A method and system for assessing the risk of sternal fixation surgery

Abstract

The application discloses a risk assessment method and system for sternal fixation surgery, and relates to the technical field of model simulation; the method comprises the following steps: acquiring image data of a target patient, and performing image preprocessing and key feature extraction; based on the extracted key features, a specific three-dimensional finite element model of the sternum of the target patient is constructed, and the assignment and fracture damage setting are integrated; the technical key points are as follows: the precise quantitative early warning of the specific complication of the postoperative cutting pad increase is realized, the technical problem that the micro-failure risk of the bone-implant interface is easily neglected in the traditional internal fixation design is solved, the overall scheme innovatively defines the cutting surface risk assessment region RAZ, the focus of the assessment is contracted from the macro fracture stability to the micro edge of the contact between the fixation and the bone, and the local stress concentration caused by osteoporosis is captured; the risk of the bone cutting surface collapse caused by osteoporosis is also quantified, and the intuitive effect that the fixation is prone to loosening at this position can be identified in the preoperative rehearsal.

Description

Technical Field

[0001] This invention relates to the field of model simulation technology, specifically to a risk assessment method and system for sternal fixation surgery. Background Technology

[0002] Current sternal fixation surgery mainly relies on the surgeon's clinical experience, with some using 3D visualization technology to aid in understanding anatomical structures. Fixation methods mainly include traditional stainless steel wire cerclage and various rigid internal fixation systems, such as titanium plates and nickel-titanium shape memory alloy cerclage devices. Preoperative risk assessment usually uses standards such as the ASA classification, focusing on evaluating the patient's overall physiological tolerance to surgery. In the field of computer simulation, there are already methods to design simulation models to simulate surgery, thereby virtually assessing preoperative risks and surgical planning.

[0003] However, existing surgical planning for sternal fractures has the following main technical shortcomings:

[0004] First, traditional approaches rely heavily on physician experience and two-dimensional imaging, lacking quantitative evidence. For example, judging bone quality solely by visual inspection makes it difficult to translate poor bone quality into specific material parameters, resulting in insufficient precision in the chosen fixation strategy. Second, existing assessment methods are often limited to single mechanical indicators, neglecting the impact of the patient's overall condition on local stability. Clinically, fixation methods that are tolerable for younger patients often lead to fixation component cutting or pull-out in elderly patients with osteoporosis. This one-size-fits-all approach to assessment easily causes fixation failure. Furthermore, traditional designs often focus only on macroscopic reduction, ignoring the microscopic failure risks at the bone-implant interface, especially in osteoporotic areas where increased cutting pads are prone to occur, such as micro-bone disintegration at the edges of the wires in the fixation component. If such risks cannot be predicted preoperatively, postoperative internal fixation loosening is highly likely, posing a risk of surgical failure. Summary of the Invention

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A risk assessment method for sternal fixation surgery, comprising:

[0007] Acquire image data of the target patient and perform image preprocessing and key feature extraction;

[0008] Based on the extracted key features, a specific three-dimensional finite element model of the sternum for the target patient is constructed, integrating assignment and fracture damage settings, and the osteoporosis level is determined according to the preset rule engine.

[0009] On a specific three-dimensional finite element model of the sternum, preliminary risk correction actions are performed to determine the preliminary fixation plan and complete the virtual surgical simulation. Simultaneously, the cutting risk analysis strategy is executed to perform the first and second corrections of surgical risks and generate secondary risk correction factors.

[0010] By integrating the results of initial risk correction actions with secondary risk correction factors, a comprehensive surgical risk assessment value corresponding to the target patient is defined. Based on the comprehensive surgical risk assessment value, a decision is made on whether to initiate an expert consultation mechanism to obtain a final fixed surgical plan.

[0011] Furthermore, the image data consisted of a sequence of chest computed tomography (CT) images. Image preprocessing involved using a deep learning-driven image segmentation network to automatically process the CT images, segmenting the overall contour of the sternum, the fracture line region, and adjacent key anatomical structures of the target patient. The segmentation result was output as a binary mask, where regions with a pixel value of 1 represented the sternum entity and regions with a pixel value of 0 represented the background. Key feature extraction involved applying an edge detection algorithm combined with morphological operations based on the binary mask and the grayscale gradient information of the original CT images to locate the spatial orientation of the fracture line and extracting the proposed cutting surface in three-dimensional space based on the spatial orientation. For the sternal region of the target patient, the average HU value was calculated, and an osteoporosis index (OI) was extracted and introduced. Key features included the cutting surface and the osteoporosis index (OI).

[0012] Furthermore, the construction process of the specific sternal 3D finite element model is as follows: Model reconstruction: The binary mask is imported into reverse engineering software, and a target patient-specific sternal 3D surface mesh is generated through the isosurface extraction algorithm; Subsequently, the surface mesh is post-processed, including smoothing and hole repair; Finite element mesh generation: The post-processed surface mesh is imported into finite element pre-processing software to generate a 3D solid mesh; Local mesh refinement technology is used near the fracture line of the target patient and in the planned cutting surface area to capture the stress concentration phenomenon in the cutting surface area.

[0013] Furthermore, the integrated assignment and fracture damage setting are as follows: based on the Osteoporosis Index (OI), non-uniform material properties are assigned to the elements in the specific sternal 3D finite element model; in the specific sternal 3D finite element model, the fracture line region is defined as the initial crack or weak connection interface; for the cutting surface region, a virtual cutting gap is pre-created in the specific sternal 3D finite element model, and its contact properties are defined. Fixed constraints are applied to the distal end of the specific sternal 3D finite element model to complete the fracture damage setting; the process of determining the osteoporosis level according to the rule engine is as follows: when the Osteoporosis Index (OI) < 0.7, the osteoporosis level corresponds to severe osteoporosis; when 0.7 ≤ OI < 0.85, the osteoporosis level corresponds to moderate osteoporosis; when the OI ≥ 0.85, the osteoporosis level corresponds to normal.

[0014] Furthermore, the initial risk correction action is based on the fracture condition and proceeds as follows: Without applying any internal fixation, a standardized physiological load F_physio is applied to a specific sternal 3D finite element model; simulation calculations are performed to obtain the relative displacement D_fracture and maximum principal stress σ_max_fracture of the fracture ends under the physiological load F_physio; based on the simulation results, an initial risk correction factor PRCF is defined; the osteoporosis grade of the target patient is retrieved, and if the osteoporosis grade is determined to be abnormal, a correction action is triggered, generating a corrected risk correction factor PRCF_adj based on the initial risk correction factor PRCF; and the initial fixation plan is determined based on the corrected risk correction factor PRCF_adj.

[0015] Furthermore, the initial risk correction factor PRCF is defined as follows: PRCF = δ × (D_fracture / D_crit) + ε × (σ_max_fracture / σ_crit); where δ and ε are both weighting coefficients, and the sum of δ and ε is 1; D_crit and σ_crit are the clinically acceptable critical displacement threshold and critical stress threshold, respectively; the triggering correction action is as follows: introducing the osteoporosis correction factor OMF, where OMF = 1.5Q when the osteoporosis grade corresponds to severe osteoporosis; and OMF = 1.2Q when the osteoporosis grade corresponds to moderate osteoporosis; Q represents the defining index, with a value range of (0, 1]; the corrected risk correction factor PRCF_adj is generated as follows: PRCF_adj = PRCF × OMF; the determination of the initial fixation plan is based on... The method is as follows: The corrected risk correction factor PRCF_adj is compared with the preset risk threshold. When the corrected risk correction factor PRCF_adj is less than the lower limit of the risk threshold, the first fixed strategy is retrieved from the pre-configured solution library and used as the initial fixed strategy. When the lower limit of the risk threshold is less than or equal to the corrected risk correction factor PRCF_adj and less than the upper limit of the risk threshold, the second fixed strategy is retrieved from the pre-configured solution library and used as the initial fixed strategy. When the corrected risk correction factor PRCF_adj is greater than or equal to the upper limit of the risk threshold, the third fixed strategy is retrieved from the pre-configured solution library and used as the initial fixed strategy.

[0016] Wherein, risk threshold = [lower value of risk threshold, upper value of risk threshold].

[0017] Furthermore, the basis for implementing the cutting risk analysis strategy is as follows: During the virtual surgical simulation, the same physiological load F_physio is applied again, and the mechanical response near the cutting surface area is analyzed. The cutting risk is quantified by monitoring the maximum equivalent stress σ_vM_cut of the corresponding element at the cutting surface, and a secondary risk correction factor SRCF is derived. The specific process is as follows: Define the cutting surface risk assessment region RAZ; extract the target stress σ_vM of all finite element elements within the RAZ; determine the local yield strength σ_yield(OI) of the bone tissue; calculate the maximum equivalent stress ratio and quantify the cutting risk: find the maximum value of all single target stresses of finite element elements within the RAZ, denoted as σ_vM_cut; then calculate the ratio of σ_vM_cut to the local yield strength, which is the maximum equivalent stress ratio SR; map the maximum equivalent stress ratio to the secondary risk correction factor: map the maximum equivalent stress ratio SR to the secondary risk correction factor SRCF using a piecewise function.

[0018] Furthermore, the comprehensive surgical risk assessment value is defined using a multiplicative coupling model, and the specific operating principle of the multiplicative coupling model is: CSRS=BR×(1+PRCF_adj)×(1+SRCF); where CSRS is the comprehensive surgical risk assessment value, BR is the basic risk score, and the ASA classification is used for assignment; when ASA I-II, BR=1, as the baseline; when ASA III, BR=1.5; when ASA IV and above, BR=2.

[0019] Furthermore, the decision on whether to initiate the expert consultation mechanism is based on the following: a preset risk threshold CSRS_threshold is set. If the comprehensive surgical risk assessment value CSRS ≤ the risk threshold CSRS_threshold, the patient directly enters the surgical preparation stage. Otherwise, the expert consultation mechanism is initiated: a digital medical record containing a three-dimensional sternal model of the target patient, fracture details, osteoporosis index, virtual surgical simulation animation, stress cloud map, and a detailed CSRS breakdown report is automatically generated and pushed to a multi-disciplinary expert group. Based on the digital medical record, multiple alternative plans are derived. For each alternative plan, the process is repeated on a specific three-dimensional finite element model of the sternum until a target that meets the criteria for directly entering the surgical preparation stage is found, and this is taken as the final fixed surgical plan.

[0020] A risk assessment system for sternal fixation surgery, comprising a data acquisition module for acquiring image data of the target patient and performing image preprocessing and key feature extraction;

[0021] Model building module: Based on the extracted key features, a specific three-dimensional finite element model of the sternum for the target patient is constructed, integrating assignment and fracture damage settings, and determining the osteoporosis level according to the preset rule engine;

[0022] Analysis and Correction Module: On a specific sternal 3D finite element model, preliminary risk correction actions are performed to determine the preliminary fixation plan and complete the virtual surgical simulation. Simultaneously, the cutting risk analysis strategy is executed to perform the first and second corrections of surgical risks and generate secondary risk correction factors.

[0023] The decision-making module integrates the results of the initial risk correction actions with the secondary risk correction factors, defines the comprehensive surgical risk assessment value corresponding to the target patient, and decides whether to initiate the expert consultation mechanism based on the comprehensive surgical risk assessment value to obtain the final fixed surgical plan.

[0024] This invention provides a method and system for risk assessment in sternal fixation surgery, which has the following beneficial effects:

[0025] (1) This scheme uses deep learning segmentation combined with finite element modeling to construct a patient-specific three-dimensional mechanical model of the sternum; introduces osteoporosis index to assign non-uniform values ​​to material properties, the model can truly reflect the weakness of the patient's bone, and achieves the effect of transforming the abstract concept of poor bone quality into a specific Young's modulus value, so that the subsequent mechanical simulation is more in line with the actual physiological state of the patient, providing a solid data foundation for formulating scientific fixation strategies, solving the technical problems of strong subjectivity and lack of quantitative basis in traditional surgical planning, and realizing the individualization and precision of surgical plan formulation.

[0026] (2) On the one hand, this scheme comprehensively considers the relative displacement D_fracture and the maximum principal stress σ_max_fracture of the fracture ends under load by defining the initial risk correction factor PRCF, and uses weight to balance the influence of the two failure modes; on the other hand, for osteoporosis patients, the osteoporosis correction coefficient is introduced to generate the corrected risk correction factor PRCF_adj, which amplifies the risk level of high-risk groups, achieves the effect of stratified early warning, solves the technical problem that a single mechanical index cannot fully assess the stability of fracture, and constructs a dual risk assessment mechanism that takes into account both micromotion and stress.

[0027] Meanwhile, the corrected risk correction factor PRCF_adj is not only the basis for risk stratification, but also the core hub connecting the patient's biological characteristics and biomechanical stability. It not only transforms simple mechanical indicators into a comprehensive score that integrates the patient's bone quality by introducing the osteoporosis correction factor OMF, solving the technical problem that a mechanical environment that young patients can tolerate may fail in older osteoporotic patients; it also directly determines the strength setting of subsequent fixation schemes, achieving a dual logical effect of identifying the stability of the fracture itself and assessing whether the patient's bone can withstand the force. Ultimately, it achieves the precise treatment goal of avoiding excessive surgical damage to low-risk patients while ensuring the fixation reliability of high-risk patients.

[0028] (3) This solution solves the technical problem that traditional internal fixation design easily overlooks the risk of microscopic failure at the bone-implant interface, and realizes accurate quantitative early warning of the specific complication of increased cutting pads after surgery; the overall solution innovatively defines the risk assessment area RAZ of the cutting surface, narrowing the assessment focus from macroscopic fracture stability to the microscopic edge of the wire-bone contact, capturing local stress concentration caused by osteoporosis; it also quantifies the risk of bone cutting surface fracture caused by osteoporosis, achieving the intuitive effect of identifying, for example, the wire is easy to loosen in the preoperative rehearsal. If the secondary risk correction factor SRCF shows that the risk is too high, it will guide medical staff to make adjustments to the wire position or increase bone cement reinforcement, thereby significantly improving the long-term stability of internal fixation and the success rate of surgery; in addition, the secondary risk correction factor SRCF is also integrated into the multiplicative coupling model of the comprehensive surgical risk assessment value CSRS. When the comprehensive surgical risk assessment value CSRS exceeds the risk threshold, it will automatically trigger the expert group consultation mechanism and push digital medical records containing the risk attribution corresponding to the secondary risk correction factor SRCF to assist multidisciplinary experts in formulating a safer final surgical plan. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the method flow in Embodiment 1 of the present invention;

[0030] Figure 2 This is a schematic diagram of the system operation in Embodiment 2 of the present invention. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0032] Example 1:

[0033] Please see Figure 1This embodiment provides a risk assessment method for sternal fixation surgery. The core idea of ​​this solution is as follows: using model simulation as the foundation, a three-dimensional model of the sternum and surrounding tissues is constructed with the help of computer technology to simulate the surgical procedure; then, combined with image analysis, the sternal fracture image is deeply analyzed to obtain the necessary information related to the fracture, providing accurate data support for model simulation, and an initial risk assessment is made for the fracture to assist medical personnel in determining the surgical plan for the target patient. At the same time, the risk level is further quantified by combining the patient's bone quality factors and the stress factors of the virtual surgical cutting, so as to determine the risk of the surgical plan. The initial risk assessment and the secondary quantification of the risk level are comprehensively evaluated to generate a final risk score suitable for the target patient, thereby further assisting medical personnel in changing or adjusting the surgical plan for the target patient, realizing a comprehensive assessment of surgical risks, and providing scientific guidance for clinical decision-making.

[0034] The specific steps of this assessment method are explained below:

[0035] S1. Acquire the image data of the target patient and perform image preprocessing and key feature extraction;

[0036] The target patients refer to those who require sternal fixation surgery. This step aims to provide the necessary quality of data input for subsequent model building and simulation analysis, and its specific content includes:

[0037] S1.1 Data Acquisition: Acquire the image data of the target patient. In this embodiment, the image data refers to the chest computed tomography (CT) image sequence, i.e., CT images. The CT images should cover the entire sternal region of the target patient, with a slice thickness of no more than 1.0 mm, to ensure that the details of the fracture line and the trabecular structure can be clearly distinguished.

[0038] S1.2 Image Preprocessing: A deep learning-driven image segmentation network is used to automate the processing of CT images. Specifically, a 3D U-Net convolutional neural network model trained on a large labeled sternal CT dataset is used to perform pixel-level classification of the original CT images, segmenting the overall contour of the sternum, fracture line regions, and adjacent key anatomical structures. Among these, adjacent key anatomical structures include at least the costal cartilage junctions. The 3D U-Net model, through its configured encoder-decoder structure, effectively captures the 3D spatial context information of the sternum, achieving high-precision segmentation. The segmentation result is output as a binary mask; where regions with a pixel value of 1 represent the sternal entity, and regions with a pixel value of 0 represent the background.

[0039] It should be noted that the training of a large number of labeled sternal CT datasets refers to using hundreds to thousands of 3D CT images of the sternum, fracture lines, and other structures delineated by radiologists as supervisory signals to train the 3D U-Net model end-to-end, enabling it to learn the mapping relationship from the original CT images to the segmentation labels. The 3D spatial context information of the sternum refers to the continuous morphology, adjacency relationships, and anatomical location features of the sternum in the coronal, sagittal, and axial dimensions. The encoder-decoder structure generates a binary mask representation: the encoder extracts high-dimensional semantic features through layer-by-layer convolution and downsampling, and compresses the spatial size; the decoder gradually restores the spatial resolution through upsampling and skip connections, and finally outputs a voxel-level probability map of the same size as the input CT, which is thresholded (e.g., >0.5) and converted into a binary mask to obtain 1 corresponding to the sternum and 0 corresponding to the background.

[0040] S1.3 Key Feature Extraction: Based on the binary mask of the sternum corresponding to the target patient obtained in S1.2 and the grayscale gradient information of the original CT image, an edge detection algorithm combined with morphological operations is applied to locate the spatial direction of the fracture line. In this embodiment, the edge detection algorithm can specifically use the Canny operator. The cutting surface to be performed is extracted in three-dimensional space according to the spatial direction of the fracture line. At the same time, the average HU value, i.e., Hounsfield Unit, is calculated for the sternal region of the target patient. The HU value is a physical quantity that represents tissue density in CT images, and it is positively correlated with bone mineral density. In order to more accurately reflect the degree of osteoporosis of the target patient, an osteoporosis index OI is extracted and introduced. Therefore, the key features include: the cutting surface and the osteoporosis index OI.

[0041] The gray-level gradient information of the original CT image refers to the rate of change of HU values ​​between adjacent voxels, reflecting the tissue density boundary. The specific process of locating the spatial direction of the fracture line is as follows: high gradient regions are detected by the Canny operator as candidate edges, and then the direction of continuous fracture lines is extracted by morphological connection and thinning operations. The high gradient region is the neighborhood of voxels in the CT image whose rate of change of HU values ​​exceeds the preset value, which usually corresponds to the tissue boundary or fracture line. The preset value is set in advance and can be set according to actual needs. The specific value is not fixed, so it will not be elaborated here.

[0042] Specifically, when extracting the intended cutting surface in three-dimensional space based on the spatial orientation of the fracture line, it is necessary to determine the final required cutting surface in advance according to clinical surgical guidelines, either by the system or under the guidance of the corresponding physician. This cutting surface aims to remove sclerotic or irregular tissue at the fracture ends to create a good contact surface for subsequent internal fixation.

[0043] The derivation of the Osteoporosis Index (OI) is based on the following:

[0044] The sternal region of the target patient is the Region of Interest (ROI) of the sternum. The average Hounsfield unit (HU) value described above is the average HU value corresponding to the ROI. Additionally, reference HU values ​​for the sternum of healthy adults need to be retrieved from a pre-calibrated database of existing large-scale population studies. Then, the following formula is used: OI = a_1 × (μ_HU / HU_ref) + a_2; where a_1 and a_2 are calibration coefficients, with values ​​greater than 0, used to map the HU value to a clinically interpretable osteoporosis grade. Specific values ​​are set according to actual needs. μ_HU represents the average HU value of the ROI corresponding to the target patient, and HU_ref represents the reference average Hounsfield unit value of the sternum in CT images of healthy adults. This is typically obtained through large-scale clinical trials. The data is obtained through clinical research statistics. For example, CT scans of the sternum were performed on hundreds of adults without a history of osteoporosis, and their average HU value was calculated. This value represents the baseline density level under normal bone density and is used to standardize HU measurements among different patients, eliminating the influence of differences in equipment or scanning parameters. The derived formula is normalized to obtain a dimensionless ratio (μ_HU / HU_ref). The normalized ratio is mapped to a preset OI scale, such as 0~1. This mapping can be determined by fitting clinical data. For example, the osteoporosis grade diagnosed by DXA (dual-energy X-ray absorptiometry) can be used as the gold standard, and a_1 and a_2 can be calibrated in reverse to make the OI highly consistent with the clinical diagnosis. The core logic is: normalize the HU of healthy individuals as a benchmark, and then establish a quantitative index consistent with the clinical osteoporosis grading through linear calibration.

[0045] S2. Based on the extracted key features, construct a specific three-dimensional finite element model of the sternum for the target patient, integrate assignment and fracture damage settings, and determine the osteoporosis level according to the preset rule engine;

[0046] The construction process of the specific sternal finite element model is as follows:

[0047] S2.1 Model Reconstruction: The binary mask obtained in S1.2 is imported into reverse engineering software, such as Mimics or 3-Matic, and a target patient-specific sternal 3D surface mesh is generated using an isosurface extraction algorithm. Subsequently, the surface mesh is post-processed. The post-processing includes smoothing and hole repair to eliminate geometric defects caused by image noise. In this embodiment, the specific post-processing is smoothing. The isosurface extraction algorithm is based on the binary mask to extract isosurfaces with HU values ​​equal to the sternal threshold, generating a closed 3D surface mesh.

[0048] S2.2 Finite Element Mesh Generation: The post-processed surface mesh is imported into finite element preprocessing software, such as Hypermesh or Abaqus / CAE, to generate a three-dimensional solid mesh dominated by tetrahedrons or hexahedrons. The size of the three-dimensional solid mesh needs to undergo convergence analysis to ensure a balance between the accuracy of the simulation results and computational efficiency. Local mesh refinement technology is used near the fracture line of the target patient and in the planned cutting surface area to accurately capture the stress concentration phenomenon in the cutting surface area. The convergence analysis process is based on: progressively refining the mesh size, repeating the simulation and monitoring key outputs, such as maximum stress and displacement. In this embodiment, the maximum stress is taken as an example. When the change in the result is less than a preset threshold, the mesh is considered to have converged. The size that meets the preset accuracy requirements and has the lowest computational cost is selected. Near the fracture line of the target patient can refer to a range of 1 cm around the fracture line or defined according to actual needs. The application of local mesh refinement technology is to set smaller element sizes or use local refinement algorithms, such as h-refinement, in the neighborhood of the fracture line and cutting surface to ensure that the elements in these high gradient areas are dense enough to accurately analyze stress concentration and deformation behavior.

[0049] The integrated assignment and fracture damage setting are as follows: Based on the osteoporosis index (OI) calculated in S1.3, non-uniform material properties are assigned to the elements in the specific sternal 3D finite element model; in the specific sternal 3D finite element model, the fracture line region is defined as the initial crack or weak connection interface; for the cutting surface, a virtual cutting gap is pre-created in the specific sternal 3D finite element model and its contact properties are defined, and a fixed constraint is applied to the distal end of the specific sternal 3D finite element model, thereby completing the fracture damage setting.

[0050] Here, "element" refers to each volume element in the finite element mesh, such as a tetrahedral or hexahedral element. Each element represents a small volume of the sternum, and its Young's modulus E is independently assigned based on its location's OI value, achieving a non-uniform spatial distribution of material properties. "Assignment" refers to assigning values ​​to material properties. The sternum is considered an isotropic linear elastic material, and its Young's modulus E and Poisson's ratio ν are determined using the empirical formula: E = E_0 × (OI). γ And ν=0.3; where E_0 represents the reference Young's modulus of healthy bone, for example, E_0=10 GPa; γ represents an empirical index greater than 1, the specific value of which can be set according to actual needs. The empirical index γ is used to amplify the nonlinear weakening effect of osteoporosis on bone strength; ν=0.3 is the commonly used empirical value of isotropic linear elastic Poisson's ratio for bone tissue, which will not be elaborated here. This assignment strategy enables the model to truly reflect the mechanical weakness of the target patient's bone.

[0051] In this embodiment, the contact attribute specifically refers to the friction coefficient. The definition of the contact attribute is as follows: a surface-to-surface contact pair is set on both sides of the cutting surface. The normal direction adopts hard contact to prevent penetration, and the tangential direction adopts the existing Coulomb friction model. The friction coefficient is referenced from literature values ​​for bone-to-bone or surgical instrument-to-bone interfaces, and is usually selected in the range of 0.2 to 0.6 to ensure reasonable slippage behavior. Applying a fixed constraint means that all degrees of freedom are completely constrained at the node at the distal end of the sternum where it connects with the rib / cartilage, i.e., the node at the distal end mentioned above. This corresponds to UX=UY=UZ=0, simulating that it is firmly fixed by the thoracic structure, thereby providing stable mechanical boundary conditions. UX=UY=UZ=0 reflects that in the specific sternal three-dimensional finite element model, the displacement of this node in the three spatial directions of X, Y, and Z is constrained to zero, i.e., completely fixed, and no translation is allowed. This is used to simulate the boundary conditions where the structure is rigidly supported or firmly connected.

[0052] The process of determining the osteoporosis level based on the rule engine is as follows: when the Osteoporosis Index (OI) < 0.7, the osteoporosis level corresponds to severe osteoporosis; when 0.7 ≤ OI < 0.85, the osteoporosis level corresponds to moderate osteoporosis; when the OI ≥ 0.85, the osteoporosis level corresponds to normal or mild osteoporosis.

[0053] By adopting the above technical solution, a patient-specific three-dimensional mechanical model of the sternum was constructed using deep learning segmentation in S1 and finite element modeling in S2. By introducing the osteoporosis index to non-uniformly assign values ​​to material properties, the model can realistically reflect the weakness of the patient's bone, achieving the effect of transforming the abstract concept of poor bone quality into a specific Young's modulus value. This makes the subsequent mechanical simulation more in line with the patient's actual physiological state, providing a solid data foundation for formulating scientific fixation strategies. It solves the technical problems of traditional surgical planning being highly subjective and lacking quantitative basis, and realizes the individualization and precision of surgical planning.

[0054] S3. On the specific sternal three-dimensional finite element model, perform preliminary risk correction actions to determine the preliminary fixation plan, complete the virtual surgical simulation, simultaneously execute the cutting risk analysis strategy, perform the first and second corrections of surgical risks, and generate secondary risk correction factors.

[0055] The initial risk correction actions are based on the fracture condition, and the specific process is as follows:

[0056] S3.1. Without applying any internal fixation, apply a standardized physiological load to the specific sternal 3D finite element model. In this embodiment, this refers to the chest wall expansion force simulating coughing or deep breathing, with a magnitude of F_physio. S3.2. Through simulation calculation, obtain the simulation results: the relative displacement D_fracture of the fracture ends under load and the maximum principal stress σ_max_fracture. S3.3. Based on the simulation calculation results, define the initial risk correction factor PRCF. S3.4. Retrieve the osteoporosis grade of the target patient. If the osteoporosis grade is determined to be abnormal or mild, trigger the correction action. Generate a corrected risk correction factor PRCF_adj based on the initial risk correction factor PRCF, and use PRCF_adj as the standard in subsequent steps. S3.5. Determine the initial fixation plan based on the corrected risk correction factor PRCF_adj.

[0057] In S3.1, the standardized physiological load specifically refers to F_physio. F_physio is a standardized physiological load applied to the sternal region to simulate coughing or deep breathing. It is usually applied as distributed pressure in kPa or equivalent concentrated force in N, and its value is derived based on the following physiological basis:

[0058] 1. Range of intrathoracic pressure changes:

[0059] During coughing, the positive pressure inside the chest can reach +30~50 cmH2O (≈ 3–5 kPa);

[0060] During deep inhalation, the negative pressure inside the chest is approximately −8 to −10 cmH2O (≈ −0.8 to −1.0 kPa).

[0061] 2. Estimation of the stress area of ​​the sternum:

[0062] The surface area of ​​the sternum in adults is approximately 50–70 cm² (0.005–0.007 m²).

[0063] 3. Load Calculation:

[0064] Taking coughing as an example: F = pressure × area ≈ 4 kPa × 0.006 m² = 24 N;

[0065] Considering the leverage effect of the ribs and the transmission through muscles, the equivalent force acting on the sternum is usually taken as 100~300 N. For specific details, please refer to existing relevant literature on chest biomechanics and collision / breathing simulation studies. It should be noted that this load range has been adopted by many finite element studies, such as pectus excavatum correction and rib injury simulation, and its biomechanical representativeness has been verified through animal experiments or cadaver specimens. Using this standardized load can ensure the comparability of risk assessment between different patient models, while reflecting the impact of real physiological activities on fracture stability.

[0066] The simulation calculation in S3.2 is explained as follows: In the finite element software, the displacement values ​​u1 and u2 of the corresponding nodes on both sides of the fracture line in the load direction are extracted respectively. Then the relative displacement D_fracture = |u1-u2|. In the elements near the fracture end, i.e., within 5mm of the fracture line, the first principal stress of each integration point is extracted, and the maximum value among all elements is taken as the maximum principal stress σ_max_fracture. The derivation is explained as follows: The finite element solution is based on the equilibrium equation K×U=F, where K is the stiffness matrix, U is the nodal displacement vector, and F is the external load vector. After solving U, the strain ε and stress σ are calculated through the constitutive relation, and then the required index is output. This process can accurately reflect the mechanical response of the fracture area under physiological load. Specifically, the above equilibrium equation expresses that in the static equilibrium state of the structure, F is equal to the internal force (K×U) caused by the deformation U. Solving this linear equation system can obtain the displacement U of all nodes. The subsequent supplementary explanation is: Calculate the element strain from U: ε = B×U, where B is the strain-displacement matrix; Calculate the stress from the constitutive relation: σ = D×U, where D is the material elasticity matrix; perform principal stress decomposition on σ to obtain σ1, σ2, σ3; extract the required results, namely the required relative displacement D_fracture and the maximum principal stress σ_max_fracture.

[0067] The finite element software used in this embodiment refers to commercial open-source software for structural mechanics simulation, such as ANSYS, Abaqus, and LS-DYNA, which can be selected according to the needs. This embodiment addresses orthopedic biomechanics research, and Abaqus is chosen because it supports complex material models, such as anisotropic bone tissue, contact nonlinearity, and large deformation analysis. When establishing a specific 3D finite element model of the sternum, the fracture surface is explicitly modeled as two separate bone segments, namely the proximal and distal ends. Nodes are discrete points in the finite element mesh, and displacement is their basic degree of freedom. Corresponding nodes refer to a pair of nodes that are geometrically opposite and functionally symmetrical on the fracture surface; for example, matching points at the upper and lower edges or anterior and posterior edges of the fracture surface, or the motion of the entire bone segment represented by a rigid reference point (RRP). In practice, coupling sets or reference points are often created on both sides of the fracture to average the local node displacements, and then the motion of both is calculated. The displacement difference in the principal directions of the load yields the required relative displacement D_fracture. For the integration point, it represents the Gaussian point within the finite element used for numerical integration, precisely calculating the element stiffness matrix and stress-strain field. Since the displacement field is interpolated within the element, and stress / strain is calculated from the displacement derivative, direct calculation at the nodes would lead to discontinuities or errors. Therefore, stress is usually most accurate at the integration point. The data for the integration point depends on the element type; for example, an 8-node hexahedral element commonly uses 2×2×2=8 integration points. Because stress concentration often occurs at the fracture tip or edge, and this peak often appears within the element rather than at the nodes, the maximum value of the first principal stress is taken from all elements adjacent to the fracture, such as those ≤5mm from the fracture line. This more realistically captures potential micro-damage or crack propagation driving forces, thus yielding the required maximum principal stress σ_max_fracture.

[0068] In S3.3, the initial risk adjustment factor PRCF is defined as follows:

[0069] PRCF = δ × (D_fracture / D_crit) + ε × (σ_max_fracture / σ_crit); where δ and ε are weighting coefficients that can be adjusted based on clinical experience, ensuring that the sum of δ and ε is 1; D_crit and σ_crit are the clinically acceptable critical displacement threshold and critical stress threshold, respectively, set according to actual needs, and will not be elaborated here; (D_fracture / D_crit) and (σ_max_fracture / σ_crit) are all dimensionless; Definition: A comprehensive assessment using a two-dimensional failure mechanism is adopted. Fracture healing failure or internal fixation failure is usually dominated by two mechanisms: excessive micromotion and local stress concentration. Excessive micromotion corresponds to excessive relative displacement of the fracture ends, i.e., >D_crit, which disrupts the formation of new callus and leads to delayed healing or nonunion. Local stress concentration corresponds to excessive stress on the bone tissue or internal fixation material, i.e., >σ_crit, which causes microcrack propagation, bone resorption, or wire loosening. Therefore, focusing only on displacement or only on stress is insufficient to comprehensively assess the risk. The definition formula of the initial risk correction factor PRCF incorporates these two independent but complementary failure paths through linear combination to achieve a more robust risk assessment.

[0070] In S3.4, under the condition that the osteoporosis grade is determined to be abnormal or mild, the content that triggers the correction action is: introducing the osteoporosis correction factor OMF. When the osteoporosis grade corresponds to severe osteoporosis, then OMF=1.5Q; when the osteoporosis grade corresponds to moderate osteoporosis, then OMF=1.2Q.

[0071] Q represents the defining index, with a value range of (0, 1]. In this embodiment, Q is usually set to 1.

[0072] The corrected risk correction factor PRCF_adj is generated based on the following formula: PRCF_adj = PRCF × OMF. The logical explanation is as follows: Osteoporosis significantly reduces bone strength and stiffness, amplifying displacement and stress responses under the same load. Clinical studies show that patients with severe osteoporosis have low fracture healing rates and a high risk of internal fixation failure. Therefore, based on the mechanical simulation reflecting the impact of the osteoporosis index (OI) on material properties, further nonlinear weighting of the overall risk using the osteoporosis correction factor (OMF) more realistically reflects the patho-mechanical coupling mechanism from poor bone quality → poorer biomechanical stability → higher clinical risk. This correction action ensures that risk assessment not only relies on instantaneous mechanical responses, i.e., D_fracture and σ_max_fracture, but also incorporates the long-term prognostic impact of bone quality, avoiding underestimation of the fixation failure risk in osteoporotic patients, thereby guiding the selection of stronger internal fixation schemes, such as double plates, locking wires, or bone cement reinforcement.

[0073] In S3.5, the initial fixed scheme is determined as follows: the corrected risk correction factor PRCF_adj is compared with the preset risk threshold. When the corrected risk correction factor PRCF_adj < the lower limit of the risk threshold, the first fixed strategy is retrieved from the pre-configured scheme library and used as the initial fixed scheme. When the lower limit of the risk threshold ≤ the corrected risk correction factor PRCF_adj < the upper limit of the risk threshold, the second fixed strategy is retrieved from the pre-configured scheme library and used as the initial fixed scheme. When the corrected risk correction factor PRCF_adj ≥ the upper limit of the risk threshold, the third fixed strategy is retrieved from the pre-configured scheme library and used as the initial fixed scheme.

[0074] Wherein, the risk threshold = [lower risk threshold value, upper risk threshold value]; in this embodiment, the lower risk threshold value is greater than or equal to 1, and 1 can be taken as the baseline. Therefore, the upper risk threshold value exceeds 1 and can be taken as 1.5; if OMF = 1.3 and PRCF = 0.8, then PRCF_adj = 1.04. Although the original mechanical risk is not high, the overall risk has reached the critical point due to the patient's fragility, and further treatment is still required. Therefore, the second fixed strategy needs to be selected for interventional treatment.

[0075] Specifically, the first fixation strategy corresponds to observation / elastic fixation / single shape memory alloy circumferential device, the purpose of which is to avoid excessive surgery and reduce complications; the second fixation strategy is: standard titanium plate internal fixation, that is, ≥4 holes using multi-wire fixation, used to resist rotational torsion and strengthen fixation; the third fixation strategy is: enhanced fixation, that is, double plate cross fixation, combined with locking wire + bone cement, and postoperative chest strap immobilization combined with delayed weight-bearing, which can resist high micromotion and stress concentration and compensate for insufficient bone quality; regardless of which fixation strategy is ultimately selected as the initial fixation plan, it is only to provide a reference for medical staff.

[0076] The content of the virtual surgical simulation is described below:

[0077] The implantation procedure is performed in a specific sternal finite element model according to the preliminary fixation plan. For example, if the preliminary fixation plan adopts the second fixation strategy, the implantation procedure is based on standard titanium plate internal fixation + multi-wire fixation, and virtual cutting is performed according to the cutting surface planned in S1.3.

[0078] The basis for implementing the risk analysis strategy for cutting off the product is as follows:

[0079] During the virtual surgical simulation, the same physiological load F_physio was applied again, and the mechanical response near the cutting surface was analyzed. Due to osteoporosis, micro-bone fractures or increased cutting pads are prone to occur at the cutting surface edge during or after the procedure, resulting in an uneven cutting surface and additional bone fragments. This significantly affects fixation stability. This risk was quantified by monitoring the maximum equivalent stress σ_vM_cut of the cutting surface unit, and a secondary risk correction factor SRCF was derived. It is important to clarify the biomechanical nature of increased cutting pads: it does not refer to the actual amount of physical debris generated during surgery, but rather to the tendency of bone tissue in the cutting surface edge area to locally yield, propagate microcracks, or even micro-disintegrate when compressive or shear stress is applied by the internal fixation device due to insufficient bone strength caused by osteoporosis. This tendency directly weakens the contact integrity and load-bearing capacity of the bone-internal fixation interface, thereby reducing fixation stability. Therefore, the key to quantifying this risk is to assess whether the stress state of the bone tissue in the cutting surface area is close to or exceeds its local yield limit. The specific implementation process is as follows:

[0080] The first step is to define the Risk Assessment Zone (RAZ) for the cut surface:

[0081] After completing the virtual cutting, the focus is on a local area highly relevant to clinical significance. This local area is defined as an annular zone extending outward from the boundary of the cutting surface by a characteristic distance d, i.e., the cutting surface risk assessment zone (RAZ). The characteristic distance d is set empirically based on the size of the internal fixation device used in the initial fixation plan. The internal fixation device, also known as the fixation element, usually refers to steel wire. The characteristic distance d is taken as 2-3 times the diameter of the steel wire, usually 1.5mm to 3mm. This value has a clear biomechanical basis: as a flexible fixation material, the steel wire has a narrower contact area with bone tissue, and the stress concentration effect is more significant. Taking an annular zone of 2-3 times the diameter of the steel wire can cover the direct stress area of ​​the steel wire contacting the bone surface, and also include the surrounding transition area where microcracks are prone to propagate. For example, the clinically commonly used 0.8mm stainless steel wire has a d value of 2.0mm, which corresponds exactly to the effective compressive stress distribution range formed by the steel wire on the bone surface. This is consistent with the anatomical characteristics of the cortical bone thickness of the sternum, which is about 1.5 to 2.5mm. This ensures that the assessment area is neither too large to dilute the signal, nor too small to miss key risk points.

[0082] The second step is to extract the target stress of all finite element elements within the RAZ:

[0083] After applying the physiological load F_physio and completing the finite element solution, each finite element in the risk assessment region RAZ of the cutting surface is traversed, and the calculated target stress is extracted. In this embodiment, the target stress refers to the von Mises stress. The von Mises stress is a scalar value used to measure the distortion energy density of a material under complex multiaxial stress conditions. It is a classic criterion for judging whether ductile materials, such as cortical bone, will yield. Its mathematical expression is: σ_vM=√(0.5×[(σ1 - σ2)² + (σ2 - σ3)² + (σ3 - σ1)²]).

[0084] Wherein, σ_vM represents the target stress, and σ1, σ2, and σ3 are the three principal stresses of the finite element element, representing the normal stress in the characteristic direction of the three-dimensional stress state at that point. This mathematical expression simplifies the complex three-dimensional stress state into an equivalent value that can be directly compared with the uniaxial yield strength of the material. The target stress σ_vM is based on the distortion energy theory, i.e., the shape change energy density, and is used to measure whether a material is close to yielding under complex multiaxial stress. Its logic is that no matter how complex the actual stress state is, as long as σ_vM reaches the yield strength σ_y of the material in the uniaxial tensile test, it is considered that the material has begun to undergo plastic deformation. This criterion has high accuracy in predicting yield behavior for ductile materials, such as cortical bone, and is independent of hydrostatic pressure, reflecting only the distortion effect caused by shear. Therefore, in this embodiment, the target stress σ_vM can be used to evaluate the mechanical safety of bone or implants in biomechanics.

[0085] The third step is to determine the local yield strength σ_yield(OI) of the bone tissue:

[0086] The yield strength of bone tissue is not a constant value, but depends on its degree of osteoporosis. Using the osteoporosis index OI obtained in S1.3, the effective yield strength of bone tissue within the risk assessment zone RAZ of the cut surface is determined by a pre-calibrated empirical formula. This empirical formula can be expressed as: σ_yield(OI) = σ0×(OI)^η; where σ0 is the reference yield strength of healthy bone, for example, 150 MPa; η is a decay index greater than 1, used to reflect the nonlinear, accelerated weakening effect of osteoporosis on bone strength. The specific values ​​in this embodiment can be set according to requirements, which will not be elaborated here. The parameters of this empirical formula can be obtained by fitting in vitro bone biomechanical experimental data.

[0087] Step 4: Calculate the maximum equivalent stress ratio and quantify the risk (corresponding to the initial step):

[0088] Find the maximum value of all single-objective stresses in the RAZ, which is the maximum equivalent stress given above, and denote it as σ_vM_cut; then calculate the ratio of σ_vM_cut to the local yield strength, which is the maximum equivalent stress ratio SR: SR = σ_vM_cut / σ_yield (OI); this SR is the core risk quantification index; if SR does not exceed 1, it indicates that even under the maximum load, the bone stress in the cutting area has not reached the yield point, and the risk of increased cutting pad is low; if SR exceeds 1, it indicates that the local bone tissue has entered the plastic deformation or even failure stage, and the risk is extremely high.

[0089] The fifth step is to map the maximum equivalent efficiency ratio to a secondary risk correction factor (corresponding to secondary correction).

[0090] To seamlessly integrate the above mechanical analysis results into the overall risk assessment framework, the maximum equivalent stress ratio SR is mapped to a quadratic risk correction factor SRCF through a monotonically increasing function; the mapping function used in this embodiment can be a piecewise linear or a sigmoid function; for example, the following piecewise function is used:

[0091] When SR≤0.8Qt, SRCF=0, indicating that the risk is negligible;

[0092] When 0.8Qt < SR < 1.2Qt, SRCF = k × (SR - 0.8), which indicates that the risk increases linearly with stress.

[0093] When SR≥1.2Qt, SRCF=k×0.4Qt+m×(SR - 1.2), indicating that the risk increases sharply;

[0094] Where Qt represents the critical reference value, with a range exceeding 0, typically set to 1. It defines a dynamic safety range; for example, 0.8Qt implies an 80% safety margin, a common and standard technique in engineering design, and therefore will not be elaborated upon here. k and m are both adjustable weighting coefficients, with m > k, used to amplify the risk penalty in high-stress areas. Specific values ​​can be set according to actual needs. Furthermore, the aforementioned 0.4Qt serves as the cumulative risk base at the end of the middle interval of the piecewise function, derived from 1.2Qt - 0.8Qt = 0. 0.4Qt ensures that the starting point of the third stage, the section corresponding to a sharp increase in risk, does not start from zero, but rather follows the second stage, which is the section where risk increases linearly with stress, representing the accumulated risk value and the transition bandwidth between the safe zone and the dangerous zone. When Qt=1, then 0.4Qt=0.4, which means that a fluctuation of 20% above and below the critical reference value Qt is allowed. This 0.4 range is usually regarded as the elastoplastic transition zone or the micro-damage accumulation zone. Overall, the above method transforms the abstract mechanical simulation results into an intuitive secondary risk correction factor that can participate in the final decision.

[0095] In summary, the method described above does not simply read a stress value, but rather defines risk areas, applies mature yield criteria, combines individualized bone strength models, and ultimately transforms the results into clinically actionable risk factors, thereby scientifically and rigorously quantifying the risk of increased cutting pads caused by osteoporosis.

[0096] On the one hand, this scheme comprehensively considers the relative displacement D_fracture and the maximum principal stress σ_max_fracture of the fracture ends under load by defining an initial risk correction factor PRCF, and uses weights to balance the effects of the two failure modes. On the other hand, for osteoporosis patients, an osteoporosis correction factor OMF is introduced to generate a corrected risk correction factor PRCF_adj, which amplifies the risk level of high-risk groups. This achieves a stratified early warning effect, that is, low-risk patients can use conventional fixation, while high-risk patients automatically trigger a reinforced fixation strategy, realizing a shift from a one-size-fits-all approach to a step-by-step precision treatment, solving the technical problem that a single mechanical index cannot comprehensively assess fracture stability, and constructing a system that takes into account both micromotion and stress. The dual risk assessment mechanism, along with the corrected risk correction factor PRCF_adj, serves not only as the basis for risk stratification but also as a core link between patient biological characteristics and biomechanical stability. By introducing the osteoporosis correction factor OMF, it transforms simple mechanical indicators into a comprehensive score integrating patient bone quality, solving the technical challenge of mechanical environments that younger patients can tolerate but which may fail in older patients with osteoporosis. Furthermore, it directly determines the strength setting of subsequent fixation plans, achieving a dual logical effect of identifying the stability of the fracture itself and assessing whether the patient's bone can withstand the stress. Ultimately, this achieves the precise treatment goal of avoiding excessive surgical damage to low-risk patients while ensuring the reliability of fixation for high-risk patients.

[0097] S4. Integrate the results of the initial risk correction actions with the secondary risk correction factors, define the comprehensive surgical risk assessment value corresponding to the target patient, and decide whether to initiate the expert consultation mechanism based on the comprehensive surgical risk assessment value to obtain the final fixed surgical plan.

[0098] The comprehensive surgical risk assessment value is defined using a multiplicative coupling model, which is based on the following formula: CSRS = BR × (1 + PRCF_adj) × (1 + SRCF). Here, CSRS represents the comprehensive surgical risk assessment value, and BR is the basic risk score. BR does not require complex finite element analysis and primarily reflects the target patient's physiological reserve and overall health. In this embodiment, a predefined rule engine is typically used for setting the score. Specifically, the widely accepted Association of Anesthesiologists (ASA) classification is used for assignment. For ASA I-II, indicating health or mild illness, BR = 1 as the baseline. For ASA III, indicating severe illness and functional limitations, BR = 1.5. For ASA IV and above, indicating life-threatening severe illness, BR = 2.

[0099] The core logic of the above multiplicative coupling model lies in the cascading amplification effect of risk. The logic is based on the following: using the baseline risk score (BR) as a physiological foundation, a physically weak patient, even with a minor fracture, has poor baseline tolerance; conversely, a physically strong patient has high risk tolerance. The combination of (1+PRCF_adj) achieves anatomical / mechanical amplification; PRCF_adj can reflect the stability of the fracture itself to some extent. If the fracture is extremely unstable, it will directly amplify the baseline risk. For example, the same osteoporosis might be a minor issue in a stable fracture, but catastrophic in a comminuted fracture. (1+SRCF) achieves implantation / secondary amplification: SRCF reflects the appropriateness of the initial fixation plan, such as stress shielding or stress concentration; if the initial fixation plan is inadequate, further adjustments will be made based on the first two. In the multiplicative coupling model, any extremely high value will lead to a surge in the final result, consistent with the clinical "barrel effect," where the shortest plank determines the final outcome.

[0100] The basis for deciding whether to initiate the expert panel consultation mechanism is as follows:

[0101] S4.1. A risk threshold CSRS_threshold is preset. If the comprehensive surgical risk assessment value CSRS ≤ the risk threshold CSRS_threshold, the risk of the current preliminary fixation plan is considered controllable, and the surgical preparation stage can be directly entered. S4.2. If the comprehensive surgical risk assessment value CSRS > the risk threshold CSRS_threshold, the expert group consultation mechanism is activated: a digital medical record containing a three-dimensional sternal model of the target patient, fracture details, osteoporosis index, virtual surgical simulation animation, stress cloud map, and a detailed CSRS breakdown report is automatically generated and pushed to a multi-domain expert group. This multi-domain expert group consists of thoracic surgeons, orthopedic surgeons, radiology specialists, and biomechanics experts, which facilitates subsequent online or offline consultations. Based on the digital medical record, multiple alternative plans are discussed. For each alternative plan, the process can return to S3 to re-perform virtual simulation and risk assessment until an alternative plan that can directly enter the surgical preparation stage, that is, a plan that meets the comprehensive surgical risk assessment value CSRS ≤ the risk threshold CSRS_threshold, is selected as the final fixation surgical plan.

[0102] In S4.1, the risk threshold CSRS_threshold is set based on the following: it should be determined by retrospective clinical data calibration or expert Delphi method, usually selecting the inflection point where the incidence of complications increases significantly in historical cases, such as a predicted complication rate >15% or 20% as the boundary; for example, if the baseline risk score BR is 1, the risk threshold CSRS_threshold can be set to 2.5, which means that when the comprehensive surgical risk assessment value CSRS exceeds 2.5 times the baseline risk score BR, it is considered to exceed the safety boundary of routine surgery and manual intervention is necessary;

[0103] Therefore, the fracture details mentioned in S4.2 can be based on the quantitative morphological report generated in stages S1 and S2, specifically including fracture classification and location: clearly defining the location of the fracture line, such as the middle of the sternal body, and the type, such as transverse, oblique, or comminuted fracture; displacement parameters: the specific relative displacement and angular angle of the fracture ends, which is the key basis for determining whether reduction is necessary, and others, which will not be elaborated here; the stress cloud map can be a visual mechanical distribution map generated based on the above virtual simulation, which can intuitively show where damage will occur, including bone tissue stress distribution: using Von Mises stress cloud map to show the stress on the sternum under cough / deep breathing load, specifically: the red highlighted area indicates the stress concentration point, if the area exceeds the bone yield strength, it indicates that secondary fracture or bone resorption may occur; implant stress distribution: It displays the stress state of the fixing plate and steel wire; if high stress appears around the steel wire, it will be displayed in red, indicating a high risk of the steel wire loosening or breaking; if the stress of the fixing plate is too high, it indicates insufficient rigidity; and so on, which will not be elaborated here; for the CSRS detailed breakdown report, it can be regarded as a risk attribution analysis table, which helps the subsequent multi-domain expert group to quickly locate why the risk is high, rather than just looking at a total score.

[0104] The detailed CSRS breakdown report should include: First, the breakdown of the baseline risk score (BR): listing the specific factors leading to a high baseline score, such as a history of diabetes; Second, the contribution of PRCF_adj: indicating whether excessive relative displacement or high maximum principal stress caused mechanical risk, for example, a high PRCF_adj is mainly caused by micromotion at the fracture ends exceeding 2mm; Third, the contribution of SRCF: pointing out the deficiencies of the current fixation scheme; for example, a high SRCF is due to the use of ordinary steel wire in osteoporotic areas, resulting in a maximum equivalent stress ratio (SR) of 1.3. Through the combination of the above three points, the subsequent multi-disciplinary expert group will no longer be blindly speculating, but will formulate alternative solutions based on a precise chain of evidence of morphology, mechanics, and data, such as changing to double fixation plates, and return to S3 for verification.

[0105] The final alternative solutions may include replacing with a stronger internal fixation device, changing the direction / angle of the wire, or using bone cement reinforcement, etc., subject to the actual discussion results.

[0106] This solution addresses the technical problem of traditional internal fixation designs easily overlooking the risk of microscopic failure at the bone-implant interface, achieving precise quantitative early warning for the specific complication of postoperative cut pad increase. The overall solution innovatively defines the RAZ (Reducing Area of ​​Incision) risk assessment zone, narrowing the assessment focus from macroscopic fracture stability to the microscopic edge of the fixation device in contact with bone, capturing local stress concentration caused by osteoporosis. (Note: In this embodiment, the fixation device is typically a steel wire.) Furthermore, by introducing a secondary risk correction factor (SRCF), the von [risk assessment zone] [is further refined]. The maximum equivalent stress ratio (SR) is calculated by comparing the Mises stress with the local yield strength calibrated by the Osteoporosis Index (OI). This ratio is then mapped to an intuitive risk value using a piecewise function, quantifying the risk of bone fracture due to osteoporosis. This allows for the identification of areas prone to wire loosening during preoperative simulations. If the secondary risk correction factor (SRCF) indicates an excessively high risk, it guides medical staff to adjust the wire angle / position or increase bone cement reinforcement, significantly improving the long-term stability of internal fixation and the success rate of surgery. Furthermore, the SRCF is integrated into a multiplicative coupling model of the comprehensive surgical risk assessment value (CSRS). When the CSRS exceeds the risk threshold, an expert consultation mechanism is automatically triggered, pushing a digital medical record containing the risk attribution corresponding to the SRCF to assist multidisciplinary experts in developing a safer final surgical plan.

[0107] Example 2:

[0108] Based on Example 1, this embodiment also provides a risk assessment system for sternal fixation surgery. The system includes: a data acquisition module: acquiring image data of the target patient and performing image preprocessing and key feature extraction; a model construction module: constructing a specific three-dimensional finite element model of the sternum for the target patient based on the extracted key features, integrating assignment and fracture damage settings, and determining the osteoporosis level according to a preset rule engine; an analysis and correction module: performing preliminary risk correction actions on the specific sternal three-dimensional finite element model to determine a preliminary fixation plan, completing virtual surgical simulation, simultaneously executing a cutting risk analysis strategy, performing initial and secondary corrections of surgical risks, and generating a secondary risk correction factor; and a decision-making module: integrating the results of the preliminary risk correction actions and the secondary risk correction factor, defining a comprehensive surgical risk assessment value corresponding to the target patient, and deciding whether to initiate an expert consultation mechanism based on the comprehensive surgical risk assessment value to obtain a final fixation surgical plan.

[0109] In summary, this invention addresses two major technical challenges in existing methods for assessing the risk of sternal fixation surgery: first, the assessment process relies too heavily on the doctor's subjective experience, lacking objective and quantitative evidence; second, it fails to integrate the patient's individual bone condition with the specific surgical procedure, making it impossible to accurately predict the risk of structural complications caused by bone fragility during and after surgery. This invention, by integrating medical image analysis, patient-specific finite element modeling, and virtual surgical simulation technology, achieves a comprehensive, multi-dimensional, and quantitative assessment of surgical risk. This method objectively reflects the instability of the fracture itself and the weakening effect of osteoporosis on the reliability of internal fixation, thus providing a solid scientific basis for clinical decision-making and significantly improving the accuracy and safety of surgical planning. Furthermore, by establishing a risk assessment-expert consultation feedback mechanism, this invention automatically generates detailed and visualized digital medical records when high-risk cases are identified, greatly improving the efficiency and quality of multidisciplinary expert consultations. This not only optimizes the treatment outcome for individual patients but also helps to accumulate and standardize experience in handling complex cases, promoting the overall improvement of sternal surgery diagnosis and treatment.

[0110] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0111] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0112] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A risk assessment method for sternal fixation surgery, characterized in that, The method includes: Acquire image data of the target patient and perform image preprocessing and key feature extraction; Based on the extracted key features, a specific three-dimensional finite element model of the sternum for the target patient is constructed, integrating assignment and fracture damage settings, and the osteoporosis level is determined according to the preset rule engine. On a specific three-dimensional finite element model of the sternum, preliminary risk correction actions are performed to determine the preliminary fixation plan and complete the virtual surgical simulation. Simultaneously, the cutting risk analysis strategy is executed to perform the first and second corrections of surgical risks and generate secondary risk correction factors. By integrating the results of initial risk correction actions with secondary risk correction factors, a comprehensive surgical risk assessment value corresponding to the target patient is defined. Based on the comprehensive surgical risk assessment value, a decision is made on whether to initiate an expert consultation mechanism to obtain a final fixed surgical plan.

2. The risk assessment method for sternal fixation surgery according to claim 1, characterized in that: The image data consisted of a sequence of chest computed tomography (CT) images. Image preprocessing involved using a deep learning-driven image segmentation network to automatically process the CT images, segmenting the overall contour of the sternum, the fracture line region, and adjacent key anatomical structures of the target patient. The segmentation result was output as a binary mask, where regions with a pixel value of 1 represented the sternum entity and regions with a pixel value of 0 represented the background. Key feature extraction involved applying an edge detection algorithm combined with morphological operations based on the grayscale gradient information of the binary mask and the original CT images to locate the spatial orientation of the fracture line and extract the proposed cutting surface in three-dimensional space according to the spatial orientation. For the sternal region of the target patient, the average HU value was calculated, and an osteoporosis index (OI) was extracted and introduced. Key features included the cutting surface and the osteoporosis index (OI).

3. The risk assessment method for sternal fixation surgery according to claim 2, characterized in that: The construction process of the specific sternal three-dimensional finite element model is as follows: Model reconstruction: The binary mask is imported into the reverse engineering software, and the target patient-specific sternal three-dimensional surface mesh is generated through the isosurface extraction algorithm. Subsequently, the surface mesh is post-processed, including smoothing and hole repair; finite element mesh generation: the post-processed surface mesh is imported into finite element preprocessing software to generate a three-dimensional solid mesh; in the vicinity of the fracture line of the target patient and the planned cutting surface area, local mesh refinement technology is used to capture the stress concentration phenomenon in the cutting surface area.

4. The risk assessment method for sternal fixation surgery according to claim 2, characterized in that: The integrated assignment and fracture damage setting are as follows: Based on the Osteoporosis Index (OI), non-uniform material properties are assigned to the elements in the specific sternal 3D finite element model; in the specific sternal 3D finite element model, the fracture line region is defined as the initial crack or weak connection interface; for the cutting surface region, a virtual cutting gap is pre-created in the specific sternal 3D finite element model, and its contact properties are defined. Fixed constraints are applied to the distal end of the specific sternal 3D finite element model to complete the fracture damage setting; the process of determining the osteoporosis level according to the rule engine is as follows: when the Osteoporosis Index (OI) < 0.7, the osteoporosis level corresponds to severe osteoporosis; when 0.7 ≤ OI < 0.85, the osteoporosis level corresponds to moderate osteoporosis; when the OI ≥ 0.85, the osteoporosis level corresponds to normal.

5. The risk assessment method for sternal fixation surgery according to claim 4, characterized in that: The initial risk correction action is based on the fracture condition and proceeds as follows: A standardized physiological load F_physio is applied to a specific sternal 3D finite element model without any internal fixation; simulation calculations are performed to obtain the relative displacement D_fracture and maximum principal stress σ_max_fracture of the fracture ends under the physiological load F_physio; based on the simulation results, an initial risk correction factor PRCF is defined; the osteoporosis grade of the target patient is retrieved, and if the osteoporosis grade is determined to be abnormal, a correction action is triggered, generating a corrected risk correction factor PRCF_adj based on the initial risk correction factor PRCF; the initial fixation plan is determined based on the corrected risk correction factor PRCF_adj.

6. The risk assessment method for sternal fixation surgery according to claim 5, characterized in that: The initial risk correction factor (PRCF) is defined as follows: PRCF = δ × (D_fracture / D_crit) + ε × (σ_max_fracture / σ_crit); where δ and ε are weighting coefficients, and the sum of δ and ε is 1; D_crit and σ_crit are the clinically acceptable critical displacement threshold and critical stress threshold, respectively. The triggering correction action involves introducing an osteoporosis correction factor (OMF). When the osteoporosis grade corresponds to severe osteoporosis, OMF = 1.5Q; when the osteoporosis grade corresponds to moderate osteoporosis, OMF = 1.2Q; Q represents the defining index, with a value range of (0, 1]. The corrected risk correction factor (PRCF_adj) is generated as follows: PRCF_adj = PRCF × OMF. The initial fixation plan is determined based on... The method is as follows: The corrected risk correction factor PRCF_adj is compared with the preset risk threshold. When the corrected risk correction factor PRCF_adj is less than the lower limit of the risk threshold, the first fixed strategy is retrieved from the pre-configured solution library and used as the initial fixed strategy. When the lower limit of the risk threshold is less than or equal to the corrected risk correction factor PRCF_adj and less than the upper limit of the risk threshold, the second fixed strategy is retrieved from the pre-configured solution library and used as the initial fixed strategy. When the corrected risk correction factor PRCF_adj is greater than or equal to the upper limit of the risk threshold, the third fixed strategy is retrieved from the pre-configured solution library and used as the initial fixed strategy. Wherein, risk threshold = [lower value of risk threshold, upper value of risk threshold].

7. The risk assessment method for sternal fixation surgery according to claim 5, characterized in that: The basis for implementing the cutting risk analysis strategy is as follows: During the virtual surgical simulation, the same physiological load F_physio is applied again, and the mechanical response near the cutting surface area is analyzed. The cutting risk is quantified by monitoring the maximum equivalent stress σ_vM_cut of the corresponding element at the cutting surface, and a secondary risk correction factor SRCF is derived. The specific process is as follows: Define the cutting surface risk assessment region RAZ; extract the target stress σ_vM of all finite element elements within the RAZ; determine the local yield strength σ_yield(OI) of the bone tissue; calculate the maximum equivalent stress ratio and quantify the cutting risk: find the maximum value of all single target stresses of finite element elements within the RAZ, denoted as σ_vM_cut; then calculate the ratio of σ_vM_cut to the local yield strength, which is the maximum equivalent stress ratio SR; map the maximum equivalent stress ratio to the secondary risk correction factor: map the maximum equivalent stress ratio SR to the secondary risk correction factor SRCF using a piecewise function.

8. The risk assessment method for sternal fixation surgery according to claim 7, characterized in that: The comprehensive surgical risk assessment value is defined using a multiplicative coupling model, and the specific operating principle of the multiplicative coupling model is: CSRS=BR×(1+PRCF_adj)×(1+SRCF); where CSRS is the comprehensive surgical risk assessment value, BR is the basic risk score, and the ASA classification is used for assignment; when ASA I-II, BR=1, as the baseline; when ASA III, BR=1.5; when ASA IV and above, BR=2.

9. The risk assessment method for sternal fixation surgery according to claim 8, characterized in that: The decision on whether to initiate the expert consultation mechanism is based on the following: a preset risk threshold CSRS_threshold is used. If the comprehensive surgical risk assessment value CSRS ≤ the risk threshold CSRS_threshold, the patient directly enters the surgical preparation stage. Otherwise, the expert consultation mechanism is initiated: a digital medical record is automatically generated, which includes a three-dimensional sternal model of the target patient, fracture details, osteoporosis index, virtual surgical simulation animation, stress cloud map, and a detailed CSRS breakdown report. This record is then sent to a multi-disciplinary expert group. Based on the digital medical record, multiple alternative plans are derived. For each alternative plan, the process is repeated on a specific three-dimensional sternal finite element model until a target that meets the criteria for directly entering the surgical preparation stage is found. This target plan is then used as the final fixed surgical plan.

10. A risk assessment system for sternal fixation surgery, applied to a risk assessment method for sternal fixation surgery as described in any one of claims 1-9, characterized in that, The system includes a data acquisition module: acquiring image data of the target patient and performing image preprocessing and key feature extraction; Model building module: Based on the extracted key features, a specific three-dimensional finite element model of the sternum for the target patient is constructed, integrating assignment and fracture damage settings, and determining the osteoporosis level according to the preset rule engine; Analysis and Correction Module: On a specific sternal 3D finite element model, preliminary risk correction actions are performed to determine the preliminary fixation plan and complete the virtual surgical simulation. Simultaneously, the cutting risk analysis strategy is executed to perform the first and second corrections of surgical risks and generate secondary risk correction factors. The decision-making module integrates the results of the initial risk correction actions with the secondary risk correction factors, defines the comprehensive surgical risk assessment value corresponding to the target patient, and decides whether to initiate the expert consultation mechanism based on the comprehensive surgical risk assessment value to obtain the final fixed surgical plan.

Images