step described in this specification may be broken down into multiple steps in other embodiments; and multiple steps described in this specification may be combined into a single step in other embodiments.

[0031] Example 1

[0032] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with one or more embodiments of this specification. Rather, they are merely examples of apparatuses and methods consistent with some aspects of one or more embodiments of this specification as detailed in the appended claims.

[0033] It should be noted that the steps of the corresponding methods are not necessarily performed in the order shown and described in this specification in other embodiments. In some other embodiments, the methods may include more or fewer steps than described in this specification. Furthermore, a single step described in this specification may be broken down into multiple steps in other embodiments; and multiple steps described in this specification may be combined into a single step in other embodiments.

[0034] Example 1

[0035] This application provides a method for registering pathological sections and MRI images based on biomechanical elastic deformation. It utilizes an in vitro dynamic finite element simulation technique based on a volume contraction function, and performs local deformation using the edge information of the pathological section as a constraint. This quantifies and restores the global non-uniform contraction of the ex vivo tissue and the differentiated local deformation of tumor and normal tissues. Based on the global biomechanical deformation simulation, it achieves refined local nonlinear adaptation of the section plane, improving the pixel-level registration accuracy between the pathological section and the image section. Specifically, refer to... Figure 1 The method includes:

[0036] Obtain 3D MRI images of the tumor region, construct a 3D digital twin corresponding to the 3D MRI images based on image segmentation algorithms, and obtain the digital resection area in the 3D digital twin based on the preset surgical margins;

[0037] A volume shrinkage function was constructed based on the fixative concentration and tissue ex vivo time. Based on the volume shrinkage function, the tissue deformation results were obtained by performing ex vivo dynamic simulation of the digitally excised area.

[0038] Pathological sections were obtained from the ex vivo tissue after surgery. The two-dimensional section with the largest mutual information with the pathological section in the tissue deformation results was selected as the optimal section. The edge information of the pathological section was used as a constraint to perform local deformation on the optimal section until the mutual information between the optimal section and the pathological section was maximized, thus obtaining the optimal deformation section.

[0039] The inverse transformation deformation field is obtained based on the mapping relationship between the optimal deformation section and the three-dimensional digital twin. The edge of the tumor region of the pathological slice is projected into the three-dimensional digital twin based on the inverse transformation deformation field.

[0040] In the current embodiment, a deep learning algorithm is used to segment the organ region and tumor region in the MRI three-dimensional image, and the organ region and tumor region are converted into tetrahedral finite element mesh and assigned preset physical properties to obtain a three-dimensional digital twin.

[0041] Specifically, a three-dimensional MRI image of a tumor region is a complete image that includes both the organ and the tumor. For example, a three-dimensional MRI image of a liver cancer patient is a complete three-dimensional MRI image that includes the liver and the corresponding tumor, while a three-dimensional MRI image of a gastric cancer patient is a complete three-dimensional image that includes the stomach and the corresponding tumor.

[0042] Specifically, a pre-trained 3D U-Net algorithm is used to automatically segment MRI three-dimensional images to achieve automated segmentation of organ regions and tumor regions.

[0043] Specifically, the organ region and tumor region are transformed into tetrahedral finite element meshes using finite element analysis software. Preset physical properties include Young's modulus, elastic coefficient, and Poisson's ratio. For example, using historical data and physiological characteristics of different patients, machine learning is used to predict the physical properties of the organ region and tumor region to obtain the preset physical properties.

[0044] For example, since the stiffness of the tumor region is significantly greater than that of organ tissue, the tumor region is given a higher Young's modulus.

[0045] In this approach, the preset surgical margin refers to the resection location set by the doctor based on the patient's actual condition. During the operation, the tumor will be removed at the preset surgical margin. This approach registers the pathological slides back into the three-dimensional digital twin, so that the precise spatial location of the tumor infiltration boundary in the pathological slides in the body before the operation can be clearly viewed, thereby determining whether the tumor resection is complete.

[0046] In the current embodiment, after surgical resection, the excised tissue loses its support, leading to the release of internal pressure and causing changes in tissue morphology, typically manifested as contraction or distortion. A schematic diagram of postoperative excised tissue contraction is shown below. Figure 2 As shown, to simulate this process in the digitally removed region, the formula for the volume shrinkage function is expressed as:

[0047]

[0048] in, It is a volume contraction function. It is a constant used to represent the maximum rate of contraction. For the maximum tissue time outside the body, To organize the time outside the body, The concentration of the fixative is given.

[0049] Specifically, the output of the volume shrinkage function is a proportional value used to represent the volume shrinkage ratio of the ex vivo tissue after surgery. For example, if the output of the volume shrinkage function is 50%, it means that the ex vivo tissue will shrink to 50% of its original volume after surgery.

[0050] Specifically, as time goes on, the excised tissue after surgery, which has lost its blood supply, will gradually shrink. The excised tissue during surgery is usually preserved using a high-concentration fixative such as formaldehyde. The fixative will accelerate the shrinkage of the excised tissue and increase its rigidity. Therefore, this method constructs a volume shrinkage function by using the tissue excision time and fixative concentration to simulate the shrinkage of the excised tissue after surgery.

[0051] For example, since different Young's modulus, Poisson's ratio, and elastic coefficients have been assigned to the organ region and tumor region in the 3D digital twin, a continuum mechanical equilibrium equation is established in the finite element simulation to ensure that the tumor region and normal tissue in the 3D digital twin have different deformation responses. The formula is expressed as:

[0052]

[0053] in, For Hamiltonian differential operators, For stress tensor, It is a volume force vector.

[0054] For example, an energy term is constructed based on the volume shrinkage function. Then, in the finite element simulation, the deformation displacement field is solved by minimizing the partial energy of the energy term to perform an out-of-body dynamics simulation of the digitally cut region. The formula for the energy term is expressed as:

[0055]

[0056] in, For energy terms, For deformation displacement field Jacobian determinant, For the time of tissue ex vivo stationary phase concentration The volume contraction function under the given conditions.

[0057] Specifically, the specific technical implementation of finite element analysis is a well-known existing technology in this field, and will not be elaborated upon in this solution.

[0058] Furthermore, local areas of the excised tissue after surgery may be distorted due to irregular shape. Therefore, in order to take into account the non-uniform deformation of local areas, a preset local correction factor is introduced into the volume shrinkage function.

[0059] Specifically, considering the differences in tissue type, boundary curvature, and elastic parameters among different local regions of excised tissue after surgery, this scheme presets a local correction factor for each grid cell or node in the digital resection area. The local correction factor is determined based on at least one of the following: tissue region type, local geometric feature parameters, and preset physical properties. The formula for introducing the preset local correction factor into the volume shrinkage function is expressed as follows:

[0060]

[0061] in, Let x be the local correction factor at position x. It is a volume contraction function. This is the volume shrinkage function after local correction factor.

[0062] Furthermore, in the finite element analysis, the displacement of each node in the digitally excised region is performed in real time based on the real-time acquired fixative concentration, tissue excision time, and volume shrinkage function results to complete the excision dynamics simulation.

[0063] Specifically, by displacing each node of the digitally excised region in finite element analysis, the contraction and local deformation of the excised tissue after surgery are simulated, thereby accurately reflecting the biomechanical changes after surgery and optimizing the simulation of postoperative tissue recovery.

[0064] Specifically, before the excised tissue is removed after surgery, the excised tissue is under pressure. In the three-dimensional digital twin, this is manifested as stress at all nodes. This stress is generated by factors such as physiological fluid and vascular pressure within the tissue. After surgical removal, these internal pressures and supports disappear, causing the stress within the tissue to be released. This process is simulated in finite element analysis by calculating the change in stress state at each node.

[0065] Specifically, in finite element analysis, the shrinkage ratio of the ex vivo tissue after surgery is determined by the volume shrinkage function result; the shrinkage rate is simulated by the concentration of fixative. Generally speaking, the higher the concentration of fixative, the faster the shrinkage rate of the ex vivo tissue after surgery; as for the tissue ex vivo time, in the finite element simulation, it represents that as time goes by, the ex vivo tissue after surgery gradually shrinks, the distance between nodes decreases, and the digital resection area becomes smaller.

[0066] Specifically, for hard tissues (such as tumors) in ex vivo tissues after surgery, the contraction is smaller, while for soft tissues (such as organs), the degree of contraction may be greater.

[0067] In the current embodiment, multiple two-dimensional sections are obtained from the tissue deformation results to form a two-dimensional section set. The two-dimensional section with the largest mutual information with the pathological slide is selected by traversal and selected as the optimal section. The mutual information is used to measure the overlap of the gray-level distribution of two two-dimensional sections.

[0068] Specifically, mutual information is an indicator that measures the similarity between two images. The larger the value, the more overlapping information there is between the images. Therefore, this scheme obtains the optimal cross section by maximizing mutual information, thereby finding the two-dimensional image cross section that best matches the pathological section in the tissue deformation results.

[0069] In the current embodiment, a uniformly distributed free-deformation mesh is generated on the optimal section. The free-deformation mesh includes multiple freely displaceable nodes. The node coordinates of each node within the free-deformation mesh are obtained, and the gradient information of the mutual information relative to the node coordinates is calculated. Based on the gradient information, the nodes within the free-deformation mesh are adjusted until the mutual information between the optimal section and the pathological slide is maximized. A schematic diagram of adjusting the nodes within the free-deformation mesh based on the gradient information is shown below. Figure 3 As shown.

[0070] Specifically, each node in the free-deformation mesh can move freely along the X, Y, and Z axes, thereby achieving local deformation of the optimal cross-section by moving the nodes.

[0071] Specifically, an edge extraction algorithm is used to obtain the edge information of the pathological slices, and the edge information of the pathological slices is used as a constraint to ensure that local deformation of the optimal cut surface will not deviate from the range of the pathological slices.

[0072] For example, the gradient information is used to represent the optimal direction for node coordinate movement. When the gradient information is positive, the mutual information can be increased by moving the node to the left, so the node is moved to the left. When the gradient information is negative, the mutual information can be increased by moving the node to the right, so the node is moved to the right. In the process of adjusting the replacement of nodes in the free deformation mesh, the registration accuracy between the optimal section and the pathological slice will be continuously improved until the mutual information reaches the maximum.

[0073] Specifically, the objective function for adjusting the nodes within the free-deformation mesh based on gradient information until the mutual information between the optimal section and the pathological slide is maximized is:

[0074]

[0075] This is the optimal spatial deformation field, used to directly adjust the nodes within the free-deformation mesh. The operator for minimizing the independent variable is used to find the independent variable that minimizes the function within the brackets. For the optimal cross section, For pathological sections, Represents mutual information computation. For deformation smoothing constraints, The weights are for deformation smoothing constraints.

[0076] In other embodiments, the nodes in the free-deformation mesh are adjusted based on gradient information until the mutual information between the optimal section and the pathological slice reaches a set threshold, or the iteration stops when the change in mutual information is less than a preset threshold.

[0077] In this scheme, the formula for calculating mutual information is:

[0078]

[0079] Where A and B are two different two-dimensional cross-sections. For mutual information between A and B, Let A be the entropy. Let B be the entropy. Let be the joint entropy of A and B.

[0080] In the current embodiment, the forward deformation field from the three-dimensional digital twin to the optimal deformation section is obtained. Based on the correspondence of discrete control points of the forward deformation field, an initial inverse mapping is obtained by interpolation reconstruction. The forward deformation field is then processed by reprojection error iterative correction to obtain the inverse transformation deformation field.

[0081] The transformation formula for the inverse transformation deformation field is expressed as follows:

[0082]

[0083] in, These are the coordinates of the target point in the optimal deformation section. These are the spatial coordinates of the target point in the three-dimensional digital twin. It is the inverse transformation deformation field.

[0084] In the current embodiment, the inverse transformation deformation field is obtained based on the mapping relationship between the optimal deformation section and the three-dimensional digital twin. .

[0085] Specifically, based on the pre-established mapping relationship between the optimal deformation section and the three-dimensional digital twin, the inverse transformation deformation field can be obtained. The inverse transformation deformation field describes the spatial transformation relationship from the optimal deformation section to the corresponding position in the three-dimensional digital twin.

[0086] In the current embodiment, after projecting the edge of the tumor region of the pathological slide onto the three-dimensional digital twin, the Euclidean distance between the edge of the tumor region of the pathological slide and the preset surgical margin is calculated. If the Euclidean distance is less than the clinical safety threshold, it indicates that the tumor has not been completely removed. If the Euclidean distance is not less than the clinical safety threshold, it indicates that the tumor has been completely removed.

[0087] Specifically, the Euclidean distance between the edge of the tumor region on the pathological section and the pre-designated surgical margin is calculated using the following formula:

[0088]

[0089] in, This represents the Euclidean distance, and min indicates taking the minimum value. The edge of the tumor area, Pre-set surgical margins.

[0090] For example, a set of feature points at the edge of the tumor region and a set of feature points at the preset surgical resection margin are obtained, and the Euclidean distance is calculated based on the set of feature points at the edge of the tumor region and the set of feature points at the preset surgical resection margin.

[0091] Specifically, the clinical safety thresholds in this protocol are adaptively set based on different tumor types. For example, the safe margin for breast cancer is usually greater than 2mm. Therefore, when the Euclidean distance is less than 2mm, it is considered a positive margin, indicating that there are still tumor cells remaining at the surgical cutting boundary, the resection is incomplete, and subsequent treatment measures need to be taken in time. When the Euclidean distance is greater than or equal to 2mm, it indicates that there is sufficient normal tissue space between the surgical margin and the tumor edge, the tumor is completely removed, and there is no risk of residual tumor.

[0092] Example 2

[0093] Based on the same concept, referencing Figure 4This application also proposes a biomechanical elastic deformation-based registration device for pathological sections and MRI images, comprising:

[0094] The acquisition module is used to acquire MRI three-dimensional images of the tumor region, construct a three-dimensional digital twin corresponding to the MRI three-dimensional images based on image segmentation algorithms, and acquire the digital resection area in the three-dimensional digital twin based on the preset surgical margin.

[0095] The in vitro deformation module constructs a volume shrinkage function based on the concentration of the fixative and the time of tissue ex vivo, and performs in vitro dynamics simulation on the digitally excised area based on the volume shrinkage function to obtain the tissue deformation results.

[0096] The local deformation module is used to obtain pathological sections from ex vivo tissues after surgery. The optimal section is the two-dimensional section with the largest mutual information with the pathological section in the tissue deformation result. The optimal section is locally deformed with the edge information of the pathological section as a constraint until the mutual information between the optimal section and the pathological section is maximized.

[0097] The registration module obtains the inverse transformation deformation field based on the mapping relationship between the optimal deformation section and the three-dimensional digital twin, and projects the edge of the tumor region of the pathological slice into the three-dimensional digital twin based on the inverse transformation deformation field.

[0098] Example 3

[0099] This embodiment also provides an electronic device, see reference. Figure 5 It includes a memory 404 and a processor 402, wherein the memory 404 stores a computer program and the processor 402 is configured to run the computer program to perform the steps in any of the above method embodiments.

[0100] Specifically, the processor 402 may include a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0101] Memory 404 may include a mass storage device for data or instructions. For example, and not limitingly, memory 404 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, memory 404 may include removable or non-removable (or fixed) media. Where appropriate, memory 404 may be internal or external to a data processing device. In a particular embodiment, memory 404 is non-volatile memory. In a particular embodiment, memory 404 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable read-only memory (PROM), an erasable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), an electrically alterable read-only memory (EAROM), or flash memory, or a combination of two or more of these. Where appropriate, the RAM can be Static Random-Access Memory (SRAM) or Dynamic Random-Access Memory (DRAM). DRAM can be Fast Page Mode Dynamic Random-Access Memory (FPMDRAM), Extended Data Out Dynamic Random-Access Memory (EDODRAM), Synchronous Dynamic Random-Access Memory (SDRAM), etc.

[0102] The memory 404 can be used to store or cache various data files that need to be processed and / or communicated, as well as possible computer program instructions executed by the processor 402.

[0103] The processor 402 reads and executes computer program instructions stored in the memory 404 to implement any of the biomechanical elastic deformation-based pathological slide and MRI image registration methods in the above embodiments.

[0104] Optionally, the electronic device may further include a transmission device 406 and an input / output device 408, wherein the transmission device 406 is connected to the processor 402, and the input / output device 408 is connected to the processor 402.

[0105] The transmission device 406 can be used to receive or send data via a network. Specific examples of the network described above may include wired or wireless networks provided by the communication provider of the electronic device. In one example, the transmission device includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 406 may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0106] The input / output device 408 is used to input or output information. In this embodiment, the input information may be MRI three-dimensional images, etc., and the output information may be the registration results of the tumor region edge of a pathological section and a three-dimensional digital twin, etc.

[0107] Optionally, in this embodiment, the processor 402 can be configured to perform the following steps via a computer program:

[0108] Obtain 3D MRI images of the tumor region, construct a 3D digital twin corresponding to the 3D MRI images based on image segmentation algorithms, and obtain the digital resection area in the 3D digital twin based on the preset surgical margins;

[0109] A volume shrinkage function was constructed based on the fixative concentration and tissue ex vivo time. Based on the volume shrinkage function, the tissue deformation results were obtained by performing ex vivo dynamic simulation of the digitally excised area.

[0110] Pathological sections were obtained from the ex vivo tissue after surgery. The two-dimensional section with the largest mutual information with the pathological section in the tissue deformation results was selected as the optimal section. The edge information of the pathological section was used as a constraint to perform local deformation on the optimal section until the mutual information between the optimal section and the pathological section was maximized, thus obtaining the optimal deformation section.

[0111] The inverse transformation deformation field is obtained based on the mapping relationship between the optimal deformation section and the three-dimensional digital twin. The edge of the tumor region of the pathological slice is projected into the three-dimensional digital twin based on the inverse transformation deformation field.

[0112] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0113] Generally, various embodiments can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects of the invention can be implemented in hardware, while others can be implemented by firmware or software executed by a controller, microprocessor, or other computing device, but the invention is not limited thereto. Although various aspects of the invention may be shown and described as block diagrams, flowcharts, or using some other graphical representation, it should be understood that, by way of non-limiting example, these blocks, apparatuses, systems, techniques, or methods described herein can be implemented in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0114] Embodiments of the present invention can be implemented by computer software, which may be executable by a data processor of a mobile device, such as a processor entity, or by hardware, or by a combination of software and hardware. Computer software or programs (also referred to as program products) including software routines, applets, and / or macros can be stored in any device-readable data storage medium, and they include program instructions for performing specific tasks. The computer program product may include one or more computer-executable components configured to perform the embodiments when the program is run. The one or more computer-executable components may be at least one piece of software code or a portion thereof. Additionally, it should be noted in this respect that, as Figure 5 Any box in the logical flow can represent a program step, or interconnected logic circuits, boxes and functions, or a combination of program steps and logic circuits, boxes and functions. Software can be stored on physical media such as memory chips or blocks of storage implemented within a processor, magnetic media such as hard disks or floppy disks, and optical media such as DVDs and their data variants, CDs, etc. The physical medium is a non-transient medium.

[0115] Those skilled in the art should understand that the technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments have been described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0116] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for registering pathological sections and MRI images based on biomechanical elastic deformation, characterized in that, Includes the following steps: Obtain 3D MRI images of the tumor region, construct a 3D digital twin corresponding to the 3D MRI images based on image segmentation algorithms, and obtain the digital resection area in the 3D digital twin based on the preset surgical margins; A volume shrinkage function is constructed based on the fixative concentration and the time of tissue ex vivo. Based on this volume shrinkage function, in vitro dynamic simulations of the digitally excised region are performed to obtain tissue deformation results. The formula for the volume shrinkage function is as follows: in, It is a volume contraction function. It is a constant used to represent the maximum rate of contraction. For the maximum tissue time outside the body, To organize the time outside the body, The concentration of the stationary phase; Pathological sections are obtained from ex vivo tissue after surgery. The two-dimensional section with the largest mutual information with the pathological section in the tissue deformation results is selected as the optimal section. Multiple two-dimensional sections are obtained from the tissue deformation results to form a set of two-dimensional sections. The two-dimensional section with the largest mutual information with the pathological section is selected by traversal. The mutual information is used to measure the overlap of the gray-level distribution of the two two-dimensional sections. The edge information of the pathological section is used as a constraint to locally deform the optimal section until the mutual information between the optimal section and the pathological section is maximized, thus obtaining the optimal deformed section. A uniformly distributed free deformation mesh is generated on the optimal section. The free deformation mesh includes multiple freely displaceable nodes. The node coordinates of each node in the free deformation mesh are obtained. The gradient information of the mutual information relative to the coordinates of each node is calculated. The nodes in the free deformation mesh are adjusted based on the gradient information until the mutual information between the optimal section and the pathological section is maximized. The inverse transformation deformation field is obtained based on the mapping relationship between the optimal deformation section and the three-dimensional digital twin. The edge of the tumor region of the pathological slice is projected into the three-dimensional digital twin based on the inverse transformation deformation field.

2. The method for registering pathological sections and MRI images based on biomechanical elastic deformation according to claim 1, characterized in that, Deep learning algorithms are used to segment organ regions and tumor regions in MRI 3D images. The organ regions and tumor regions are then converted into tetrahedral finite element meshes and assigned preset physical properties to obtain 3D digital twins.

3. The method for registering pathological sections and MRI images based on biomechanical elastic deformation according to claim 1, characterized in that, In finite element analysis, the displacement of each node in the digitally excised region is performed in real time based on the real-time obtained fixative concentration, tissue excision time, and volume shrinkage function results to complete the excision dynamics simulation.

4. The method for registering pathological sections and MRI images based on biomechanical elastic deformation according to claim 1, characterized in that, After projecting the edge of the tumor region of the pathological slide onto a three-dimensional digital twin, the Euclidean distance between the edge of the tumor region of the pathological slide and the preset surgical margin is calculated. If the Euclidean distance is less than the clinical safety threshold, it indicates that the tumor has not been completely removed. If the Euclidean distance is not less than the clinical safety threshold, it indicates that the tumor has been completely removed.

5. A device for registering pathological sections and MRI images based on biomechanical elastic deformation, characterized in that, include: The acquisition module is used to acquire MRI three-dimensional images of the tumor region, construct a three-dimensional digital twin corresponding to the MRI three-dimensional images based on image segmentation algorithms, and acquire the digital resection area in the three-dimensional digital twin based on the preset surgical margin. The in vitro deformation module constructs a volume shrinkage function based on the fixative concentration and the tissue ex vivo time. Based on this volume shrinkage function, in vitro dynamic simulations of the digitally excised region are performed to obtain the tissue deformation results. The formula for the volume shrinkage function is as follows: in, It is a volume contraction function. It is a constant used to represent the maximum rate of contraction. For the maximum tissue time outside the body, To organize the time outside the body, The concentration of the stationary phase; The local deformation module is used to obtain pathological sections from ex vivo tissue after surgery. The optimal section is the two-dimensional cross-section with the highest mutual information between the tissue deformation result and the pathological section. Multiple two-dimensional cross-sections are obtained from the tissue deformation result to form a set of two-dimensional cross-sections. The optimal section is obtained by iterating through the set of two-dimensional cross-sections and finding the one with the highest mutual information with the pathological section. The mutual information is used to measure the overlap of the gray-level distribution of two two-dimensional cross-sections. The optimal section is locally deformed using the edge information of the pathological section as a constraint until the mutual information between the optimal section and the pathological section is maximized. A uniformly distributed free deformation mesh is generated on the optimal section. The free deformation mesh includes multiple freely displaceable nodes. The node coordinates of each node in the free deformation mesh are obtained, and the gradient information of the mutual information relative to the node coordinates is calculated. Based on the gradient information, the nodes in the free deformation mesh are adjusted until the mutual information between the optimal section and the pathological section is maximized. The registration module obtains the inverse transformation deformation field based on the mapping relationship between the optimal deformation section and the three-dimensional digital twin, and projects the edge of the tumor region of the pathological slice into the three-dimensional digital twin based on the inverse transformation deformation field.

6. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform a biomechanical elastic deformation-based method for registering pathological sections and MRI images as described in any one of claims 1-4.

7. A readable storage medium, characterized in that, The readable storage medium stores a computer program that, when executed by a processor, implements a method for registering pathological sections and MRI images based on biomechanical elastic deformation as described in any one of claims 1-4.

Images