A perspective image analysis method for bionic repair of bone tumor

Abstract

The application discloses a perspective image analysis method for bionic repair of bone tumors, comprising the following steps: acquiring perspective image data of a bone tumor area; performing cross-modal feature alignment on the data, constructing a three-dimensional bone tumor image model in a three-dimensional voxel space coordinate system, and eliminating image artifacts; performing segmentation processing on the bone tumor through a deep learning segmentation network, identifying a bone tumor infiltration boundary, and acquiring three-dimensional morphological parameters and bone tumor characteristic parameters of the bone tumor area; reconstructing a bionic scaffold structure of a bone defect area through biomechanical modeling, and implanting the bionic scaffold into the bone tumor area; and acquiring bone-scaffold interface fusion rate, stress distribution uniformity and new bone growth depth indexes based on postoperative image data. The application has the advantages that: through the combination of accurate segmentation at the voxel level and a convolutional neural network, the bone tumor area and morphology can be accurately identified and divided, thereby providing reliable support for subsequent tumor infiltration evaluation and bionic scaffold design.

Description

Technical Field

[0001] This invention relates to dynamic adjustment technology, and in particular to a perspective image analysis method for biomimetic repair of bone tumors. Background Technology

[0002] Fluoroscopy image analysis for bone tumor repair has become a widely used technique in medical imaging in recent years. It aims to assist physicians in accurately assessing the detection, diagnosis, treatment, and repair process of bone tumors by analyzing fluoroscopic images such as X-rays or CT scans. With the continuous advancement of medical imaging technology, traditional manual diagnostic methods are facing challenges in terms of efficiency and accuracy, especially in the identification and analysis of complex lesions. Fluoroscopy images can provide detailed information on the internal structure and growth morphology of bone tumors, thus holding significant value in the early detection and monitoring of bone tumors.

[0003] Most current methods for fluoroscopic image analysis in bone tumor repair rely on traditional image processing techniques, such as threshold-based segmentation, simple region growing algorithms, or manual annotation. These methods often have limitations when dealing with complex tumor morphologies and minute structures. Due to the lack of voxel-level fine segmentation, they cannot accurately distinguish subtle differences between the tumor and surrounding healthy tissue, potentially leading to inaccurate segmentation results, misjudgment of tumor boundaries, or even omission of tumor areas. Furthermore, traditional methods are generally ineffective at processing multimodal imaging data (such as CT, MRI, and C-arm images), exhibiting significant shortcomings in cross-modal feature alignment and artifact removal. These limitations make it difficult to achieve high precision in tumor morphology assessment, invasive boundary identification, and subsequent biomimetic scaffold design, impacting repair outcomes and patient prognosis prediction. Summary of the Invention

[0004] To improve existing methods for perspective image analysis in bone tumor repair, this paper proposes a perspective image analysis method for biomimetic bone tumor repair. This method improves the accuracy of bone tumor images through precise voxel-level segmentation, accurately identifies tumor regions and morphology, and significantly enhances segmentation accuracy by combining convolutional neural networks and morphological operations, providing reliable support for subsequent tumor invasion assessment and biomimetic scaffold design.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A perspective image analysis method for biomimetic repair of bone tumors, comprising:

[0007] Acquire fluoroscopic imaging data of the bone tumor region, including X-ray computed tomography, magnetic resonance imaging, and intraoperative C-arm fluoroscopic images;

[0008] Cross-modal feature alignment is performed based on the acquired multi-source image data to construct a three-dimensional bone tumor image model in a three-dimensional voxel space coordinate system and eliminate image artifacts.

[0009] Based on the acquired three-dimensional bone tumor image data, the bone tumor is segmented using a deep learning segmentation network, and the invasion boundary of the bone tumor is identified to obtain the three-dimensional morphological parameters and bone tumor feature parameters of the bone tumor region.

[0010] Based on the acquired bone tumor region data, a biomimetic scaffold structure in the bone defect area was reconstructed through biomechanical modeling, and the biomimetic scaffold was implanted into the bone tumor region.

[0011] Based on postoperative imaging data, the bone-scaffold interface fusion rate, stress distribution uniformity, and new bone ingrowth depth were obtained.

[0012] Preferably, the acquisition of fluoroscopic image data of the bone tumor region, including X-ray computed tomography, magnetic resonance imaging, and intraoperative C-arm fluoroscopic images, specifically includes:

[0013] High-resolution spiral CT is used to acquire the fine structure and features of bone tissue and tumors, generating axial tomographic images in DICOM format.

[0014] The contrast between tumors and soft tissues is highlighted through T1-weighted, T2-weighted, and enhanced sequences of MRI scans;

[0015] Two-dimensional X-ray images are acquired in real time during the operation and the data is transmitted through the DICOM interface of the C-arm.

[0016] Preferably, the step of performing cross-modal feature alignment based on the acquired multi-source image data, constructing a three-dimensional bone tumor image model in a three-dimensional voxel space coordinate system, and eliminating image artifacts specifically includes:

[0017] Global alignment of multi-source image data based on landmarks in CT image data and soft tissue contours in MRI;

[0018] Local alignment of multi-source image data is achieved by compensating for the differences in soft tissue deformation between CT and MRI.

[0019] Using the DICOM coordinate system of preoperative CT as a reference, MRI and C-arm data are mapped to the CT coordinate system through a registration matrix to construct a three-dimensional voxel space coordinate system.

[0020] Voxel-weighted fusion was performed based on the proportion of bone tumor structures and soft tissue in bone tumors.

[0021] Based on the fused image data, it is matched with the intraoperative C-arm data, and image artifacts are eliminated through deep learning;

[0022] A three-dimensional bone tumor imaging model was constructed in a three-dimensional voxel space coordinate system based on the fused image data after artifact removal processing.

[0023] Preferably, the step of segmenting the bone tumor using a deep learning segmentation network based on the acquired three-dimensional bone tumor image data, identifying the bone tumor invasion boundary, and obtaining the three-dimensional morphological parameters and bone tumor feature parameters of the bone tumor region specifically includes:

[0024] Based on the acquired three-dimensional bone tumor image data, the bone tumor region is initially segmented using an image segmentation algorithm to obtain the image portion containing the bone tumor.

[0025] Based on morphological operations, the image portion containing bone tumors is further segmented and refined. The segmentation accuracy is improved by using a convolutional neural network to obtain voxel-level bone tumor image data.

[0026] Based on the characteristics of bone tumors, each voxel is determined to be a bone tumor voxel and classified accordingly;

[0027] Based on the voxel determination results of all image portions containing bone tumors, the portions determined to be non-bone tumor voxels are removed to obtain image data containing only bone tumors.

[0028] Based on the acquired image data containing only bone tumors, the invasion boundary of bone tumors is identified, and the smoothness of the boundary is analyzed to assess whether the tumor has invaded the adjacent area.

[0029] Based on the identified bone tumor invasion boundaries, the invasion boundaries are annotated in the three-dimensional bone tumor image model;

[0030] Based on the labeled bone tumor image data, morphological parameters and characteristic parameters of bone tumors were extracted.

[0031] Preferably, the step of determining whether each voxel is a bone tumor voxel based on bone tumor characteristics and classifying it specifically includes:

[0032] Calculate the spatial consistency of each voxel and its neighboring voxels. If the spatial consistency between voxels varies greatly, the voxel is located at the boundary of the bone tumor. If the spatial consistency varies little, the voxel is located inside the bone tumor.

[0033] Based on voxels located at the boundary of bone tumors, the alignment neighborhood is analyzed. The determination of whether a voxel belongs to a tumor is made by comparing its attribute characteristics with those of surrounding voxels, including grayscale difference, texture contrast, and uniformity.

[0034] Preferably, the step of reconstructing a biomimetic scaffold structure in the bone defect area based on the acquired bone tumor region data through biomechanical modeling, and implanting the biomimetic scaffold into the bone tumor region specifically includes:

[0035] Based on the obtained morphological and characteristic parameters of bone tumors, the mechanical properties of the bone defect area are obtained through biomechanical modeling.

[0036] Based on the mechanical properties and mechanical requirements of the bone tumor region, biomimetic materials were selected and a biomimetic scaffold structure was designed.

[0037] Mechanical simulations are performed using finite element analysis to simulate the load on the stent within the human body, optimizing stress and strain distribution and stent stability.

[0038] Preferably, the acquisition of bone-scaffold interface fusion rate, stress distribution uniformity, and new bone ingrowth depth indicators based on postoperative imaging data specifically includes:

[0039] By acquiring bone-scaffold interface fusion image data, the bone-scaffold interface is extracted, and the area where bone and scaffold are in contact is identified.

[0040] The bone-scaffold interface fusion rate is calculated based on the ratio of the bone-scaffold contact area to the total bone defect area.

[0041] The finite element method was used to simulate the skeletal mechanics after surgery and calculate the uniformity index of stress distribution, including stress standard deviation and stress concentration index.

[0042] Compared with the prior art, the advantages of the present invention are:

[0043] Precise voxel-level segmentation enables high-resolution identification and classification of bone tumor regions. By combining multi-source imaging data (such as CT, MRI, and C-arm fluoroscopy), this method first constructs a precise bone tumor imaging model in three-dimensional voxel space, effectively eliminating image artifacts. Based on this, morphological operations and convolutional neural networks (CNNs) are used to further optimize the segmentation accuracy of the tumor region, achieving accurate classification of bone tumor voxels. This voxel-level segmentation accurately distinguishes bone tumors from surrounding healthy tissue, eliminating irrelevant background information, thereby improving the accuracy of subsequent analysis and treatment planning. Furthermore, meticulous identification of infiltration boundaries allows for the assessment of whether the tumor has invaded adjacent bone regions, providing strong support for developing personalized surgical plans and implantable scaffold designs. Attached Figure Description

[0044] Figure 1 This is a schematic diagram of the method proposed in this invention;

[0045] Figure 2This is a schematic diagram illustrating the acquisition of bone tumor imaging data proposed in this invention;

[0046] Figure 3 This is a schematic diagram illustrating the construction of the three-dimensional bone tumor imaging model proposed in this invention;

[0047] Figure 4 This is a schematic diagram of the bone tumor image processing proposed in this invention;

[0048] Figure 5 This is a schematic diagram of the bone tumor voxel determination proposed in this invention;

[0049] Figure 6 This is a schematic diagram of bone loss area reconstruction proposed in this invention;

[0050] Figure 7 This is a schematic diagram of postoperative image analysis proposed in this invention;

[0051] Figure 8 This is an architecture diagram of the electronic devices in this solution;

[0052] Figure 9 This is a schematic diagram of the computer-readable storage medium structure in this scheme. Detailed Implementation

[0053] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0054] See Figure 1 As shown, a perspective image analysis method for biomimetic repair of bone tumors includes:

[0055] Step 1: Obtain fluoroscopic imaging data of the bone tumor area, including X-ray computed tomography, magnetic resonance imaging, and intraoperative C-arm fluoroscopic images;

[0056] Step 2: Based on the acquired multi-source image data, perform cross-modal feature alignment, construct a three-dimensional bone tumor image model in the three-dimensional voxel space coordinate system, and eliminate image artifacts;

[0057] Step 3: Based on the acquired 3D bone tumor image data, the bone tumor is segmented using a deep learning segmentation network, and the invasion boundary of the bone tumor is identified to obtain the 3D morphological parameters and bone tumor feature parameters of the bone tumor region.

[0058] Step 4: Based on the acquired bone tumor region data, reconstruct the biomimetic scaffold structure of the bone defect area through biomechanical modeling, and implant the biomimetic scaffold into the bone tumor region;

[0059] Step 5: Based on postoperative imaging data, obtain indicators such as bone-scaffold interface fusion rate, stress distribution uniformity, and new bone ingrowth depth.

[0060] See Figure 2 As shown, fluoroscopic imaging data of the bone tumor region was acquired, including X-ray computed tomography, magnetic resonance imaging, and intraoperative C-arm fluoroscopic images. Specifically, this included:

[0061] High-resolution spiral CT is used to acquire the fine structure and features of bone tissue and tumors, generating axial tomographic images in DICOM format.

[0062] The contrast between tumors and soft tissues is highlighted through T1-weighted, T2-weighted, and enhanced sequences of MRI scans;

[0063] Two-dimensional X-ray images are acquired in real time during the operation and the data is transmitted through the DICOM interface of the C-arm.

[0064] Specifically, based on density differences in CT images, Hounsfield units (HU) are used to distinguish different tissues: bone tissue typically exhibits high HU values ​​(e.g., above 1000 HU); tumors (such as malignant tumors) may have low HU values ​​(typically between 0 and 20 HU, depending on the tumor type and its internal composition).

[0065] MRI uses different weighted imaging methods to highlight the contrast between tumors and soft tissues, including T1-weighted imaging, T2-weighted imaging, and enhanced sequences;

[0066] Among them, T1-weighted imaging can display the details of anatomical structures and soft tissues with good contrast. T1-weighted images are suitable for observing the size, margins, and relationship of tumors with surrounding tissues. T2-weighted imaging has good contrast for tissues with high water content (such as tumors), and tumors usually appear as high signal (bright white) in T2 images. Enhanced sequences enhance the signal in the tumor area by injecting contrast agents (such as Gd-DTPA), and are particularly suitable for observing tumor vascularization.

[0067] Intraoperative real-time imaging is used for tumor localization and navigation during the operation, and the C-arm device displays the patient's internal structure in real time through two-dimensional X-ray images.

[0068] See Figure 3 As shown, cross-modal feature alignment is performed based on the acquired multi-source image data to construct a three-dimensional bone tumor image model in a three-dimensional voxel space coordinate system, and image artifacts are eliminated. Specifically, this includes:

[0069] Global alignment of multi-source image data based on landmarks in CT image data and soft tissue contours in MRI;

[0070] Local alignment of multi-source image data is achieved by compensating for the differences in soft tissue deformation between CT and MRI.

[0071] Using the DICOM coordinate system of preoperative CT as a reference, MRI and C-arm data are mapped to the CT coordinate system through a registration matrix to construct a three-dimensional voxel space coordinate system.

[0072] Voxel-weighted fusion was performed based on the proportion of bone tumor structures and soft tissue in bone tumors.

[0073] Based on the fused image data, it is matched with the intraoperative C-arm data, and image artifacts are eliminated through deep learning;

[0074] A three-dimensional bone tumor imaging model was constructed in a three-dimensional voxel space coordinate system based on the fused image data after artifact removal processing.

[0075] Specifically, in CT images, the selection of landmarks mainly involves the edges of bone structures or tumor boundaries. During global alignment, rigid registration is used, employing rotation and translation to achieve global alignment while maintaining the positional relationship of the landmarks between the two images. Affine registration can also be used, employing scaling and shearing transformations to increase the flexibility of the registration.

[0076] During local alignment, a non-rigid registration method is used to compensate for the soft tissue deformation differences between CT and MRI images. The deformation field between the images is calculated, and the alignment of the MRI and CT images is locally adjusted. The calculation of the deformation field is typically based on image similarity metrics (such as mutual information and mean squared error), and the formula is as follows:

[0077] Y = X + D(X)

[0078] Where Y is the deformed point, X is the original point, and D(X) is the deformation field;

[0079] Through the aforementioned rigid and non-rigid registration steps, a registration matrix is ​​obtained that maps MRI and C-arm data to the CT coordinate system, and the MRI and C-arm images are transformed from their original coordinate system to the coordinate system of the preoperative CT.

[0080] Weighting coefficients for different tissue types are defined based on the ratio of bone tissue to tumor soft tissue in CT images.

[0081] For bone tissue, high-density areas on CT are given higher weight; for soft tissue, high-signal areas on MRI are given higher weight. The weighted fusion formula is as follows:

[0082]

[0083] in, , The weighting coefficients for bone tissue and tumor soft tissue are... , Voxel values ​​in CT and MRI images;

[0084] Based on the processed fused image data, a three-dimensional model is constructed using a three-dimensional reconstruction algorithm, and the reconstructed three-dimensional bone tumor image model is further analyzed.

[0085] See Figure 4 As shown, based on the acquired three-dimensional bone tumor image data, a deep learning segmentation network is used to segment the bone tumor and identify the invasion boundary of the bone tumor. Specifically, the three-dimensional morphological parameters and feature parameters of the bone tumor region are obtained, including:

[0086] Based on the acquired three-dimensional bone tumor image data, the bone tumor region is initially segmented using an image segmentation algorithm to obtain the image portion containing the bone tumor.

[0087] Based on morphological operations, the image portion containing bone tumors is further segmented and refined. The segmentation accuracy is improved by using a convolutional neural network to obtain voxel-level bone tumor image data.

[0088] Based on the characteristics of bone tumors, each voxel is determined to be a bone tumor voxel and classified accordingly;

[0089] Based on the voxel determination results of all image portions containing bone tumors, the portions determined to be non-bone tumor voxels are removed to obtain image data containing only bone tumors.

[0090] Based on the acquired image data containing only bone tumors, the invasion boundary of bone tumors is identified, and the smoothness of the boundary is analyzed to assess whether the tumor has invaded the adjacent area.

[0091] Based on the identified bone tumor invasion boundaries, the invasion boundaries are annotated in the three-dimensional bone tumor image model;

[0092] Based on the labeled bone tumor image data, morphological parameters and characteristic parameters of bone tumors were extracted.

[0093] Specifically, image data containing bone tumors are initially segmented through image segmentation algorithms. Then, morphological operations such as dilation, erosion, opening and closing operations are performed to remove small artifacts and noise, smooth the boundaries, and improve the segmentation quality. The 3D U-Net model is trained for refined segmentation of bone tumor image data through deep learning methods, and the data is output as voxel-level bone tumor image segmentation data.

[0094] Based on spatial consistency, each voxel is used to determine whether it is a bone tumor voxel. The invasion boundary of the tumor is used to determine whether the tumor has invaded adjacent tissues. The edge detection algorithm of the Sobel operator is used to extract the tumor boundary. The formula is as follows:

[0095]

[0096] in, The gradient of the image in the horizontal and vertical directions;

[0097] The smoothness of the tumor boundary is used to assess whether it has invaded adjacent areas. Smoothness can be represented by the boundary curvature, as shown in the formula:

[0098]

[0099] in, For position functions on the boundary, For arc length parameters;

[0100] Based on the labeled bone tumor image data, morphological parameter extraction includes extracting morphological features such as tumor volume, surface area, aspect ratio, roundness, and surface complexity. Feature parameter extraction includes extracting tumor texture features, density distribution, heterogeneity, etc., to further analyze the invasiveness and malignancy of the tumor.

[0101] See Figure 5 As shown, the process of determining and classifying each voxel based on bone tumor characteristics to identify whether it is a bone tumor voxel specifically includes:

[0102] Calculate the spatial consistency of each voxel and its neighboring voxels. If the spatial consistency between voxels varies greatly, the voxel is located at the boundary of the bone tumor. If the spatial consistency varies little, the voxel is located inside the bone tumor.

[0103] Based on voxels located at the boundary of bone tumors, the alignment neighborhood is analyzed. The determination of whether a voxel belongs to a tumor is made by comparing its attribute characteristics with those of surrounding voxels, including grayscale difference, texture contrast, and uniformity.

[0104] Specifically, when determining the location of a voxel within a bone tumor, for each voxel, the grayscale difference between it and its neighboring voxels is calculated. Grayscale values ​​can be represented by the pixel values ​​of the image (such as Hounsfield units in CT or signal intensity in MRI), and the grayscale difference... Defined as:

[0105]

[0106] in, voxels grayscale value, voxels The neighborhood, The number of voxels in the neighborhood. For neighboring voxels;

[0107] Spatial consistency measures the grayscale consistency between a voxel and its neighboring voxels. It can be represented by the mean or variance of the grayscale difference, as shown in the formula:

[0108]

[0109] in, The maximum grayscale difference between all voxels and their neighborhoods;

[0110] A higher spatial consistency (close to 1) means that the voxel is located in the interior region of the tumor, while a lower spatial consistency (close to 0) means that the voxel is located at the boundary of the tumor.

[0111] Texture contrast is obtained from the contrast index in the gray-level co-occurrence matrix, using the following formula:

[0112]

[0113] in, represents the elements in the gray-level co-occurrence matrix, where i and j are the gray levels, measuring the intensity of gray level changes;

[0114] Uniformity refers to the evenness of gray-level distribution, representing the degree of concentration of pixel gray-level values ​​within an image region. It is calculated using the uniformity in the gray-level co-occurrence matrix, with the following formula:

[0115]

[0116] By combining features such as grayscale difference, texture contrast, and uniformity, machine learning classification methods (such as support vector machines and decision trees) are used to determine voxels based on these features. Is it a tumor?

[0117] See Figure 6 As shown, based on the acquired bone tumor region data, a biomimetic scaffold structure for the bone defect area is reconstructed through biomechanical modeling, and the biomimetic scaffold is implanted into the bone tumor region, specifically including:

[0118] Based on the obtained morphological and characteristic parameters of bone tumors, the mechanical properties of the bone defect area are obtained through biomechanical modeling.

[0119] Based on the mechanical properties and mechanical requirements of the bone tumor region, biomimetic materials were selected and a biomimetic scaffold structure was designed.

[0120] Mechanical simulations are performed using finite element analysis to simulate the load on the stent within the human body, optimizing stress and strain distribution and stent stability.

[0121] Specifically, the mechanical properties of bone tissue are usually represented by elastic modulus and Poisson's ratio. Assuming the elastic modulus of the tumor region is derived through a biomechanical model, a commonly used model is estimated based on tissue density, with the following formula:

[0122]

[0123] in, For elastic modulus, The density of the tumor region, , These are empirical parameters;

[0124] Finite element analysis was used to transform the geometry of the bone tumor into a finite element mesh. Material properties were defined, and necessary boundary conditions and external loads were applied to the model to simulate the actual usage environment. The finite element method was used to solve for the displacement, stress, and strain distribution of the structure under loading conditions. The stress of each voxel can be calculated using the stress-strain relationship, as shown in the formula:

[0125]

[0126] in, For strain, it can be derived from the displacement field and geometric relationships: , For voxel displacement, The length of the voxel;

[0127] Based on the model, the stress, strain, displacement and other results are evaluated to analyze the strength and stability of the structure.

[0128] See Figure 7 As shown, based on postoperative imaging data, the following parameters were obtained: bone-scaffold interface fusion rate, stress distribution uniformity, and new bone ingrowth depth.

[0129] By acquiring bone-scaffold interface fusion image data, the bone-scaffold interface is extracted, and the area where bone and scaffold are in contact is identified.

[0130] The bone-scaffold interface fusion rate is calculated based on the ratio of the bone-scaffold contact area to the total bone defect area.

[0131] The finite element method was used to simulate the skeletal mechanics after surgery and calculate the uniformity index of stress distribution, including stress standard deviation and stress concentration index.

[0132] Specifically, based on the segmentation results, the bone-scaffold contact area is extracted. In the 3D image, by constructing the bone-scaffold contact surface, the part where the bone tissue intersects with the scaffold surface is extracted. Based on the extracted bone-scaffold contact area and the total bone defect area, the bone-scaffold interface fusion rate is calculated.

[0133] The stress standard deviation, a metric for stress distribution uniformity, describes the uniformity of stress distribution. The calculation formula is as follows:

[0134]

[0135] in, Let be the stress value of the i-th voxel. This represents the average stress. The total number of prime numbers;

[0136] The stress concentration index, a measure of stress concentration phenomena in the uniformity index of stress distribution, is usually defined as the ratio of maximum stress to average stress, and the formula is:

[0137]

[0138] in, This represents the maximum stress value.

[0139] The uniformity of stress distribution is assessed using the stress standard deviation and stress concentration index. Ideally, a smaller stress standard deviation and a lower stress concentration index indicate a more uniform load distribution.

[0140] Furthermore, the method according to the embodiments of this application can also be achieved by means of... Figure 8 The architecture of the electronic device shown is used to implement this. For example... Figure 8 As shown, the electronic device 500 may include a bus 501, one or more CPUs 502, a read-only memory (ROM) 503, a random access memory (RAM) 504, a communication port 505 connected to a network, an input / output component 506, a hard disk 507, etc. The storage device in the electronic device 500, such as the ROM 503 or the hard disk 507, may store a perspective image analysis method for biomimetic repair of bone tumors provided in this application. The electronic device 500 may also include a terminal interface 508. Of course, Figure 8 The architecture shown is merely exemplary and can be omitted as needed when implementing different devices. Figure 8 One or more components in the illustrated electronic device.

[0141] Figure 9 This is a schematic diagram of a computer-readable storage medium structure provided in one embodiment of this application. Figure 9The diagram illustrates a computer-readable storage medium 600 according to one embodiment of this application. The computer-readable storage medium 600 stores computer-readable instructions. When executed by a processor, the computer-readable instructions can perform a perspective image analysis method for biomimetic repair of bone tumors according to an embodiment of this application, as described with reference to the above figures. The storage medium 600 includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0142] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0143] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0144] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for perspective image analysis for biomimetic repair of bone tumors, characterized in that, include: Acquire fluoroscopic imaging data of the bone tumor region, including X-ray computed tomography, magnetic resonance imaging, and intraoperative C-arm fluoroscopic images; Cross-modal feature alignment is performed based on the acquired multi-source image data to construct a three-dimensional bone tumor image model in a three-dimensional voxel space coordinate system and eliminate image artifacts. Based on the acquired three-dimensional bone tumor image data, the bone tumor region is initially segmented using an image segmentation algorithm to obtain the image portion containing the bone tumor. Based on morphological operations, the image portion containing bone tumors is further segmented and refined. The segmentation accuracy is improved by using a convolutional neural network to obtain voxel-level bone tumor image data. Calculate the spatial consistency of each voxel and its neighboring voxels. If the spatial consistency between voxels varies greatly, the voxel is located at the boundary of the bone tumor. If the spatial consistency varies little, the voxel is located inside the bone tumor. Based on voxels located at the boundary of bone tumors, the alignment neighborhood is analyzed, and whether a voxel belongs to a tumor is determined by comparing its attribute characteristics with those of surrounding voxels. These characteristics include: grayscale difference, texture contrast, and uniformity. Based on the voxel determination results of all image portions containing bone tumors, the portions determined to be non-bone tumor voxels are removed to obtain image data containing only bone tumors. Based on the acquired image data containing only bone tumors, the invasion boundary of bone tumors is identified, and the smoothness of the boundary is analyzed to assess whether the tumor has invaded the adjacent area. Based on the identified bone tumor invasion boundaries, the invasion boundaries are annotated in the three-dimensional bone tumor image model; Based on the labeled bone tumor imaging data, morphological parameters and characteristic parameters of bone tumors were extracted. Based on the acquired bone tumor region data, a biomimetic scaffold structure in the bone defect area was reconstructed through biomechanical modeling, and the biomimetic scaffold was implanted into the bone tumor region. Based on postoperative imaging data, the bone-scaffold interface fusion rate, stress distribution uniformity, and new bone ingrowth depth were obtained.

2. The method for perspective image analysis for biomimetic repair of bone tumors according to claim 1, characterized in that, The acquisition of fluoroscopic image data of the bone tumor region, including X-ray computed tomography, magnetic resonance imaging, and intraoperative C-arm fluoroscopic images, specifically includes: High-resolution spiral CT is used to acquire the fine structure and features of bone tissue and tumors, generating axial tomographic images in DICOM format. The contrast between tumors and soft tissues is highlighted through T1-weighted, T2-weighted, and enhanced sequences of MRI scans; Two-dimensional X-ray images are acquired in real time during the operation and the data is transmitted through the DICOM interface of the C-arm.

3. The method for perspective image analysis for biomimetic repair of bone tumors according to claim 1, characterized in that, The process of performing cross-modal feature alignment based on acquired multi-source image data, constructing a three-dimensional bone tumor image model in a three-dimensional voxel space coordinate system, and eliminating image artifacts specifically includes: Global alignment of multi-source image data based on landmarks in CT image data and soft tissue contours in MRI; Local alignment of multi-source image data is achieved by compensating for the differences in soft tissue deformation between CT and MRI. Using the DICOM coordinate system of preoperative CT as a reference, MRI and C-arm data are mapped to the CT coordinate system through a registration matrix to construct a three-dimensional voxel space coordinate system. Voxel-weighted fusion was performed based on the proportion of bone tumor structures and soft tissue in bone tumors. Based on the fused image data, it is matched with the intraoperative C-arm data, and image artifacts are eliminated through deep learning; A three-dimensional bone tumor imaging model was constructed in a three-dimensional voxel space coordinate system based on the fused image data after artifact removal processing.

4. The method for perspective image analysis for biomimetic repair of bone tumors according to claim 1, characterized in that, The process of reconstructing a biomimetic scaffold structure in the bone defect area based on the acquired bone tumor region data through biomechanical modeling, and then implanting the biomimetic scaffold into the bone tumor region specifically includes: Based on the obtained morphological and characteristic parameters of bone tumors, the mechanical properties of the bone defect area are obtained through biomechanical modeling. Based on the mechanical properties and mechanical requirements of the bone tumor region, biomimetic materials were selected and a biomimetic scaffold structure was designed. Mechanical simulations are performed using finite element analysis to simulate the load on the stent within the human body, optimizing stress and strain distribution and stent stability.

5. The method for perspective image analysis for biomimetic repair of bone tumors according to claim 1, characterized in that, The acquisition of bone-scaffold interface fusion rate, stress distribution uniformity, and new bone ingrowth depth based on postoperative imaging data specifically includes: By acquiring bone-scaffold interface fusion image data, the bone-scaffold interface is extracted, and the area where bone and scaffold are in contact is identified. The bone-scaffold interface fusion rate is calculated based on the ratio of the bone-scaffold contact area to the total bone defect area. The finite element method was used to simulate the skeletal mechanics after surgery and calculate the uniformity index of stress distribution, including stress standard deviation and stress concentration index.

6. An electronic device, characterized in that, include: At least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enables the at least one processor to perform a perspective image analysis method for biomimetic repair of bone tumors as described in any one of claims 1-5.

7. A computer-readable storage medium storing computer-readable instructions, characterized in that, When the computer-readable instructions are executed by a processor, they implement the perspective image analysis method for biomimetic repair of bone tumors as described in any one of claims 1-5.

Images