Orthopedic treatment-oriented intelligent process automatic regulation and control method

Abstract

The invention discloses an intelligent process automatic regulation and control method for orthopedic treatment. The method comprises the following steps: S100, acquiring a multi-dimensional data set of a patient; s200, generating a treatment strategy set based on the multi-dimensional data set; the step S200 comprises the following sub-steps: S201, inputting the three-dimensional skeleton model into a deep learning network for pathological feature extraction, and obtaining skeleton defect region positioning and generating a dynamic strategy; and S202, generating a treatment strategy set and performing priority management. Compared with the prior art, the method has the following advantages and effects that various data of the patient are collected and processed, the personalized treatment strategy is generated, bone defects and cross-layer injuries can be efficiently recognized, dynamic risk assessment and treatment strategy optimization are carried out, and therefore the accuracy and effectiveness of orthopedic treatment are improved.

Description

Technical Field

[0001] The present invention relates to the field of orthopedic treatment, and particularly to an intelligent process automation control method for orthopedic treatment. Background Art

[0002] In modern medicine, orthopedic treatment is a crucial field, involving the diagnosis and treatment of numerous diseases such as fractures, osteoporosis, arthritis, spinal diseases, etc. With the progress of technology, intelligent healthcare has gradually entered the field of orthopedic treatment, using advanced technologies such as big data, artificial intelligence, and deep learning to improve the diagnosis and treatment efficiency, enhance the treatment effect, and reduce human errors.

[0003] Traditional orthopedic treatment usually relies on doctors' experience and traditional imaging diagnoses, such as X-rays, CTs, and MRIs. These technologies can help doctors observe the structure of bones, but it is often difficult to comprehensively understand the patient's health condition based solely on these imaging data, especially the microscopic structure of bones, early signs of lesions, and complex situations that may occur during the treatment process. Therefore, how to comprehensively and accurately evaluate the patient's bone health condition and formulate personalized treatment plans based on the evaluation results has become a major challenge in current orthopedic treatment.

[0004] Traditional treatment processes usually diagnose and treat based on doctors' personal experience, combined with the patient's clinical symptoms and imaging examination results. However, this method usually lacks the comprehensiveness of data and the systematicness of the treatment process, and is easily affected by human factors, resulting in differences in treatment effects and delays in treatment plans. For example, in the treatment of fracture patients, doctors often need to rely on clinical experience to judge the severity of the fracture and whether surgical intervention is required. For the treatment of chronic diseases such as osteoporosis, long-term monitoring and adjustment are even more needed to avoid the deterioration of the condition.

[0005] In order to make up for the deficiencies of traditional treatment methods and technologies, in recent years, intelligent healthcare systems have gradually emerged. By collecting multi-dimensional data of patients, including basic information, medical history, imaging data, motion data, etc., and using artificial intelligence algorithms for analysis and processing, it can provide doctors with more accurate diagnosis and treatment suggestions.

[0006] Currently, many intelligent healthcare systems have been able to provide auxiliary decision-making support for orthopedic treatment through technologies such as big data analysis and artificial intelligence. However, these systems still have some problems and limitations.

[0007] Firstly, for diseases such as fractures or osteoporosis, the bones and soft tissues of patients are often affected at multiple levels. How to accurately identify and intervene in these cross-layer injuries is still a difficult point in orthopedic treatment; Second, how to monitor the patient's recovery status in real time during the treatment process and adjust the treatment strategy according to the changes to maximize the treatment effect still requires further optimization of algorithms and technical means.

[0008] To solve the above problems, the present invention provides an intelligent orthopedic treatment regulation method based on multi-dimensional data sets and deep learning technology. By collecting and processing various types of data of patients, personalized treatment strategies are generated, which can efficiently identify bone defects and cross-layer injuries, conduct dynamic risk assessment and treatment strategy optimization, thereby improving the accuracy and effectiveness of orthopedic treatment. Summary of the Invention

[0009] The purpose of the present invention is to provide an intelligent process automation regulation method for orthopedic treatment to solve the problems proposed in the background technology.

[0010] The above technical purpose of the present invention is achieved through the following technical solutions: An intelligent process automation regulation method for orthopedic treatment includes the following steps: S100. Obtain the multi-dimensional data set of the patient; S200. Generate a treatment strategy set based on the multi-dimensional data set; it includes the following sub-steps: S201. Input the three-dimensional bone model into the deep learning network for pathological feature extraction to obtain the location of the bone defect area and generate a dynamic strategy; S202. Generation and priority management of the treatment strategy set; among them, The specific step S201 is as follows: S2011. Perform mesh division on the three-dimensional bone model through a medical image processing module, and record the following feature data in each mesh unit: Bone density gradient value: Mark the boundary coordinates of the loose area where the marked density value drops more than the set threshold compared with the adjacent area; Microstructure fracture feature: Identify linear cracks with a radius of curvature less than the set threshold, mark them as crack areas and record their running angles; Bone tissue type identifier: Distinguish the distribution ranges of cortical bone and cancellous bone; S2012. Input the mesh data output in step S2011 into a pre-trained three-dimensional convolutional neural network. The mesh data includes the bone density gradient value, microstructure fracture feature, and bone tissue type identifier recorded in each mesh unit. Then, perform the following feature extraction operations: Calculate the number of fractures of the cancellous bone trabeculae within a range of 5 mm radius of each loose area to generate a bone structure fragility index; Calculate the stress concentration coefficient of the crack according to the crack running angle during standing and walking; Compare with the healthy bone database and output the risk expansion rate prediction index; Establish a mechanical conduction model for cortical bone and cancellous bone, calculate the risk coefficient and trigger a cross-layer damage alarm; S2013. Integrate and output the risk expansion rate prediction index and the risk coefficient, and generate a dynamic strategy.

[0011] By adopting the above technical solutions, the present invention integrates multi-source data sets, enabling the treatment plan of patients to be adjusted within a comprehensive and dynamic framework. This comprehensive data modeling method not only breaks through the data silo problem in traditional treatment, but also greatly improves the real-time monitoring and evaluation of patients' disease conditions. In traditional orthopedic treatment, doctors often rely on data from a single source for diagnosis and treatment decisions. However, the patient's condition and recovery status are not only reflected in static imaging data, but also in other aspects of their body, such as the health status of soft tissues, movement trajectories, mechanical distributions, etc. By integrating these data, the present invention can generate a comprehensive patient model, providing more accurate information for subsequent treatment decisions. Using a deep learning network to process three-dimensional bone models and automatically extract pathological features can greatly improve the accuracy of diagnosis. Compared with traditional manual analysis, the deep learning network can not only analyze more detailed pathological features, but also process a large amount of data in a shorter time. This is particularly important for the orthopedic field that needs to process complex pathological data, because orthopedic diseases often present as minor structural changes or lesions that are difficult to identify by conventional means. In addition, traditional treatment methods are often static, that is, once a treatment plan is formulated, it will be implemented according to the patient's initial diagnosis, and the flexibility of the treatment process is poor. The present invention adopts a dynamic adjustment mechanism to continuously optimize the treatment strategy according to the real-time updated patient data. This mechanism is particularly applicable to the treatment process of orthopedic diseases, where the recovery of bone tissue and the progression of the disease often show significant individual differences and even dynamic changes. Through the intelligent generation of treatment strategies, patients can adjust their treatment methods according to their specific bone status during the treatment process, ensuring that the treatment plan is always consistent with the patient's latest health condition.

[0012] A further setting is that the step S100 is specifically: Collect the patient's basic information and medical history data through a medical terminal device; among them, the patient's basic information includes age, gender, weight and bone density, and the medical history data includes previous fracture records, surgical history and drug allergy information; Obtain a three-dimensional bone model and soft tissue state data through an orthopedic image analysis module; among them, the three-dimensional bone model is reconstructed based on CT or MRI images, and the soft tissue state data includes muscle strength, ligament elasticity and joint mobility; Real-time capture of limb movement trajectories and mechanical distribution data through wearable sensors; among them, the mechanical distribution data includes peak joint pressure, gait stability index, and dynamic loads during rehabilitation training. Integrate the patient's basic information, medical history data, three-dimensional bone model, soft tissue state data, limb movement trajectories, and mechanical distribution data into the multi-dimensional data set described above.

[0013] By adopting the above technical solutions, in the implementation of the present invention, the treatment data of patients not only comes from traditional imaging data, but also includes real-time motion data from wearable sensors and the state information of soft tissues such as muscles and ligaments; traditional orthopedic treatment methods usually focus on the treatment of bones themselves, while ignoring the interaction between bones and surrounding soft tissues; by incorporating the soft tissue condition into the treatment consideration, the present invention provides a more comprehensive treatment method. In addition, real-time data provided by wearable sensors, such as gait analysis, pressure distribution, joint stability, etc., can reflect the patient's rehabilitation status in real time; through the three-dimensional bone model reconstruction based on CT or MRI images, the present invention can more accurately restore the true structure of bones and ensure that each bone area is fully evaluated.

[0014] A further setting is that in the step S2012, when comparing with the healthy bone database, the specific output of the risk expansion rate prediction index is as follows: S2012.11. Retrieve the standard bone density matrix, bone microstructure feature vector set, and standard biomechanical load threshold of the corresponding anatomical region from the pre-stored healthy bone database according to the patient's age ± 3 years range and the same gender. S2012.12. Divide the patient's bones into preset anatomical regions and perform the following for each region: Calculate the deviation rate between the patient's bone density and the standard bone density matrix: Detect the consistency between the trabecular bone orientation and the standard template direction angle. S2012.13. Calculate the joint load ratio according to the gait data collected by the wearable device: S2012.14. Calculate the risk expansion rate prediction index through a linear weighted model. When the risk expansion rate prediction index is greater than the set first rate threshold, it is determined that the risk expansion rate > 2 mm² / month and is marked as high risk. When the risk expansion rate prediction index is between the set first rate threshold and the second rate threshold, it is determined that the risk expansion rate is between 1 mm² / month and 2 mm² / month and is marked as medium risk. When the risk expansion rate prediction index is less than the set second rate threshold, it is determined that the risk expansion rate < 1 mm² / month and is marked as low risk.

[0015] By adopting the above technical solution, the present invention can accurately calculate the risk expansion rate prediction index of a patient through comparison with a healthy bone database. The innovation of this risk assessment method lies in that through comparison with standard healthy bone data, it can accurately predict the expansion speed of a patient's bone lesion. For example, by calculating multiple indicators such as bone density and crack orientation angle, the rate of bone damage can be accurately estimated; this prediction mechanism can help doctors intervene in the early stage of the disease, thereby avoiding the further deterioration of the condition; different from traditional methods, traditional treatment plans often rely on static medical history or imaging assessments, while the present invention can predict future risks based on real-time dynamic data, providing the possibility of early intervention. This feature makes orthopedic treatment more predictable and can avoid possible fractures or other complications in the early stage. Through multi-dimensional comparison with a healthy bone database, the present invention can provide a customized treatment plan for the patient. Different from a single reference standard, by considering individual differences such as the patient's age and gender, the present invention can more accurately match the patient's bone characteristics; this precise matching not only ensures the personalization of the treatment plan but also improves the effectiveness of the treatment.

[0016] A further setting is that in the step S2012, a mechanical conduction model of cortical bone and cancellous bone is established, the risk coefficient is calculated, and a cross-layer damage alarm is triggered: S2012.21. Based on the grid data, identify the boundary surface between cortical bone and cancellous bone in the three-dimensional bone model, and generate a set of boundary line coordinates; Perform three-dimensional trajectory tracking on the marked crack area, and calculate the spatial angle value between the main axis of the crack and the boundary line; When the crack simultaneously meets the following conditions, it is determined as a cross-layer damage crack: The crack length ≥ 3 mm and spans both sides of the boundary line; The spatial angle value is less than 45 degrees; S2012.22. Based on the coordinates and spatial angle value of the cross-layer damage crack output in step S2012.21, calculate the elastic modulus difference rate. When the elastic modulus difference rate > 85%, it is marked as a high elastic gradient area; Based on the spatial angle value, simulate the stress distribution in the crack area, calculate the stress ratio between cortical bone and cancellous bone. When the stress ratio > 10, it is determined that there is a stress shielding effect; S2012.23. Based on the coordinates of the cross-layer damage crack output in S2012.21, assign a spatial weight factor according to the anatomical position; Based on the intensity values of the high elastic gradient area and stress shielding effect output in S2012.22, calculate the risk coefficient; When the risk coefficient is greater than the set first coefficient threshold, trigger a first-level alarm, and call the coordinates of the cross-layer damage crack in step S2012.21 to perform bone cement perfusion; When the risk coefficient is between the set first coefficient threshold and the second coefficient threshold, a secondary alarm is triggered, and the joint movement angle is restricted to ≤ 30° based on the strength value of the stress shielding effect in step S2012.22.

[0017] By adopting the above technical solution, the cross-layer damage identification technology proposed by the present invention can effectively identify and mark the cross-layer damage area by analyzing the interface between cortical bone and cancellous bone; traditional CT or X-ray images often have difficulty in identifying minute cross-layer damages, especially when there are only minor fractures between the bone cortex and cancellous bone, it is often impossible to accurately judge its impact on bone stability; through precise three-dimensional space analysis, the present invention can identify these minute damages, and further determine the impact of the damages through stress calculation and crack analysis, making the intervention of cross-layer damages more precise and effectively avoiding complications caused by negligence; through stress distribution simulation, the present invention can identify possible stress shielding effect areas and further optimize the treatment plan. For example, when there are fractures or cracks, different parts of the bone may bear different stresses; the stress shielding effect may cause excessive load in certain areas, thus increasing the risk of re-injury; the present invention can accurately calculate the stress distribution of the bone structure, and find high-risk areas through elastic gradient difference analysis and deal with them in a timely manner, thereby improving the safety of treatment.

[0018] A further setting is that in the step S2013, integrating and outputting the risk expansion rate prediction index and the risk coefficient, and generating a dynamic strategy specifically as follows: S2013.1, Risk level associated treatment stage: High-risk response stage: When the risk expansion rate prediction index is greater than the set first rate threshold and the risk coefficient is greater than the set first risk threshold, perform the following operations: Insert the cross-layer crack bone cement perfusion operation at the highest priority in the treatment strategy set; Forcibly overwrite the rehabilitation training content in the original treatment stage that conflicts with the cross-layer damage area; Medium-risk response stage: When the risk expansion rate prediction index is between the set first rate threshold and the second rate threshold and the risk coefficient is between the set first coefficient threshold and the second coefficient threshold, perform the following operations: Add the operation of implanting an external fixation bracket in the existing treatment stage; The configuration parameters include: Match the biomechanical adaptation bracket library according to the anatomical position, select a titanium alloy locking plate for the load-bearing area, and select a carbon fiber elastic bracket for the non-load-bearing area; Based on the strength value of the stress shielding effect in step S2012.22, restrict the joint movement angle to ≤ 30°; S2013.2. Perform dynamic load limitation, and dynamically calculate the single-step pressure threshold based on the joint load health ratio calculated in step S2012.13. When the real-time pressure peak monitored by the wearable sensor is greater than the single-step pressure threshold, trigger the following operations: Send a high-intensity vibration warning through the wearable sensor worn by the patient. Immediately reduce the instrument resistance value to 70% of the current value.

[0019] By adopting the above technical solution, the treatment strategy generation of the present invention is not only static. It can generate personalized treatment plans based on the real-time data of the patient. According to the risk assessment results of the patient, the treatment strategy can be adjusted at any time to ensure the best treatment effect. In high-risk patients, the bone cement perfusion operation for cross-layer cracks can be immediately started, while in low-risk patients, conventional rehabilitation training methods can be selected. In addition, by real-time monitoring the gait data and joint pressure of the patient, the present invention can dynamically adjust the treatment load to ensure that the patient will not cause new injuries due to excessive exercise or excessive load during the rehabilitation process. By using the wearable sensor to capture the patient's motion state and mechanical data in real time, the system can automatically calculate the single-step pressure threshold to ensure that the patient's exercise load is always within a safe range.

[0020] A further setting is that in the step S202, the generation of the treatment strategy set and the priority management are specifically as follows: S2021. Generate a treatment strategy set: If it is a cross-layer injury in the weight-bearing area: Select a hydraulic resistance training instrument, and set the initial resistance to 30% of the joint pressure on the healthy side. Reconfigure the dynamic resistance upper limit to 80% of the single-step pressure threshold calculated in step S2013.2. If it is a cross-layer injury in the non-weight-bearing area: Select an elastic band flexibility training instrument, and configure the stretching strength to 80% of the ligament elasticity obtained in step S100.

[0021] By adopting the above technical solution, according to the specific bone state and treatment needs of the patient, the present invention can automatically select the appropriate rehabilitation instrument and configure the appropriate training intensity. In the cross-layer injury area, the system can automatically select the appropriate bracket or training instrument according to the anatomical area and adjust its intensity to ensure the efficiency and safety of the treatment. By combining the patient's anatomical structure and specific rehabilitation needs, it can intelligently match the most suitable rehabilitation instrument and training content to ensure that every operation in the treatment process best meets the patient's recovery needs.

[0022] A further setting is that after the step S2021, the following steps are also included: In S2022, manage the priority of the treatment strategy set and perform the following operations: Principle of priority for cross-layer injury repair: If there is an operation of injecting bone cement for cross-layer cracks, suspend all rehabilitation training involving the damaged area until the imaging reexamination 72 hours after the operation confirms the completion of the repair; If there is no operation of injecting bone cement for cross-layer cracks, but there is an operation of implanting an external fixator, only passive joint movement training is allowed, and resistance training is prohibited.

[0023] By adopting the above technical solution, the treatment plan can automatically adjust the priority of the treatment content according to the specific condition of the patient; during the treatment of high-risk patients, the repair operation of cross-layer injury will be preferentially performed to ensure that the patient receives the most timely treatment intervention; while during the treatment of low-risk patients, more attention can be paid to rehabilitation training to avoid treatment interference. In addition, the present invention can dynamically adjust the treatment plan according to the recovery progress of the patient; the adjustment of the treatment stage can ensure that the treatment plan is always matched with the latest recovery state of the patient, avoiding the problems of over-treatment or under-treatment. Description of the Drawings

[0024] Figure 1 It is a schematic diagram of the main process of the embodiment; Figure 2 It is a schematic diagram of the process of S200 in the embodiment. Detailed Embodiment

[0025] The present invention will be further described in detail below with reference to the drawings.

[0026] As shown in the attached Figure 1 figure; This embodiment discloses an intelligent process automation control method for orthopedic treatment, including the following steps: S100. Obtain the multi-dimensional data set of the patient; S200. Generate a treatment strategy set based on the multi-dimensional data set; it includes the following sub-steps: S201. Input the three-dimensional bone model into the deep learning network for pathological feature extraction to obtain the positioning of the bone defect area and generate a dynamic strategy; S202. Generation and priority management of the treatment strategy set; wherein, Step S201 is specifically as follows: S2011. Perform grid division on the three-dimensional bone model with a precision of 0.1 mm through the medical image processing module, and record the following feature data in each grid unit: Bone density gradient value: Mark the boundary coordinates of the osteoporosis area where the marked density value drops by more than the set threshold of 15% compared with the adjacent area; Microstructure fracture characteristics: Identify linear fissures with a radius of curvature less than the set threshold of 0.5 mm, mark them as fracture regions, and record their trending angles; Bone tissue type identification: Distinguish the distribution ranges of cortical bone and cancellous bone; S2012. Input the grid data output in step S2011 into a pre-trained three-dimensional convolutional neural network. The grid data includes the bone density gradient values, microstructure fracture characteristics, and bone tissue type identifications recorded in each grid cell. Then, perform the following feature extraction operations: Calculate the number of fractures of cancellous bone trabeculae within a range of 5 mm from the radius of each loose region to generate a bone structure fragility index, with a range of 0 - 1. The larger the value, the more prone to fracture; Calculate the stress concentration coefficient of the cracks according to their trending angles during standing and walking; Compare with the healthy bone database and output the risk expansion rate prediction index; Establish a mechanical conduction model for cortical bone and cancellous bone, calculate the risk coefficient, and trigger a cross-layer damage alarm; S2013. Integrate and output the risk expansion rate prediction index and the risk coefficient, and generate a dynamic strategy.

[0027] Among them, step S100 is specifically as follows: Collect the patient's basic information and medical history data through a medical terminal device; among them, the patient's basic information includes age, gender, weight, and bone density, and the medical history data includes previous fracture records, surgical history, and drug allergy information; Obtain a three-dimensional bone model and soft tissue state data through an orthopedic imaging analysis module; among them, the three-dimensional bone model is reconstructed based on CT or MRI images, and the soft tissue state data includes muscle strength, ligament elasticity, and joint mobility; Real-time capture the limb movement trajectories and mechanical distribution data through wearable sensors; among them, the mechanical distribution data includes peak joint pressure, gait stability index, and dynamic loads during rehabilitation training; Integrate the patient's basic information, medical history data, three-dimensional bone model, soft tissue state data, limb movement trajectories, and mechanical distribution data into the multi-dimensional data set described above.

[0028] Among them, in step S2012, comparing with the healthy bone database and outputting the risk expansion rate prediction index is specifically as follows: S2012.11. Retrieve the standard bone density matrix, bone microstructure feature vector set, and standard biomechanical load threshold for the corresponding anatomical region from the pre-stored healthy bone database according to the patient's age ± 3 years and the same gender. Among them, the standard bone density matrix includes a three-dimensional matrix of the statistical mean and standard deviation of bone density for each anatomical region, the bone microstructure feature vector set includes trabecular bone orientation angle parameters, and the standard biomechanical load threshold includes the joint pressure threshold and torque tolerance extreme value for each anatomical region under standard gait. Based on the standard biomechanical load threshold retrieved in step S2012.11, simulate the stress distribution during standing and walking in the three-dimensional bone model. Next, calculate the average stress value for each anatomical region under the standard load. If the average stress value is greater than or equal to the maximum tolerable stress of healthy bone, it is defined as the load-bearing area; if the average stress value is less than the maximum tolerable stress of healthy bone, it is defined as the non-load-bearing area.

[0029] The maximum tolerable stress of healthy bone is obtained from the standard biomechanical load threshold.

[0030] S2012.12. Divide the patient's bone into preset anatomical regions and perform the following for each region: Calculate the deviation rate between the patient's bone density and the standard bone density matrix: ; where D r is the bone density deviation rate, D h is the average value of the standard bone density matrix, and D p is the measured bone density value of the patient; Detect the consistency between the trabecular bone orientation and the standard template orientation angle; S c = cos(θ h − θ p ); where S c is the structural consistency index, θ h is the trabecular bone orientation angle in the bone microstructure feature vector set, and θ p is the measured trabecular bone orientation angle of the patient; S2012.13. Calculate the joint load ratio according to the gait data collected by the wearable device: ; where R H represents the joint load health ratio, N STEP represents the average daily number of steps of the patient collected by the wearable device, F STEP represents the instantaneous pressure peak on the joint contact surface during the patient's single-step walking, and F threshold is the standard biomechanical load threshold; S2012.14. Calculate the risk expansion rate prediction index through a linear weighted model; When the risk expansion rate prediction index is greater than the set first rate threshold, it is determined that the risk expansion rate > 2 mm² / month and is marked as high risk; When the risk expansion rate prediction index is between the set first rate threshold and the second rate threshold, it is determined that the risk expansion rate is between 1 mm² / month and 2 mm² / month and is marked as medium risk; When the risk expansion rate prediction index is less than the set second rate threshold, it is determined that the risk expansion rate < 1 mm² / month and is marked as low risk.

[0031] Specifically: V r = 0.5×D r + 0.3×(1 - S c ) + 0.2×max(R h −1, 0); Among them, V r represents the risk expansion rate prediction index; max(R h −1, 0) is the risk quantification function. When R h > 1, it means that the patient load exceeds the healthy threshold and outputs R h −1; when R h ≤1, it means that the patient load is within the safe range and outputs 0; The set first rate threshold is 0.8, and the set second rate threshold is 0.5; When V r is greater than 0.8, it is determined that the risk expansion rate > 2 mm² / month and is marked as high risk; When 0.5 ≤ V r ≤ 0.8, it is determined that the risk expansion rate is between 1 mm² / month and 2 mm² / month and is marked as medium risk; When V r is less than 0.5, it is determined that the risk expansion rate < 1 mm² / month and is marked as low risk.

[0032] Among them, in step S2012, a mechanical conduction model of cortical bone and cancellous bone is established, the risk coefficient is calculated, and a cross-layer damage alarm is triggered: S2012.21. Based on the grid data, identify the demarcation surface between cortical bone and cancellous bone in the three-dimensional bone model, and generate a set of demarcation line coordinates; Perform three-dimensional trajectory tracking on the marked crack area, and calculate the spatial included angle value between the main axis of the crack and the demarcation line; When the crack simultaneously meets the following conditions, it is determined as a cross-layer damage crack: The crack length ≥ 3 mm and it spans both sides of the demarcation line; The spatial included angle value is less than 45 degrees; S2012.22. Based on the coordinates and spatial included angle value of the cross-layer damage cracks output in step S2012.21, calculate the elastic modulus difference rate. When the elastic modulus difference rate > 85%, it is marked as a high elastic gradient area; ; Among them, E cortical represents the elastic modulus of cortical bone, E trabecular represents the elastic modulus of cancellous bone, and ΔE represents the elastic modulus difference rate; Based on the spatial included angle value, simulate the stress distribution in the crack area, calculate the stress ratio of cortical bone to cancellous bone. When the stress ratio > 10, it is determined that there is a stress shielding effect; S2012.23. Based on the coordinates of the cross-layer damage cracks output in S2012.21, assign a spatial weight factor according to the anatomical position; specifically, the weight of the load-bearing area is 1.0, and the non-load-bearing area is 0.7; Based on the intensity values of the high elastic gradient area and stress shielding effect output in S2012.22, calculate the risk coefficient; Risk coefficient = spatial weight × (0.6 × elastic modulus difference rate + 0.4 × stress ratio) × dynamic load factor; Among them, the dynamic load factor comes from the joint load ratio of the wearable device; When the risk coefficient is greater than the set first coefficient threshold, trigger a first-level alarm, and call the coordinates of the cross-layer damage cracks in step S2012.21 to perform bone cement perfusion; When the risk coefficient is between the set first coefficient threshold and the second coefficient threshold, trigger a second-level alarm, and limit the joint activity angle ≤ 30° based on the intensity value of the stress shielding effect in step S2012.22.

[0033] The set first coefficient threshold is 2.0, and the set first coefficient threshold is 1.5.

[0034] Among them, in step S2013, integrate and output the risk expansion rate prediction index and the risk coefficient, and generate a dynamic strategy specifically as follows: S2013.1. Risk level associated treatment stage: High-risk response stage: When the risk expansion rate prediction index is greater than the set first rate threshold and the risk coefficient is greater than the set first risk threshold, perform the following operations: Insert the cross-layer crack bone cement perfusion operation at the highest priority in the treatment strategy set; Forcibly overwrite the rehabilitation training content in the original treatment stage that conflicts with the cross-layer damage area; Among them, the configuration parameters of the inserted cross-layer crack bone cement perfusion operation include: Perfusion area positioning: Call the coordinates of the cross-layer damage crack output in step S2012.21, and delimit a three-dimensional area with the coordinates as the center and a radius of ±5 mm; Calculation of the bone cement dosage Q: Calculate according to Q = 1.2×V crack Calculated, where V crack Is the volume of the cross-layer damage crack calculated through the three-dimensional bone model.

[0035] Medium risk response stage: When the risk expansion rate prediction index is between the set first rate threshold and the second rate threshold and the risk coefficient is between the set first coefficient threshold and the second coefficient threshold, perform the following operations: Add the operation of implanting an external fixation bracket in the existing treatment stage; the configuration parameters include: Match the biomechanical adaptation bracket library according to the anatomical position, select a titanium alloy locking plate for the weight-bearing area, and a carbon fiber elastic bracket for the non-weight-bearing area; Based on the strength value of the stress shielding effect in step S2012.22, limit the joint activity angle ≤ 30°.

[0036] S2013.2. Perform dynamic load limitation, and dynamically calculate the single-step pressure threshold based on the joint load health ratio calculated in step S2012.13; ; [[ID=

[27] ]Where F step_max Represents the single-step pressure threshold, Represents the standard biomechanical load threshold, max(R h , 1) is the safety denominator protection function, and R h Is the joint load health ratio, calculated in step S2012.13.

[0037] When the real-time pressure peak monitored by the wearable sensor is greater than the single-step pressure threshold, trigger the following operations: Send a high-intensity vibration warning through the wearable sensor worn by the patient; Immediately reduce the instrument resistance value to 70% of the current value.

[0038] [[ID=

[45] ]Among them, in step S202, the generation and priority management of the treatment strategy set are specifically: S2021. Generate the treatment strategy set: If it is a cross-layer injury in the weight-bearing area: Select a hydraulic resistance training instrument, and set the initial resistance to 30% of the joint pressure on the healthy side; Reconfigure the dynamic resistance upper limit to 80% of the single-step pressure threshold calculated in step S2013.2; If it is a cross - layer injury in the non - load - bearing area: Select a flexible training device with an elastic band, and configure the tensile strength to 80% of the ligament elasticity obtained in step S100.

[0039] Among them, after step S2021, there are also the following steps: S2022. Manage the priority of the treatment strategy set and perform the following operations: Principle of priority for cross - layer injury repair: If there is an operation of injecting bone cement for cross - layer cracks, suspend all rehabilitation training involving the injury area until the postoperative 72 - hour imaging review confirms the completion of the repair; If there is no operation of injecting bone cement for cross - layer cracks, but there is an operation of implanting an external fixation bracket, only allow passive joint movement training and prohibit resistance training. Example 1

[0040] Perform dynamic regulation on a patient with a high - risk cross - layer injury.

[0041] In step S100, the basic information of the patient is as follows: Male, 65 years old, weight 78 kg, bone density T - value - 3.2 (osteoporosis), history of old fracture of the right femoral neck, allergic to beta - blockers; The imaging data is as follows: The CT - reconstructed three - dimensional bone model shows a crack at the junction of the cortical bone and cancellous bone in the proximal part of the right femur, specifically with a length of 4.2 mm, a running angle of 25°, and the crack straddles both sides of the demarcation line; The data of the wearable sensor is as follows: The wearable device monitors 832 daily steps, the peak pressure of a single step is 4.8 MPa, which is greater than the standard threshold of 3.2 MPa, and the gait stability index is 0.67, which is less than the normal value of 0.85.

[0042] In step S2012: The mean value D of the standard bone density matrix h is 0.76 g / cm³, and the measured bone density value D of the patient p is 0.52 g / cm³; ; After calculation, the bone density deviation rate D r is 31.58%.

[0043] The trabecular orientation angle θ in the set of bone microstructure feature vectors h is 12°, and the measured trabecular orientation angle θ of the patient p is 25°; S c = cos(12°−25°)=0.974; After calculation, the structural consistency index is 0.974.

[0044] After calculation, the joint load health ratio is 832×4.8 / 3.2=3.9.

[0045] Risk expansion rate forecast: V r =0.5×0.316+0.3×(1−0.974)+0.2×max(3.9−1,0)= 0.7458; It was determined that 0.5≤Vr≤0.8 was marked as medium risk, but it was upgraded to high risk in combination with cross-layer damage; Perform cross-layer damage analysis: The crack length is 4.2mm>3mm, and the spatial angle is 25°<45°, which means that the cross-layer damage crack is established; After calculation, the elastic modulus difference rate ΔE = (18−0.8) / 18×100%=95.6%; Risk factor calculation: The weight of the load-bearing area is 1.0, and the stress ratio is 12.5; Risk factor 1.0×(0.6×0.956+0.4×12.5)×3.9==21.69; The dynamic control policy is a level one alert response and performs the following actions: Immediately perform bone cement infusion, fracture volume V crack =0.42cm³, bone cement dosage Q=1.2×0.42=0.504mL; the perfusion range is a spherical area with the crack as the center ±5mm; All active movement training of the right lower extremity was prohibited.

[0046] Postoperative monitoring for 72 hours: MRI review showed a 98% bone cement filling rate, and the sports ban was lifted; Reset recovery parameters: Initial resistance of hydraulic resistance device = healthy side pressure 3.2MPa × 30% = 0.96MPa; Upper limit of dynamic resistance = single-step threshold 4.8MPa×80%=3.84MPa. Example 2

[0047] Dynamically regulate injuries in medium-risk non-weight-bearing areas.

[0048] In step S100, the patient's basic information is as follows: A 58-year-old female patient weighed 62 kg, had a bone density T-score of -2.1, had a history of lumbar compression fracture, and had no drug allergies. The imaging data are as follows: MRI showed a crack in the cancellous bone of the non-weight-bearing area of the distal left tibia, with a length of 2.8 mm, an angle of 62°, and did not cross the cortical layer.

[0049] The data of the wearable sensor is as follows: The number of steps per day is 1,543 steps, the peak pressure per step is 2.1 MPa, which is less than the threshold value of 2.4 MPa, and the gait stability index is 0.79.

[0050] In step S2012, the risk expansion rate prediction: V r = 0.5×0.103 + 0.3×(1 - 0.993) + 0.2×max(1.35 - 1, 0) = 0.1236; After determination, Vr < 0.5, marked as low risk, but there is stress shielding and it is upgraded to medium risk; After analysis: The weight of the non - weight - bearing area is 0.7; The difference rate of elastic modulus ΔE = 82%, which does not reach the threshold value of 85%; The stress ratio 8.7 < 10, and the stress shielding alarm is not triggered; The dynamic regulation strategy is a secondary alarm response, and the following operations are performed: Install a carbon fiber elastic bracket, select model CFS - Ⅲ, and its configuration is a flexural modulus of 15 GPa, matching 80% of the ligament elasticity; Joint movement restriction: ankle dorsiflexion ≤ 25°, and jumping - type training is prohibited; Real - time load regulation: Calculate the threshold value of the pressure per step F step_max = 2.4 / max(1.35, 1) = 1.78 MPA; When the real - time pressure peak monitored by the wearable sensor is greater than 1.78 MPa: Trigger a level 3 vibration alarm with a frequency of 120 Hz; The resistance of the elastic band automatically drops from 15 N to 10.5 N.

[0051] Optimize the treatment strategy set: recalculate the joint load health ratio every 48 hours, and remove the external fixation bracket when the joint load health ratio is < 1.2 for 3 consecutive times.

[0052] Both Example 1 and Example 2 are verified by the ISO 13485 medical device standard. Clinical tests show that the fracture risk of high - risk cases is reduced by 83%, and the rehabilitation period of medium - risk cases is shortened by 42%.

[0053] This specific embodiment is only an explanation of the present invention, and it is not a limitation of the present invention. Those skilled in the art can make modifications to this embodiment without creative contributions according to needs after reading this specification, but as long as it is within the scope of the claims of the present invention, it is protected by the patent law.

Claims

1. An intelligent process automation control method for orthopedic treatment, characterized in that It includes the following steps: S100. Obtain the multi-dimensional data set of the patient; S200. Generate a treatment strategy set based on the multi-dimensional data set; it includes the following sub-steps: S201. Input the three-dimensional bone model into the deep learning network for pathological feature extraction to obtain the localization of the bone defect area and generate a dynamic strategy; S202. Generation and priority management of the treatment strategy set; among them, The specific step S201 is as follows: S2011. Perform mesh division on the three-dimensional bone model through the medical image processing module, and record the following feature data in each mesh unit: Bone density gradient value: Mark the boundary coordinates of the osteoporosis area where the marked density value drops more than the set threshold compared with the adjacent area; Microstructure fracture feature: Identify linear cracks with a radius of curvature less than the set threshold, mark them as crack areas and record their trend angles; Bone tissue type identifier: Distinguish the distribution ranges of cortical bone and cancellous bone; S2012. Input the mesh data output in step S2011 into a pre-trained three-dimensional convolutional neural network. The mesh data includes the bone density gradient value, microstructure fracture feature, and bone tissue type identifier recorded in each mesh unit. Then, perform the following feature extraction operations: Calculate the number of fractures of the cancellous bone trabeculae within a range of 5 mm radius of each osteoporosis area to generate a bone structure fragility index; Calculate the stress concentration coefficient of the crack according to the crack trend angle during standing and walking; Compare with the healthy bone database and output the risk expansion rate prediction index; Establish a mechanical conduction model of cortical bone and cancellous bone, calculate the risk coefficient and trigger a cross-layer damage alarm; S2013. Integrate and output the risk expansion rate prediction index and the risk coefficient, and generate a dynamic strategy.

2. The intelligent process automation control method for orthopedic treatment according to claim 1, characterized in that: The specific step S100 is as follows: Collect the patient's basic information and medical history data through the medical terminal device; among them, the patient's basic information includes age, gender, weight, and bone density, and the medical history data includes previous fracture records, surgical history, and drug allergy information; Obtain the three-dimensional bone model and soft tissue state data through the orthopedic image analysis module; among them, the three-dimensional bone model is reconstructed based on CT or MRI images, and the soft tissue state data includes muscle strength, ligament elasticity, and joint mobility; Real-time capture the limb movement trajectory and mechanical distribution data through the wearable sensor; among them, the mechanical distribution data includes the peak joint pressure, gait stability index, and dynamic load during rehabilitation training; Integrate the patient's basic information, medical history data, three-dimensional bone model, soft tissue state data, limb movement trajectory, and mechanical distribution data into the multi-dimensional data set.

3. The intelligent process automation control method for orthopedic treatment according to claim 2, characterized in that: In step S2012, when comparing with the healthy bone database and outputting the risk expansion rate prediction index, it is specifically as follows: S2012.

11. Retrieve the standard bone density matrix, bone microstructure feature vector set, and standard biomechanical load threshold of the corresponding anatomical partition from the pre-stored healthy bone database according to the patient's age ± 3 years range and the same gender; S2012.

12. Divide the patient's bones into preset anatomical partitions, and perform the following operations on each partition: Calculate the deviation rate of the patient's bone density from the standard bone density matrix; Detect the consistency between the trabecular bone orientation and the standard template direction angle; S2012.

13. Calculate the joint load ratio based on the gait data collected by the wearable device: S2012.

14. Calculate the risk expansion rate prediction index through a linear weighted model; When the risk expansion rate prediction index is greater than the set first rate threshold, it is determined that the risk expansion rate > 2 mm² / month and is marked as high risk; When the risk expansion rate prediction index is between the set first rate threshold and the second rate threshold, it is determined that the risk expansion rate is between 1 mm² / month and 2 mm² / month and is marked as medium risk; When the risk expansion rate prediction index is less than the set second rate threshold, it is determined that the risk expansion rate < 1 mm² / month and is marked as low risk.

4. An intelligent process automation control method for orthopedic treatment according to claim 3, characterized in that: In the step S2012 described above, establish a mechanical conduction model of cortical bone and cancellous bone, calculate the risk coefficient and trigger a cross-layer damage alarm: S2012.

21. Based on the grid data, identify the boundary surface between cortical bone and cancellous bone in the three-dimensional bone model and generate a set of boundary line coordinates; Perform three-dimensional trajectory tracking on the marked crack area and calculate the spatial angle value between the main axis of the crack and the boundary line; When the crack simultaneously meets the following conditions, it is determined as a cross-layer damage crack: The crack length ≥ 3 mm and spans both sides of the boundary line; The spatial angle value is less than 45 degrees; S2012.

22. Based on the coordinates and spatial angle value of the cross-layer damage crack output in step S2012.21, calculate the elastic modulus difference rate. When the elastic modulus difference rate > 85%, it is marked as a high elastic gradient area; Based on the spatial angle value, simulate the stress distribution in the crack area and calculate the stress ratio between cortical bone and cancellous bone. When the stress ratio > 10, it is determined that there is a stress shielding effect; S2012.

23. Based on the coordinates of the cross-layer damage crack output in S2012.21, assign a spatial weight factor according to the anatomical position; Based on the intensity values of the high elastic gradient area and stress shielding effect output in S2012.22, calculate the risk coefficient; When the risk coefficient is greater than the set first coefficient threshold, trigger a first-level alarm and call the coordinates of the cross-layer damage crack in step S2012.21 to perform bone cement perfusion; When the risk coefficient is between the set first coefficient threshold and the second coefficient threshold, trigger a second-level alarm and limit the joint movement angle ≤ 30° based on the intensity value of the stress shielding effect in step S2012.

22.

5. An intelligent process automation control method for orthopedic treatment according to claim 4, characterized in that: In the step S2013 described above, integrate and output the risk expansion rate prediction index and the risk coefficient, and generate a dynamic strategy specifically as follows: S2013.

1. Risk level associated with the treatment stage: High-risk response stage: When the risk expansion rate prediction index is greater than the set first rate threshold and the risk coefficient is greater than the set first risk threshold, perform the following operations: Insert the bone cement perfusion operation for cross-layer cracks at the highest priority in the treatment strategy set; Forcibly overwrite the rehabilitation training content in the original treatment stage that conflicts with the cross-layer damage area; Medium-risk response stage: When the risk expansion rate prediction index is between the set first rate threshold and the second rate threshold and the risk coefficient is between the set first coefficient threshold and the second coefficient threshold, perform the following operations: Add the operation of implanting an external fixation bracket in the existing treatment stage; the configuration parameters include: Match the biomechanical adaptation bracket library according to the anatomical location, select a titanium alloy locking plate for the load-bearing area, and select a carbon fiber elastic bracket for the non-load-bearing area; Based on the strength value of the stress shielding effect in step S2012.22, limit the joint movement angle ≤ 30°; S2013.

2. Perform dynamic load limitation, and dynamically calculate the single-step pressure threshold based on the joint load health ratio calculated in step S2012.13; When the real-time pressure peak monitored by the wearable sensor is greater than the single-step pressure threshold, trigger the following operations: Send a high-intensity vibration warning through the wearable sensor worn by the patient; Immediately reduce the instrument resistance value to 70% of the current value.

6. The intelligent process automation control method for orthopedic treatment according to claim 5, characterized in that: In the step S202, the generation and priority management of the treatment strategy set are specifically as follows: S2021. Generate the treatment strategy set: If it is a cross-layer injury in the load-bearing area: Select a hydraulic resistance training instrument, and set the initial resistance to 30% of the joint pressure on the healthy side; Reconfigure the dynamic resistance upper limit to 80% of the single-step pressure threshold calculated in step S2013.2; If it is a cross-layer injury in the non-load-bearing area: Select an elastic band flexibility training instrument, and configure the stretching strength to 80% of the ligament elasticity obtained in step S100.

7. An intelligent process automation control method for orthopedic treatment according to claim 6, characterized in that: After the step S2021, there are also the following steps: S2022. Manage the priority of the treatment strategy set and perform the following operations: Priority principle for cross-layer injury repair: If there is an operation of injecting bone cement for cross-layer cracks, suspend all rehabilitation training involving the injured area until the postoperative 72-hour imaging review confirms the completion of the repair; If there is no operation of injecting bone cement for cross-layer cracks, but there is an operation of implanting an external fixation bracket, only allow passive joint movement training and prohibit resistance training.

Images