Design method and system of spinal surgery preset screw rod based on big data

Abstract

The invention discloses a big-data-based design method and system for a preset screw rod in spinal surgery, and particularly relates to the technical field of preset screw rod design, and the method comprises the following steps: collecting preoperative medical image data of a target patient, extracting anatomical feature parameters, and establishing a surgery fixing instrument database containing historical patient data and implanted instrument parameters; evaluating the degree of adaptation between the standard nail rod and the anatomical structure of the patient and the postoperative fixation stability, dividing the patient matching grade based on the evaluation result, screening patient groups with similar anatomical structures and biomechanical characteristics by adopting a clustering analysis method, and calculating a personalized customization data set in combination with the individual parameters of the target patient; according to the method, the structural adaptation index and the biomechanical adaptation index are adopted, it is ensured that the matching performance of fixing instruments can be judged before an operation, when the standard screw rod is not suitable, the personalized customization process can be automatically triggered, the judgment deviation of a doctor is reduced, and the accuracy of customized screw design is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pre-set nail rod design, and more specifically, to a design method and system for pre-set nail rods for spinal surgery based on big data. Background Art

[0002] Spinal surgery is widely used in spinal deformity correction, trauma repair and degenerative disease treatment. In traditional surgery, doctors usually select standardized screws and rods for implantation based on the patient's imaging results and personal clinical experience, and adjust them according to actual conditions during the operation. However, this method has many limitations: Standard screws and rods have limited adaptability to individual anatomical structures: Due to individual differences in the spinal anatomy of each patient, such as significant variations in pedicle diameter, length, implantation angle, and bone density, the standardized design of fixation devices cannot fully meet the needs of all patients. Standard screws may be too long or too short, have mismatched diameters, or have angle deviations, leading to unstable fixation after surgery and even increasing the risk of screw loosening or fracture.

[0003] Surgery relies on the doctor's experience, and intraoperative adjustments are difficult: Existing surgical planning methods are mainly based on the doctor's subjective judgment. It is difficult to accurately predict the screw specifications before surgery, resulting in the need to constantly adjust the screw implantation method during surgery, or even replace screws of different specifications, which prolongs the operation time. For complex cases such as scoliosis and vertebral slippage, traditional methods are difficult to provide the best preoperative fixation solution, resulting in unsatisfactory postoperative correction results.

[0004] There are risks in postoperative stability, which may lead to complications: if the implanted screw does not match the patient's bone, it may cause postoperative stress concentration, resulting in screw loosening, displacement or breakage, increasing the possibility of secondary surgery. During the postoperative recovery process, if the biomechanical adaptability is insufficient, the fixation system may fail, affecting the patient's long-term prognosis.

[0005] The existing fixation device design fails to combine personalized biomechanical assessment and lacks intelligent matching and recommendation mechanisms, resulting in the inability to fully meet personalized needs. Therefore, a design method and system for pre-set screws and rods for spinal surgery based on big data is proposed to solve the above problems. Summary of the invention

[0006] To achieve the above object, the present invention provides the following technical solutions: The design method of pre-set screw rods for spinal surgery based on big data includes the following steps: Collect the preoperative medical imaging data of the target patient, extract the patient's anatomical characteristic parameter group, and establish a surgical fixation device database, including historical patient data and implant device parameters; The degree of fit between the standard nail rod and the patient's anatomical structure and the degree of postoperative fixation stability are evaluated respectively, and then the patient's matching level is divided into high matching level and low matching level based on the evaluation results. The personalized customization mechanism is triggered when the matching level is low; When the personalized customization mechanism is triggered, the cluster analysis method is used to cluster the patient data in the historical surgery database, select patient groups with similar anatomical structures and biomechanical characteristics, and combine the individual parameters of the target patients to obtain a personalized customization data group.

[0007] In a preferred embodiment, the degree of fit between the standard nail rod and the patient's anatomical structure is evaluated to generate a structural fit index, and the degree of postoperative fixation stability between the standard nail rod and the patient is evaluated to generate a biomechanical fit index.

[0008] In a preferred embodiment, the logic for obtaining the structural adaptation index is: The patient's pedicle diameter is obtained from the patient's anatomical feature parameter group and marked as , standardized screw diameters matched to the patient and marked , the straight-line distance from the pedicle entrance to the exit of the patient after the safety margin adjustment is marked as , standard screw length and marked as , obtain the patient's pedicle axial angle and mark it as , and the standard screw axial angles are marked as , and then substitute it into the following calculation formula: ; , , and are all preset non-zero scale factors, is the structural fit index.

[0009] In a preferred embodiment, the logic for obtaining the biomechanical fitness index is: The optimal fixation force of the screw in the patient's spine obtained by finite element analysis is obtained and marked as , the fixing capacity of standard specification screws and marked as , the optimal moment stability index of the patient's spine and is marked as , the torque stability of the standard screw and is marked as , the screw shear strength requirement adapted to the patient's bone density and marked as , the shear strength of standard screws and is marked as , the patient's bone density is marked as , and the minimum bone density required for standard screws to fit and are marked as , and then substitute it into the following calculation formula: ; ; , , and are all preset non-zero influence coefficients, is the biomechanical fit index.

[0010] In a preferred embodiment, dividing the patient matching level into a high matching level and a low matching level based on the evaluation result means: The structural fit index and the biomechanical fit index are substituted into the pre-trained machine learning model. If the output result of the machine learning model is 1, the patient's matching level is classified as a high matching level. If the output result of the machine learning model is 0, the patient's matching level is classified as a low matching level.

[0011] In a preferred embodiment, the machine learning model is a convolutional neural network model.

[0012] In a preferred embodiment, the cluster analysis method uses the DBSCAN algorithm for clustering: The neighborhood radius and the minimum number of samples are set. The neighborhood radius controls the similarity threshold, and the minimum number of samples controls the clustering density requirement. All patients belonging to the same cluster as the current patient are found to obtain a group of patients with similar anatomical structures and biomechanical characteristics.

[0013] In a preferred embodiment, the logic for obtaining the personalized customized data set is: ; is the similarity distance between the current patient and the i-th patient, N represents the number of patients belonging to the same cluster as the current patient, is the nail rod parameter data set of the i-th patient, A personalized data set corresponding to the current patient.

[0014] In a preferred embodiment, the similarity distance is Euclidean distance or cosine similarity.

[0015] In a preferred embodiment, a design system for pre-set screw rods for spinal surgery based on big data includes: A data acquisition module collects the preoperative medical imaging data of the target patient and extracts the patient's anatomical feature parameter group; Database construction module, which builds a surgical fixation device database, including historical patient data and implant device parameters; Patient assessment module, which evaluates the degree of fit between the standard nail rod and the patient's anatomical structure and the degree of postoperative fixation stability; The classification module divides the patient's matching level into high matching level and low matching level based on the evaluation results, and triggers the personalized customization mechanism when the matching level is low; The personalized customization module, when the personalized customization mechanism is triggered, uses the cluster analysis method to cluster the patient data in the historical surgery database, selects patient groups with similar anatomical structures and biomechanical characteristics, and combines the individual parameters of the target patients to obtain a personalized customization data group.

[0016] Technical effects and advantages of the present invention: The present invention uses the structural fit index and biomechanical fit index to quantitatively evaluate the anatomical characteristics and biomechanical fit between the standard nail rod and the individual patient, ensuring that the compatibility of the fixation device can be determined before surgery. The matching level is automatically divided through a machine learning model to avoid doctors relying solely on experience to select the screw model, reducing repeated adjustments caused by improper matching during surgery. This makes the selection of screw specifications more accurate, improves intraoperative stability, and reduces surgical difficulty and time consumption.

[0017] The present invention screens the cases most similar to the target patients through density cluster analysis, and obtains personalized parameters such as screw length, diameter, implantation angle, etc. that are closer to individual needs based on big data calculation, thereby improving matching accuracy. The exponential weighted calculation method is adopted to ensure that the patient data with higher similarity contributes more to the customized plan, reduce the error caused by individual data fluctuations, and make the personalized plan more stable and reliable. The personalized parameters are optimized through the historical surgery database, making the personalized customization data more scientific, rather than simply relying on individual case analysis, to improve the rationality of personalized surgical planning.

[0018] The present invention adopts an automatic matching evaluation model based on a convolutional neural network, which can quickly determine whether the standard nail rod is applicable before surgery, avoiding the reliance of traditional surgery on the subjective decision of the doctor and shortening the preoperative planning time. The intelligent screening algorithm is used to automatically extract the individual that best matches the target patient from the massive historical surgical data, reducing the time for the doctor to manually screen the data and improving the efficiency of surgical decision-making. When the standard nail rod is not applicable, the personalized customization process can be automatically triggered to reduce the doctor's judgment bias and improve the accuracy of the custom screw design.

[0019] The present invention combines big data, deep learning and cluster analysis to make spinal surgery more precise and intelligent, improve patient matching, reduce reliance on experience, and increase the success rate of surgery. Promote the digitization and intelligence of spinal surgery, provide accurate matching data for new technologies such as robot-assisted spinal surgery and intraoperative navigation, and improve the application effect of future intelligent medical equipment. In traditional surgical procedures, multiple replacements of implants may be required due to mismatched screws, increasing surgical costs. The present invention reduces material waste and medical expenses through precise preoperative matching calculations. Optimize surgical plans through big data, reduce the need for secondary surgery due to loosening and breakage of implants after surgery, improve hospital resource utilization, reduce patient hospitalization time, and improve medical efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings; Figure 1 This is a schematic diagram of the design method of the pre-set screw rod for spinal surgery based on big data in the present invention.

[0021] Figure 2 This is a schematic diagram of the design system for pre-setting screw rods for spinal surgery based on big data in the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Reference Figure 1 - Figure 2 The following embodiments are obtained: Example 1: In spinal surgery, pedicle screws and connecting rod systems are key devices for spinal fixation and are widely used in spinal deformity correction, degenerative spinal lesions, spinal trauma repair and other surgeries. However, the currently commonly used standardized fixation devices in clinical practice have the following problems: Insufficient personalized matching: The specifications of traditional screws and connecting rods are relatively fixed. Doctors need to adjust the fixation plan according to the patient's anatomical characteristics during surgery. However, due to the large individual differences in the spine, standard screws may not match the patient's pedicle diameter, angle, bone density, etc., resulting in poor fixation effect.

[0024] The surgery depends on experience, and intraoperative adjustments are difficult: Currently, surgical plans mainly rely on the doctor's clinical experience. During the operation, it may be necessary to repeatedly adjust the screw diameter, length, and implantation angle, which increases the operation time and may affect the fixation stability.

[0025] The risk of postoperative complications is high: If the screw fixation is unstable, it may lead to screw loosening, fracture, implant failure, affecting postoperative recovery, and even requiring a second operation.

[0026] Lack of data-driven intelligent optimization: Existing spinal fixation device designs do not fully utilize big data analysis technology and are unable to optimize personalized solutions based on patient historical data, resulting in poor adaptability of implants.

[0027] Based on the above problems, the present invention proposes a spinal surgery pre-set screw rod design method based on big data to improve screw matching, optimize the personalized customization process, and enhance postoperative stability.

[0028] The present invention realizes personalized customization of screws by combining image analysis, index calculation, fuzzy reasoning and cluster analysis in a data-driven manner, which has the following important significance: Improve the matching degree between screws and patients: By calculating the structural matching index and biomechanical adaptation index, the matching degree between standard screws and patients can be accurately evaluated to avoid blind adjustments during surgery and reduce the risk of mismatch.

[0029] Reduce the difficulty of intraoperative adjustments and improve surgical accuracy: Through big data analysis, the optimal screw specifications and implantation angles can be predicted before surgery, reducing the time doctors spend on repeated adjustments during surgery and improving surgical efficiency and accuracy.

[0030] Reduce postoperative complications and improve fixation stability: Optimize personalized parameters through biomechanical simulation to ensure reasonable postoperative screw fixation torque, reduce risks such as loosening and breakage, and improve the long-term stability of the implant.

[0031] Utilize big data and artificial intelligence to optimize surgical plans: Use DBSCAN cluster analysis to screen similar patient groups, and use their historical surgical data to optimize personalized customized plans, making personalized plans more scientific and reliable.

[0032] Promote the intelligent development of spinal surgery: This invention combines artificial intelligence, image analysis and big data computing to build an intelligent surgical planning system, providing a foundation for future intelligent surgical navigation and robot-assisted surgery in spinal surgery.

[0033] The present invention proposes a design method for pre-set screw rods for spinal surgery based on big data, comprising the following steps: Collect the preoperative medical imaging data of the target patient, extract the patient's anatomical feature parameter group, and establish a surgical fixation device database, including historical patient data and implant device parameters; specifically: Medical imaging data acquisition: Use computed tomography (CT) or magnetic resonance imaging (MRI) to perform high-precision scanning of the target patient's spine to obtain complete preoperative imaging data. Use medical image processing software (such as Mimics, 3DSlicer) to convert CT or MRI data into a three-dimensional model to accurately reconstruct the spinal anatomical structure. Anatomical feature extraction: Pedicle diameter: affects the selection of screw diameter to avoid fractures caused by too large screws or loosening caused by too small screws. Pedicle length: used to optimize screw length and ensure fixation stability. Pedicle axial angle: determines the direction of screw implantation and affects biomechanical adaptability. Spinal curvature (Cobb angle): used for corrective surgery planning. Bone density (BMD): affects screw fixation force and ensures postoperative stability. Database construction: Historical patient data: stores anatomical parameters of patients who have completed surgery and postoperative effect feedback data. Implant device parameters: includes data such as screw length, diameter, implantation angle, and biomechanical properties of different specifications. It is conducive to realizing precise data analysis, ensuring that personalized design is based on real data rather than relying on the doctor's experience and judgment, and building a dynamic database. It can continuously update and optimize the personalized fixation device recommendation system, improve matching accuracy, provide data support for subsequent evaluation and optimization, avoid manual adjustments during surgery, and improve surgical efficiency and accuracy.

[0034] The degree of fit between the standard screw rod and the patient's anatomical structure and the degree of postoperative fixation stability are evaluated respectively, and then the patient's matching level is divided into high matching level and low matching level based on the evaluation results. The personalized customization mechanism is triggered when the matching level is low; specifically: Structural fit evaluation (structural matching index SMI): Calculate the matching degree between the standard screw rod and the patient's anatomical characteristics, including key parameters such as diameter, length, and implantation angle, and use nonlinear weighted calculation to ensure the reasonable distribution of the importance of different parameters. Biomechanical fit evaluation (biomechanical fit index BMI): Calculate the fixation strength, torque stability, shear force bearing capacity and bone density fit after screw implantation. Finite element analysis can also be used to verify the postoperative fixation effect, optimize the screw length and angle, ensure fixation stability, reduce the risk of postoperative loosening and fracture, and improve the life of the implant. Data-based evaluation replaces empirical judgment, reduces intraoperative adjustments, improves the success rate of surgery, supports personalized solution decision-making, and provides a scientific basis for subsequent matching level division. With the structural matching index and biomechanical fit index as input and the matching level as output, accurate classification is carried out to reduce the cost of personalized customization, avoid unnecessary 3D printing or processing, automate decision-making, and reduce the subjective judgment error of doctors.

[0035] When the personalized customization mechanism is triggered, the cluster analysis method is used to cluster the patient data in the historical surgery database, select patient groups with similar anatomical structures and biomechanical characteristics, and combine the individual parameters of the target patients to obtain a personalized customization data group. Specifically: DBSCAN cluster analysis is used: key features such as pedicle diameter, length, angle, Cobb angle, and bone density are selected for clustering. Highly similar patient groups are identified by density threshold and minimum sample number. Outliers are automatically removed to ensure the reliability of personalized customization data. Based on the statistical data of similar patient groups, personalized customization errors are reduced, and personalized screw parameters are optimized in combination with historical data to improve matching accuracy. Combining the individual characteristics of patients with the clustering results, the following parameters are finally determined: personalized screw diameter, length, angle and other parameters, 3D printing personalized implant data or CNC processing parameters, and it is conducive to preoperative simulation evaluation and improve postoperative stability. CNC (computer numerical control) is a technology that uses computer-controlled machine tools to perform high-precision automatic processing, which can be used to manufacture personalized spinal fixation devices, such as customized screws and implants.

[0036] The degree of fit between the standard nail rod and the patient's anatomical structure is evaluated to generate a structural fit index, and the degree of postoperative fixation stability between the standard nail rod and the patient is evaluated to generate a biomechanical fit index.

[0037] The structural fit index is used to evaluate the degree of match between the standard screw rod and the individual spinal anatomy of the patient, including key parameters such as screw diameter, screw length, and implantation angle, to ensure that the screw can be accurately embedded in the pedicle without causing the risk of fracture, loosening or penetration. Clinical effect: Optimize screw selection, reduce intraoperative adjustments, and improve surgical accuracy. Reduce the risk of misimplantation and prevent screw loosening or damage to nerves and blood vessels due to improper matching. Improve implant stability, ensure that the screw can provide sufficient mechanical support, and improve postoperative healing.

[0038] The biomechanical fit index is used to evaluate the stability of standard screws and rods after surgery in patients, including factors such as fixation force, torque distribution, shear strength, and biomechanical stability, to ensure that the implant will not loosen, break, or cause mechanical stress concentration after surgery. Clinical effect: Improve postoperative stability and reduce the need for secondary surgery due to mismatch. Optimize biomechanical stress distribution to ensure that the screws will not fatigue and break due to stress concentration. Adapt to the patient's bone density to ensure that the screws will not fail to fix due to osteoporosis after implantation.

[0039] The dual index evaluation system, the structural fit index focuses on preoperative anatomical matching, and the biomechanical fit index focuses on postoperative fixation stability. The combination of the two can comprehensively evaluate the adaptability of standard nail rods to patients.

[0040] The logic for obtaining the structural adaptation index is: The patient's pedicle diameter is obtained from the patient's anatomical feature parameter group and marked as , standardized screw diameters matched to the patient and marked , the straight-line distance from the pedicle entrance to the exit of the patient after the safety margin adjustment is marked as , standard screw length and marked as , obtain the patient's pedicle axial angle and mark it as , and the standard screw axial angles are marked as , and then substitute it into the following calculation formula: ; ; , and are all preset non-zero scale factors, is the structural fit index.

[0041] Patient pedicle diameter: measured by the patient's preoperative computed tomography data, indicating the maximum width of the cross section of the pedicle. This parameter determines the diameter of the screw. If the screw diameter is too large, it may cause fracture or penetration, and if it is too small, the fixation effect is poor. Standard screw diameter: select a standardized screw diameter that matches the target patient. Patient pedicle adaptation screw length: measure the straight-line distance from the patient's pedicle entrance to the exit through imaging data, and adjust the safety margin. This parameter affects the fixation depth of the screw. Too long may damage the nerve, and too short will result in insufficient fixation. Standard screw length: select the standard screw length for spinal fixation, which is usually preset based on clinical experience. The formula uses an exponential decay function, so that smaller deviations have less impact on the matching degree, while larger deviations significantly reduce the matching degree. Patient pedicle axial angle: the pedicle axial angle measured by computed tomography, that is, the angle between the screw implantation direction and the sagittal plane of the spine. This parameter affects the stability and biomechanical adaptability of the screw. Standard screw axial angle: The preset standard screw insertion angle uses the cosine function to calculate the matching degree. When the angle deviation is small, the matching degree is high, and when the deviation is too large, the matching degree drops sharply. The non-zero proportional coefficient reflects the degree of influence of different parameters on the matching degree, which is obtained based on clinical experience or machine learning optimization. SMI is close to 1 (such as above 0.85): the screw matching degree is high, and the standard specification nail rod can be used directly; SMI is between 0.6 and 0.85: the screw matching degree is general, and the doctor can fine-tune it during the operation or consider personalized adjustment; SMI is lower than 0.6: the matching degree is poor, and personalized customization is required to improve surgical stability.

[0042] The logic for obtaining the biomechanical fitness index is: The optimal fixation force of the screw in the patient's spine obtained by finite element analysis is obtained and marked as , the fixing capacity of standard specification screws and marked as , the optimal moment stability index of the patient's spine and is marked as , the torque stability of the standard screw and is marked as , the screw shear strength requirement adapted to the patient's bone density and marked as , the shear strength of standard screws and is marked as , the patient's bone density is marked as , and the minimum bone density required for standard screws to fit and are marked as , and then substitute it into the following calculation formula: ; ; , , and are all preset non-zero influence coefficients, is the biomechanical fit index.

[0043] The optimal fixation force of the patient's spine to the screw: calculated by finite element analysis in the prior art, indicating the optimal fixation force of the screw in the patient's spine to provide maximum stability. This value usually depends on factors such as the patient's bone density, pedicle size, and screw length. Standard screw fixation force: obtained from biomechanical experiments or databases, representing the fixation capacity of standard specification screws. The calculation method uses a normalized ratio, so that it has a greater impact on the matching degree when the fixation force is small, and has a smaller impact when the fixation force is close to the optimal value. The optimal moment stability index of the patient's spine: calculates the moment (bending moment) of the spine under postoperative load, which is used to measure postoperative stability. This value depends on the position of the implant, bone density, and spinal load mode. The torque stability of the standard screw: obtained from a database or mechanical experiment, representing the torque stability capacity of the standard specification screw. The calculation method uses a square ratio, so that when the torque difference is small, the matching degree is still high, and when the difference is large, the matching degree drops rapidly. The shear strength requirement of the screw adapted to the patient's bone density: the maximum shear force that the screw can withstand in the pedicle under the patient's individual bone density. Shear strength of standard screws: obtained from material properties and experimental data, representing the shear bearing capacity of standard screws in spinal fixation. The calculation method uses the shear force ratio, so that the matching degree is high when the shear strength requirement is small, and the matching degree drops sharply when the requirement far exceeds the standard screw bearing capacity. Patient bone density: converted from Hounsfield units (HU) measured in computed tomography (CT) images, representing the mineral density (BMD) of the patient's bones. Minimum bone density requirement for standard screw adaptation: obtained from databases or clinical trials, representing the minimum bone density value at which standard screws can be stably implanted. The calculation method uses an exponential decay function, so that when the bone density is close to the screw requirement, the matching degree is high; when the bone density is far below the standard value, the matching degree drops rapidly. The non-zero influence coefficient represents the influence weight of different factors on the biomechanical adaptation index, which can be optimized based on clinical experience or machine learning methods. BMI is close to 1 (such as above 0.85): biomechanical matching is high, and standard nail rods can be used directly. BMI is between 0.6 and 0.85: the matching degree is good, and appropriate fine-tuning can be made during surgery. BMI below 0.6: The match is poor and personalized customization is recommended to optimize postoperative stability.

[0044] Classifying the patient matching level into a high matching level and a low matching level based on the evaluation results means that the structural fit index and the biomechanical fit index are substituted into a pre-trained machine learning model. If the output result of the machine learning model is 1, the patient matching level is classified as a high matching level. If the output result of the machine learning model is 0, the patient matching level is classified as a low matching level. The machine learning model is a convolutional neural network model.

[0045] Based on the convolutional neural network model, the structural fit index and biomechanical fit index are used as input, and the matching level is judged through the pre-trained model. The specific process is as follows: The input layer performs data preprocessing: collects the patient's preoperative imaging data, calculates the structural fit index and biomechanical fit index, and uses the two indices as feature inputs to construct a data set for the matching grade prediction model.

[0046] Convolutional computing layer for feature extraction: The convolutional neural network extracts potential rules in the index data through multi-layer convolution operations, such as which numerical ranges of matching index are more likely to have a good match for the fixed device. This process is equivalent to letting the model learn which anatomical features and biomechanical conditions are more suitable for standard fixed device implantation.

[0047] The fully connected layer determines the matching level: After deep feature extraction, the model passes the data to the fully connected layer for final classification calculation. If the neural network output is "matching success", the matching level is classified as a high matching level; if the neural network output is "matching failure", the matching level is classified as a low matching level, triggering the personalized customization mechanism.

[0048] Training and optimization: The model is trained based on a large amount of historical patient data to learn matching rules and ensure accurate classification results. The loss function is used for optimization to ensure that the model can accurately distinguish which patients are suitable for standard nail rods and which patients need personalized customization.

[0049] Through machine learning methods, the errors that may be caused by traditional manual threshold division are avoided, and the accuracy and robustness of classification are improved. Matching decisions are made more data-based and intelligent, reducing the subjective errors of doctors in matching judgments. Matching judgments can be made based on the evaluation results before surgery to avoid temporary adjustments to the surgical plan due to mismatches during surgery, thereby improving the stability and success rate of the surgery. Only patients with a low matching index enter the personalized customization process to avoid unnecessary customized production, reduce surgical costs and time, improve the utilization of medical resources, and ensure that personalized customization is only used for patients who really need it. This method combines big data and artificial intelligence technology to achieve precision medicine, so that each patient can obtain the best fixation device solution and improve postoperative rehabilitation effects.

[0050] Cluster analysis method uses DBSCAN algorithm for clustering: The neighborhood radius and the minimum number of samples are set. The neighborhood radius controls the similarity threshold, and the minimum number of samples controls the clustering density requirement. All patients belonging to the same cluster as the current patient are found to obtain a group of patients with similar anatomical structures and biomechanical characteristics.

[0051] Using a density-based clustering analysis algorithm, by setting the neighborhood radius and the minimum number of samples, we can screen out a group of patients with similar anatomical structures and biomechanical characteristics to the target patient. The core steps are as follows: Determine the neighborhood radius: The neighborhood radius is a threshold used to measure the similarity of patient data and defines the maximum allowable difference between patients. Calculate the distance between each patient's anatomical features and the target patient, such as pedicle diameter, length, implant angle, spinal curvature, bone density, etc. Set a reasonable neighborhood radius to ensure that similar patients can be classified into the same category, and dissimilar patients will not be clustered incorrectly.

[0052] Set the minimum sample size: The minimum sample size is used to control the cluster density and ensure that each group contains at least enough patients to ensure the reliability of statistical data. If the number of patients in a cluster is less than the minimum sample size, the cluster may be judged as invalid, thereby avoiding overfitting or outliers affecting the analysis results.

[0053] Perform density clustering: With the target patient (the current patient) as the center, search for similar patients that meet the minimum sample number within the set neighborhood radius. Patients that meet the conditions are grouped into the same cluster to form a group of similar patients. If the data of some patients does not match any known cluster, they are marked as abnormal data and automatically removed to ensure the accuracy of personalized customized data.

[0054] Obtain similar patient groups: After cluster analysis, the group closest to the target patient in terms of anatomical structure and biomechanical characteristics is screened out. By screening similar patient groups, the impact of individual data errors on the customization plan can be reduced, ensuring that the parameters of personalized nail rods are more scientific and reasonable. Avoid using a single patient's data directly, which leads to deviations in personalized customization and improves matching stability. Traditional classification methods require doctors to manually set different categories, while density clustering algorithms can automatically discover the natural distribution patterns in the data without human intervention, thus improving efficiency. If the anatomical features of some patients are too different from those of the target patients, or the data is abnormal, the algorithm will automatically mark them as outliers to avoid affecting the accuracy of personalized customization plans. Reduce the interference of extreme cases on customization parameters to ensure that the final generated fixation device is more in line with the needs of most similar patients. Using the successful surgical data of similar patients as a reference can effectively reduce the risk of mismatching fixation devices and reduce problems such as postoperative screw loosening and fixation failure. Through a data-driven approach, optimize preoperative planning, improve implant stability, and improve patient postoperative recovery.

[0055] The logic for obtaining personalized customized data groups is: ; is the similarity distance between the current patient and the i-th patient, which can be calculated by Euclidean distance or cosine similarity, indicating the differences between the two in anatomical structure and biomechanical characteristics. N represents the number of all patients belonging to the same cluster as the current patient (screened by cluster analysis), is the nail rod parameter data set of the i-th patient, A personalized data set corresponding to the current patient.

[0056] The purpose of obtaining personalized customized data is to calculate the personalized nail rod parameter data set that best meets the needs of the target patient among similar patient groups. The core idea is to calculate the weighted average of similar patients, so that more similar patients contribute higher weights to ensure the accuracy of the personalized customization plan. Through cluster analysis, historical patients with similar anatomical structures and biomechanical characteristics to the target patients are found to avoid the errors that may be caused by single patient data and improve the accuracy of the customized plan. The nail rod parameters of the target patients are more in line with the requirements of biomechanical stability and improve the postoperative effect. Similarity weighted calculation is introduced, and exponential decay weights are used to make the weights of similar patients closer to the target patients higher, reducing the impact of cases far away from the target patients on the customized data, ensuring that the data is more in line with individual needs, avoiding the errors that may be caused by directly calculating the mean, and making the customized data more accurate. Moreover, since the weights of patients with low similarity are small, the data of abnormal patients will not significantly affect the customization results of the target patients, thereby enhancing the robustness of personalized customization.

[0057] Embodiment 2: A design system for pre-set screw rods for spinal surgery based on big data, comprising: A data acquisition module collects the preoperative medical imaging data of the target patient and extracts the patient's anatomical feature parameter group; Database construction module, which builds a surgical fixation device database, including historical patient data and implant device parameters; Patient assessment module, which evaluates the degree of fit between the standard nail rod and the patient's anatomical structure and the degree of postoperative fixation stability; The classification module divides the patient's matching level into high matching level and low matching level based on the evaluation results, and triggers the personalized customization mechanism when the matching level is low; The personalized customization module, when the personalized customization mechanism is triggered, uses the cluster analysis method to cluster the patient data in the historical surgery database, selects patient groups with similar anatomical structures and biomechanical characteristics, and combines the individual parameters of the target patients to obtain a personalized customization data group.

[0058] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0059] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0060] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example 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 performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0061] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0062] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A design method for pre-set screw rods for spinal surgery based on big data, characterized in that: The following steps are involved: Collect the preoperative medical imaging data of the target patient, extract the patient's anatomical characteristic parameter group, and establish a surgical fixation device database, including historical patient data and implant device parameters; The degree of fit between the standard nail rod and the patient's anatomical structure and the degree of postoperative fixation stability are evaluated respectively, and then the patient's matching level is divided into high matching level and low matching level based on the evaluation results. The personalized customization mechanism is triggered when the matching level is low; When the personalized customization mechanism is triggered, the cluster analysis method is used to cluster the patient data in the historical surgery database, select patient groups with similar anatomical structures and biomechanical characteristics, and combine the individual parameters of the target patients to obtain a personalized customization data group.

2. The design method for pre-set nail rods for spinal surgery based on big data according to claim 1 is characterized in that: The degree of fit between the standard nail rod and the patient's anatomical structure is evaluated to generate a structural fit index, and the degree of postoperative fixation stability between the standard nail rod and the patient is evaluated to generate a biomechanical fit index.

3. The design method for pre-set nail rods for spinal surgery based on big data according to claim 2 is characterized in that: The logic for obtaining the structural adaptation index is: The patient's pedicle diameter is obtained from the patient's anatomical feature parameter group and marked as , standardized screw diameters matched to the patient and marked , the straight-line distance from the pedicle entrance to the exit of the patient after the safety margin adjustment is marked as , standard screw length and marked as , obtain the patient's pedicle axial angle and mark it as , and the standard screw axial angles are marked as , and then substitute it into the following calculation formula: ； ; , and are all preset non-zero scale factors, is the structural fit index.

4. The design method for pre-set screw rods for spinal surgery based on big data according to claim 3 is characterized in that: The logic for obtaining the biomechanical fitness index is: The optimal fixation force of the screw in the patient's spine obtained by finite element analysis is obtained and marked as , the fixing capacity of standard specification screws and marked as , the optimal moment stability index of the patient's spine and is marked as , the torque stability of the standard screw and is marked as , the screw shear strength requirement adapted to the patient's bone density and marked as , the shear strength of standard screws and is marked as , the patient's bone density is marked as , and the minimum bone density required for standard screws to fit and are marked as , and then substitute it into the following calculation formula: ; ; , , and are all preset non-zero influence coefficients, is the biomechanical fit index.

5. The design method for pre-set screw rods for spinal surgery based on big data according to claim 4, characterized in that: The patient matching level is divided into high matching level and low matching level based on the evaluation results, which means: The structural fit index and the biomechanical fit index are substituted into the pre-trained machine learning model. If the output result of the machine learning model is 1, the patient's matching level is classified as a high matching level. If the output result of the machine learning model is 0, the patient's matching level is classified as a low matching level.

6. The design method for pre-set screw rods for spinal surgery based on big data according to claim 5, characterized in that: The machine learning model is a convolutional neural network model.

7. The design method for pre-set screw rods for spinal surgery based on big data according to claim 6, characterized in that: Cluster analysis method uses DBSCAN algorithm for clustering: The neighborhood radius and the minimum number of samples are set. The neighborhood radius controls the similarity threshold, and the minimum number of samples controls the clustering density requirement. All patients belonging to the same cluster as the current patient are found to obtain a group of patients with similar anatomical structures and biomechanical characteristics.

8. The design method for pre-set screw rods for spinal surgery based on big data according to claim 7, characterized in that: The logic for obtaining personalized customized data groups is: ; is the similarity distance between the current patient and the i-th patient, N represents the number of patients belonging to the same cluster as the current patient, is the nail rod parameter data set of the i-th patient, A personalized data set corresponding to the current patient.

9. The design method for pre-set screw rods for spinal surgery based on big data according to claim 8, characterized in that: The similarity distance is either Euclidean distance or cosine similarity.

10. A design system for pre-set nail rods for spinal surgery based on big data, used to implement the design method for pre-set nail rods for spinal surgery based on big data according to any one of claims 1 to 9, characterized in that: include: A data acquisition module collects the preoperative medical imaging data of the target patient and extracts the patient's anatomical feature parameter group; Database construction module, which builds a surgical fixation device database, including historical patient data and implant device parameters; Patient assessment module, which evaluates the degree of fit between the standard nail rod and the patient's anatomical structure and the degree of postoperative fixation stability; The classification module divides the patient's matching level into high matching level and low matching level based on the evaluation results, and triggers the personalized customization mechanism when the matching level is low; The personalized customization module, when the personalized customization mechanism is triggered, uses the cluster analysis method to cluster the patient data in the historical surgery database, selects patient groups with similar anatomical structures and biomechanical characteristics, and combines the individual parameters of the target patients to obtain a personalized customization data group.

Images