A method and system for risk assessment in orthopedic surgery

Abstract

This invention relates to the field of orthopedic surgical risk assessment, and in particular to a method and system for orthopedic surgical risk assessment. By acquiring preoperative and intraoperative data of a patient's orthopedic surgery, the method preprocesses the preoperative and intraoperative data, extracts the spatial features of bone microstructure from the initial orthopedic surgical data using a cascaded feature extraction network, calculates the correlation weights between imaging data and biological signals through a cross-modal attention mechanism, and constructs a patient-specific bone model based on finite element analysis to simulate the characteristic orthopedic surgical data in real time. The method simulates the distribution of joint contact stress at different osteotomy angles, the impact of micromotion thresholds at the bone implant interface and hemodynamic changes on the bone healing microenvironment after internal fixation device implantation, significantly reducing complications caused by improper operation and improving the surgical success rate.

Description

Technical Field

[0001] This invention relates to the field of medical security technology, and in particular to a method and system for risk assessment in orthopedic surgery. Background Technology

[0002] Traditional orthopedic surgical risk assessment mainly relies on static clinical indicators and two-dimensional imaging features, which has significant limitations. Existing technologies are mostly based on isolated data sources, lacking effective integration of multimodal data, and are unable to fully reflect individual patient differences. Traditional assessment models are only run once before surgery, and cannot perceive changes in bone-instrument interaction mechanics and fluctuations in vital signs in real time during surgery, resulting in delayed response to sudden risks. Existing finite element simulations mostly use population average parameters, ignoring key features such as patient-specific bone density distribution and microstructural anisotropy. Furthermore, postoperative complication prediction and long-term follow-up assessment are independent of each other, lacking a full-cycle risk management framework covering "intraoperative-perioperative-long-term". Summary of the Invention

[0003] The purpose of this invention is to solve the above-mentioned problems by designing a method and system for risk assessment in orthopedic surgery.

[0004] The technical solution of the present invention to achieve the above objectives is as follows: In an orthopedic surgery risk assessment method, the orthopedic surgery risk assessment method includes the following steps:

[0005] Acquire preoperative and intraoperative data of the patient's orthopedic surgery, and perform data preprocessing on the preoperative and intraoperative data to obtain initial orthopedic surgery data;

[0006] A cascaded feature extraction network is used to extract the spatial features of bone microstructure, the temporal signal of instrument stress, and the topological relationship of biomechanical conduction paths from the initial orthopedic surgical data; the association weights between image data and biological signals are calculated through a cross-modal attention mechanism to obtain characteristic orthopedic surgical data;

[0007] Based on finite element analysis, a patient-specific bone model was constructed to simulate the characteristic orthopedic surgical data in real time.

[0008] Real-time simulation of joint contact stress distribution under different osteotomy angles, the influence of bone implant interface micromotion threshold and hemodynamic changes on the bone healing microenvironment after internal fixation device implantation, and the risk status of orthopedic surgery are obtained.

[0009] Furthermore, in the aforementioned method for assessing the risk of orthopedic surgery, the acquisition of preoperative and intraoperative data of the patient's orthopedic surgery, and the preprocessing of the preoperative and intraoperative data to obtain initial orthopedic surgical data, include:

[0010] Acquire preoperative and intraoperative data of patients undergoing orthopedic surgery. The preoperative data includes at least: high-resolution CT images, MRI images, gait analysis data, serum bone metabolism markers, and susceptibility gene SNP loci; the intraoperative data includes at least: bone mineral density impedance signal, stress distribution of internal fixation devices, and time series of patient vital signs.

[0011] The CT and MRI images in the preoperative data were aligned spatially and temporally with the intraoperative data using an affine transformation matrix. Outliers in the temporal data of the intraoperative data were detected using the Grubbs test and deleted. The genetic data were processed using hierarchical standardization to obtain the first orthopedic surgical data.

[0012] After enhancing the imaging data in the first orthopedic surgery data, the gene data and phenotypic data will be corrected using KL divergence distribution to obtain the initial orthopedic surgery data.

[0013] Furthermore, in the aforementioned orthopedic surgical risk assessment method, the extraction of spatial features of bone microstructure, temporal signals of instrument stress, and topological relationships of biomechanical conduction paths from the initial orthopedic surgical data using a cascaded feature extraction network includes:

[0014] Define the cascaded feature extraction network as a triple. Its output characteristics satisfy the following formula:

[0015]

[0016] Where, Φ fuse Let X represent the cross-modal fusion function, S represent the image data, S represent the stress time series signal, and A represent the biomechanical topological adjacency matrix. Represents a 3D convolutional neural network; Indicates an LSTM network; Describing a GNN (Graph Neural Network), F total Indicates output features;

[0017] The spatial features of bone microstructure in the initial orthopedic surgical data are extracted using the 3D convolutional neural network; the temporal features of instrument stress in the initial orthopedic surgical data are extracted using an LSTM network; thus, spatiotemporal orthopedic surgical data are obtained.

[0018] The topological relationships of biomechanical conduction paths are established based on graph neural networks (GNNs), and a biomechanical conduction graph is defined. node This represents the bone substructure segmented from CT images, with edge e. ij ∈ε represents the force transmission path determined based on finite element simulation, with edge weight ω. ij =exp(-||fi -f j || 2 ), f i and f j Represents the mechanical eigenvectors of nodes i and j;

[0019] Using the RGCN relational graph convolutional network to convolve anisotropic graphs:

[0020]

[0021] in, This represents the hidden state of node i at level l+1, where l represents the hidden level number and R represents a predefined set of biomechanical relation types. This represents the set of neighboring nodes of node i under relation r. Indicates the size of the neighbor node set, used for normalization; This represents the weight matrix of relation r at layer l, used to transform the hidden states of neighboring nodes. ReLU represents the activation function, which performs a nonlinear transformation on the weighted summation result.

[0022] Furthermore, in the aforementioned orthopedic surgery risk assessment method, the step of calculating the association weights between image data and biological signals through a cross-modal attention mechanism to obtain characteristic orthopedic surgery data includes:

[0023] Define the joint embedding mapping of image modalities and biological signal modalities in the feature extraction of the initial orthopedic surgical data in the cascaded feature extraction network;

[0024] A bidirectional query-key-value interaction structure is established, an intermodal weight matrix is ​​generated through a gating network, Kullback-Leibler divergence loss is added to promote attention sparsity, and the dynamic association weights of the cascaded feature extraction network are optimized.

[0025] By adjusting the correlation strength of the cascaded feature extraction network in extracting initial orthopedic surgical data through real-time biosignal change rate, characteristic orthopedic surgical data is obtained.

[0026] Furthermore, in the aforementioned orthopedic surgery risk assessment method, the step of constructing a patient-specific bone model based on finite element analysis to perform real-time scenario simulation of the characteristic orthopedic surgery data includes:

[0027] A patient-specific bone model was constructed based on finite element analysis. This model includes a spatially variable material model established based on the nonlinear relationship between CT values ​​and bone mineral density.

[0028] E(x)=α·ρ(x) β +γ·softplus(GSNP );

[0029] Where E(x) represents the material strength at voxel x, ρ(x) = a·HU(x) + b represents the apparent density at voxel x, a and b represent the correlation coefficients, and HU(x) represents the CT value of voxel x; G SNP The term represents the correction term for the effect of genetic traits on material strength. α = 2045, β = 1.55, and γ = 0.3 represent clinically calibrated parameters used to quantify the relationship between material strength and other variables. Softplus represents a nonlinear function used to adjust G... SNP Perform the transformation;

[0030] The patient-specific bone model also includes generating a conformal tetrahedral mesh using an improved Marching Cubes algorithm:

[0031]

[0032] in, The generated conformal tetrahedral mesh region is a set of 3D points that satisfy specific conditions; N represents the number of elements in the set, and V... i φ represents a point in three-dimensional space, which is the basic element that constitutes the grid. φ represents the signed distance function of the CT image. τ represents the threshold parameter. δ represents the adaptive offset based on the orientation of the trabecular bone. This represents the signed distance function, used to adjust the threshold based on the gradient direction of Φ;

[0033] GPU acceleration based on the preconditional conjugate gradient method is used to obtain the finite element solver for the specific skeleton model by constructing a block-diagonal preprocessing matrix of sparse stiffness matrix and performing parallel optimization of matrix-vector multiplication under the CUDA architecture.

[0034] Furthermore, in the aforementioned orthopedic surgical risk assessment method, the real-time simulation of joint contact stress distribution at different osteotomy angles, the influence of micromotion threshold and hemodynamic changes at the bone implant interface after internal fixation device implantation on the bone healing microenvironment, yields the orthopedic surgical risk status, including:

[0035] Real-time simulation of joint contact stress distribution at different osteotomy angles includes calculating the angle variation range with a resolution of 0.1° based on the aforementioned characteristic orthopedic surgical data, determining the peak value and distribution pattern of joint contact stress at each angle, and obtaining a risk outcome.

[0036] The micromotion threshold of the bone implant interface after internal fixation device implantation includes the strain signal of the piezoelectric film array on the implant surface, the change of bone acoustic wave propagation characteristics, and the interface gap data of intraoperative OCT imaging in the implantation characteristic orthopedic surgical data. A three-dimensional feature vector of micromotion amplitude, frequency, and spatial distribution is established to determine whether the current micromotion mode belongs to the high-risk type and obtain a Class II risk result.

[0037] The impact of hemodynamic changes on the bone healing microenvironment includes generating a four-dimensional bone healing progress map based on blood flow simulation results, quantifying local repair capacity using the MAI metabolic activity index, determining the warning line of MAI in a region for 5 consecutive minutes, and obtaining three types of risk results;

[0038] The risk status of orthopedic surgery is obtained by dynamically assigning weights to the three types of risk assessment results.

[0039] To achieve the above objectives, the technical solution of the present invention further includes an orthopedic surgery risk assessment system, comprising:

[0040] The surgical data acquisition module is used to acquire preoperative and intraoperative data of the patient's orthopedic surgery, and to perform data preprocessing on the preoperative and intraoperative data to obtain initial orthopedic surgical data.

[0041] The data feature fusion module is used to extract the spatial features of bone microstructure, the temporal signal of instrument stress, and the topological relationship of biomechanical conduction paths from the initial orthopedic surgical data using a cascaded feature extraction network; and to calculate the association weight between image data and biological signals through a cross-modal attention mechanism to obtain characteristic orthopedic surgical data.

[0042] The real-time scene simulation module is used to construct a patient-specific bone model based on finite element analysis to perform real-time scene simulation of the characteristic orthopedic surgical data.

[0043] The risk status assessment module is used to simulate the joint contact stress distribution under different osteotomy angles in real time, the impact of bone implant interface micromotion threshold and hemodynamic changes on the bone healing microenvironment after internal fixation device implantation, and obtain the risk status of orthopedic surgery.

[0044] Furthermore, in the aforementioned orthopedic surgical risk assessment system, the surgical data acquisition module includes the following sub-modules:

[0045] The acquisition submodule is used to acquire preoperative and intraoperative data of patients undergoing orthopedic surgery. The preoperative data includes at least: high-resolution CT images, MRI images, gait analysis data, serum bone metabolism markers, and susceptibility gene SNP loci; the intraoperative data includes at least: bone mineral density impedance signal, stress distribution of internal fixation devices, and time series of patients' vital signs.

[0046] The alignment submodule is used to align the CT and MRI images in the preoperative data with the intraoperative data in terms of space and time using an affine transformation matrix, to detect outliers in the temporal data of the intraoperative data using the Grubbs test and to delete the detected outliers, and to process the gene data using hierarchical standardization to obtain the first orthopedic surgery data.

[0047] A submodule is obtained, which enhances the image data in the first orthopedic surgery data, and then corrects the gene data and phenotypic data through KL divergence distribution to obtain the initial orthopedic surgery data.

[0048] Furthermore, in the aforementioned orthopedic surgical risk assessment system, the data feature fusion module comprises the following sub-modules:

[0049] Define a submodule for defining the joint embedding mapping of image modalities and biological signal modalities in the feature extraction of the initial orthopedic surgical data in the cascaded feature extraction network;

[0050] A submodule is established to build a bidirectional query-key value interaction structure. An intermodal weight matrix is ​​generated through a gating network. Kullback-Leibler divergence loss is added to promote attention sparsity and optimize the dynamic association weights of the cascaded feature extraction network.

[0051] The extraction submodule is used to adjust the correlation strength of the cascaded feature extraction network in extracting the initial orthopedic surgical data by adjusting the real-time biosignal change rate, so as to obtain feature orthopedic surgical data.

[0052] Its beneficial effects lie in effectively integrating fragmented data and eliminating noise interference by fusing preoperative imaging data with intraoperative biomechanical signals, forming structured initial orthopedic surgical data, and solving the problems of isolated data and inconsistent formats in traditional surgery. The cascaded feature extraction network overcomes the limitations of single-modal data; simultaneously, the cross-modal attention mechanism strengthens the semantic association between different modalities by calculating the correlation weights between imaging data and biosignals, achieving in-depth analysis of skeletal biomechanical behavior and instrument-bone interface interaction in the surgical scenario, avoiding the feature omission problem caused by modal fragmentation in traditional methods. The patient-specific bone model constructed based on finite element analysis can accurately reflect the geometric morphology, material properties, and biomechanical characteristics of individual bones, solving the problem of the disconnect between traditional general models and the actual patient situation. By simulating in real time the effects of different osteotomy angles and internal fixation implantation schemes on joint contact stress, interface micromotion threshold, and hemodynamics, it can provide surgeons with a pre-assessment of surgical outcomes and support dynamic adjustment of surgical plans during surgery. By quantitatively analyzing the impact of surgical procedures on the skeletal biomechanical environment, empirical surgical decisions are transformed into data-driven precision control, significantly reducing complications caused by improper operation and improving surgical success rates. Attached Figure Description

[0053] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention.

[0054] Figure 1 This is a schematic diagram of the first embodiment of an orthopedic surgery risk assessment method according to the present invention;

[0055] Figure 2 This is a schematic diagram of a second embodiment of an orthopedic surgery risk assessment method according to the present invention;

[0056] Figure 3 This is a schematic diagram of a third embodiment of an orthopedic surgery risk assessment method according to the present invention;

[0057] Figure 4 This is a schematic diagram of the first embodiment of an orthopedic surgery risk assessment system according to the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0059] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the word “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0060] The present invention will now be described in detail with reference to the accompanying drawings, such as... Figure 1 As shown, an orthopedic surgery risk assessment method includes the following steps:

[0061] Step 101: Obtain the preoperative and intraoperative data of the patient's orthopedic surgery, perform data preprocessing on the preoperative and intraoperative data to obtain the initial orthopedic surgery data;

[0062] Specifically, in this embodiment, preoperative and intraoperative data of the patient's orthopedic surgery are acquired. The preoperative data includes at least: high-resolution CT images, MRI images, gait analysis data, serum bone metabolism markers and susceptibility gene SNP loci; the intraoperative data includes at least: bone mineral density impedance signal, stress distribution of internal fixation devices, and time series of the patient's vital signs.

[0063] The spatial and temporal alignment of CT and MRI images in preoperative data with intraoperative data was performed using an affine transformation matrix. Grubbs' test was used to detect outliers in the temporal data of intraoperative data and to remove the detected outliers. Stratified standardization was used to process the genetic data to obtain the first orthopedic surgical data.

[0064] After enhancing the imaging data in the first orthopedic surgery data, the genetic and phenotypic data will be corrected using KL divergence distribution to obtain the initial orthopedic surgery data.

[0065] Step 102: Use a cascaded feature extraction network to extract the spatial features of bone microstructure, the temporal signal of instrument stress, and the topological relationship of biomechanical conduction paths from the initial orthopedic surgical data; calculate the association weight between image data and biological signals through a cross-modal attention mechanism to obtain characteristic orthopedic surgical data;

[0066] Specifically, in this embodiment,

[0067] Define the cascaded feature extraction network as a triple. Its output characteristics satisfy the following formula:

[0068]

[0069] Where, Φ fuseLet X represent the cross-modal fusion function, S represent the image data, S represent the stress time series signal, and A represent the biomechanical topological adjacency matrix. Represents a 3D convolutional neural network; Indicates an LSTM network; Describing a GNN (Graph Neural Network), F total Indicates output features;

[0070] Spatial features of bone microstructure in initial orthopedic surgical data were extracted using a 3D convolutional neural network; temporal features of instrument stress in the initial orthopedic surgical data were extracted using an LSTM network; thus obtaining spatiotemporal orthopedic surgical data.

[0071] The topological relationships of biomechanical conduction paths are established based on graph neural networks (GNNs), and a biomechanical conduction graph is defined. node This represents the bone substructure segmented from CT images, with edge e. ij ∈ε represents the force transmission path determined based on finite element simulation, with edge weight ω. ij =exp(-||f i -f j || 2 ), f i and f j Represents the mechanical eigenvectors of nodes i and j;

[0072] Using the RGCN relational graph convolutional network to convolve anisotropic graphs:

[0073]

[0074] in, This represents the hidden state of node i at level l+1, where l represents the hidden level number and R represents a predefined set of biomechanical relation types. This represents the set of neighboring nodes of node i under relation r. Indicates the size of the neighbor node set, used for normalization; This represents the weight matrix of relation r at layer l, used to transform the hidden states of neighboring nodes. ReLU represents the activation function, which performs a nonlinear transformation on the weighted summation result.

[0075] Define the joint embedding mapping of image modalities and biological signal modalities in the feature extraction of initial orthopedic surgical data in a cascaded feature extraction network;

[0076] A bidirectional query-key-value interaction structure is established, an intermodal weight matrix is ​​generated through a gating network, Kullback-Leibler divergence loss is added to promote attention sparsity, and the dynamic association weights of the cascaded feature extraction network are optimized.

[0077] By adjusting the correlation strength of the cascaded feature extraction network in the extraction of initial orthopedic surgical data through real-time biological signal change rate, characteristic orthopedic surgical data is obtained.

[0078] Step 103: Construct a patient-specific skeletal model based on finite element analysis to perform real-time scene simulation of characteristic orthopedic surgical data;

[0079] Specifically, in this embodiment, a patient-specific bone model is constructed based on finite element analysis. This patient-specific bone model includes a spatially variable material model established based on the nonlinear relationship between CT values ​​and bone density.

[0080] E(x)=α·ρ(x) β +γ·softplus(G SNP );

[0081] Where E(x) represents the material strength at voxel x, ρ(x) = a·HU(x) + b represents the apparent density at voxel x, a and b represent the correlation coefficients, and HU(x) represents the CT value of voxel x; G SNP The term represents the correction term for the effect of genetic traits on material strength. α = 2045, β = 1.55, and γ = 0.3 represent clinically calibrated parameters used to quantify the relationship between material strength and other variables. Softplus represents a nonlinear function used to adjust G... SNP Perform the transformation;

[0082] The patient-specific bone model also includes the generation of conformal tetrahedral meshes using an improved Marching Cubes algorithm:

[0083]

[0084] in, The generated conformal tetrahedral mesh region is a set of 3D points that satisfy specific conditions; N represents the number of elements in the set, and V... i Φ represents a point in three-dimensional space, which is the basic element that constitutes the grid. Φ represents the signed distance function of the CT image. τ represents the threshold parameter. δ represents the adaptive offset based on the orientation of the trabecular bone. This represents the signed distance function, used to adjust the threshold based on the gradient direction of Φ;

[0085] GPU acceleration based on the preconditional conjugate gradient method is achieved by constructing a block-diagonal preprocessing matrix of sparse stiffness matrix and performing parallel optimization of matrix-vector multiplication under the CUDA architecture to obtain a finite element solver for a specific skeletal model.

[0086] Step 104: Simulate the joint contact stress distribution under different osteotomy angles in real time, and the influence of the micromotion threshold and hemodynamic changes of the bone implant interface after internal fixation device implantation on the bone healing microenvironment to obtain the risk status of orthopedic surgery.

[0087] Specifically, in this embodiment, the real-time simulation of joint contact stress distribution under different osteotomy angles includes calculating the angle change range with a resolution of 0.1° based on characteristic orthopedic surgical data, determining the peak value and distribution pattern of joint contact stress at each angle, and obtaining a risk result.

[0088] The micromotion threshold of the bone implant interface after internal fixation device implantation includes the strain signal of the piezoelectric film array on the implant surface, the change of bone acoustic wave propagation characteristics, and the interface gap data of intraoperative OCT imaging in the implantation characteristic orthopedic surgical data. A three-dimensional feature vector of micromotion amplitude, frequency, and spatial distribution is established to determine whether the current micromotion mode belongs to the high-risk type and obtain the second-class risk result.

[0089] The impact of hemodynamic changes on the bone healing microenvironment includes generating a four-dimensional bone healing progress map based on blood flow simulation results, quantifying local repair capacity using the MAI metabolic activity index, determining the warning line of MAI in a region for 5 consecutive minutes, and obtaining three types of risk results;

[0090] The risk status of orthopedic surgery is obtained by dynamically assigning weights to the three types of risk assessment results.

[0091] A three-tiered surgical risk warning mechanism should be established based on the risk status of orthopedic surgery, including at least intraoperative warning, short-term complications, and long-term failure.

[0092] Specifically, in this embodiment, the intraoperative warning includes combining intraoperative LSC I laser speckle imaging with simulation prediction values. When the local blood flow decreases by 30% from the baseline and lasts for more than 2 minutes, the hemostatic device standby mode is activated. Based on the patient's anatomical variation data, a 3mm dynamic buffer zone is generated around the blood vessels and nerve pathways. When the device touches the buffer zone, an optical flashing warning is triggered.

[0093] Short-term complications include triggering gradient compression stocking wearing reminders and anticoagulant dosage adjustment recommendations when the peak blood flow velocity decreases by 40% and D-dimer >5 mg / L, based on lower extremity venous ultrasound elastography data.

[0094] Long-term failure includes constructing a digital twin of the patient based on dynamic monitoring data from 1 year post-surgery to predict the amount of prosthesis wear, the extent of osteolysis, and the optimal time window for revision surgery 5-10 years later.

[0095] Its beneficial effects lie in effectively integrating fragmented data and eliminating noise interference by fusing preoperative imaging data with intraoperative biomechanical signals, forming structured initial orthopedic surgical data, and solving the problems of isolated data and inconsistent formats in traditional surgery. The cascaded feature extraction network overcomes the limitations of single-modal data; simultaneously, the cross-modal attention mechanism strengthens the semantic association between different modalities by calculating the correlation weights between imaging data and biosignals, achieving in-depth analysis of skeletal biomechanical behavior and instrument-bone interface interaction in the surgical scenario, avoiding the feature omission problem caused by modal fragmentation in traditional methods. The patient-specific bone model constructed based on finite element analysis can accurately reflect the geometric morphology, material properties, and biomechanical characteristics of individual bones, solving the problem of the disconnect between traditional general models and the actual patient situation. By simulating in real time the effects of different osteotomy angles and internal fixation implantation schemes on joint contact stress, interface micromotion threshold, and hemodynamics, it can provide surgeons with a pre-assessment of surgical outcomes and support dynamic adjustment of surgical plans during surgery. By quantitatively analyzing the impact of surgical procedures on the skeletal biomechanical environment, empirical surgical decisions are transformed into data-driven precision control, significantly reducing complications caused by improper operation and improving surgical success rates.

[0096] In this embodiment, please refer to Figure 2 The second embodiment of the orthopedic surgery risk assessment method in this invention involves acquiring preoperative and intraoperative data of the patient's orthopedic surgery, performing data preprocessing on the preoperative and intraoperative data to obtain initial orthopedic surgery data, and including the following steps:

[0097] Step 201: Obtain preoperative and intraoperative data for the patient's orthopedic surgery. Preoperative data should include at least: high-resolution CT images, MRI images, gait analysis data, serum bone metabolism markers, and susceptibility gene SNP loci. Intraoperative data should include at least: bone mineral density impedance signal, stress distribution of internal fixation devices, and time series of the patient's vital signs.

[0098] Step 202: Align the CT and MRI images in the preoperative data with the intraoperative data in terms of space and time using an affine transformation matrix; use the Grubbs test to detect outliers in the time series data of the intraoperative data and delete the detected outliers; use hierarchical standardization to process the genetic data to obtain the first orthopedic surgery data.

[0099] Step 203: After enhancing the imaging data in the first orthopedic surgery data, the gene data and phenotypic data will be corrected by KL divergence distribution to obtain the initial orthopedic surgery data.

[0100] Its beneficial effects lie in the fact that by fusing preoperative imaging data (CT / MRI) with intraoperative biomechanical signals (instrument stress, hemodynamic parameters) and performing standardized preprocessing, fragmented data can be effectively integrated, noise interference eliminated, and structured initial orthopedic surgical data formed. This process solves the problems of isolated and inconsistent data formats in traditional surgery, providing high-quality, multi-dimensional input for subsequent analysis and ensuring the comprehensiveness and reliability of feature extraction.

[0101] In this embodiment, please refer to Figure 3 The third embodiment of the orthopedic surgery risk assessment method and system of the present invention includes the following steps: calculating the association weight between image data and biological signals through a cross-modal attention mechanism to obtain characteristic orthopedic surgery data:

[0102] Step 301: Define the joint embedding mapping of image modalities and biological signal modalities in the feature extraction of initial orthopedic surgical data in the cascaded feature extraction network;

[0103] Step 302: Establish a bidirectional query-key value interaction structure, generate an intermodal weight matrix through a gating network, add Kullback-Leibler divergence loss to promote attention sparsity, and optimize the dynamic association weights of the cascaded feature extraction network.

[0104] Step 303: Adjust the correlation strength of the cascaded feature extraction network in the extraction of initial orthopedic surgical data by adjusting the real-time biological signal change rate to obtain feature orthopedic surgical data.

[0105] Its beneficial effect is that the cross-modal attention mechanism strengthens the semantic association between different modalities by calculating the association weight between image data and biological signals (dynamically coupling the bone morphology in X-ray images with intraoperative stress sensor data), and achieves in-depth analysis of bone mechanical behavior and instrument-bone interface interaction in the surgical scene, avoiding the feature omission problem caused by modal fragmentation in traditional methods.

[0106] The above describes a method for assessing the risk of orthopedic surgery according to an embodiment of the present invention. The following describes a system for assessing the risk of orthopedic surgery according to an embodiment of the present invention. Please refer to [link / reference]. Figure 4 One embodiment of the orthopedic surgery risk assessment system of the present invention includes:

[0107] The surgical data acquisition module is used to acquire preoperative and intraoperative data of patients' orthopedic surgeries, and to perform data preprocessing on the preoperative and intraoperative data to obtain initial orthopedic surgical data.

[0108] The data feature fusion module is used to extract the spatial features of bone microstructure, the temporal signal of instrument stress, and the topological relationship of biomechanical conduction paths from the initial orthopedic surgical data using a cascaded feature extraction network; and to calculate the association weight between image data and biological signals through a cross-modal attention mechanism to obtain characteristic orthopedic surgical data.

[0109] The real-time scene simulation module is used to construct patient-specific bone models based on finite element analysis to perform real-time scene simulations of characteristic orthopedic surgical data.

[0110] The risk status assessment module is used to simulate the joint contact stress distribution under different osteotomy angles in real time, the impact of bone implant interface micromotion threshold and hemodynamic changes on the bone healing microenvironment after internal fixation device implantation, and obtain the risk status of orthopedic surgery.

[0111] A three-tiered surgical risk warning mechanism should be established based on the risk status of orthopedic surgery, including at least intraoperative warning, short-term complications, and long-term failure.

[0112] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. An orthopedic surgery risk assessment method, characterized by, The orthopedic surgery risk assessment method comprises the following steps: Obtaining preoperative data and intraoperative data of orthopedic surgery of a patient, and pre-processing the preoperative data and intraoperative data to obtain initial orthopedic surgery data; Using a cascaded feature extraction network to extract bone microstructure spatial features, instrument stress time series signals and topological relationships of biomechanical conduction paths in the initial orthopedic surgery data; calculating the correlation weight of image data and biological signals by a cross-modal attention mechanism to obtain feature orthopedic surgery data; Based on finite element analysis, a patient-specific bone model is constructed to simulate the real-time scene of the feature orthopedic surgery data, which includes a spatial variable material model established based on the nonlinear relationship between CT value and bone density: ； wherein, represents a material strength at a voxel, represents an apparent density at a voxel, and represents a correlation coefficient, represents a CT value at a voxel; represents a correction term of a genetic feature to a material strength, , , represents a clinical calibration parameter for quantifying a relationship of a material strength to other variables, represents a non-linear function for transforming , the patient-specific bone model further comprising generating a conforming tetrahedral mesh employing a modified Marching Cubes algorithm:​​​ ； wherein, represents a generated conformal tetrahedral mesh region, is a set of three-dimensional points satisfying a specific condition, represents the number of elements in the set, represents a point in a three-dimensional space, a basic element constituting a mesh, represents a signed distance function of a CT image, represents a threshold parameter, represents an adaptive offset based on the trabecular orientation, represents a signed distance function, used to adjust the threshold value according to the gradient direction, based on the GPU acceleration of the preconditioned conjugate gradient method, by constructing a block diagonal preconditioner of a sparse stiffness matrix, parallel optimization of matrix-vector multiplication under the CUDA architecture, to obtain the finite element solver of the specific bone model; Real-time simulation of joint contact stress distribution under different osteotomy angles, bone implant interface micromotion threshold and hemodynamic changes after internal fixation instrument implantation on the bone healing microenvironment to obtain the orthopedic surgery risk state.

2. The orthopedic surgery risk assessment method of claim 1, wherein, The obtaining preoperative data and intraoperative data of orthopedic surgery of a patient, and pre-processing the preoperative data and intraoperative data to obtain initial orthopedic surgery data comprises: Obtaining preoperative data and intraoperative data of orthopedic surgery of a patient, and pre-processing the preoperative data and intraoperative data to obtain initial orthopedic surgery data, including: Obtaining preoperative data and intraoperative data of orthopedic surgery of a patient, and pre-processing the preoperative data and intraoperative data to obtain initial orthopedic surgery data, including: Aligning the CT image and MRI image in the preoperative data and the space and time of the intraoperative data by an affine transformation matrix, detecting outliers in the time series data in the intraoperative data by Grubbs test, and deleting the detected outliers, processing the gene data by hierarchical standardization to obtain first orthopedic surgery data; 3. The orthopedic surgery risk assessment method of claim 1, wherein, After enhancing the image data in the first orthopedic surgery data, the gene data and phenotype data are corrected by KL divergence distribution to obtain initial orthopedic surgery data. The cascaded feature extraction network is defined as a triplet whose output features satisfy the following equation: ； wherein, denotes a cross-modal fusion function, denotes image data, denotes a stress time series signal, denotes a biomechanical topology adjacency matrix; denotes a 3D convolutional neural network; denotes an LSTM network; denotes a GNN graph neural network, denotes output features; The using a cascaded feature extraction network to extract bone microstructure spatial features, instrument stress time series signals and topological relationships of biomechanical conduction paths in the initial orthopedic surgery data comprises: Topology relationship of biomechanical conduction path is established based on a GNN graph neural network, and a biomechanical conduction graph is defined , nodes represent bone substructures segmented from CT images, edges represent force conduction paths determined based on finite element simulation, and edge weights , and represent mechanical characteristic vectors of nodes and nodes ; Using the 3D convolutional neural network to extract the bone microstructure spatial features in the initial orthopedic surgery data; extracting the instrument stress time series features in the initial orthopedic surgery data by the LSTM network; obtaining spatiotemporal orthopedic surgery data; ； wherein, represents a node at the i-th layer, represents the number of hidden layers, represents a predefined set of biomechanical relation types, represents a node in the set of neighbor nodes under the relation represents the size of the set of neighbor nodes for normalization; represents the relation at the i-th layer, for transforming the hidden states of the neighbor nodes ; represents an activation function, for nonlinearly transforming the result of the weighted sum.

4. The orthopedic surgery risk assessment method of claim 3, wherein, Using the RGCN relational graph convolution network to perform anisotropic graph convolution: The calculating the correlation weight of image data and biological signals by a cross-modal attention mechanism to obtain feature orthopedic surgery data comprises: Defining the joint embedding mapping of the initial orthopedic surgery data in the image modal and biological signal modal in the feature extraction of the cascaded feature extraction network; Establishing a bidirectional query-key value interaction structure, generating an inter-modal weight matrix through a gating network, adding a Kullback-Leibler divergence loss to promote attention sparsity, and optimizing the dynamic correlation weight of the cascaded feature extraction network; The cascade feature extraction network is adjusted by a real-time biological signal change rate to extract the correlation strength in the initial orthopedic surgery data, and feature orthopedic surgery data is obtained.

5. The orthopedic surgery risk assessment method of claim 1, wherein, The joint contact stress distribution under different osteotomy angles, the micro-motion threshold of the bone-implant interface after the implantation of internal fixation devices, and the influence of blood flow dynamics on the bone healing microenvironment are simulated in real time to obtain the orthopedic surgery risk state, including: The joint contact stress distribution under different osteotomy angles includes calculating the angle change interval with a resolution of 0.1°, judging the joint contact stress peak value and distribution pattern at each angle, and obtaining a type of risk results based on the feature orthopedic surgery data; The micro-motion threshold of the bone-implant interface after the implantation of internal fixation devices includes the strain signal of the surface piezoelectric film array implanted in the feature orthopedic surgery data, the bone acoustic wave propagation characteristic change, and the interface gap data of intraoperative OCT imaging, establishes a three-dimensional feature vector of the micro-motion amplitude, frequency, and spatial distribution, judges whether the current micro-motion mode belongs to a high-risk type, and obtains a second type of risk results; The influence of blood flow dynamics on the bone healing microenvironment includes generating a four-dimensional bone healing progress map based on blood flow simulation results, quantifying local repair capacity with MAI metabolic activity index, and judging the regional MAI continuous 5-minute warning line to obtain a third type of risk results; The three types of risk assessment results are dynamically weighted and allocated to obtain the orthopedic surgery risk state.

6. An orthopedic surgery risk assessment system, comprising: The orthopedic surgery risk assessment system includes the following modules: A surgery data acquisition module is configured to acquire preoperative data and intraoperative data of a patient's orthopedic surgery, pre-process the preoperative data and intraoperative data, and obtain initial orthopedic surgery data. A data feature fusion module is configured to extract bone microstructure spatial features, instrument stress time series signals, and topological relationships of biomechanical conduction paths in the initial orthopedic surgery data using a cascade feature extraction network, and calculate the correlation weight of image data and biological signals through a cross-modal attention mechanism to obtain feature orthopedic surgery data. A real-time scene simulation module is configured to simulate the feature orthopedic surgery data in real time based on a patient-specific bone model constructed by finite element analysis, wherein the patient-specific bone model includes a spatial variable material model established based on the nonlinear relationship between CT values and bone density. ； wherein, represents a material strength at a voxel, represents an apparent density at a voxel, and represents a correlation coefficient, represents a CT value of a voxel; represents a correction term of a genetic feature to a material strength, , , represents a clinical calibration parameter for quantifying a relationship of a material strength to other variables, represents a non-linear function for transforming , the patient-specific bone model further comprising generating a conforming tetrahedral mesh employing a modified Marching Cubes algorithm:​​​ ； wherein, represents a generated conformal tetrahedral mesh region, is a set of three-dimensional points satisfying a specific condition, represents the number of elements in the set, represents a point in a three-dimensional space, a basic element constituting a mesh, represents a signed distance function of a CT image, represents a threshold parameter, represents an adaptive offset based on the trabecular orientation, represents a signed distance function, used to adjust the threshold according to the gradient direction, based on the preconditioned conjugate gradient method, through the construction of the block diagonal preconditioner of the sparse stiffness matrix, the parallel optimization of matrix-vector multiplication under the CUDA architecture, the finite element solver of the specific bone model is obtained; A risk state assessment module is configured to simulate the joint contact stress distribution under different osteotomy angles, the micro-motion threshold of the bone-implant interface after the implantation of internal fixation devices, and the influence of blood flow dynamics on the bone healing microenvironment in real time to obtain the orthopedic surgery risk state.

7. The orthopedic surgical risk assessment system of claim 6, wherein, The surgery data acquisition module includes the following sub-modules: An acquisition sub-module is configured to acquire preoperative data and intraoperative data of a patient's orthopedic surgery, wherein the preoperative data at least includes high-resolution CT images, MRI images, gait analysis data, serum bone metabolism markers, and susceptible gene SNP sites; and the intraoperative data at least includes bone density impedance signals, internal fixation device stress distribution, and patient vital sign time series. An alignment submodule is configured to align CT images and MRI images in the preoperative data and intraoperative data in space and time by using an affine transformation matrix, detect outliers in time-series data in the intraoperative data by using Grubbs test, delete the detected outliers, and process gene data by using hierarchical normalization to obtain first orthopedic surgery data. A obtaining submodule is configured to enhance image data in the first orthopedic surgery data, correct gene data and phenotype data by using KL divergence distribution, and obtain initial orthopedic surgery data.

8. The orthopedic surgical risk assessment system of claim 6, wherein, The data feature fusion module comprises the following submodules: A defining submodule is configured to define joint embedding mapping of image modalities and biological signal modalities in the initial orthopedic surgery data in cascade feature extraction network feature extraction. An establishing submodule is configured to establish a bidirectional query-key value interaction structure, generate an inter-modality weight matrix by using a gating network, add Kullback-Leibler divergence loss to promote attention sparsity, and optimize dynamic correlation weights of the cascade feature extraction network. An extracting submodule is configured to adjust correlation strength of the cascade feature extraction network in extracting the initial orthopedic surgery data by using real-time biological signal change rate to obtain feature orthopedic surgery data.

Images