Artificial intelligence assisted diagnosis and treatment system for bone and joint diseases based on internet hospital

Abstract

The present application relates to the technical field of bone and joint diseases, and particularly relates to an artificial intelligence assisted diagnosis and treatment system for bone and joint diseases based on an internet hospital; comprising: an information management module; an intelligent diagnosis assistance module; a remote monitoring and rehabilitation guidance module; a disease risk prediction module; a doctor-patient interaction and health popularization module, the present application builds a comprehensive system covering information collection, data integration, image analysis, rehabilitation tracking and other functions, systematically manages the whole process information of patients from pre-consultation to rehabilitation period, effectively improves the continuity of diagnosis and treatment and the level of personalized service, uses deep learning, natural language processing, knowledge graph and other artificial intelligence technologies to automatically complete bone and joint image recognition, symptom analysis and rehabilitation plan making, greatly improves the efficiency and accuracy of doctor's diagnosis, and simultaneously realizes dynamic adjustment and fine guidance of the rehabilitation scheme.

Description

Technical Field

[0001] This invention relates to the field of osteoarthritis technology, and more specifically to an artificial intelligence-assisted diagnosis and treatment system for osteoarthritis based on an internet hospital. Background Technology

[0002] Osteoarthritis is a common type of chronic disease among middle-aged and elderly people. Common types include osteoarthritis, osteoporosis, synovitis, and rheumatoid arthritis, characterized by high incidence, chronic progression, and high disability rates. With the increasing aging of the population, the demand for medical treatment for osteoarthritis continues to grow, putting significant pressure on the allocation of medical resources and the efficiency of diagnosis and treatment.

[0003] Currently, the diagnosis and rehabilitation management of musculoskeletal diseases in clinical practice mainly rely on offline hospitals. Doctors make diagnoses based on patient complaints, physical examinations, and imaging data, and then provide rehabilitation guidance based on individual circumstances. This traditional model has the following problems: uneven distribution of medical resources, lack of professional orthopedic diagnosis and treatment capabilities in primary healthcare institutions; reliance on manual interpretation of images, which is labor-intensive and easily affected by the doctor's experience level; patients' inability to receive long-term and continuous rehabilitation guidance, resulting in poor rehabilitation compliance and ineffective outcomes; lack of proactive identification and intervention for high-risk groups; insufficient disease prevention capabilities; limited channels for doctor-patient interaction; and limited health education methods, making it difficult to improve patients' self-management awareness.

[0004] In conclusion, developing an AI-assisted diagnosis and treatment system for osteoarthritis based on internet hospitals remains a critical issue that urgently needs to be addressed in the field of osteoarthritis technology. Summary of the Invention

[0005] The purpose of this invention is to address the following problems in the existing technology: primary healthcare institutions lack professional orthopedic diagnosis and treatment capabilities; image interpretation relies on manual labor, which is labor-intensive and easily affected by the doctor's experience level; patients cannot receive long-term and continuous rehabilitation guidance, resulting in poor rehabilitation compliance and ineffective results; there is a lack of proactive identification and intervention for high-risk groups; disease prevention capabilities are insufficient; and doctor-patient interaction channels are limited.

[0006] To achieve the above objectives, the present invention provides an artificial intelligence-assisted diagnosis and treatment system for bone and joint diseases based on an internet hospital, comprising: an information management module and an intelligent diagnostic assistance module;

[0007] The information management module is used to obtain bone texture fractal dimensions, dynamic change rate of joint space width, disease symptoms, and medical knowledge graphs.

[0008] The intelligent diagnostic assistance module is used to calculate diseases. of , Used to assist in diagnosis, specifically:

[0009] Based on the fractal dimension of bone texture and the dynamic change rate of mid-joint space width, a deep learning model is used to analyze medical images and obtain disease... Predicted probability ;

[0010] Based on symptom and medical knowledge graphs, calculate diseases The matching score is obtained. ;

[0011] Calculate current symptoms and disease :

[0012] ,

[0013] in, This represents the predicted probability of a disease obtained after analyzing medical images using a deep learning model. This represents the disease matching score calculated based on symptom semantic similarity and knowledge graph structure, where a and b are weight coefficients.

[0014] according to For diseases The system sorts the candidates and automatically generates diagnostic criteria for each candidate disease.

[0015] Beneficial effects

[0016] Compared with known public technologies, the technical solution provided by this invention has the following beneficial effects:

[0017] This invention constructs a comprehensive system that includes functions such as information collection, data integration, image analysis, and rehabilitation tracking. It systematically manages patient information throughout the entire process from pre-treatment to rehabilitation, effectively improving the continuity of diagnosis and treatment and the level of personalized services.

[0018] This invention utilizes artificial intelligence technologies such as deep learning, natural language processing, and knowledge graphs to automatically complete bone and joint image recognition, symptom analysis, and rehabilitation plan formulation, greatly improving the efficiency and accuracy of doctors' diagnoses, while also enabling dynamic adjustment and refined guidance of rehabilitation plans. Attached Figure Description

[0019] Figure 1 This is a system diagram of the AI-assisted diagnosis and treatment system for bone and joint diseases based on an internet hospital, as described in this invention. Detailed Implementation

[0020] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0021] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0022] The present invention will now be described in further detail with reference to the accompanying drawings:

[0023] Example:

[0024] like Figure 1 As shown, the present invention provides an artificial intelligence-assisted diagnosis and treatment system for bone and joint diseases based on an internet hospital, including: an information management module;

[0025] Furthermore, the operation process of the information management module includes:

[0026] When patients register at the internet hospital, they enter detailed personal information, past medical history, family medical history, and lifestyle habits. The system connects with the information systems of partner medical institutions through an interface to automatically obtain the patient's past imaging examination and test reports and diagnostic records. The system then structures the obtained data.

[0027] The system uses an automated image segmentation algorithm (U-Net) to identify the edges of the femur and tibia; calculates the minimum distance between the articular surfaces and outputs the JSW value; and automatically calculates ΔJSW using time labels (such as the image acquisition date) and visualizes it.

[0028] Image data feature quantization calculates the dynamic change rate of the joint space width in the image data, using the following formula:

[0029] ,

[0030] in, This indicates the width of the joint space measured in the current image. This indicates the joint space width measured in the previous imaging. Indicates time interval, This represents the error tolerance for image measurement or AI image segmentation. The dynamic rate of change of the joint space width is used as input to the AI ​​model.

[0031] Select the ROI (Region of Interest), such as the femoral condyle or tibial plateau; use the binarized results of the bone tissue image; automatically solve the fractal dimension based on the box-counting method; output the numerical value as input to the AI ​​model to help determine the disease stage.

[0032] The formula for obtaining the fractal dimension of bone texture is:

[0033] ,

[0034] in, Indicates using a side length of The minimum number of cubes required to completely cover the target bone tissue trabeculae. The side length of the cube is represented by a value that approaches 0 with microscopic precision. Fractal dimension representing the complexity of bone tissue structure in an image;

[0035] Specifically, when patients register at the internet hospital, they enter detailed personal information, past medical history, treatment experience related to bone and joint diseases, family medical history, lifestyle habits, exercise methods, and labor intensity. At the same time, the system connects with the information systems of partner medical institutions through interfaces to automatically obtain data such as patients' past imaging examinations, X-rays, CT scans, MRI scans, laboratory reports, and diagnostic records. Using structured and unstructured database technologies, the collected data is classified and stored, and data cleaning and standardization processes are used to ensure the accuracy and consistency of the data.

[0036] Intelligent diagnostic assistance module;

[0037] Furthermore, the operation process of the intelligent diagnostic assistance module includes:

[0038] The image intelligent analysis module uses a convolutional neural network (CNN) to extract features and identify lesions in user-uploaded bone and joint images. It precisely locates joint regions through an attention mechanism and automatically identifies lesion types such as fractures, bone hyperplasia, and joint space narrowing using a trained classification model. It also measures quantitative indicators such as joint space width. Precise joint region localization is achieved using a UNET network with an attention gating mechanism. The formula for calculating the attention weight of each pixel belonging to the joint region is as follows:

[0039] ,

[0040] in, Indicates at pixel point Image features With position The concatenated vector, This represents the trainable attention weight parameters. This means that all pixels are normalized to obtain a probability distribution. This indicates the degree of integration of the location with the current task of fracture identification or joint contour segmentation, and can automatically focus on image regions that are highly correlated with joint lesions, thereby improving the accuracy of lesion detection and segmentation.

[0041] Using natural language symptom descriptions, and employing pre-trained models such as BERT, the text is transformed into a feature vector model:

[0042] ,

[0043] in, The patient reported dull knee pain at night and morning stiffness lasting 60 minutes.

[0044] Using medical knowledge graphs for reasoning, the strength of the association between symptoms and diseases is calculated. The calculation formula is:

[0045]

[0046] Among them, the vector of the user's current input symptom, Indicating the first in the atlas A vector of standard symptom nodes, This indicates the relationship between the current symptom and the symptom nodes in the knowledge graph. Semantic similarity between them Indicates symptoms To disease The graph edge weights represent the strength of the correlation between the two. Indicates the relationship between the atlas and candidate diseases The number of all symptom nodes connected by edges.

[0047] The formula for calculating the semantic matching strength between the current symptom and the disease node is as follows:

[0048] ,

[0049] in, This indicates the probability of disease in the image output. Indicate candidate diseases The final score, This represents the predicted probability of a disease obtained after analyzing medical images using a deep learning model (such as CNN). This represents the disease matching score calculated based on symptom semantic similarity and knowledge graph structure.

[0050] The system ranks the diseases based on a comprehensive score. The system sorts the candidate diseases and automatically generates diagnostic criteria for each disease, including typical imaging features such as bone hyperplasia, joint space narrowing, location of cartilage degeneration areas, symptom similarity analysis of the degree of overlap between the pain location and the disease onset area, and corresponding clinical differential suggestions. It also distinguishes between osteoarthritis and rheumatoid arthritis by the duration of morning stiffness and serological indicators, assisting doctors in making accurate diagnoses and appropriate treatments.

[0051] Specifically, the intelligent image analysis uses deep learning algorithms and convolutional neural networks to automatically identify morphological and structural changes in bones and joints, detect fractures, bone hyperplasia, and joint space narrowing. The system generates detailed image analysis reports, marking the location, extent, and severity of lesions to provide doctors with diagnostic references. Patients input their current symptoms, pain location, severity, onset time, and accompanying symptoms. The system uses machine learning algorithms, combined with massive clinical case data and medical knowledge graphs, to analyze the correlation between symptoms and diseases, providing doctors with a list of possible disease diagnoses, sorted by probability, and offering corresponding diagnostic criteria and differential diagnosis suggestions.

[0052] Remote monitoring and rehabilitation guidance module

[0053] Rehabilitation equipment connection:

[0054] 1. The system provides open IoT access capabilities, supporting patients to bind and authorize various wearable devices and home smart rehabilitation devices.

[0055] 2. Collect data such as joint range of motion, muscle strength, gait parameters, subjective pain scores, and physiotherapy equipment usage in real time or at regular intervals, and automatically upload them to the system platform through an encrypted channel.

[0056] Personalized rehabilitation plan development:

[0057] 1. Based on the patient's final diagnosis, current rehabilitation stage, baseline assessment of physical function, and personal goals, the system uses rule-based reasoning combined with reinforcement learning (RL) to automatically generate personalized, step-by-step rehabilitation training plans, including specific movement diagrams / videos, number of sets, repetitions, frequency, intensity / resistance settings, and rehabilitation physiotherapy protocols.

[0058] 2. The plan clearly sets phased rehabilitation goals and expected progress timelines.

[0059] Follow-up and adjustment of the rehabilitation process:

[0060] 1. The system continuously receives and analyzes the rehabilitation monitoring data uploaded by patients, and dynamically compares the actual progress of ROM improvement and pain reduction with the preset goals.

[0061] 2. If the algorithm detects that the recovery progress is significantly lagging behind expectations, key indicators are deteriorating, or abnormal patterns appear, the system will automatically trigger an alert and notify the responsible doctor to conduct a remote or offline clinical assessment.

[0062] 3. After the doctor's assessment, the system can combine the assessment conclusions and use a feedback-based model to predict and control the MPC to intelligently adjust the rehabilitation plan, including modifying the difficulty of movements, increasing or decreasing the amount of training, and adjusting the physical therapy parameters.

[0063] To ensure patients can perform the exercises correctly, the system provides multimedia guidance and can also incorporate data from wearable devices to provide real-time motion correction feedback during patient practice.

[0064] Disease risk prediction module

[0065] Risk Factor Analysis:

[0066] 1. The system aggregates multidimensional data from patient records: demographic data, genetic information, detailed lifestyle habits, past medical history, bone density, and specific biomarkers.

[0067] 2. Using association rule mining (Apriori) and feature importance analysis (SHAP) algorithm, we identified modifiable and non-modifiable risk factors that are significantly associated with the occurrence and progression of specific osteoarthritis, including primary osteoarthritis (OA), osteoporosis (OP), and rheumatoid arthritis (RA).

[0068] Risk prediction model construction:

[0069] 1. Based on massive historical data of de-identified osteoarthritis patient cohorts, train and validate various machine learning risk prediction models:

[0070] ① Logistic Regression: Used for binary risk prediction.

[0071] ② Cox proportional hazards model: Predicts the risk of disease occurrence within a specific time window over the next 5 or 10 years.

[0072] ③ Gradient boosting machine or deep learning model: Handles more complex nonlinear relationships and high-dimensional features, improving prediction accuracy.

[0073] 2. The model output is the quantitative risk probability of a patient developing a specific target musculoskeletal disease.

[0074] Early warning and prevention recommendations:

[0075] 1. When the system predicts that a patient is at medium to high risk, it automatically generates an early warning message and pushes it to the patient and their contracted / attending physician simultaneously via platform message, SMS or email.

[0076] 2. Based on the identified dominant risk factors, the system generates personalized and actionable prevention and intervention recommendations:

[0077] ① Lifestyle interventions: such as setting weight loss goals, recommending low-impact exercises, and avoiding specific occupational postures / loads.

[0078] ② Nutritional supplementation recommendations: such as calcium supplements and vitamin D intake guidelines.

[0079] ③ Preventive exercise prescription: A training program specifically designed to strengthen the muscles around the joints and improve joint stability.

[0080] ④ Regular monitoring recommendations: It is recommended that high-risk individuals undergo regular bone density tests or specific joint imaging screenings.

[0081] Doctor-patient interaction and health science popularization module

[0082] Online consultation:

[0083] 1. Provides secure and confidential real-time online consultation channels in various formats, including text, voice, and high-definition video.

[0084] 2. When a doctor sees a patient, the system proactively pushes the patient's integrated health record and the analysis results from the intelligent diagnostic assistance module, including imaging reports and a list of diagnostic suggestions, to help the doctor quickly and comprehensively grasp the condition, improve the efficiency and accuracy of online consultations, and provide more precise treatment suggestions.

[0085] Health Science Popularization:

[0086] 1. The system maintains a structured knowledge base for musculoskeletal diseases, including popular science articles, short videos, animations, infographics, etc. from authoritative sources.

[0087] 2. The content covers the entire lifecycle of disease prevention, early identification, diagnostic methods, treatment options, postoperative care, home rehabilitation, nutrition and health care, and psychological adjustment.

[0088] 3. The application of content recommendation algorithms is based on collaborative filtering or content similarity recommendation systems. According to the patient's specific disease diagnosis, current health status, recovery stage, and historical browsing / interaction behavior, personalized and accurate popular science content is pushed to improve the patient's disease awareness and self-health management ability.

[0089] Patient feedback and evaluation:

[0090] 1. After completing online consultations, using rehabilitation guidance services, or receiving popular science information, patients can evaluate and provide feedback on the doctor's service attitude, professional level, clarity of answers, and the ease of use, stability, and helpfulness of the system functions through standardized questionnaires or open-ended comments.

[0091] 2. The system collects this feedback data in a structured manner, and uses sentiment analysis technology to mine and analyze it. The results are used to continuously improve the quality of medical services, optimize the user experience of the system, and improve the functional design.

[0092] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An AI-assisted diagnosis and treatment system for bone and joint diseases based on an internet hospital, characterized in that: include: Information management module and intelligent diagnostic assistance module; The information management module is used to obtain bone texture fractal dimensions, dynamic change rate of joint space width, disease symptoms, and medical knowledge graphs. The intelligent diagnostic assistance module is used to calculate diseases. of , Used to assist in diagnosis, specifically: Based on the fractal dimension of bone texture and the dynamic change rate of mid-joint space width, a deep learning model is used to analyze medical images and obtain disease... Predicted probability ; Based on symptom and medical knowledge graphs, calculate diseases The matching score is obtained. ; Calculate current symptoms and disease : ， in, This represents the predicted probability of a disease obtained after analyzing medical images using a deep learning model. This represents the disease matching score calculated based on symptom semantic similarity and knowledge graph structure, where a and b are weight coefficients. according to For diseases Sort them, and for each candidate disease Automatically generate diagnostic criteria.

2. The artificial intelligence-assisted diagnosis and treatment system for bone and joint diseases based on an internet hospital as described in claim 1, characterized in that, The information management module is used for: The image features are quantified to calculate the dynamic rate of change of the joint space width: , in, This indicates the width of the joint space measured in the current image. This indicates the joint space width measured in the previous imaging. Indicates time interval, This indicates the error tolerance for image measurement or AI image segmentation.

3. The artificial intelligence-assisted diagnosis and treatment system for bone and joint diseases based on an internet hospital as described in claim 2, characterized in that, In the information management module: The formula for determining the fractal dimension of bone texture is: in, Indicates using a side length of The minimum number of cubes required to completely cover the target bone tissue trabeculae. The side length of the cube is represented by a value that approaches 0 with microscopic precision. The fractal dimension represents the complexity of bone tissue structure in an image.

4. The artificial intelligence-assisted diagnosis and treatment system for bone and joint diseases based on an internet hospital as described in claim 1, characterized in that, The system also includes a remote monitoring and rehabilitation guidance module, used for: Receive joint range of motion, muscle strength, gait parameters, subjective pain scores, and physiotherapy equipment usage data uploaded from wearable devices or home smart rehabilitation devices bound to the patient; Based on the patient's diagnosis, rehabilitation stage, baseline assessment of physical function, and personal goals, a personalized, step-by-step rehabilitation training plan is automatically generated using rule-based reasoning combined with a reinforcement learning model. The plan includes specific movement instructions, number of sets, number of repetitions, frequency, intensity settings, and phased rehabilitation goals. The system continuously analyzes the rehabilitation monitoring data uploaded by patients, dynamically compares the actual rehabilitation progress with the preset goals, and automatically triggers an alert and notifies the responsible doctor when it detects a lag in rehabilitation progress, a deterioration of key indicators, or an abnormal pattern. Receive assessment feedback from doctors and dynamically adjust the rehabilitation plan based on the feedback using model predictive control algorithms.

5. The artificial intelligence-assisted diagnosis and treatment system for bone and joint diseases based on an internet hospital as described in claim 1, characterized in that, The system also includes a disease risk prediction module, used for: Aggregate multidimensional data from patient records, including demographic data, genetic information, lifestyle habits, medical history, bone density, and specific biomarkers; Using association rule mining and feature importance analysis algorithms, risk factors significantly associated with the occurrence and progression of specific osteoarthritis diseases were identified; Based on historical de-identified patient cohort data, a machine learning risk prediction model is trained and validated. The model includes at least one of logistic regression, Cox proportional hazards model, gradient boosting machine or deep learning model, and is used to output the quantitative risk probability of a patient having a specific musculoskeletal disease. When the predicted risk reaches the medium-to-high risk threshold, an early warning message is automatically generated and pushed to patients and doctors, and personalized prevention and intervention suggestions are generated based on the dominant risk factors.

6. The artificial intelligence-assisted diagnosis and treatment system for bone and joint diseases based on an internet hospital as described in claim 1, characterized in that, The system also includes a doctor-patient interaction and health education module, used for: It provides online consultation channels in the form of text, voice or video, and proactively pushes the patient's health records and the analysis results of the intelligent diagnostic assistance module to the doctor during the consultation process; Maintain a structured knowledge base for musculoskeletal diseases, including popular science content related to disease prevention, diagnosis, treatment, rehabilitation, and health care; Based on the patient's specific disease diagnosis, health status, recovery stage, and historical behavior, recommendation algorithms are used to achieve personalized and accurate delivery of popular science content. We receive patient evaluations and feedback on consultation services, rehabilitation guidance, and science popularization content, and use sentiment analysis technology to mine and analyze the feedback data in order to continuously improve service quality and system functions.

Images