Method for dynamically predicting mechanical properties of implanted stent

Abstract

The invention provides a mechanical property dynamic prediction method after stent implantation, and the method combines CT image three-dimensional reconstruction, finite element simulation analysis and data-driven modeling to construct a whole-process mapping relation from'stent model structure parameters-new bone growth characteristics-mechanical response '; a training data set is constructed through a large amount of simulation data, a proper analysis method is introduced to establish a prediction model, the performance change trend under different structural designs can be efficiently predicted, stent parameters can be reversely optimized, the individualized adaptability and the function-oriented design capability of the stent are improved, and the method is suitable for large-scale popularization and application. Meanwhile, a complex structure parameter space is covered by combining real bone growth data simulation, and the requirement for rapid optimization design of the stent is met; according to the method, the minimum repetitive unit of the model structure is improved and optimized, and meanwhile, the real growth condition of the new bone tissue is combined, so that the method can be more accurately and truly applied to predicting the mechanical property of the bone tissue in the growth process after the stent is implanted.

Description

Technical Field

[0001] This invention relates to the field of tissue engineering scaffold design and intelligent performance prediction technology, and in particular to a method for dynamic prediction of the mechanical properties of tissue engineering scaffolds that combines three-dimensional CT image reconstruction, finite element simulation and data-driven modeling. Background Technology

[0002] In the field of tissue engineering and biomaterials, scaffolds, as the core structure guiding tissue regeneration, not only affect cell adhesion, proliferation, and migration processes, but also determine whether newly formed tissue can successfully fill and repair damaged areas. Especially in bone tissue repair, scaffolds not only need to possess initial mechanical support capabilities, but also need to gradually degrade in vivo and accompany the new bone growth process to achieve a dynamic match between structure and mechanical properties, avoiding abrupt changes in the mechanical environment that could affect the quality of tissue reconstruction.

[0003] However, most scaffold designs currently remain at the stage of static modeling and initial mechanical assessment, neglecting the dynamic evolution process over time, such as the gradual growth of new bone and the gradual degradation of the scaffold. This static approach may lead to two types of clinical problems: on the one hand, insufficient scaffold stiffness leads to early structural instability and fails to provide effective support; on the other hand, an overly rigid scaffold structure inhibits inward growth of new bone, which is detrimental to osseointegration and remodeling.

[0004] Furthermore, most mechanical performance assessments in existing studies are based on initial scaffold performance evaluations, lacking input information on real anatomical structures and biological processes, making it difficult to accurately reflect the true mechanical behavior of the scaffold-bone composite system. Although animal experiments can provide some validation data, they are time-consuming, have scarce samples, and cannot cover complex structural parameter spaces, thus failing to meet the needs of rapid scaffold optimization design.

[0005] Therefore, this invention proposes a dynamic prediction method for stent performance that combines 3D reconstruction of CT images, finite element simulation analysis, and data-driven modeling to construct a full-process mapping relationship from "stent structural parameters → new bone growth → mechanical response". By constructing a training dataset using a large amount of simulation data and introducing appropriate analysis methods to establish a prediction model, this method can not only efficiently predict performance change trends under different structural designs but also optimize stent parameters in reverse, improving its individualized adaptability and function-oriented design capabilities. Summary of the Invention

[0006] In view of the shortcomings of existing technologies, this invention proposes a dynamic prediction method for the mechanical properties of stents after implantation. It combines 3D reconstruction of CT images, finite element simulation analysis and data-driven modeling to construct a full-process mapping relationship from "stent structural parameters → new bone growth → mechanical response".

[0007] In a first aspect, the present invention proposes a method for dynamic prediction of the mechanical properties after stent implantation, characterized in that the method includes the following steps:

[0008] Step S1: Based on the pre-set model structure type of the support and the vertex and rod connection information of the smallest repeating unit, generate node spheres and connecting columns in the 3D modeling environment. Iterate the rod radius by the target density and the side length of the smallest unit. After iteration, the generated structure is merged into an integrated support structure through Boolean operations.

[0009] Step S2: The integrated stent structure is array-replicated, and a Boolean cropping operation is performed using a cropping tool to generate a porous stent model that conforms to the preset implantation area size;

[0010] Step S3: Reconstruct the volume of the new bone in three dimensions based on the CT images, calculate the thickness of the new bone, and add the thickness to the radius of the strut in step S1. The increased thickness is used to generate the three-dimensional model structure of the new bone.

[0011] Step S4: Repeat steps S1–S2, using S3 as the radius of the member to construct the outer new bone model;

[0012] Step S5: Import the inner porous scaffold model and the outer new bone model into the three-dimensional modeling environment, and generate a porous scaffold model-new bone bilayer composite model structure for finite element analysis through Boolean operations;

[0013] Step S6: Import the double-layer composite model into the Mechanical module of the Ansys Workbench software platform, perform static compression simulation analysis, and calculate the elastic modulus of the porous scaffold model-new bone double-layer composite model structure.

[0014] Step S7: Acquire the CT images at different time points, and repeat steps S3–S6 to generate mechanical simulation data for the corresponding time points.

[0015] Step S8: Update the preset model structure type, repeat steps S1-S7 to obtain more mechanical simulation data corresponding to model structure parameters and new bone growth characteristic parameters, construct a complete dataset with all the above model structure parameters and corresponding mechanical simulation data, and establish a mapping relationship model between model structure parameters and mechanical performance response. This model realizes dynamic prediction of mechanical performance after stent implantation.

[0016] Further, specifically, step S1 includes the following steps:

[0017] S11, determine the type of the smallest repeating unit of the selected preset model structure, define the unit side length parameter lside, and export the node vertex information and rod connection information in the smallest repeating unit;

[0018] S12, Based on the connection information of the rods, determine the connection relationship between each pair of nodes to generate an initial index list of connecting rods;

[0019] S13, traverse the above node information, use modeling tools to generate spherical node spheres and rod cylinders, and set the node radius and rod radius to a preset rc;

[0020] S14, merge the node sphere and the rod cylinder into a complete three-dimensional lattice structure through a Boolean fusion operation to obtain the structural model of the minimum repeating unit;

[0021] S15, Obtain the solid volume M of the structural model. lattice Simultaneously calculate the volume M of the minimum repeating element bounding box. box Then, the current stent density is calculated;

[0022] S16, compare the current density with the preset target density error. If the error does not meet the tolerance condition, adjust the radius of the rod according to the error direction and repeat S13-S15 using the new radius.

[0023] S17. When the error meets the preset tolerance or reaches the maximum number of iterations, the current integrated support structure is saved.

[0024] Preferably, the step of obtaining the solid volume M of the structural model is... lattice Specifically: Use gmsh.model.occ.getMass to obtain the solid volume M of the structural model. lattice ;

[0025] Calculate the volume M of the minimum repeating element bounding box. box Specifically:

[0026] Calculate the current stent density Where nx, ny, and nz represent the number of arrays of the smallest repeating units in the X, Y, and Z directions, which are determined by the target size of the support and the unit side length lside, density current This represents the density of the smallest repeating unit.

[0027] Specifically, the calculation of the new bone thickness based on the three-dimensional reconstruction of the new bone volume from CT images includes:

[0028] S31, a threshold segmentation method is used to separate tissues in CT images. The segmentation threshold is set according to the density difference between the porous scaffold model and the new bone, and the newly formed bone tissue region attached to the outer surface of the porous scaffold model and the porous scaffold model region are extracted.

[0029] S32, 3D calculations are performed on the segmented new bone tissue region using Mimics software to obtain the volume (BV) of the new bone region; the total volume MV of the porous scaffold model structure is extracted, and the volume ratio (BV / MV) of the new bone to the porous scaffold model is calculated.

[0030] S33, the new bone is pre-installed uniformly on the outer surface of the porous scaffold model, and the BV / MV is converted into the equivalent thickness rb of the new bone wrapping layer.

[0031] Specifically, converting BV / MV into the equivalent thickness rb of the new bone wrapping layer includes: calculating the volume percentage: rv = MV / BV;

[0032] The equivalent thickness is calculated using the formula:

[0033] Specifically, step S6 includes:

[0034] S61, import the double-layer composite model structure into the Ansys Mechanical module, and assign different material properties to the support part and the new bone part in the structure according to their geometric partitions.

[0035] S62, a fixed support is applied to the bottom surface of the stent, and an axial forced displacement is applied to the top surface of the stent to simulate the working condition of the implanted structure being subjected to compressive load;

[0036] S63, calculate and solve, output the stress and strain data of the overall structure, and obtain mechanical simulation data including the elastic modulus.

[0037] Secondly, the present invention proposes a device for dynamic prediction of mechanical properties after stent implantation, characterized in that it includes a memory and one or more programs, wherein one or more programs are stored in the memory and configured to be executed by one or more processors, wherein the one or more programs include the steps of the method for dynamic prediction of mechanical properties after stent implantation.

[0038] Thirdly, the present invention provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the dynamic prediction method for mechanical properties after stent implantation as described in any one of the claims.

[0039] The present invention provides a method for dynamically predicting the mechanical properties of a stent after implantation, which achieves the following beneficial technical effects:

[0040] This invention proposes a dynamic prediction method for scaffold performance that combines 3D reconstruction from CT images, finite element simulation analysis, and data-driven modeling. It constructs a complete mapping relationship from "scaffold model structural parameters → new bone growth characteristics → mechanical response." By building a training dataset using extensive simulation data and introducing appropriate analysis methods to establish a prediction model, it can not only efficiently predict performance trends under different structural designs but also optimize scaffold parameters in reverse, improving its individualized adaptability and function-oriented design capabilities. Furthermore, by incorporating real bone growth data into the simulation, it covers complex structural parameter spaces, meeting the needs for rapid scaffold optimization design. This invention improves and optimizes the minimum repeating unit of the model structure and, combined with the actual growth of new bone tissue, can be more accurately and realistically applied to predict the mechanical properties of bone tissue growth after scaffold implantation. Finally, extensive experiments demonstrate the effectiveness and superiority of our algorithm, showcasing its broad application prospects. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a schematic diagram illustrating the process of dynamically predicting the mechanical properties of the stent after implantation in an embodiment of the present invention.

[0043] Figure 2 This is a schematic diagram of the smallest repeating unit of the octahedral structure support in an embodiment of the present invention.

[0044] Figure 3 This is a schematic diagram of the smallest repeating unit of the diamond structure support in an embodiment of the present invention.

[0045] Figure 4 This is a schematic diagram of the solid model after the three-dimensional array of the smallest repeating unit in an embodiment of the present invention;

[0046] Figure 5 This is a model diagram of a cylindrical support with a height of 1.5mm and a diameter of 5mm generated in an embodiment of the present invention;

[0047] Figure 6 This is a CT image taken at a certain time point after the stent was implanted in the skull defect area of ​​a rat in an embodiment of the present invention;

[0048] Figure 7 This is a schematic diagram of a 3D reconstruction support for CT images in an embodiment of the present invention;

[0049] Figure 8 This is a model diagram of the porous scaffold-new bone double-layer structure established in an embodiment of the present invention;

[0050] Figure 9 This is a schematic diagram of the boundary conditions set for the two-layer model in Ansys for finite element analysis in an embodiment of the present invention;

[0051] Figure 10 This is a stress diagram of the double-layer model after static compression simulation in an embodiment of the present invention;

[0052] Figure 11 This is a stress-strain curve extracted from static compression simulation in an embodiment of the present invention. Detailed Implementation

[0053] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0054] In the following embodiments of the present invention, the specific exemplary technical solutions are for illustrative purposes only, and the implementation of the technical solutions is not limited thereto.

[0055] First Embodiment

[0056] like Figure 1 As shown, this invention provides a dynamic prediction modeling method for the mechanical properties of scaffolds used in bone tissue repair, specifically including the following steps:

[0057] In step S1, the model of the minimum repeating unit of the stent structure is completed according to the pre-set model structure type of the stent to obtain the integrated stent structure of the minimum repeating unit.

[0058] Specifically, bone tissue can have different pore structures. For scaffolds with different pore types within the bone tissue, there are corresponding different types of minimum repeating unit structures. The structure type can be selected or adjusted according to actual needs; for example... Figure 2 and 3As shown, specifically, for example, octahedral or diamond structures, the smallest repeating unit is the smallest geometric unit constituting the three-dimensional periodic structure of the porous scaffold—that is, this unit possesses "repeatability": by array replication in the X, Y, and Z directions, replication operations including, for example, translation and stacking, without changing the structure of the smallest repeating unit itself, a large-scale scaffold of arbitrary size and uniform structure can be generated. The core characteristics of the smallest repeating unit are "minimum" and "periodicity". Therefore, based on the structural characteristics of the smallest repeating unit, in one embodiment, by selecting a preset structure type and performing replication operations on the smallest repeating unit, a preliminary integrated scaffold structure of bone tissue can be obtained by modeling the bone tissue structure.

[0059] In another embodiment, in order to obtain an integrated scaffold structure of the minimum repeating unit that is closer to the bone tissue type, the present invention further includes preprocessing the modeling of the minimum repeating unit. The preprocessing process includes: generating node spheres and connecting cylinders in a three-dimensional modeling environment according to the preset model structure type and the vertex and rod connection information of the minimum repeating unit; iterating the rod radius by using the target density and the side length of the minimum unit; and merging the generated structure into an integrated scaffold structure by Boolean operation.

[0060] Specifically, the node spheres and connecting cylinders together constitute the geometric entity of the minimum repeating unit: the node spheres correspond to the vertices defined in the minimum unit vertex file, representing the positions of the lattice nodes; the connecting cylinders correspond to the vertex index pairs defined in the rod connection file, connecting adjacent nodes to form rods. Together, they constitute the complete spatial model structure of the minimum repeating unit. Specifically, both the vertex coordinate file and the rod connection relationship file of the minimum repeating unit can be set to plain text format. The former defines standardized node coordinates, with a file format of xyz; the latter defines the connection indexes between nodes, with a file format of node1 node2. The files themselves do not contain physical quantities; the actual structural size and density are determined by the parameters of the set minimum repeating unit—the minimum repeating unit structure side length lside and the rod radius rc. In step S1, the vertex and rod connection information required for the structure type can be read from a standard structure library or customized by the researcher.

[0061] Furthermore, for the selected minimum repeating unit, the vertex coordinate file information and rod connection file information of the initial minimum repeating unit are first read, and the vertex coordinates are scaled to actual spatial coordinates using the actual unit side length lside. Using the initial value rc0 (rc0 = lside / 6) as the node radius and rod radius, the final rod radius is determined according to the following iterative process: For the current radius rc, the Gmsh OpenCASCADE interface gmsh.model.occ.addSphere(x,y,z,rc) is called point by point to generate a spherical node at each vertex, and gmsh.model.occ.addCylinder(x,y,z,dx,dy,dz,rc) is called pair by pair to generate a cylindrical rod between each pair of connecting nodes. Secondly, after completing the geometric generation of the node spheres and connecting rod cylinders, the Boolean union function gmsh.model.occ.fuse() provided by the Gmsh–OpenCASCADE kernel is called to perform a fusion operation on all generated 3D entities.

[0062] Furthermore, the input object for this fusion operation is a set of three-dimensional entities composed of spheres at each node and cylinders at each member. Each entity is recognized by the system as an independent three-dimensional body in the form of (3, tag). During the fusion process, the OpenCASCADE kernel automatically identifies the intersecting or adjacent regions between these geometries and performs geometric union and topological reconstruction: that is, merging the parts of the originally independent entities that overlap or contact in space into a unified body, deleting overlapping areas of internal repeated faces, edges, and volumes, and maintaining the continuous and smooth outer boundary, thereby obtaining a topologically connected overall structure. After fusion, all nodes and members are geometrically seamlessly connected, and there are no gaps or repeated volumes in the overall three-dimensional body. This continuous structure is the complete support model of the minimum repeating unit, which is a closed single entity both geometrically and topologically. It can be directly exported to STEP format and used for subsequent array copying, trimming, and finite element analysis. The geometric structure of the minimum repeating unit is as follows: Figure 2 , Figure 3 As shown. Figure 2 It has an octahedral structure. Figure 3 It has a diamond structure. Further, the volume M of this solid support model is obtained using gmsh.model.occ.getMass(3,tag). lattice `getMass()` is a method of the OpenCASCADE shape object used to retrieve the "mass attribute" of the geometry. When the object is a 3D solid, it retrieves its volume. This is then expressed as the volume enveloped by the unit cells. Calculate the current relative density Where nx, ny, and nz represent the number of arrays of the smallest repeating units in the X, Y, and Z directions, which are determined by the target size of the support and the unit side length lside, density current This indicates the density of the current scaffold (or smallest repeating unit). It is then compared to the target density. current The difference tol = density target -density current， The density of the simulated porous scaffold is determined by the design accuracy requirements; firstly, based on the porosity of the scaffold actually implanted in animal experiments, the target density is set. target The value is 0.35 (i.e., 35% solid fill rate; scaffolds of this density can better induce new bone tissue growth). If |tol| exceeds the preset tolerance, the rod radius is adjusted according to the error direction to update rc, and the update formula is as follows:

[0063]

[0064] Because the rods and nodes intersect at the ends, the sum of the total volume of the solid and the volume of a single rod is not linearly added (there is overlap). Therefore, it is impossible to accurately solve for the required rc using simple algebraic formulas. Through a closed-loop iteration of "modeling → calculating the actual volume → comparing density → updating rc", the actual rc that meets the target density can be found after considering geometric overlap and Boolean union. Therefore, the above modeling-fusion-volume measurement process is repeated until the density error meets the tolerance or the maximum number of iterations is reached; finally, the converged solid is exported (STEP) and used as the minimum repeating unit prototype for array replication and trimming. This iterative method can ensure that the obtained rod diameter meets the porosity target of the design under the consideration of node-rod overlap and Boolean union deduplication. After iterating to obtain the rc value under the final target density, the final three-dimensional integrated support structure is generated. It should be noted that the target porosity can be set independently according to actual needs and adjusted according to the actual application scenario; its set value is not limited to this.

[0065] Through this embodiment of the invention, the smallest repeatable unit that conforms to bone tissue and the bone tissue growth process can be obtained, thereby improving the accuracy of mechanical property prediction in subsequent steps.

[0066] In step S2, the integrated stent structure is array-replicated, and a Boolean cropping operation is performed using a cropping tool to generate a porous stent model that conforms to the preset implantation area size.

[0067] Specifically, in order to obtain a porous scaffold model that meets the target scaffold size requirements, the smallest repeating unit is copied and trimmed in a three-dimensional array, such as... Figure 4As shown, in order to construct a complete scaffold structure that conforms to the size of the target implantation area, the translate_lattice() method is used to perform array replication of the smallest repeating unit in the X, Y, and Z directions, and Boolean operations are used for cropping and fusion.

[0068] Specifically, in step S2, the array is copied to the smallest repeating unit of the fused state obtained in step S1, which corresponds to the entity of the integrated scaffold structure containing node spheres and connecting pillars. The copying process uses this unit as a template, performing translational copying in the X, Y, and Z directions according to the array parameters [nx, ny, nz], with a translation step size of the unit's side length lside. The copied content includes the unit's complete geometry, topological relationships, and physical properties, ensuring the periodicity and continuity of the overall structure after arraying. After arraying, a general scaffold with an uncut shape is formed, with dimensions of (nx × lside) × (ny × lside) × (nz × lside). This general scaffold itself does not yet meet the requirements of the target scaffold size, and is subsequently cut to the target implantation area size through Boolean operations.

[0069] In one specific embodiment, the Boolean operation clipping and fusion process in step S2 includes the following.

[0070] (1) The composition of the trimmed object: the general support formed by the smallest repeating unit of the fusion state generated in step S1, which is copied in the X, Y and Z directions according to the array parameters [nx,ny,nz].

[0071] Cutting tool: A geometric entity generated by pre-setting the target implantation area size. For example, a cylindrical entity is used as the cutting tool. In this embodiment, it is a solid cylinder with a diameter of 5mm and a height of 1.5mm. It can be directly generated by the addCylinder() function of Gmsh (parameters: bottom center coordinates, axial vector, radius, height. In this embodiment, it is a cylinder with "bottom center (3mm, 3mm, 0mm), axial direction along the Z-axis, radius 2.5mm (diameter 5mm), and height 1.5mm".

[0072] (2) Input and Execution of Boolean Operations: The inputs to the operation are the 3D entity label list `arrayed_entities` of the overall scaffold and the entity label list `tool_entity` of the trimming tool. Specifically, a Boolean intersection operation (`intersect`) is used to perform a Boolean operation on the overall scaffold structure and the cylindrical trimming body, retaining the structural regions within the cylindrical area, thereby obtaining a porous scaffold model that matches the target implantation area. In practice, `gmsh.model.occ.intersect(arrayed_entities,tool_entity)` is called to calculate the spatial intersection of the two, retaining only the overlapping portion, and returning a new entity label representing the trimmed scaffold. This `intersect` operation can implement the Boolean intersection operation, thus achieving the trimming function. A schematic diagram of the trimming result is shown below. Figure 5 As shown.

[0073] In one embodiment, before performing Boolean operation trimming, it is necessary to align the general scaffold structure (after array copying and before trimming) with the trimming tool. The purpose of this alignment is to ensure that the trimming tool perfectly covers the material being trimmed, ensuring that the center of the trimmed scaffold coincides with the center of the target implantation area. The Boolean operation is performed after aligning the axis of the trimmed cylinder with the axis of symmetry of the general scaffold. This operation improves the accuracy of obtaining the target porous scaffold model.

[0074] In step S3, the equivalent thickness of the new bone is calculated based on the volume of the new bone reconstructed from the CT image in three dimensions. This thickness is added to the radius of the support rod mentioned in step S1, and the increased thickness is used to generate the three-dimensional model structure of the new bone part.

[0075] Specifically, this invention uses 3D reconstruction of CT images from different time periods to obtain the volume of newly grown bone tissue. Specifically, the 3D reconstruction of the new bone tissue region is performed using Mimics software. First, 3D calculations are automatically performed on the segmented new bone region to calculate the volume (BV) of the new bone region. Specifically, voxel labeling is performed first, then isosurface extraction of triangular patches is used to generate the new bone volume (BV): BV = number of voxels × voxel size 3. The volume of the porous scaffold model region is obtained using the same method.

[0076] To obtain data on new bone growth at different time points, this embodiment performed CT scans on the scaffold area at weeks 4, 6, 8, 10, and 12 after scaffold implantation in rat skulls. Figure 6 As shown, the data is then imported into Mimics software for processing, specifically including the following steps:

[0077] S31, a threshold segmentation method is used to separate tissues in the CT image. The segmentation threshold is set based on the density difference between the porous scaffold model and the new bone, extracting the newly formed bone tissue region attached to the outer surface of the porous scaffold model and the porous scaffold model region. Figure 7 As shown, the set HU value range for the new bone tissue region is 6600–14775. After segmentation, three-dimensional reconstruction calculations are performed on the new bone region to obtain its volume data, denoted as BV (Bone Volume). Specifically, tissue separation is performed using the different density characteristics (manifested as HU value differences) of the scaffold material and new bone in CT images. The histograms of HU value distributions in the scaffold and new bone regions are analyzed to determine the optimal segmentation threshold range: generally, the HU value range of new bone is 6600–14775 HU (high-density cortical bone), while the scaffold is usually higher than 15000 HU. The segmentation threshold is manually adjusted by manually adjusting the HU value to ensure separation between the two tissues.

[0078] S32, extract the total volume of the porous scaffold model structure, denoted as MV (Material Volume). By calculating the volume ratio of new bone to the porous scaffold model (i.e., BV / MV), the degree of new bone growth on the scaffold surface can be preliminarily reflected.

[0079] S33, assuming the new bone is uniformly attached to the outer surface of the scaffold, convert BV / MV into the equivalent thickness of the new bone wrapping layer, denoted as rb. This thickness value rb will be used as an input parameter in the subsequent modeling of the new bone outer layer structure (step S4) to set the radius increment of the new bone members. Specifically, calculate the volume ratio: rv = MV / BV. Calculate the equivalent thickness using the formula: The radius increment of the new bone outer layer is passed to step S4 for geometric modeling (adding rb to the radius of the original scaffold member in Gmsh). Here, r0 refers to the "radius of the original member of the porous scaffold model", and is the same parameter as the member radius "rc" mentioned in the previous steps.

[0080] In step S4, steps S1–S2 are repeated, and the outer new bone model is constructed using the increased thickness described in step S3 as the updated rod radius.

[0081] Specifically, in step S4, based on the new bone growth thickness information obtained from the animal experiment CT images reconstructed in step S3, the radius parameter r of each rod in the basic scaffold structure is... c Increase new bone thickness b The updated radius r of the new bone layer member is obtained. bone :

[0082] r bone =r c +r b

[0083] Based on this, the structural modeling process described in steps S1 to S2 is repeated, that is, the structural nodes and connection relationships are retraced to generate a new bone member model with updated rod diameters, and a new three-dimensional bone model structure is constructed to cover the porous scaffold model. The final structure is exported in STEP file format for subsequent generation of the porous scaffold model-new bone double-layer composite structure.

[0084] In a specific embodiment, a new bone thickness r is generated for each time point in step S3 (weeks 4, 6, 8, 10, and 12). b The radius r of each member of the basic porous support is respectively... c Increase r b This generates a porous scaffold model—a new bone bilayer composite structure model—corresponding to each time point. Therefore, for example, this embodiment yields a total of 5 composite structure models, with a new bone thickness r generated weekly. b It is used to dynamically analyze the impact of new bone growth on the morphology and mechanical properties of the scaffold.

[0085] In step S5, the inner porous scaffold model and the outer new bone model are imported into a 3D modeling environment, and Boolean operations are used to generate a porous scaffold model-new bone bilayer composite model structure for finite element analysis. Specifically, the porous scaffold model constructed in step S2 and the new bone outer layer structure model constructed in step S4 are jointly imported into the Python-Gmsh modeling environment, and Boolean intersection and union operations are performed to construct the porous scaffold model-new bone bilayer composite model, as follows. Figure 8 As shown.

[0086] Specifically, within Gmsh's Open CASCADE geometry kernel, Boolean fusion operations (such as fuse() or intersect()) are performed to integrate the inner scaffold and the outer neobone structure into a two-layer composite model with a mechanical contact interface. This composite structure exhibits distinct material layering characteristics, allowing different material parameters to be assigned to different regions in subsequent finite element analysis (e.g., setting the elastic modulus and density for the scaffold body and the neobone overlay layer, respectively). The final fused 3D geometry is exported in standard STEP file format for subsequent finite element simulation analysis.

[0087] In step S6, the bilayer composite model is imported into the Mechanical module of the Ansys Workbench software platform, static compression simulation analysis is performed, stress-strain data is output, stress-strain curves are extracted, and the elastic modulus of the scaffold-new bone bilayer structure is calculated.

[0088] Specifically, and understandably, the Ansys Workbench platform is a multiphysics simulation integration platform that can uniformly manage the entire process of "model import - mesh generation - load setting - solution - post-processing". It can realize the conventional parameter setting and mechanical simulation process of mechanical simulation. Among them, the Mechanical module is the core module in Workbench that focuses on structural mechanics simulation. It can realize geometric partitioning, material assignment, boundary condition application, mesh generation, and mechanical result (stress, strain) calculation and post-processing. Through this module, the static compression simulation of this invention can be directly completed.

[0089] Specifically, the double-layer composite model is imported into finite element software, and different material properties are set according to the porous scaffold model and the new bone part model. The plane where the scaffold bottom is located is fixed, and a 1% axial forced displacement in the negative Z-axis direction is applied as a boundary condition. Static compression simulation is performed, stress and strain data are output, stress-strain curves are extracted, and the elastic modulus of the scaffold-new bone double-layer structure is calculated.

[0090] In one embodiment, performing a static compression simulation analysis specifically includes the following steps:

[0091] S61 imports the two-layer composite model geometry exported in STEP format into the Ansys Mechanical module. By distinguishing between the two parts of the geometric model, different material parameters are assigned to the two regions. Based on the material properties of different regions in the structure, the parameters are set as follows:

[0092] Scaffold region: Young's modulus 3.6 GPa; Poisson's ratio 0.28; density 3.14 g / cm³ 3 .

[0093] New bone tissue region: Young's modulus 1 GPa; Poisson's ratio 0.3; density 1.36 g / cm³ 3 .

[0094] The geometric model differentiation is achieved through the "Geometric Selection Function" of the Mechanical module. After importing the STEP model, the module preserves the original boundaries between the porous scaffold model and the new bone model. It uses the "Volume Selection Tool" to identify and select the inner porous scaffold model and the outer new bone, creating "Geometric Groups" (such as "Stent" and "Bone") respectively to ensure the partitioning function.

[0095] Furthermore, the material parameters are manually set based on experimental / literature data (mechanical properties of scaffolds and newborn bone). Using the "Material Library" function in the Mechanical module, corresponding materials are first created (by inputting Young's modulus, Poisson's ratio, etc.), and then the materials are assigned to the corresponding "Geometry Group." The module automatically associates the mesh elements, ensuring that different regions use different mechanical properties during simulation. The numerical values ​​of the material parameters are not limited to these; any parameter settings that can distinguish different regions from different mechanical properties can be used.

[0096] S62, to simulate the vertical compressive load scenario, the following boundary conditions are set:

[0097] The bottom surface of the support is set as a fixed support, constraining the degree of freedom in the Z-axis direction;

[0098] A 1% axial forced displacement boundary condition is applied to the top surface of the support, in the negative Z-axis direction, to simulate the axial compression process, such as... Figure 9 As shown.

[0099] S63 employs a tetrahedral meshing method for the two-layer composite model, further refining the mesh size to 0.05mm to accommodate complex curved surfaces and interface transition regions, thereby improving simulation accuracy. The simulation type is selected as Static Structural, and the solver is automatically selected and controlled by Ansys.

[0100] S64: After setting the boundary conditions, material properties, and mesh generation, start the simulation solution process.

[0101] The simulation calculation records the mechanical response data at each time step, including information such as overall reaction force, structural deformation, and stress distribution. Figure 10 As shown. After the simulation, the relationship between the total reaction force on the loading surface and the applied displacement is extracted using the post-processing module, and the stress and strain data corresponding to the compressive load are calculated. The obtained data is further exported as a stress-strain table, and stress-strain curves (σ–ε) are plotted to evaluate the structural mechanical properties, such as... Figure 11 As shown.

[0102] In step S7, the CT images at different time points are acquired, and steps S3–S6 are repeated to generate mechanical simulation data for the corresponding time points.

[0103] Specifically, the above steps utilize the reconstruction results of CT images of rat skulls implanted with stents to construct a two-layer composite model of porous stent model-new bone model at different time periods, and perform finite element static simulations to obtain the elastic modulus at different time periods.

[0104] In step S8, the preset model structure type is updated, and steps S1-S7 are repeated to obtain more mechanical simulation data corresponding to model structure parameters and new bone growth characteristic parameters. All the above model structure parameters and corresponding mechanical simulation data are used to construct a complete dataset and establish a mapping relationship model between model structure parameters and mechanical performance response. This model enables dynamic prediction of mechanical performance after stent implantation.

[0105] Specifically, to establish a predictive model with adaptability and generalization ability, multiple sets of bilayer composite scaffold models with different combinations of structural parameters (including new bone thickness, scaffold density, and member dimensions) were generated. Finite element simulations were performed in batches to obtain stress-strain response data and effective elastic modulus for each structure. All the simulation data were then compiled into a complete dataset, including input features (combinations of structural parameters) and output targets (elastic modulus). Various data analysis methods were employed for modeling, such as multiple linear regression, support vector regression (SVR), or random forest. Finally, a mapping model from "structural parameters → new bone growth thickness → mechanical performance response" was established, enabling dynamic prediction of the mechanical performance after scaffold implantation. This predictive model can be used to assist in optimizing scaffold design and individualized parameter tuning strategies.

[0106] In one embodiment, the prediction model specifically includes:

[0107] 1. Input: To improve the correlation with actual bone tissue growth and enhance the accuracy of predicting the elastic modulus of real bone tissue, the model structural parameters in this invention encompass the new bone tissue growth parameters. Therefore, the model structural parameters include a combination of "structural design parameters + new bone growth characteristic parameters," specifically comprising two types of parameters:

[0108] Scaffold model structural design parameters: minimum repeating unit side length (lside), original radius of scaffold member (rc), target density of scaffold, number of arrays (nx / ny / nz);

[0109] New bone growth characteristic parameters: new bone equivalent thickness (rb, obtained by converting BV / MV from CT reconstruction at different time points), and new bone growth time nodes (e.g., 4 weeks / 6 weeks, indirectly reflecting the change pattern of rb).

[0110] 2. Output: The output is the mechanical performance index of the scaffold-new bone double-layer composite structure. The core index in this invention is the equivalent elastic modulus (calculated from the stress-strain curve extracted by finite element simulation).

[0111] 3. The acquisition of the training set consists of two parts: "experimental data supplementation" and "parametric simulation extension," ensuring both data volume and coverage.

[0112] Supplementary experimental data: Based on five time points (4 / 6 / 8 / 10 / 12 weeks) of rat skull implantation experiments, a two-layer composite model was constructed at each time point, and finite element simulation was performed to obtain five sets of "parameter-elastic modulus" data, which served as verification data in real biological scenarios.

[0113] Parametric simulation extension: Using Ansys Workbench's parametric design function, multiple sets of models with different parameter combinations can be generated in batches (e.g., rc = 0.07-0.11 mm, rb = 0.01-0.05 mm, density = 0.3-0.4). Each set of models automatically performs static simulation and extracts the elastic modulus, generating hundreds to thousands of sets of data (e.g., 500 sets) as the main data of the training set.

[0114] Data cleaning: Remove data that contains simulation errors (such as poor mesh quality) or outliers (such as elastic modulus far exceeding the theoretical range of materials) to form a complete training set.

[0115] 4. Learn the mapping pattern of "input parameters → output elastic modulus" in the training set. In actual modeling, compare the prediction errors (such as root mean square error RMSE) of the three algorithms through "cross-validation" and select the algorithm with the smallest error to finally build the prediction model.

[0116] 5. The input for using the model is the same as the input for the training set, which is "scaffold design parameters + new bone growth parameters", and no additional data is required.

[0117] The output during use directly outputs the predicted mechanical properties of the double-layer structure under the corresponding input parameters, with the core being the equivalent elastic modulus.

[0118] The output is "Equivalent elastic modulus under this parameter combination = 1.8 GPa". If this value meets clinical requirements (e.g., close to the host bone elastic modulus of 1.5-2.0 GPa), then the scaffold design is feasible. If it does not meet the requirements, the input parameters can be adjusted in reverse (e.g., increase the target density), and the model can be re-entered for verification to achieve rapid optimization.

[0119] This invention proposes a dynamic prediction method for scaffold performance that combines 3D reconstruction from CT images, finite element simulation analysis, and data-driven modeling. It constructs a complete mapping relationship from "scaffold model structural parameters → new bone growth characteristics → mechanical response." By building a training dataset using extensive simulation data and introducing appropriate analysis methods to establish a prediction model, it can not only efficiently predict performance trends under different structural designs but also optimize scaffold parameters in reverse, improving its individualized adaptability and function-oriented design capabilities. Furthermore, by incorporating real bone growth data into the simulation, it covers complex structural parameter spaces, meeting the needs for rapid scaffold optimization design. This invention improves and optimizes the minimum repeating unit of the model structure and, combined with the actual growth of new bone tissue, can be more accurately and realistically applied to predict the mechanical properties of bone tissue growth after scaffold implantation. Finally, extensive experiments demonstrate the effectiveness and superiority of our algorithm, showcasing its broad application prospects.

[0120] Example 2

[0121] The present invention also provides another embodiment, which proposes a dynamic prediction device for mechanical properties after stent implantation, comprising: a processor 1 and a memory 2.

[0122] The memory 2 is used to store computer programs.

[0123] The memory 2 includes various media capable of storing program code, such as ROM, RAM, magnetic disk, USB flash drive, memory card, or optical disk.

[0124] The processor 1 is connected to the memory 2 and is used to execute the computer program stored in the memory 2 so that the processor can implement the above-mentioned method for dynamic prediction of mechanical properties after stent implantation when executing the computer program.

[0125] Preferably, the processor 1 may be a central processing unit (CPU); or it may be an application-specific integrated circuit (ASIC).

[0126] Example 3

[0127] The present invention also provides another embodiment, namely, a computer-readable storage medium storing a computer program that can be executed by at least one processor to cause the at least one processor to perform the steps of the dynamic prediction method for mechanical properties after stent implantation as described above.

[0128] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0129] Those skilled in the art will recognize that the modules, units, and / or method steps of the various embodiments 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. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0130] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another device or system, or some features may be ignored or not executed.

[0131] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated unit described above can be implemented in hardware or as a software functional unit.

[0132] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for dynamically predicting the mechanical properties of a stent after implantation, characterized in that, The method includes the following steps: Step S1: Based on the pre-set model structure type of the support and the vertex and rod connection information of the smallest repeating unit, generate node spheres and connecting columns in the 3D modeling environment. Iterate the rod radius by the target density and the side length of the smallest unit. After iteration, the generated structure is merged into an integrated support structure through Boolean operations. Step S2: The integrated stent structure is array-replicated, and a Boolean cropping operation is performed using a cropping tool to generate a porous stent model that conforms to the preset implantation area size; Step S3: Reconstruct the volume of the new bone in three dimensions based on the CT images, calculate the equivalent thickness of the new bone, and add the thickness to the radius of the strut in the scaffold described in step S1. The increased thickness is used to generate the three-dimensional model structure of the new bone portion. Step S4: Repeat steps S1–S2, using the increased thickness described in step S3 as the updated rod radius to construct the outer new bone model; Step S5: Import the inner porous scaffold model and the outer new bone model into the three-dimensional modeling environment, and generate a porous scaffold model-new bone bilayer composite model structure for finite element analysis through Boolean operations; Step S6: Import the double-layer composite model into the Mechanical module of the Ansys Workbench software platform, perform static compression simulation analysis, and calculate the elastic modulus of the porous scaffold model-new bone double-layer composite model structure. Step S7: Acquire the CT images at different time points, and repeat steps S3–S6 to generate mechanical simulation data for the corresponding time points. Step S8: Update the preset model structure type, repeat steps S1-S7 to obtain more mechanical simulation data corresponding to model structure parameters and new bone growth characteristic parameters, construct a complete dataset with all the above model structure parameters and corresponding mechanical simulation data, and establish a mapping relationship model between model structure parameters and mechanical performance response. This model realizes dynamic prediction of mechanical performance after stent implantation.

2. The method for dynamic prediction of mechanical properties after stent implantation according to claim 1, characterized in that, Step S1 includes the following steps: S11, determine the type of the smallest repeating unit of the selected preset model structure, define the unit side length parameter lside, and export the node vertex information and rod connection information in the smallest repeating unit; S12, Based on the connection information of the rods, determine the connection relationship between each pair of nodes to generate an initial index list of connecting rods; S13, traverse the above node information, use modeling tools to generate spherical node spheres and rod cylinders, and set the node radius and rod radius to a preset rc; S14, merge the node sphere and the rod cylinder into a complete three-dimensional lattice structure through a Boolean fusion operation to obtain the structural model of the minimum repeating unit; S15, Obtain the solid volume M of the structural model. lattice Simultaneously calculate the volume M of the minimum repeating element bounding box. box Then, the current stent density is calculated; S16, compare the current density with the preset target density error. If the error does not meet the tolerance condition, adjust the radius of the rod according to the error direction and repeat S13-S15 using the new radius. S17. When the error meets the preset tolerance or reaches the maximum number of iterations, the current integrated support structure is saved.

3. The method for dynamic prediction of mechanical properties after stent implantation according to claim 2, characterized in that, The solid volume M of the structural model is obtained. lattice Specifically: Use gmsh.model.occ.getMass to obtain the solid volume M of the structural model. lattice ; Calculate the volume M of the minimum repeating element bounding box. box Specifically: Calculate the current stent density Where nx, ny, and nz represent the number of arrays of the smallest repeating units in the X, Y, and Z directions, which are determined by the target size of the support and the unit side length lside, density current This represents the density of the smallest repeating unit.

4. The method for dynamic prediction of mechanical properties after stent implantation according to any one of claims 1-3, characterized in that, The calculation of new bone thickness based on three-dimensional reconstruction of new bone volume from CT images specifically includes: S31, a threshold segmentation method is used to separate tissues in CT images. The segmentation threshold is set according to the density difference between the porous scaffold model and the new bone, and the newly formed bone tissue region attached to the outer surface of the porous scaffold model and the porous scaffold model region are extracted. S32, 3D calculations are performed on the segmented new bone tissue region using Mimics software to obtain the volume (BV) of the new bone region; the total volume MV of the porous scaffold model structure is extracted, and the volume ratio (BV / MV) of the new bone to the porous scaffold model is calculated. S33, the new bone is pre-attached uniformly to the outer surface of the porous scaffold model, and the BV / MV is converted into the equivalent thickness rb of the new bone wrapping layer.

5. The method for dynamic prediction of mechanical properties after stent implantation according to claim 4, characterized in that, The process of converting BV / MV into the equivalent thickness rb of the new bone wrapping layer specifically includes: calculating the volume percentage: rv = MV / BV; The equivalent thickness is calculated using the formula:

6. The method for dynamic prediction of mechanical properties after stent implantation according to any one of claims 1-3, characterized in that, Step S6 specifically includes: S61, import the double-layer composite model structure into the Ansys Mechanical module, and assign different material properties to the support part and the new bone part in the structure according to their geometric partitions. S62, a fixed support is applied to the bottom surface of the stent, and an axial forced displacement is applied to the top surface of the stent to simulate the working condition of the implanted structure being subjected to compressive load; S63, calculate and solve, output the stress and strain data of the overall structure, and obtain mechanical simulation data including the elastic modulus.

7. A device for dynamically predicting the mechanical properties after stent implantation, characterized in that, It includes a memory and one or more programs, wherein one or more programs are stored in the memory and configured to be executed by one or more processors, wherein the one or more programs include steps for performing the method for dynamic prediction of mechanical properties after stent implantation as described in any one of claims 1-6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method for dynamically predicting the mechanical properties after stent implantation as described in any one of claims 1-6.

Images