A preoperative integrated planning system and method for proximal humerus fracture locking plate surgery

Abstract

This invention discloses an integrated preoperative planning system and method for locking plates in proximal humeral fractures, comprising: a data preprocessing module for processing CT data; a fracture segmentation and 3D reconstruction module for segmenting and generating fracture fragment models; a target anatomy prediction and fragment disposal module for predicting target anatomical morphology and classifying fracture fragments; a surgical feasibility-constrained virtual reduction module for virtual reduction under anatomical and clinical constraints; a plate fitting and fixation planning module for optimizing plate fit, screw configuration, and generating equivalent fixation parameters; a finite element modeling and solving module for automatic mechanical simulation; and a risk assessment and closed-loop optimization module for assessing risks and triggering replanning to achieve closed-loop optimization. This invention significantly reduces preoperative planning time, improves the mechanical reliability and clinical fit of the plan, and provides a quantitative basis for surgical decisions.

Description

Technical Field

[0001] This invention belongs to the field of digital orthopedics and computer-aided medical technology, and particularly relates to an integrated preoperative planning system and method for locking plates in proximal humeral fractures. Background Technology

[0002] Proximal humeral fractures are common in clinical practice, especially three- or four-part fractures, which are often accompanied by humeral head collapse, valgus fractures, loss of medial support, and multiple small fracture fragments. Current preoperative planning usually relies on surgeons' experience in reduction and plate / screw selection based on 2D CT or 3D reconstruction models. Furthermore, significant manual interaction is required in stages such as 3D segmentation, fracture fragment separation, virtual reduction, and screw length and orientation selection, which is time-consuming and the results are heavily influenced by the surgeon's experience.

[0003] In addition, common complications of locking plate fixation include screw perforation of the articular surface, varus collapse, internal fixation failure, nonunion, and humeral head necrosis. Existing finite element analysis and biomechanical assessments are mostly used for research validation; modeling and condition setting are cumbersome, making it difficult to form a closed-loop optimization plan in a short time before clinical surgery. Some studies have attempted to use artificial intelligence for automatic segmentation and virtual reduction of fracture fragments, but these typically do not further integrate the complete process of plate fitting, screw planning, finite element solution, complication risk prediction and optimization, nor do they fully consider real surgical strategies such as the removal / disposal of small fragments during surgery and the use of sutures or Kirschner wires for the fixation of smaller but important fracture fragments.

[0004] Therefore, there is an urgent need for an integrated system and method that can automatically complete unreduced modeling, virtual reduction that meets surgical feasibility, planning of locking plates and screws / sutures / Kirschner wires, finite element analysis and complication risk prediction, and automatically iteratively optimize the fixation scheme when the risk is high, starting from CT data. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention proposes an integrated preoperative planning system and method for locking plates in proximal humeral fractures, thereby resolving the issues present in the prior art.

[0006] To achieve the above objectives, the present invention provides an integrated preoperative planning system for locking plates in proximal humeral fractures, comprising: The data preprocessing module is used to preprocess the patient's proximal humerus CT data and output standardized volume data. The fracture segmentation and 3D reconstruction module is used to segment bone tissue and fracture sutures based on normalized volume data and generate a 3D model of the fracture fragments. The target anatomy prediction and fragment disposal module is used to predict the target anatomical morphology based on the three-dimensional model of the fracture fragments through a statistical shape model, and to classify and dispose of the fracture fragments according to their fixability, outputting a set of preserved fracture fragments. The surgical feasibility constraint virtual reduction module is used to virtually reduce the set of preserved fracture fragments under the constraints of target anatomical morphology and clinical constraints, and output the reduced proximal humeral bone model; The plate fitting and fixation planning module is used to plan the plate fitting and internal fixation on the proximal humerus bone model after reduction, and output the internal fixation configuration. The finite element modeling and solving module is used to automatically generate and solve finite element models based on the repositioned proximal humerus bone model and internal fixation configuration, and output mechanical results. The risk assessment and closed-loop optimization module is used to perform risk assessment and robustness assessment based on the mechanical results, and to trigger the steel plate bonding and fixing planning module to replan when the assessment results do not meet the threshold.

[0007] Optionally, the target anatomy prediction and fragment disposal module classifies fracture fragments according to fixability discrimination rules; the fixability discrimination rules include the following parameterized threshold judgment conditions: the maximum axial continuous length along the candidate screw direction. Minimum internal diameter of fracture fragments and the effective cross-sectional area of ​​the fracture fragments The following conditions must be met respectively: ; ; ; in, The nominal diameter of the selected screw. For safety margin, This is the thread engagement multiple factor. This is the minimum usable screw length.

[0008] Optionally, the surgical feasibility-constrained virtual repositioning module obtains the rigid body transformation of each preserved fracture fragment by solving a constraint optimization problem. The objective function of the constrained optimization problem for: in, The symbolic distance field for the target anatomical model; Penalty for bone fragments penetrating or colliding; Penalty for inner support / neck angle constraint; For the initial pose translation; For the first Rigid body transformation of individual fracture fragments; These are the three-dimensional coordinates of the fracture fragment surface or voxel points. For anatomical matching weights; Weights for inter-block collision penalties; The penalty weight for pose offset; For the first The amount of translation of each fracture fragment.

[0009] Optionally, the steel plate bonding and fixing planning module optimizes screw configuration with the goal of maximizing the sum of the holding force indices of multiple screws; the holding force index for: ; in, This is the screw insertion point. The unit vector in the screw direction. Where HU is the screw length and HU is the CT grayscale value. This is the mapping function from HU to bone mineral density. Let be the integral variable along the screw axis, with the following range of values: .

[0010] Optionally, the constraints to be met by the optimization include: the minimum distance from the screw tip to the articular surface is not less than a preset safety distance, the screw does not penetrate the articular surface, the screw path does not cross the fracture suture area, and the minimum distance between any two screws is not less than a preset spacing threshold.

[0011] Optionally, the risk assessment and closed-loop optimization module obtains a score set by perturbing and sampling the material mapping parameters, friction coefficient, and load amplitude. And calculate robustness score When robust rating When the value falls below a preset threshold, the steel plate bonding and fixing planning module is triggered to replan; the formula for calculating the robustness score is as follows: ; in, [·] represents the expected value, and Std[·] represents the standard deviation. For weighting coefficients; when When the value is below a preset threshold, the steel plate bonding and fixing planning module is triggered to replan.

[0012] This invention also provides an integrated preoperative planning method for locking plates in proximal humeral fractures, implemented based on the above system, including: Acquire and preprocess proximal humeral CT data from patients to obtain standardized volume data; Based on the standardized volume data, bone tissue and fracture sutures are segmented to generate a three-dimensional model of the fracture fragments; Based on the three-dimensional model of the fracture fragments, the target anatomical morphology is predicted, and the fracture fragments are classified and treated according to their fixability to determine the set of fracture fragments to be retained. Under the constraints of target anatomical morphology and clinical conditions, the preserved fracture fragments were virtually reduced to generate a repositioned proximal humerus bone model. The plate fitting and internal fixation plan are performed on the repositioned proximal humerus bone model, and the internal fixation configuration is output. Based on the repositioned proximal humerus bone model and the internal fixation configuration, a finite element model is automatically generated and solved to obtain mechanical results. Based on the mechanical results, risk assessment and robustness assessment are performed. If the assessment results do not meet the threshold, the process returns to the internal fixed planning step to replan and iterates to output the final solution.

[0013] Optionally, the process of acquiring and preprocessing the patient's proximal humerus CT data includes: isotropic voxel resampling of the original CT data, three-dimensional nonlocal mean denoising, and grayscale normalization within the bone tissue region to obtain the normalized voxel data.

[0014] Optionally, the process of segmenting bone tissue and fracture sutures based on the normalized volume data to generate a three-dimensional model of the fracture fragment includes: The normalized volume data is processed using a three-dimensional convolutional neural network framework, and bone tissue probability map and fracture suture probability map are output simultaneously. Based on the fracture suture probability map, the bone tissue is divided and segmented into connected components to generate a three-dimensional model of the fracture block.

[0015] Optionally, the process of fitting a plate and planning internal fixation on the repositioned proximal humerus bone model, and outputting the internal fixation configuration, includes: An anatomical coordinate system was established on the repositioned proximal humerus bone model, and key anatomical landmarks for guiding plate placement were identified based on the anatomical coordinate system. Based on the standard positioning constraints defined by the key anatomical landmarks, the locking plate is initialized in position and optimized in fit. Within the hole angle sector of the locking plate, the combination optimization of screw direction and length is performed. The optimization aims to maximize the sum of screw holding force indices and meets the constraints of joint surface safety distance, no fracture suture crossing, and minimum screw spacing. For fracture fragments requiring equivalent fixation, the system automatically generates equivalent fixation parameters for suture bands or Kirschner wires. The optimized locking plate position, the optimized screw configuration, and the equivalent fixing parameters together form the complete internal fixing configuration.

[0016] Compared with the prior art, the present invention has the following advantages and technical effects: (1) Integrated process and plan rollback: The “CT preprocessing - fracture fragment segmentation - fragment disposal - surgical feasibility reduction - internal fixation planning - finite element evaluation - robust risk prediction - plan rollback and redesign” is organized into an automatically executed closed-loop process, reducing the number of manual interaction steps and improving the consistency and repeatability of the plan.

[0017] (2) Fragmentation management mechanism for complex fractures: By “judging the functional importance and fixability”, fracture fragments are divided into three categories: those that can be fixed with screws, those that require equivalent fixation (sutures / Kirschner wires), and those that can be removed. This makes the planning process conform to the actual surgical logic of removing small fragments and fixing key fragments, thereby reducing invalid planning of “reduction but not fixation” from the source and making virtual planning more in line with the actual surgical strategy.

[0018] (3) Solution of surgical feasibility constraints for repositioning: The repositioning is written as an optimization problem with collision penalty and inner support / neck angle constraints, and a feasible region is set for translation / rotation, so as to avoid the repositioning result that is geometrically perfect but clinically unrealizable or leads to failure of inner support.

[0019] (4) Joint optimization of screw planning “holding force-safety”: The integral of HU / bone density along the screw trajectory is used as the holding force objective function. At the same time, constraints such as joint penetration, fracture suture crossing, screw non-interference and medial strut support are applied so that the screw layout can not only avoid risks, but also actively select a trajectory with more bone mass, thereby improving the fixation reliability in cases of low bone quality and comminuted fractures.

[0020] (5) Robust finite element evaluation reduces sensitivity to uncertainty: The robust score of material mapping parameters, friction coefficient and load amplitude disturbance sampling is introduced, which can conservatively check the stability of the scheme under the conditions of CT noise, material mapping error and daily load uncertainty; when the robust score is lower than the threshold, it automatically backs down and redesigns, thereby reducing the risk that it only holds true under "ideal parameters".

[0021] (6) Explainable quantitative decision output: The system outputs the reset pose, plate / screw parameters, finite element cloud diagram and key index scores, and gives the reasons for the difference between the recommended and alternative solutions, which facilitates doctors to review and make necessary manual corrections, thereby improving clinical usability and traceability. Attached Figure Description

[0022] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a structural block diagram of the preoperative integrated planning system for locking plates in proximal humeral fractures according to an embodiment of the present invention. Figure 2 This is a flowchart illustrating the preoperative integrated planning method for locking plates in proximal humeral fractures according to an embodiment of the present invention. Detailed Implementation

[0023] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0024] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0025] Example 1 like Figure 1 As shown, this embodiment provides an integrated preoperative planning system for locking plates in proximal humeral fractures, comprising: Data preprocessing module: Performs voxel resampling, noise reduction filtering, and intensity standardization on the patient's proximal humerus CT data, and outputs standardized volume data; Fracture segmentation and 3D reconstruction module: Based on a 3D convolutional neural network, the module performs fine segmentation of bone tissue and identifies fracture suture regions, generating bone tissue masks, fracture suture masks and 3D models of fracture fragments; Target Anatomy Prediction and Fragment Disposal Module: Based on the Statistical Shape Model (SSM), the module predicts the target anatomical morphology and classifies fracture fragments into repositionable fragments and small fragments according to the functional importance and fixability criteria. Small fragments are removed and the volume fraction of bone defects is recorded. Important but unfixable fragments are marked for subsequent equivalent fixation. Surgical feasibility constraint virtual reduction module: Under the target anatomical constraints, the reduction pose of the preserved fracture fragments is solved, and clinical feasibility constraints such as inter-fracture collision constraints and medial support / neck-shaft angle are applied, and the reduction proximal humerus bone model is output; Plate fitting and fixation planning module: Establishes the anatomical coordinate system of the humerus and identifies features such as the apex of the greater tubercle, the intertubercular groove, and the axis of the bone shaft; performs standard plate placement and fitting optimization on the bone model after reduction; performs screw ray tracing within the hole position angle sector, and optimizes the combination of screw direction and length by combining the articular surface safety distance and gripping force index; generates suture bands and / or Kirschner wire equivalent fixation parameters for important fragments that cannot be fixed with screws; Finite element modeling and solution module: Automatic mesh generation, material mapping, contact / connection relationship and load case setting and solution, outputting results such as displacement, stress-strain and contact slip; Risk assessment and closed-loop optimization module: Extract stability and complication-related indicators from finite element results, construct normalized weighted scores and conduct robustness assessment; when the robustness score or risk score does not meet the threshold, trigger a rollback to the fixed planning module to redesign the scheme and iterate until the recommended scheme is output.

[0026] The target anatomy prediction and fragment disposal module calculates the screw fixation index for each fracture fragment and classifies them into three categories according to rules: screw fixable, requiring equivalent fixation, and removable defects. The screw fixation index includes at least the following: the distance field within the fracture fragment. The above satisfies The connected region; the maximum axial continuous length along the candidate screw direction. Minimum inscribed diameter Effective cross-sectional area And a parameterized threshold is used: in, The nominal diameter of the selected screw. For safety margin, This is the thread engagement multiple factor. This is the minimum usable screw length.

[0027] The surgical feasibility-constrained virtual reduction module obtains the rigid body transformations of each preserved fracture fragment by solving the following constrained optimization problem. : in, The symbolic distance field for the target anatomical model; Penalty for bone fragments penetrating or colliding; Penalty for inner support / neck angle constraint; For the initial pose translation; and for An upper limit is applied to the rotation angle to ensure clinical feasibility of reduction. For the first Rigid body transformation of a fracture fragment; representing the spatial transformation of a fracture fragment from its "current unreduced position" to its "reduced position"; These are the three-dimensional coordinate points on the surface of the fracture fragment or voxel points. These points are used to calculate the geometric deviation from the target anatomy after reduction. The anatomical matching weight is determined by the weight; the greater the weight, the more emphasis is placed on "the repositioning must conform to the target anatomical shape". Apply a penalty weight to inter-fracture collisions to prevent fracture fragments from penetrating each other during reduction. The penalty weight for positional deviation is used to prevent fracture fragments from "moving too far," which is within the clinically feasible range. For the first The translational amount of each fracture fragment is the actual spatial distance that the fracture fragment moves; This limits excessive traction and avoids theoretical repositioning that cannot be achieved clinically.

[0028] The steel plate bonding and fixing planning module defines the holding force index for each candidate screw. : in, This is the screw insertion point. The unit vector in the screw direction. Where HU is the screw length and HU is the CT grayscale value. Let HU be the mapping function to bone mineral density; and satisfy the constraints Under the given conditions, the screw combination is determined with the objective of maximizing the sum of the holding force indices of multiple screws, where This is the minimum distance from the end of the screw to the joint surface. Indicates whether the joint surface has been penetrated. Indicates whether the fracture line has been crossed. Screw spacing; Let be the integral variable along the screw axis, with the following range of values: .

[0029] The risk assessment and closed-loop optimization module performs perturbation sampling on material mapping parameters, friction coefficient, and load amplitude to obtain a score set. The robustness score is obtained as follows: when When the threshold is below, the process is triggered to return to the steel plate bonding and screw planning module of claim 1 to replan the internal fixation scheme.

[0030] like Figure 2 As shown in this embodiment, an integrated preoperative planning method for locking plates in proximal humeral fractures is provided, including the following steps: Step S1: Acquire CT data of the patient's proximal humerus fracture and construct three-dimensional volumetric data. The volume data was resampled to the isotropic voxel spacing. Trilinear interpolation is used to obtain : right Three-dimensional nonlocal means denoising was used to obtain : Within the bone foreground region, HU was cropped at the 0.5 and 99.5 percentiles and standardized using the foreground mean / variance of the training set, resulting in... : An initial mask of bone tissue was obtained through threshold segmentation and morphological processing, and an unreplaced three-dimensional solid model was obtained through three-dimensional surface reconstruction. .

[0031] Step S2, a 3D encoder-decoder network based on the nnU-Net framework is used. Output bone tissue probability map Probability diagram of fracture sutures And it was trained using a combination of cross-entropy and Dice loss: right Thresholds were automatically determined within the bone mask using the Otsu method. Obtain the fracture suture mask The fracture block voxel set is obtained by suture excision and connected component labeling, and the fracture block model is reconstructed. If a significant suture probability still exists within a connected component, a weighted graph is constructed within that block and subdivided using minimum cut subdivision, where the edge weights are: Statistical Shape Model (SSM) based on a normal proximal humerus training set: The target anatomical model was obtained by fitting the observable bone surface point set of the patient. .

[0032] The functional importance and fixability of the fracture fragments were assessed to obtain a set of preserved fracture fragments. Collection of tiny fragments Construct a distance field for each block. And fixability is determined using a threshold parameterized by screw specifications: For sets Perform bone defect removal and record the volume fraction of the bone defect; mark important but unscrew-fixable fragments.

[0033] Set under the constraints of the target anatomical model Solve the reset pose , minimize: in, for The sign distance field, This is a penalty for inter-block collisions. The clinical feasibility constraints and penalties include factors such as medial support / neck-shaft angle. For the initial pose translation; and for An upper limit is applied to the rotation angle to ensure clinical feasibility of reduction. A three-dimensional bone model is output after reduction. .

[0034] Step S3: Obtain the humeral shaft axis by taking the PCA of the diaphysis region on the model after repositioning. The center of the humeral head is obtained by fitting the points of the articular surface of the humeral head together. : An anatomical coordinate system was established based on this; the apex of the greater tubercle was located by the peak curvature. The internodal groove is located by the curvature trough line reference. .

[0035] Select a matching steel plate from the steel plate template library and initialize and optimize it according to the following standard placement constraints: the upper edge of the steel plate is 5–8 mm from the distal end of the apex of the greater tubercle, aligned along the axis of the humeral shaft, and located 2–4 mm behind the intertubercular groove; perform fit optimization within the allowable range to minimize the distance between the back of the steel plate and the bone surface.

[0036] Discrete sampling directions are used within the hole position angle sector. Ray tracing is used to obtain the distance to the joint surface and determine the maximum feasible screw length. A safe distance to the joint surface is then set. Set the minimum distance from the screw tip to the joint surface. satisfy .

[0037] Define screw holding force index and optimize hole position-direction-length combination: in, The starting point of the screw. The unit vector in the screw direction. The length of the screw. Here is the HU-to-bone mineral density mapping function. HU is the CT grayscale value under the following conditions: safe distance between articular surfaces, screw non-interference, fracture suture crossing penalty, and medial support constraint. The screw design aims to maximize the sum of the holding force indices of multiple screws; for important fragments that cannot be fixed with screws, equivalent fixation parameters of suture bands and / or Kirschner wires are generated to form a complete internal fixation configuration. This is the minimum distance from the end of the screw to the joint surface. Indicates whether the joint surface has been penetrated. Indicates whether the fracture line has been crossed. This refers to the screw spacing.

[0038] Step S4: Based on the bone model after reduction and the internal fixation configuration, mesh generation is performed, with preferred local densification at the fracture surface, screw tips, and medial support area; material mapping is then performed according to HU. Establish equivalent elements for fracture surface frictional contact, screw-plate locking connection, screw-bone connection, and suture / Kirschner wire; set up and solve load conditions that include at least one of abduction, flexion, rotation, or lifting loads, and output displacement, von Mises stress, equivalent strain, and contact slip.

[0039] Step S5: Extract indicators such as relative slippage of the fracture surface, changes in humeral head pose / neck-shaft angle, screw perforation risk, maximum equivalent stress of the implant, and high strain volume fraction of bone tissue from the finite element results.

[0040] Using piecewise normalization function Map each indicator to [0,1] and weight them to obtain the single-condition score and the total multi-condition score: This was obtained by sampling the material mapping coefficient, friction coefficient, and load amplitude disturbance. And calculate the robustness score: when If the risk score does not meet the threshold, the process reverts to step S3 to automatically adjust the steel plate pose, screw layout and length, inner support screw configuration and enhancement strategy parameters. S4–S5 are repeated until the threshold is met or the iteration limit is reached, and a recommended solution and quantitative report are output.

[0041] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations 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. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An integrated preoperative planning system for locking plates in proximal humeral fractures, characterized in that, include: The data preprocessing module is used to preprocess the patient's proximal humerus CT data and output standardized volume data. The fracture segmentation and 3D reconstruction module is used to segment bone tissue and fracture sutures based on normalized volume data and generate a 3D model of the fracture fragments. The target anatomy prediction and fragment disposal module is used to predict the target anatomical morphology based on the three-dimensional model of the fracture fragments through a statistical shape model, and to classify and dispose of the fracture fragments according to their fixability, outputting a set of preserved fracture fragments. The surgical feasibility constraint virtual reduction module is used to virtually reduce the set of preserved fracture fragments under the constraints of target anatomical morphology and clinical constraints, and output the reduced proximal humeral bone model. The plate fitting and fixation planning module is used to plan the plate fitting and internal fixation on the proximal humerus bone model after reduction, and output the internal fixation configuration. The finite element modeling and solving module is used to automatically generate and solve finite element models based on the repositioned proximal humerus bone model and internal fixation configuration, and output mechanical results. The risk assessment and closed-loop optimization module is used to perform risk assessment and robustness assessment based on the mechanical results, and to trigger the steel plate bonding and fixing planning module to replan when the assessment results do not meet the threshold.

2. The integrated preoperative planning system for proximal humeral fracture locking plates according to claim 1, characterized in that, The target anatomy prediction and fragment disposal module classifies fracture fragments according to fixability discrimination rules; the fixability discrimination rules include the following parameterized threshold judgment conditions: the maximum axial continuous length along the candidate screw direction. Minimum internal diameter of fracture fragments and the effective cross-sectional area of ​​the fracture fragments The following conditions must be met respectively: ； ； ； in, The nominal diameter of the selected screw. For safety margin, This is the thread engagement multiple factor. This is the minimum usable screw length.

3. The integrated preoperative planning system for locking plates in proximal humeral fractures according to claim 1, characterized in that, The surgical feasibility constraint virtual repositioning module obtains the rigid body transformation of each preserved fracture fragment by solving a constraint optimization problem. The objective function of the constrained optimization problem for: in, The symbolic distance field for the target anatomical model; Penalty for bone fragments penetrating or colliding; Penalty for inner support / neck angle constraint; For the initial pose translation; For the first Rigid body transformation of individual fracture fragments; These are the three-dimensional coordinates of the fracture fragment surface or voxel points. For anatomical matching weights; Weights for inter-block collision penalties; The penalty weight for pose offset; For the first The amount of translation of each fracture fragment.

4. The integrated preoperative planning system for locking plates in proximal humeral fractures according to claim 1, characterized in that, The steel plate bonding and fixing planning module optimizes screw configuration with the goal of maximizing the sum of the holding force indices of multiple screws; the holding force index for: ； in, This is the screw insertion point. The unit vector in the screw direction. Where HU is the screw length and HU is the CT grayscale value. This is the mapping function from HU to bone mineral density. Let be the integral variable along the screw axis, with the following range of values: .

5. The integrated preoperative planning system for proximal humeral fracture locking plates according to claim 4, characterized in that, The constraints to be met for the optimization include: the minimum distance from the screw tip to the articular surface is not less than a preset safety distance; the screw does not penetrate the articular surface; the screw path does not cross the fracture suture zone; and the minimum distance between any two screws is not less than a preset spacing threshold.

6. The integrated preoperative planning system for proximal humeral fracture locking plates according to claim 1, characterized in that, The risk assessment and closed-loop optimization module obtains a score set by perturbating and sampling the material mapping parameters, friction coefficient, and load amplitude. And calculate robustness score When robustness rating When the value falls below a preset threshold, the steel plate bonding and fixing planning module is triggered to replan; the formula for calculating the robustness score is as follows: ； in, [·] represents the expected value, and Std[·] represents the standard deviation. For weighting coefficients; when When the value is below a preset threshold, the steel plate bonding and fixing planning module is triggered to replan.

7. A preoperative integrated planning method for locking plates in proximal humeral fractures, implemented based on the system described in any one of claims 1-6, characterized in that, include: Acquire and preprocess proximal humeral CT data from patients to obtain standardized volume data; Based on the standardized volume data, bone tissue and fracture sutures are segmented to generate a three-dimensional model of the fracture fragments; Based on the three-dimensional model of the fracture fragments, the target anatomical morphology is predicted, and the fracture fragments are classified and treated according to their fixability to determine the set of fracture fragments to be retained. Under the constraints of target anatomical morphology and clinical conditions, the preserved fracture fragments were virtually reduced to generate a repositioned proximal humerus bone model. Plate fitting and internal fixation planning are performed on the repositioned proximal humerus bone model, and the internal fixation configuration is output. Based on the repositioned proximal humerus bone model and the internal fixation configuration, a finite element model is automatically generated and solved to obtain mechanical results. Based on the mechanical results, risk assessment and robustness assessment are performed. If the assessment results do not meet the threshold, the process returns to the internal fixed planning step to replan and iteratively outputs the final solution.

8. The preoperative integrated planning method for locking plates in proximal humeral fractures according to claim 7, characterized in that, The process of acquiring and preprocessing CT data of the proximal humerus of a patient includes: isotropic voxel resampling of the original CT data, three-dimensional nonlocal mean denoising, and grayscale normalization in the bone tissue region to obtain the normalized voxel data.

9. The preoperative integrated planning method for locking plates in proximal humeral fractures according to claim 7, characterized in that, The process of segmenting bone tissue and fracture sutures based on the normalized volume data to generate a three-dimensional model of the fracture fragments includes: The normalized volume data is processed using a three-dimensional convolutional neural network framework, and bone tissue probability map and fracture suture probability map are output simultaneously. Based on the fracture suture probability map, the bone tissue is divided and segmented into connected components to generate a three-dimensional model of the fracture block.

10. The integrated preoperative planning method for locking plates in proximal humeral fractures according to claim 7, characterized in that, The process of fitting plates and planning internal fixation on the repositioned proximal humerus bone model, and outputting the internal fixation configuration, includes: An anatomical coordinate system was established on the repositioned proximal humerus bone model, and key anatomical landmarks for guiding plate placement were identified based on the anatomical coordinate system. Based on the standard positioning constraints defined by the key anatomical landmarks, the locking plate is initialized in position and optimized in fit. Within the hole angle sector of the locking plate, the combination optimization of screw direction and length is performed. The optimization aims to maximize the sum of screw holding force indices and meets the constraints of joint surface safety distance, no fracture suture crossing, and minimum screw spacing. For fracture fragments requiring equivalent fixation, the system automatically generates equivalent fixation parameters for suture bands or Kirschner wires. The optimized locking plate position, the optimized screw configuration, and the equivalent fixing parameters together form the complete internal fixing configuration.

Images