Orthopedic treatment decision support system based on data driving

By analyzing fracture density and local stress distribution in bone imaging data, and combining this with changes in joint spacing, a lesion priority path is established. This addresses the lack of holistic understanding in existing orthopedic treatment systems and enables refined orthopedic treatment decision support.

CN121122672APending Publication Date: 2025-12-12THE 971ST HOSPITAL OF THE CHINESE PEOPLES LIBERATION ARMY NAVY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202511301011.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-09-12
Publication Date
2025-12-12

AI Technical Summary

Technical Problem

Existing orthopedic treatment decision support systems rely on static rule matching and single-condition judgment, which cannot simultaneously analyze changes in bone microstructure and mechanical state. This leads to treatment order relying on experience, and when multiple lesions are involved, there is a lack of priority ranking of multidimensional progression factors, making it impossible to form a holistic understanding.

Method used

By analyzing fracture density and local stress distribution based on three-dimensional bone imaging data, combined with the trend of joint spacing changes, the differences between lesions are calculated, a lesion priority path is established, and multi-source data is integrated for standardized processing to form a holistic treatment recommendation.

Benefits of technology

It enables a refined classification of skeletal risks, improves the accuracy of case matching, establishes a treatment sequence based on spatial location, size, and rate of progression, and forms treatment recommendations that take into account both the extent of lesion impact and the timing of intervention.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121122672A_ABST
    Figure CN121122672A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of orthopedic treatment, in particular to an orthopedic treatment decision support system based on data driving, which comprises a risk division module, a case matching module, a treatment decision generation module, a multi-source data fusion module and an intervention path optimization module. According to the method, a joint judgment means of fracture density and local stress distribution is introduced into bone image analysis, the microscopic damage information and the stress state of a bone structure are comprehensively measured, fine division of different risk levels is achieved, and a joint multi-measurement-point interval change trend is combined with a vectorization comparison mode; in case retrieval, the double matching accuracy of space and trend is improved, an interference relation sorting mechanism based on space position, size and progress rate is established for a multi-focus situation, so that the treatment sequence has integrity and relevance, a unified structure, focus and intervention information link is formed, and the accuracy of medical treatment is improved. And a treatment suggestion considering the focus influence range, the intervention opportunity and resource allocation is formed.
Need to check novelty before this filing date? Find Prior Art

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of orthopedic treatment, in particular to an orthopedic treatment decision support system based on data driving. BACKGROUND

[0002] The technical field of orthopedic treatment involves related technical content of diagnosing and analyzing diseases and injuries of the human skeletal system and formulating corresponding treatment plans. This technical field covers the comprehensive application of image diagnosis technology, skeletal disease analysis technology, and clinical data and medical record management technology in the diagnosis and treatment process of orthopedic diseases. Among them, the traditional orthopedic treatment decision support system refers to a system that uses past clinical experience, image diagnosis results, and patient medical record data to provide treatment plan suggestions for doctors through fixed condition matching and rule reasoning. This system usually relies on structured case information and image feature data to identify the category of the patient's current orthopedic condition and provide corresponding treatment plan recommendations through rule set comparison and condition determination.

[0003] The existing technology relies on static rule matching and single condition determination in orthopedic diagnosis and treatment. In the processing process, it can only identify the category of past cases and image features, lacks simultaneous analysis of changes in the microscopic structure of the skeleton and the mechanical state, cannot capture the distance change trend between multiple measurement points in joint lesion matching, resulting in a mismatch between the search results and the dynamic condition. When facing multiple lesion interference, it lacks priority ranking based on multiple dimensional progress factors, making the formulation of treatment order too dependent on experience, and the correlation between multiple data sources is insufficient, which cannot form a comprehensive understanding of lesion characteristics and intervention information. SUMMARY

[0004] To solve the technical problems existing in the prior art, the embodiments of the present application provide an orthopedic treatment decision support system based on data driving. The technical solution is as follows: On the one hand, an orthopedic treatment decision support system based on data driving is provided, comprising: The risk division module analyzes the trabecular bone fracture number and distance based on three-dimensional bone image data, calculates the fracture density, and combines the local stress distribution to call the principal stress and average stress, analyzes the ratio of the two and compares the fracture density and stress concentration degree to obtain the risk stratification mark; The case matching module filters a plurality of distance measurement points of the joint image corresponding to the risk area indicated in the risk stratification mark, calculates the distance difference of adjacent image frames, determines the distance change trend of each measurement point, calls the corresponding measurement point in the case data, performs difference calculation and sorting, selects the case that meets the condition as a matching candidate, and obtains the case matching result; The treatment decision generation module determines spatial position, size and progression rate of the lesions, calculates difference between the lesions, analyzes interference relationship and sorts the interference relationship according to influence degree, determines sequence of the lesions to be processed in priority, and obtains a lesion priority path based on the case matching result. The multi-source data fusion module collects orthopedic images and clinical records of each device or source based on the lesion priority path, performs standardization processing, and analyzes correlation between trabecular bone structure, joint spacing and lesion position parameters to obtain a data fusion log.

[0005] In another aspect, the risk stratification identification includes a risk category number, a risk level description and a structural damage feature, the case matching result includes a candidate case number, a matching degree classification and a matching priority order, the lesion priority path includes a lesion processing sequence number, a lesion level grouping and an interference influence category, and the data fusion log includes a fusion data source category, a fusion parameter type and a parameter correlation mode.

[0006] In another aspect, the risk stratification module includes: The fracture density calculation sub-module analyzes the number of fractures and the fracture spacing of each trabecular bone unit based on three-dimensional bone image data, compares the proportional relationship between the number of fractures and the spacing within the unit, calculates the weighted combination of the proportion and the total number of units, determines the difference pattern of the proportion distribution on the structure, and obtains a bone structure fracture distribution rate. The stress ratio determination sub-module calculates the maximum principal stress ratio and the average stress ratio of the trabecular bone unit based on the bone structure fracture distribution rate, analyzes the distribution law of the ratio in each structural region, compares the difference degree of the distribution and the stress concentration benchmark, and obtains a bone unit stress concentration degree. The risk stratification determination sub-module determines the combination feature of the same trabecular bone unit based on the bone unit stress concentration degree, compares the difference level of the combination feature and the risk stratification benchmark, calculates the structure risk state corresponding to each level, and obtains a risk stratification identification.

[0007] In another aspect, the case matching module includes: The joint spacing screening sub-module locates and analyzes a plurality of joint images corresponding to the structural damage feature in space based on the structural damage feature contained in the risk stratification identification, screens a measurement point in the images that has joint spacing measurement significance, calculates the spacing difference of the measurement point in adjacent image frames, determines the change pattern of the difference between the image frames, and obtains a joint spacing fluctuation amplitude. The change trend calculation sub-module calculates the continuous change amplitude of the adjacent frame spacing difference in the time dimension according to the order of the measurement points in the image sequence, and compares the distribution characteristics of the change amplitudes between the measurement points to obtain a joint spacing trend factor based on the joint spacing fluctuation amplitude. The case priority matching submodule performs difference calculation on corresponding measurement points in case data based on the joint space trend factor, compares the closeness of the difference to the matching benchmark, sorts the closeness, screens cases meeting the condition in the sorting result, and obtains a case matching result.

[0008] In another aspect, the treatment decision generation module comprises: The lesion feature determination submodule analyzes the spatial position coordinates, contour size, and progression rate of each lesion based on the case matching result, calculates the proportion of position difference and size change, judges the distribution of the proportion among the lesions, and obtains a lesion feature difference degree; The interference relationship analysis submodule compares the overlap degree of lesions in spatial position and the consistency of size change direction based on the lesion feature difference degree, calculates the mutual influence proportion between lesions, judges the interaction mode of the influence proportion in the whole lesion set, and obtains lesion interaction data; The priority order determination submodule calls the lesion interaction data, sorts according to the interaction mode of each lesion, compares the matching degree of the sorting result and the lesion difference distribution, judges the sorting order meeting the matching condition, and obtains a lesion priority path.

[0009] In another aspect, the matching degree of the sorting result and the lesion difference distribution is compared using the formula: ; Calculate the matching degree parameter , judge the sorting order meeting the matching condition, and obtain the lesion priority path, wherein, represents the sequence number difference of the th lesion in the sorting result, represents the sequence number difference scale coefficient of the th lesion, represents the difference amount of the th lesion in the difference distribution, represents the difference weight coefficient of the th lesion, represents the difference increment of adjacent lesions in the difference distribution of the th lesion, represents the total number of lesions.

[0010] In another aspect, the multi-source data fusion module comprises: The image and record acquisition submodule acquires orthopedic image data and clinical record data from each image device based on the lesion priority path, analyzes the distribution form of trabecular bone structure and the position parameters of joint space measurement points in the image, judges the difference of data format and unit standard, and performs uniformization processing to obtain an orthopedic image uniform data set; The parameter correlation analysis submodule calculates a proportional relationship between the trabecular bone structure density distribution and the joint spacing change amplitude based on the orthopedic image unified dataset, analyzes a corresponding mode of the proportional relationship and the lesion spatial position, compares the change trends of each parameter under the same spatial coordinates, and obtains a structure parameter correlation ratio; The fusion log generation submodule judges the matching degree of the difference source data in space and structure based on the structure parameter correlation ratio, filters parameter combinations that meet the matching degree requirements, records the corresponding image numbers and measurement point identifiers in the combination order, and obtains a data fusion log.

[0011] On the other hand, the proportional relationship between the trabecular bone structure density distribution and the joint spacing change amplitude is calculated using the formula: ; Obtain structure proportional difference parameters , analyze the corresponding mode of the proportional relationship and the lesion spatial position, compare the change trends of each parameter under the same spatial coordinates, and obtain a structure parameter correlation ratio, wherein, represents the average of the relative change rate of the trabecular bone structure density, represents the average of the relative change rate of the joint spacing change amplitude.

[0012] On the other hand, the system further comprises: The intervention path optimization module judges the lesion clinical treatment record and the intervention effect based on the data fusion log, identifies the lesion interference relationship and the treatment effect parameter, analyzes the improvement, optimizes the intervention order, and obtains a treatment intervention suggestion configuration; The treatment intervention suggestion configuration includes an intervention method category, an execution order, and an adaptation condition category.

[0013] On the other hand, the intervention path optimization module comprises: The treatment effect determination submodule analyzes the lesion change data corresponding to each treatment in the clinical treatment record of the lesion based on the data fusion log, calculates the shape difference proportion and the functional recovery amplitude proportion of the lesion at each treatment stage, judges the trend mode of the proportion change, and obtains a treatment progress improvement ratio; The interference relationship quantification submodule compares the change trends and the spatial position overlap of each lesion on the time axis based on the treatment progress improvement ratio, calculates the proportion distribution of mutual influence, judges the concentration mode of the influence distribution, and obtains a lesion interference coefficient; The path priority configuration submodule prioritizes the lesions according to the interference concentration mode based on the lesion interference coefficient, compares the matching degree of the sorting result and the treatment improvement trend, filters the lesion order that meets the conditions, and records the corresponding intervention method category, and obtains a treatment intervention suggestion configuration.

[0014] The technical scheme provided by the embodiment of the present application has at least the following beneficial effects: By introducing the joint determination means of fracture density and local stress distribution in bone image analysis, the micro-damage information and stress state of the bone structure are comprehensively measured to realize the fine division of different risk levels. In case retrieval, the vectorization comparison method of joint multi-point spacing trend is used to improve the spatial and trend matching accuracy. For multiple lesion cases, a disturbance relationship ranking mechanism based on spatial position, size and progression rate is established to make the treatment sequence have integrity and correlation. In the fusion analysis of multi-source images and clinical data, a unified structure, lesion and intervention information link is formed to form a treatment suggestion considering the lesion influence range, intervention timing and resource allocation. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0016] Figure 1 The system schematic diagram of the present application is shown in the figure. Figure 2 The system framework schematic diagram of the present application is shown in the figure. Figure 3 The flowchart of the risk division module of the present application is shown in the figure. Figure 4 The flowchart of the case matching module of the present application is shown in the figure. Figure 5 The flowchart of the treatment decision generation module of the present application is shown in the figure. Figure 6 The flowchart of the multi-source data fusion module of the present application is shown in the figure. Figure 7 The flowchart of the intervention path optimization module of the present application is shown in the figure. DETAILED DESCRIPTION

[0017] The technical solutions in the present application will be described below with reference to the drawings.

[0018] In the embodiments of the present application, the words such as "example", "for example" are used to represent as an example, illustration or description. Any embodiment or design scheme described as "example" in the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. Rather, the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present application, the meaning expressed by "and / or" can be both, or can be one of the two.

[0019] In the embodiments of the present application, the terms "image" and "picture" can be used interchangeably, and it should be noted that the meanings expressed are consistent when the distinction is not emphasized.

[0020] In the embodiments of the present application, sometimes the subscript such as W1 can be written in the form of non-subscript such as W1, and the meanings expressed are consistent when the distinction is not emphasized.

[0021] To make the technical problems, technical solutions and advantages to be solved by the present application more clear, the following will be described in detail with reference to the drawings and specific embodiments.

[0022] The embodiments of the present application provide an orthopedic treatment decision support system based on data driving, as shown in the figure, the system comprises: Figure 1 The risk division module analyzes the number of fractures and the fracture interval of each trabecular bone unit based on three-dimensional bone image data, calculates the fracture density and combines local stress simulation, calls the maximum principal stress and average stress, analyzes the ratio between the stresses, compares the fracture density and stress concentration degree, makes a combined judgment, and obtains a risk stratification mark; The case matching module filters a plurality of joint image interval measurement points based on the risk area indicated in the risk stratification mark, calculates the interval difference of adjacent image frames, determines the interval change trend of each joint measurement point, calls the corresponding measurement points in the case data, performs difference calculation and sorting, and selects the cases meeting the conditions as matching candidates to obtain a case matching result; The treatment decision generation module judges the spatial position, size and progression rate of the lesions based on the case matching result, calculates the differences between the lesions, analyzes the interference relationship between the lesions, sorts the lesions according to the interference relationship, determines the lesion order of priority, and obtains a lesion priority path; The multi-source data fusion module collects orthopedic image data and clinical record data from various devices or sources based on the lesion priority path, performs data standardization processing, analyzes the correlation between the trabecular bone structure, joint interval and lesion position parameters, and obtains a data fusion log; The intervention path optimization module judges the clinical treatment records and intervention effects of the lesions based on the data fusion log, identifies the interference relationship between the lesions and the treatment effect parameters, analyzes the improvement of the treatment effect, optimizes the intervention path, and obtains a treatment intervention suggestion configuration.

[0023] ​The risk stratification identification includes a risk category number, a risk level description, and a structural damage feature. The case matching result includes a candidate case number, a matching degree classification, and a matching priority order. The lesion priority path includes a lesion treatment order number, a lesion level grouping, and an interference influence category. The data fusion log includes a fusion data source category, a fusion parameter type, and a parameter association mode. The treatment intervention suggestion configuration includes an intervention method category, an execution order, and an adaptation condition category.

[0024] In the risk division module, the trabecular unit refers to a small bone structure unit within the skeleton, used to describe the microscopic structure of the skeleton. The fracture density refers to the number or frequency of fractures within the trabecular unit, which is an indicator of bone health. The combined local stress simulation is to simulate the stress changes in different regions by calculating the stress distribution in the bone structure to judge the bearing capacity of the skeleton under different conditions. The stress concentration degree refers to the concentration of stress in certain parts of the skeleton, which is usually a potential risk area of fracture or damage. The combined judgment is to analyze the fracture density and stress concentration of the trabecular bone, and combine the two to judge the risk level of the skeleton.

[0025] In the case matching module, the interval measurement point refers to a specific location point in the bone joint image for measuring the joint interval to evaluate the change of the joint space. The interval difference refers to the change or difference of the joint interval in different image frames, which is used to analyze the change trend of the joint. The interval change trend refers to the change trend of the joint interval over time or during treatment, reflecting the progression of the lesion. The corresponding measurement point in the case data refers to the joint measurement point extracted from the case data corresponding to the image data. The difference calculation and sorting refer to the calculation of the joint interval difference, the analysis of the difference between cases, and the sorting to select the most suitable case. The case that meets the conditions refers to those cases that meet the requirements in terms of interval change trend and other parameters in the matching process.

[0026] In the treatment decision generation module, the spatial position refers to the position of the lesion in the skeleton, which usually describes the spatial relationship of the lesion relative to other structures. The size refers to the size of the lesion, which is usually used to evaluate its impact on the skeleton or joint. The progression rate refers to the development speed of the lesion over time, which is used to predict the long-term impact of the lesion. The difference between lesions refers to the difference in position, size, progression rate, etc. between different lesions. The interference relationship refers to the influence of one lesion on another, leading to interaction between lesions or aggravation of the disease. The sorted lesions refer to the sorting of lesions according to the interference relationship or other evaluation standards to determine the priority of treatment.

[0027] In the multi-source data fusion module, each device or source refers to data from different medical devices (such as X-ray machines, CT, MRI, etc.) or data sources (such as clinical records, image databases, etc.); the association between parameters refers to the relationship and mutual influence between parameters such as trabecular bone structure, joint spacing, and lesion location, which is usually revealed by data analysis to reveal the relationship between parameters.

[0028] In the intervention path optimization module, intervention effect refers to the clinical outcome or treatment effect during the treatment process, usually including symptom improvement, functional recovery, etc.; treatment effect parameter refers to various data indicators related to treatment effect, including disease improvement rate, recurrence rate, etc.; improvement refers to the change of the lesion or the improvement degree of the patient's health status during the treatment process; intervention path refers to the steps, methods, and sequence taken during the treatment process.

[0029] As shown in Figure 2 and Figure 3 , the risk division module includes: The fracture density calculation sub-module analyzes the number of fractures and the fracture interval of each trabecular bone unit based on three-dimensional bone image data, compares the proportional relationship between the number of fractures and the interval within the unit, calculates the weighted combination of the proportion and the total number of units, judges the difference pattern of the proportion distribution on the structure, and obtains the fracture distribution rate of the bone structure. When each trabecular bone unit is processed one by one, first locate the spatial range of the unit from the three-dimensional image, read the pixel matrix information within the range, distinguish the continuous segment and the fracture segment of the trabecular bone, and count the number of fractures of the unit, for example, it is detected that there are 3 fracture positions in the unit, then measure the interval between each two fractures, for example, 1.2mm, 1.0mm and 1.5mm respectively, calculate the average value of the interval, calculate the proportional relationship between the number of fractures and the average interval of the unit, for example, 3 divided by 1.23 is about 2.44, then adjust the proportional value according to the load bearing coefficient of the region where the unit is located, multiply the proportion by 1.2 in the high load bearing area, and multiply by 0.8 in the low load bearing area, for example, the unit is located in the high load bearing area, then the proportional value becomes about 2.93, group all the proportional values of the units according to the spatial region where they are located, and calculate the global average proportional value, for example, the average proportion is 2.10, compare the difference between the proportional value of each unit and the global average value one by one, and the difference greater than or equal to 0.5 is classified as the high proportion interval, the difference less than or equal to -0.5 is classified as the low proportion interval, and the rest is the normal interval, finally map the proportional interval distribution back to the three-dimensional space of the bone structure, count the proportion of the number of units with high, low, and normal proportions in the total number of units, and obtain the fracture distribution rate of the bone structure.

[0030] The stress ratio measurement sub-module calculates the maximum principal stress and average stress ratio of the trabecular bone unit based on the bone structure fracture distribution rate, analyzes the distribution law of the ratio in each structure area, compares the difference between the distribution and the stress concentration benchmark, and obtains the stress concentration degree of the bone unit; First, the position of each trabecular bone unit is determined, and the maximum principal stress value and the average stress value at that position are read, for example, the maximum principal stress of a certain unit is 18 MPa, and the average stress is 12 MPa, and the ratio obtained by dividing the two is 1.50. The stress ratios of all units are counted according to the area, and compared with the preset stress concentration benchmark value 1.30. If the difference between the ratio and the benchmark is greater than or equal to 0.20, it is classified as a high stress concentration area, if the difference is less than or equal to -0.20, it is classified as a low stress concentration area, and if the difference is between the two, it is classified as a medium stress concentration area. For example, 1.50 is 0.20 higher than 1.30, which is determined as a high stress concentration area. The classification results are position by position corresponding to the partition of the bone structure fracture distribution rate, and the proportion of units in the same area where high fracture distribution and high stress concentration overlap is counted. For example, the proportion of such overlap in a certain area is 65%, and the stress concentration degree of the bone unit in that area is 65%.

[0031] The risk stratification judgment sub-module judges the combination characteristics of the same trabecular bone unit based on the stress concentration degree of the bone unit, compares the difference level of the combination characteristics with the risk stratification benchmark, calculates the structure risk state corresponding to each level, and obtains the risk stratification identification; In the three-dimensional bone image coordinate system, the partition level of the fracture distribution rate and the level of the stress concentration degree of each trabecular bone unit are read at the same time, for example, the fracture distribution of a certain unit is in the high area, which is recorded as level 2, and the stress concentration degree is in the high concentration area, which is recorded as level 3. The two are combined into a level pair (2, 3), and the combination level is compared with the risk stratification benchmark table. Each combination level in the benchmark table corresponds to a risk level, for example, the combination (2, 3) corresponds to the risk level 5, and the level range is 1 to 5. The higher the value, the higher the risk. The combination level of each unit is compared one by one. If it is completely consistent with the benchmark table entry, the risk level is directly taken. If there is no complete match, it is processed by reducing one level according to the principle of similarity, for example, the combination (1, 3) has no corresponding entry, and is determined as level 4. Finally, the proportion of the risk level of all units is counted, for example, the proportion of level 5 in a certain bone segment is 40%, the proportion of level 4 is 35%, the proportion of level 3 is 15%, the proportion of level 2 is 7%, and the proportion of level 1 is 3%. The statistical results are written into the risk stratification identification, which includes the risk category number "R5", the risk level description "high risk dominates", and the structure damage characteristics "trabecular bone fracture is dense and stress concentration is obvious".

[0032] As shown in Figure 2 and Figure 4 , the case matching module comprises: The joint spacing screening submodule is based on the structural damage features contained in the risk stratification identification, locates and analyzes a plurality of joint images corresponding to the damage features in space, screens the measurement points in the images that have joint spacing measurement significance, calculates the spacing difference of the measurement points in adjacent image frames, judges the change mode of the difference between the image frames, and obtains the joint spacing fluctuation amplitude; The sequence frame data of the joint image is called, and the joint region number that needs to be analyzed is obtained from the risk stratification identification. The region is located in each frame of image, then the bone end boundary position is determined by the gray contrast method of bone joint image, and the distance between the joint end points is taken as the initial data of the candidate measurement point, for example, the spacing between the medial and lateral edges of the knee joint in a frame of image is detected to be 4.8 mm, and the position is marked as measurement point 1. Then the operation is repeated for all frames to obtain the spacing data of all measurement points in different frames. Then the measurement points are compared with the set joint spacing measurement significance threshold value. The threshold value is set according to the clinical reference to be within the range of 2.0 mm to 6.0 mm, which has statistical significance. The measurement points exceeding this range are marked as invalid measurement points and are removed. In the remaining measurement points, the spacing difference between adjacent image frames is calculated one by one, for example, measurement point 1 is 4.8 mm in the first frame and 5.1 mm in the second frame, then the difference is 0.3 mm. The difference value is stored in the corresponding measurement point difference sequence. Then the change mode of the difference value sequence of the measurement point between frames is analyzed. Specifically, the difference value change direction and amplitude are compared one by one. If the difference change direction is consistent and the amplitude accumulation is greater than 0.5 mm for three consecutive frames, it is marked as fluctuation rising type. If the accumulation is less than-0.5 mm, it is marked as fluctuation falling type. The rest is marked as fluctuation stable type, for example, the difference values of measurement point 1 in three frames are 0.3 mm, 0.4 mm and 0.5 mm, the direction is consistent and the cumulative change is 1.2 mm, which is determined as fluctuation stable type. Finally, the fluctuation types of all measurement points are counted to obtain the overall fluctuation amplitude of the joint spacing.

[0033] The change trend calculation submodule is based on the joint spacing fluctuation amplitude, calculates the continuous change amplitude of the adjacent frame spacing difference in the time dimension according to the order of the measurement points in the image sequence, and compares the distribution characteristics of the change amplitude between the measurement points to obtain the joint spacing trend factor. The measurement points are sorted according to the time sequence of the image frames, and the difference values between adjacent frames are taken from the first frame, and the continuous change amplitude of the difference values in the time dimension is calculated, for example, the difference of a certain measurement point between the first frame and the second frame is 0.2 mm, and the difference between the second frame and the third frame is 0.5 mm, and the continuous change amplitude is the sum of the two, which is 0.7 mm, then the operation is repeated for all measurement points to form a list of continuous change amplitudes for each measurement point, then the distribution characteristics of the continuous change amplitudes between different measurement points are compared, and the amplitude range is divided into three intervals: less than 0.5 mm, 0.5-1.0 mm, and greater than 1.0 mm, and the proportion of the number of measurement points in each interval is counted, and the interval with the highest proportion is taken as the change trend type of the measurement point, for example, most of the continuous change amplitudes of a certain measurement point fall in the 0.5-1.0 mm interval, which is determined as a moderate change trend, and the trend is coded as trend factor 2, and the trend factors of all measurement points are generated to obtain the joint spacing trend factor.

[0034] The case priority matching submodule performs difference calculation on the corresponding measurement points in the case data based on the joint spacing trend factor, compares the closeness of the difference to the matching reference, sorts the closeness, filters the cases that meet the conditions in the sorting result, and obtains the case matching result. The case measurement point data corresponding to the position of the measurement point in the image data is extracted from the case database, and the difference value between the trend factor of the case measurement point and the trend factor of the current measurement point is calculated one by one, for example, the trend factor of the current measurement point is 2, and the trend factor of a certain case corresponding position is 3, then the difference value is 1, and the difference values of all cases are stored in the difference table, then the difference value is compared with the preset matching reference, which is set as less than or equal to 1 for matching, and greater than 1 for not matching, for example, the difference value is 1, which is marked as matching, and the matching cases are filtered out by this judgment, and sorted from small to large according to the difference value, and the cases with the same difference value are arranged in ascending order of case ID, and finally the case numbers that meet the conditions in the sorting result are output as the case matching result.

[0035] As shown in Figure 2 and Figure 5 The treatment decision generation module includes: The lesion feature determination submodule analyzes the spatial position coordinates, contour size and progression rate of each lesion based on the case matching result, calculates the proportion of position difference and size change, judges the distribution of the proportion between the lesions, and obtains the lesion feature difference degree. When analyzing each lesion, the spatial position coordinates of the lesion in the three-dimensional bone image are extracted from the matching results, the coordinate values are mapped to the skeletal anatomical reference system, and the center point positions of the lesion in the three axial directions are recorded, for example, the center point of a lesion is located at 35.2 mm in the X-axis direction, 48.6 mm in the Y-axis direction, and 12.4 mm in the Z-axis direction. Subsequently, the shape contour data of the lesion is extracted, the length, width, and height of the shape are measured along the boundary direction pixel by pixel, and the volume size is calculated, for example, the length of the lesion is 14.5 mm, the width is 9.8 mm, and the height is 7.2 mm, so the volume is about 1028.6 mm³. Then, the image data of the lesion at different time points is called, the volume change rate over time is calculated, for example, the initial volume is 1028.6 mm³, and three months later, the measurement is 1180.4 mm³, so the progression rate is about 50.6 mm³ per month. The spatial position difference between each lesion is calculated as the sum of the absolute values of the three-axis coordinate differences, for example, the center point difference between lesion A and lesion B is 3.1 mm in the X-axis direction, 5.6 mm in the Y-axis direction, and 1.2 mm in the Z-axis direction, so the position difference is 9.9 mm. At the same time, the size change proportion of the lesion is calculated, for example, the volume growth rate to the initial volume ratio is 50.6 divided by 1028.6, which is about 0.049. The position difference value and the size change proportion are compared, and classified according to the numerical interval of the proportion value, 0 to 0.05 is low proportion, 0.05 to 0.15 is medium proportion, and greater than 0.15 is high proportion, for example, 0.049 is classified as low proportion. The proportion classification results of all lesions are counted and the distribution of different categories is calculated to obtain the lesion characteristic difference degree.

[0036] The interference relationship analysis submodule compares the degree of overlap in spatial position and the consistency of size change direction between lesions based on the lesion characteristic difference degree, calculates the mutual influence proportion between each lesion, judges the interaction mode of the influence proportion in the whole lesion set, and obtains the lesion interaction data. The interference relationship analysis submodule, based on the lesion feature difference degree, when comparing the overlapping degree of the spatial position between the lesions, first extracts the shape boundary data of each lesion in the three-dimensional bone image, calculates the minimum coordinate value and the maximum coordinate value of the lesion shape in each direction one by one, and judges whether the projection of the two lesions in each direction overlaps. If there is overlap in three directions, it is marked as complete overlap, if there is overlap in one to two directions, it is marked as partial overlap, for example, the X-axis range of lesion A is 32.0 to 42.0 millimeters, and the X-axis range of lesion B is 38.0 to 45.0 millimeters. The overlapping range in the X-axis direction is 38.0 to 42.0 millimeters, which is determined as partial overlap. Then analyze the consistency of the size change direction of the lesions, judge the positive and negative values of the volume change value of each lesion in the time sequence. Positive value represents volume increase, negative value represents volume decrease. If the change direction of the two lesions is the same, it is recorded as consistent, otherwise it is recorded as inconsistent. For example, lesion A increases by 30mm³ per month, and lesion B increases by 45mm³ per month, the change direction is consistent. Then calculate the mutual influence ratio between the lesions. The lesion pair with complete overlap and consistent change direction is recorded as the influence strength 100%, the lesion pair with partial overlap and consistent change direction is recorded as the influence strength 50%, the lesion pair with complete overlap but inconsistent change direction is recorded as the influence strength 70%, and the lesion pair with partial overlap and inconsistent change direction is recorded as the influence strength 30%. For example, lesions A and B are partially overlapped and have consistent change direction, so the mutual influence ratio is 50%. Finally, all lesion pairs are analyzed and the distribution of different influence ratios is counted to obtain the lesion interaction data.

[0037] The priority order determination submodule calls the lesion interaction data, sorts according to the interaction mode of each lesion, compares the matching degree of the sorting result and the lesion difference distribution, judges the sorting order that meets the matching condition, and obtains the lesion priority path. The matching degree is compared with the sorting result and the lesion difference distribution, and the formula is: ; The matching degree parameter is calculated , the sorting order that meets the matching condition is judged, and the lesion priority path is obtained, wherein, represents the sequence number difference of the first lesion in the sorting result, represents the sequence number difference scale coefficient of the first lesion, which is used to convert the sequence number difference into a measurement consistent with the difference amount, represents the difference amount of the first lesion in the difference distribution, represents the difference weight coefficient of the first lesion, represents the difference weight coefficient of the first The incremental difference between adjacent lesions in the differential distribution of individual lesions Represents the total number of lesions; The matching degree parameter refers to a comprehensive calculation parameter used to quantify the degree of consistency between the lesion ranking results and the lesion difference distribution. It comprehensively considers the difference in the lesion's sequence number in the ranking results (after scaling factor conversion), the amount of difference in the difference distribution, the difference increment of adjacent lesions, and the corresponding weight coefficient. It is calculated by the normalized ratio of the numerator and denominator to reflect the overall matching tightness between the two.

[0038] For lesion 1, lesion 2, and lesion 3, respectively set , , Corresponding , , Through calculation The converted sorting differences were obtained as follows: lesion 1 = 2.4, lesion 2 = 1, and lesion 3 = 2.4. Then, the variance of each lesion in the differential distribution was extracted. The value is taken as lesion 1. Lesion 2 Lesion 3 ,calculate The values ​​were calculated as follows: lesion 1 = 0.1, lesion 2 = 0.8, and lesion 3 = 0.8. These values ​​were then combined with the differential weighting coefficients for each lesion. Each was designated as lesion 1. Lesion 2 Lesion 3 Multiplying the absolute value of the difference for each lesion by its corresponding weighting coefficient, we get a weighted value of 0.09 for lesion 1, 0.8 for lesion 2, and 0.88 for lesion 3. The sum of these three values ​​forms the numerator, and the result is: ; Further extract the incremental difference between adjacent lesions in the differential distribution for each lesion. Set as lesion 1 Lesion 2 Lesion 3 Perform the sum of squares calculation for each term: Lesion 1 is ; Lesion 2 is ; Lesion 3 is ; After summing the three terms, perform a square root operation. The denominator is: ; Substitute the calculated results of the numerator and denominator into the matching degree parameter formula: ; This result indicates that the current matching degree parameter value The match is already in the medium deviation range of the set matching degree range. According to the preset matching degree level division standard, the matching degree range is set as follows: High match (highly consistent priority paths): ; Medium match (some adjustments are needed): ; Low-degree matching (requires complete path reconstruction): ; Therefore, a matching value of 0.3438 falls into the second level, belonging to the "medium matching" category. This means that there is a certain degree of inconsistency between the ranking result and the lesion difference distribution. Based on this result, the current lesion ranking order is not directly adopted as the final lesion priority path. Instead, the ranking optimization process is initiated, and the ranking order is fine-tuned based on the absolute value of local differences and the influence factor of high-weight lesions. The formula unifies the calculation of the ranking adjustment amount, the actual difference amount, and the adjacent fluctuation range, and performs normalization after adding weight adjustment. This effectively completes the measurement of the tightness between the ranking path and the actual difference distribution, so that the ranking result has data support and global consistency evaluation capabilities. This result indicates that the overall matching degree level between the ranking path and the lesion difference is within the middle deviation range on the normalized deviation index.

[0039] like Figure 2 and Figure 6 As shown, the multi-source data fusion module includes: The imaging and recording acquisition submodule collects orthopedic imaging data and clinical record data from various imaging devices based on the lesion priority path, analyzes the distribution pattern of trabecular bone structure and the position parameters of joint spacing measurement points in the images, judges the differences between data format and unit standard, and performs unification processing to obtain a unified dataset of orthopedic images. During the acquisition process, the sequence number of the lesions to be processed is first read from the lesion priority path, and the corresponding imaging device identifier and acquisition time window are extracted. For example, lesion number P03 corresponds to device MRI-02, and the acquisition time range is from May 1st to May 10th, 2024. Then, the interfaces of each imaging device are called sequentially to obtain the original orthopedic image files, and the basic parameters such as image acquisition resolution, pixel spacing, and grayscale depth are recorded. At the same time, the patient's examination records, medical record summaries, and laboratory test data are retrieved from the clinical record database according to the lesion number. The image files and clinical records are associated and stored according to the lesion number. Next, in each image, the distribution pattern of trabeculae is identified. The direction of trabeculae is determined by the pixel grayscale changes of continuous slices, and spatial features such as trabeculae branching points and intersections are extracted. For example, if a trabeculae distribution density of 1 / 2000 is detected in a certain area, the trabeculae distribution density is determined by the trabeculae distribution density of 1 / 20000. Three square millimeter lines are arranged longitudinally. The location of the joint distance measurement point is identified, and the two-dimensional and three-dimensional spatial coordinates of the measurement point are recorded. For example, the measurement point on the medial side of the knee joint is located at image coordinates (320, 450) and three-dimensional coordinates (35.4 mm, 48.2 mm, 12.0 mm). Then, it is determined whether the data formats collected by different devices are consistent. For example, device A outputs DICOM format and device B outputs TIFF format. If they are inconsistent, they are uniformly converted to DICOM format. At the same time, the unit standards are compared. For example, the length of the trabecular bone is in millimeters in one image and pixels in another image. It is converted to millimeters according to the pixel pitch. For example, when the pixel pitch is 0.25 mm, the length of 100 pixels is converted to 25 mm. Finally, the images and recorded data with unified format and units are integrated into a unified orthopedic image dataset.

[0040] The parameter correlation analysis submodule is based on a unified dataset of orthopedic images. It calculates the proportional relationship between the density distribution of trabecular bone structure and the change in interarticular distance, analyzes the correspondence between the proportional relationship and the spatial location of the lesion, compares the changing trends of each parameter under the same spatial coordinates, and obtains the correlation ratio of structural parameters. The proportional relationship between the density distribution of trabecular bone structure and the variation in interarticular distance is calculated using the following formula: ; Obtain structural proportion difference parameters The correlation between proportional relationships and the spatial location of lesions was analyzed, and the changing trends of each parameter under the same spatial coordinates were compared to obtain the structural parameter correlation ratios. The average of the relative rates of change in the density of the trabecular bone structure. The average of the relative rates of change representing the magnitude of the change in joint spacing; The structural proportion difference parameter refers to the ratio of the relative difference between the average relative rate of change of trabecular structural density and the average relative rate of change of joint spacing within the spatial measurement unit of the same unified orthopedic imaging dataset. It reflects the degree of relative difference between the changes of the two types of structural features in a specific spatial range and is used to represent the relative offset between changes in bone tissue microstructure and changes in joint spatial geometry.

[0041] Using 3D spatial coordinate data from a unified orthopedic imaging dataset, within the spatial measurement range, a cubic region unit size of 5mm×5mm×5mm was set, and the bone imaging data was divided into multiple measurement units. The trabecular density index was independently extracted from each unit. This index was calculated using the number of trabeculae per unit volume of bone tissue and the average trabecular thickness. It was assumed that the trabecular density of a certain measurement unit at the initial moment was 0.68 g / cm³. 3 The density after the later stage of treatment was 0.765 g / cm³. 3 The relative rate of change is: ; That is, 12.5%. Simultaneously, for the joint spacing measurement points corresponding to the same measurement unit, the joint spacing distance in the continuous time frame images is extracted. For example, if the initial frame is 2.5mm and the final frame is 2.73mm, the change is 0.23mm, then the relative change rate is: ; That is, 9.2%. After obtaining two average rates of change, the structural proportion difference parameter is calculated: ; If the structural proportion difference threshold is set to ±0.1, then this measurement unit The region falling within the interval [0.1, 1] can be classified as the dominant region for changes in bone tissue structure. Assume the initial trabecular density of the other measurement unit is 0.58 g / cm³. 3 The final density is 0.61 g / cm³. 3 The rate of change is: ; Simultaneously, the joint spacing increased from 2.32 mm to 2.49 mm, with a change rate of: ; but: ; At this point, the measurement unit is classified as belonging to the region dominated by changes in joint spacing. Therefore, a third measurement unit is defined. , ,but: ; This can be categorized as a region dominated by joint changes. If there is a separate measurement unit, , ,but: ; Determined to be a bone-dominant region, all measurement units The calculations are completed sequentially using the above method and projected back into the three-dimensional coordinate system. Each calculation result is then bound to its spatial location and compared with the spatial points of subsequently detected lesions. When three consecutive measurement units are found in the vicinity of a lesion, the calculation is completed. The values ​​are 0.152, 0.136, and 0.103, respectively, and all fall within the dominant interval [0.1, 1]. Therefore, the spatial region where the lesion is located can be identified as the region dominated by bone tissue changes, and the corresponding structural parameter correlation ratio is determined by... The projection overlap relationship between the spatial measurement unit and the lesion point is determined, and this ratio is quantified as the spatial weighted average value at the lesion center, for example: ; The results indicate that the lesion location is associated with changes in bone tissue structure and can be classified as a region affected by bone-dominant changes.

[0042] The fusion log generation submodule determines the degree of spatial and structural matching of the source data of the differences based on the correlation ratio of structural parameters, filters parameter combinations that meet the matching requirements, and records the corresponding image number and measurement point identifier in the order of the combination to obtain the data fusion log. The generated unified dataset of orthopedic images is read, and the trabecular bone parameters, interarticular distance parameters, and lesion location parameters for each lesion are extracted. These parameters are then paired according to lesion number, and the spatial matching degree of the source data is compared. Specifically, the distance difference between each image measurement point in three-dimensional space is calculated. For example, if the coordinates of a measurement point in image A are (35.4 mm, 48.2 mm, 12.0 mm), and in image B are (35.8 mm, 48.0 mm, 12.3 mm), the three-axis differences are 0.4 mm, 0.2 mm, and 0.3 mm, respectively, with a total difference of 0.9 mm. This difference is then compared with a spatial matching baseline value set to 1.0 mm. If the total difference... If the value is less than or equal to the baseline, the spatial matching is considered acceptable. Then, the degree of structural matching is compared. The differences in trabecular bone density and interarticular distance are compared with the structural matching baseline. For example, if the difference in trabecular bone density is less than 5% and the difference in interarticular distance is less than 0.3 mm, the structural matching is considered acceptable. Parameter combinations that are acceptable in both spatial and structural matching are selected and arranged in chronological order of image acquisition time. The corresponding image number and measurement point identifier are recorded one by one. For example, the record entry is "Image ID_MRI_20240503_01, Measurement Point ID_KJ_M01". This process is repeated until all image and measurement point combinations are recorded, resulting in a data fusion log.

[0043] like Figure 2 and Figure 7 As shown, the intervention path optimization module includes: The treatment effect assessment submodule analyzes the lesion change data corresponding to each treatment in the clinical treatment record of the lesion based on the data fusion log, calculates the proportion of morphological differences and the proportion of functional recovery of the lesion at each treatment stage, judges the trend pattern of the proportion change, and obtains the treatment progress improvement rate. The treatment date, treatment method, and imaging examination time points for the corresponding lesion are retrieved, and these time points are matched with the image numbers and measurement point identifiers in the fusion log to ensure that each treatment stage has corresponding imaging data and lesion structural parameters. Subsequently, within each treatment stage, the spatial contour point set of the lesion is extracted, and its length, width, and height are measured to calculate its volume. For example, if the lesion volume is 1250 mm³ in the first stage and 1100 mm³ in the second stage, the morphological difference ratio is (1250 minus 1100) divided by 1250, resulting in 0.12, indicating a 12% reduction in volume. Then, in the functional recovery assessment, the joint range of motion values ​​or pain scores from the clinical records are retrieved, and the post-treatment values ​​are compared with the pre-treatment values. If the range of motion of the joint increases from 65 degrees before treatment to 75 degrees, the percentage of functional recovery is (75 minus 65) divided by 65, resulting in 0.154, representing an increase of 15.4%. The proportion of morphological difference and the percentage of functional recovery at each treatment stage are simultaneously stored in the time series, and the direction and magnitude of the changes in the proportions of adjacent stages are compared. If both proportions increase or decrease and the magnitude of the change exceeds 0.05, they are marked as having a consistent trend; otherwise, they are marked as having an inconsistent trend. For example, if the proportion of morphological difference increases from 0.08 to 0.12 and the percentage of functional recovery increases from 0.10 to 0.15 from the first stage to the second stage, then they have a consistent trend. Finally, the proportion of stages with a consistent trend to the total number of stages is calculated, and this proportion is the treatment progress improvement rate.

[0044] The interference relationship quantification submodule compares the trend of each lesion on the time axis and the overlap of spatial location based on the treatment progress improvement rate, calculates the proportion distribution of mutual influence, determines the concentration pattern of the influence distribution, and obtains the lesion interference coefficient. The improvement rate of all lesions is mapped one-to-one with the direction of their proportional change in the time series, and labeled as "increase," "decrease," or "no change." Then, the positional coordinates of each lesion in three-dimensional space are matched to determine the degree of spatial overlap between two lesions. If the boundary intervals in all three axes intersect, it is considered complete overlap; if there is intersection in only one or two directions, it is considered partial overlap; and if there is no intersection in any of the three directions, it is considered no overlap. For example, lesion A has a range of 30.0 to 40.0 mm on the X-axis, 50.0 to 60.0 mm on the Y-axis, and 10.0 to 20.0 mm on the Z-axis; lesion B has a range of 35.0 to 45.0 mm on the X-axis, 55.0 to 65.0 mm on the Y-axis, and 15.0 to 25.0 mm on the Z-axis. Therefore, the three... If there is an intersection in the direction, it is determined to be a complete overlap. Then, the trend type and spatial overlap are matched. If the trend is consistent and completely overlapped, the mutual influence ratio is set to 100%; if the trend is consistent and partially overlapped, it is set to 50%; if the trend is inconsistent and completely overlapped, it is set to 70%; if the trend is inconsistent and partially overlapped, it is set to 30%; and if there is no overlap, it is 0%. For example, if lesions A and B have consistent trends and completely overlap, the mutual influence ratio is 100%. The mutual influence ratio of all lesion pairs is counted and classified according to the ratio range (0%-33%, 34%-66%, 67%-100%). The number of lesion pairs in each range is calculated, and the range with the highest ratio is taken as the influence distribution concentration pattern of the lesion set, and the lesion interference coefficient is obtained.

[0045] The path priority configuration submodule prioritizes lesions based on the lesion interference coefficient and the interference concentration pattern, compares the matching degree between the ranking results and the treatment improvement trend, filters the lesion order that meets the conditions, records the corresponding intervention method category, and obtains the treatment intervention suggestion configuration. The interference coefficients are sorted from largest to smallest, and the lesions with the highest coefficients are assigned priority 1, the next highest priority 2, and so on. For example, lesion A with an interference coefficient of 95% is priority 1, and lesion B with an interference coefficient of 80% is priority 2. Then, the sorting results are compared with the treatment improvement rate of each lesion, and the difference value between the two is calculated. A difference value less than or equal to 0.1 is considered a match, and a difference value greater than 0.1 is considered a mismatch. For example, lesion A has an interference coefficient sorted as 1 and an improvement rate sorted as 2, with a difference value of 1, so it is considered a mismatch. Matching lesions are selected from the sorting list, and the corresponding intervention method category is found in the treatment record according to the lesion number. For example, lesion A corresponds to arthroscopic surgery, and lesion B corresponds to physical rehabilitation. The lesion number, priority, and intervention method category are combined into an entry and recorded in sequence to form the final treatment intervention recommendation configuration.

[0046] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A data-driven based orthopedic treatment decision support system, characterized in that, The system comprises: The risk division module analyzes the trabecular bone fracture number and spacing based on the three-dimensional bone image data, calculates the fracture density, and combines the local stress distribution to call the principal stress and average stress, analyze the ratio of the two, and compare the fracture density and stress concentration degree to obtain the risk stratification identification; The case matching module filters the joint image measurement points corresponding to the risk area indicated in the risk stratification identification, calculates the spacing difference between adjacent image frames, determines the spacing change trend of each measurement point, calls the corresponding measurement point in the case data, performs difference calculation and sorting, selects the qualified cases as matching candidates, and obtains the case matching result; The treatment decision generation module judges the lesion spatial position, size, and progression rate based on the case matching result, calculates the lesion difference, analyzes the interference relationship and sorts them according to the influence degree, determines the lesion order of priority processing, and obtains the lesion priority path; The multi-source data fusion module collects orthopedic images and clinical records from various devices or sources based on the lesion priority path, performs standardization processing, and analyzes the correlation between the trabecular bone structure, joint spacing, and lesion position parameters to obtain the data fusion log.

2. The data-driven based orthopedic treatment decision support system of claim 1, wherein, The risk stratification identification includes risk category number, risk level description, and structural damage characteristics. The case matching result includes candidate case number, matching degree classification, and matching priority order. The lesion priority path includes lesion processing order number, lesion level grouping, and interference influence category. The data fusion log includes fusion data source category, fusion parameter type, and parameter correlation mode.

3. The data-driven based orthopedic treatment decision support system of claim 1, wherein, The risk division module comprises: The fracture density calculation submodule analyzes the fracture number and spacing of each trabecular bone unit based on the three-dimensional bone image data, compares the proportion relationship of the number and spacing of the unit, calculates the weighted combination of the proportion and the total number of units, judges the difference mode of the proportion distribution on the structure, and obtains the bone structure fracture distribution rate; The stress ratio measurement submodule calculates the maximum principal stress and average stress ratio of the trabecular bone unit based on the bone structure fracture distribution rate, analyzes the distribution law of the ratio in each structural area, compares the difference degree of the distribution and the stress concentration reference, and obtains the bone unit stress concentration degree; The risk stratification determination submodule judges the combination characteristics of the same trabecular bone unit based on the bone unit stress concentration degree, compares the difference level of the combination characteristics and the risk stratification reference, calculates the structure risk state corresponding to each level, and obtains the risk stratification identification.

4. The data-driven based orthopedic treatment decision support system of claim 1, wherein, The case matching module comprises: The joint spacing screening submodule locates and analyzes a plurality of joint images corresponding to the structural damage characteristics in space based on the structural damage characteristics contained in the risk stratification identification, screens the measurement points in the images that have joint spacing measurement significance, calculates the spacing difference of the measurement points in adjacent image frames, judges the change mode of the difference between the image frames, and obtains the joint spacing fluctuation amplitude; The change trend calculation sub-module calculates the continuous change amplitude of the interval difference between adjacent frames in the time dimension according to the order of the measurement points in the image sequence based on the joint interval fluctuation amplitude, and compares the distribution characteristics of the change amplitudes between the measurement points to obtain a joint interval trend factor; The case priority matching sub-module performs difference calculation on the corresponding measurement points in the case data based on the joint interval trend factor, compares the closeness of the differences to the matching reference, sorts the closeness, filters the cases that meet the conditions in the sorting result, and obtains a case matching result.

5. The data-driven based orthopedic treatment decision support system of claim 1, wherein, The treatment decision generation module includes: The lesion feature judgment sub-module analyzes the spatial position coordinates, contour size and progression rate of each lesion based on the case matching result, calculates the proportion of position difference and size change, judges the distribution of the proportion among the lesions, and obtains a lesion feature difference degree; The interference relationship analysis sub-module compares the overlap degree of the lesions in the spatial position and the consistency of the size change direction based on the lesion feature difference degree, calculates the mutual influence proportion between the lesions, judges the interaction mode of the influence proportion in the whole lesion set, and obtains lesion interaction data; The priority order determination sub-module calls the lesion interaction data, sorts the lesions according to the interaction mode, compares the matching degree of the sorting result and the lesion difference distribution, judges the sorting order that meets the matching conditions, and obtains a lesion priority path.

6. The data-driven based orthopedic treatment decision support system of claim 5, wherein, The matching degree of the comparison sorting result and the lesion difference distribution uses the formula: ; Calculating the matching degree parameter , judging the sorting order meeting the matching condition to obtain a lesion priority path, wherein, representing the sequence number difference of the first lesion in the sorting result, representing the sequence number difference scale coefficient of the first lesion, representing the difference amount of the first lesion in the difference distribution, representing the difference weight coefficient of the first lesion, representing the difference increment of adjacent lesions in the difference distribution of the first lesion, representing the total number of lesions.

7. The data-driven based orthopedic treatment decision support system of claim 1, wherein, The multi-source data fusion module includes: The image and record acquisition sub-module acquires orthopedic image data and clinical record data from each image device based on the lesion priority path, analyzes the distribution form of the trabecular bone structure and the position parameters of the joint interval measurement points in the image, judges the difference of the data format and unit standard, and performs uniformization processing to obtain an orthopedic image unified data set; The parameter correlation analysis sub-module calculates the proportional relationship between the trabecular bone structure density distribution and the joint interval change amplitude based on the orthopedic image unified data set, analyzes the corresponding mode of the proportional relationship and the lesion spatial position, compares the change trends of each parameter at the same spatial coordinates, and obtains a structure parameter correlation ratio; The fusion log generation sub-module judges the matching degree of the difference source data in space and structure based on the structure parameter correlation ratio, filters the parameter combinations that meet the matching degree requirements, records the corresponding image number and measurement point identifier in the combination order, and obtains a data fusion log.

8. The data-driven based orthopedic treatment decision support system of claim 7, wherein, The proportional relationship between the trabecular bone structure density distribution and the joint interval change amplitude uses the formula: ; Obtaining structural proportion difference parameters , analyzing the corresponding mode of the proportion relationship and the spatial position of the lesion, comparing the change trend of each parameter under the same spatial coordinates, obtaining the structural parameter correlation ratio, wherein, the average of the relative change rate representing the trabecular bone structure density, the average of the relative change rate representing the change amplitude of the joint spacing.

9. The data-driven based orthopedic treatment decision support system of claim 1, wherein, The system further includes: The intervention path optimization module judges the lesion clinical treatment record and the intervention effect based on the data fusion log, identifies the interference relationship between the lesions and the treatment effect parameters, analyzes the improvement, optimizes the intervention order, and obtains a treatment intervention suggestion configuration; The treatment intervention suggestion configuration includes intervention method categories, execution order, and adaptation condition categories.

10. The data-driven based orthopedic treatment decision support system of claim 9, wherein, The intervention path optimization module includes: The treatment effect determination submodule analyzes the lesion change data corresponding to each treatment in the clinical treatment record of the lesion based on the data fusion log, calculates the shape difference proportion and functional recovery amplitude proportion of the lesion at each treatment stage, judges the trend mode of the proportion change, and obtains a treatment progress improvement rate; The interference relationship quantification submodule compares the change trend and spatial position overlap of each lesion on the time axis based on the treatment progress improvement rate, calculates the proportion distribution of mutual influence, judges the concentration mode of the influence distribution, and obtains a lesion interference coefficient; The path priority configuration submodule prioritizes the lesions according to the interference concentration mode based on the lesion interference coefficient, compares the matching degree of the sorting result and the treatment improvement trend, screens the lesion order that meets the conditions, records the corresponding intervention mode category, and obtains a treatment intervention suggestion configuration.