Multi-agent medical diagnosis and treatment system

Abstract

The invention relates to a multi-agent medical diagnosis and treatment system which comprises a plurality of agents. The plurality of agents comprise: a task planning agent analyzes input content and generates a corresponding task process, and the task planning agent sequentially issues task instructions to a data management agent, a tool use agent and a medical decision agent according to the task process; the task instructions comprise a data task instruction, a tool task instruction and a decision task instruction; the data management agent extracts patient data from the database and sends the patient data to the task planning agent according to a data task instruction issued by the task planning agent; the tool using agent performs feature extraction processing on the patient data according to the patient data and the tool task instruction issued by the task planning agent so as to obtain key features and send the key features to the task planning agent; and the medical decision-making agent carries out diagnosis and treatment reasoning according to the key features and the decision-making task instruction issued by the task planning agent, and a diagnosis report is obtained and displayed.

Description

Technical Field

[0001] The present application relates to the technical field of obtaining diagnosis and treatment data, and in particular to a multi-agent medical diagnosis and treatment system. Background Art

[0002] Chronic inflammatory arthritis often requires a combination of imaging data and patient-reported symptoms for diagnosis and treatment. However, there is a semantic gap between patient-reported symptoms and medical terminology, and imaging diagnosis relies on manual reading by radiologists, with an average response time of 48 to 72 hours, which is time-consuming. Therefore, there is a technical problem of inefficient acquisition of diagnostic and treatment data for chronic inflammatory arthritis. Summary of the invention

[0003] Based on this, the purpose of this application is to provide a multi-agent medical diagnosis and treatment system that can overcome the shortcomings of the existing technology.

[0004] In order to achieve the above purpose, the technical solution adopted in this application is:

[0005] A multi-agent medical diagnosis and treatment system, comprising a plurality of agents; the plurality of agents comprising: a task planning agent, a data management agent, a tool use agent and a medical decision agent;

[0006] The task planning agent analyzes the input content and generates a corresponding task flow, and the task planning agent issues task instructions to the data management agent, the tool use agent and the medical decision agent in sequence according to the task flow; the task instructions include data task instructions, tool task instructions and decision task instructions;

[0007] The data management agent extracts patient data from the database and sends it to the task planning agent according to the data task instruction issued by the task planning agent;

[0008] The tool-using agent performs feature extraction processing on the patient data according to the patient data and tool task instructions issued by the task-planning agent to obtain key features and send them to the task-planning agent;

[0009] The medical decision-making agent performs diagnosis and treatment reasoning according to the key features and decision-making task instructions issued by the task planning agent, generates and displays a diagnosis report.

[0010] As an implementation method, the task planning agent analyzes input content and generates a corresponding task process, including:

[0011] The task planning agent integrates user input and information of each agent according to a preset text template to obtain input content;

[0012] The task planning agent analyzes the input content through a large language model to obtain the task process.

[0013] As an implementation mode, the task planning agent analyzes the input content through a large language model to obtain the steps of the task flow, including:

[0014] The task planning agent analyzes the input content through the large language model to obtain the text task flow output by the large language model;

[0015] The task planning agent converts the text task flow into a list task flow, and sends task instructions to the corresponding agent for execution according to the list task flow.

[0016] As an implementation mode, the data management agent extracts patient data from a database and sends the patient data to the task planning agent according to the data task instruction issued by the task planning agent, including:

[0017] The data management agent generates a data query code according to the data task instruction;

[0018] The data management agent uses the data query code to query the corresponding patient data from the database.

[0019] As an implementation method, the patient data includes the patient's electronic medical record information, imaging data, project examination data and clinical data.

[0020] As an implementation mode, the tool using agent performs feature extraction processing on the patient data according to the patient data and tool task instructions issued by the task planning agent to obtain key features and send them to the task planning agent, including:

[0021] The tool uses an agent to preprocess the patient data sent by the task planning agent to obtain the preprocessed patient data;

[0022] The tool using agent performs feature extraction processing on the pre-processed patient data according to the tool task instruction to obtain the key features.

[0023] As an implementation mode, the patient data includes image data; the tool uses an agent to preprocess the patient data sent by the task planning agent to obtain the preprocessed patient data, including:

[0024] The tool uses an intelligent agent to perform image enhancement, normalization and resize processing on the patient data to obtain the pre-processed patient data.

[0025] As an implementation mode, the tool uses an intelligent agent to perform feature extraction processing on the pre-processed patient data according to the tool task instruction to obtain the key features, including:

[0026] The tool using agent extracts the target area corresponding to the tool task instruction from the image data through a segmentation network, and then extracts features of the target area according to the scoring model corresponding to the tool task instruction to obtain the key features;

[0027] Or the tool uses an intelligent agent to perform statistical analysis on the clinical data to identify the key features associated with axSpA.

[0028] As an implementation manner, the feature extraction of the target area according to the scoring model corresponding to the tool task instruction to obtain the key feature includes:

[0029] The tool uses an intelligent agent to perform SPARCC scoring and SSS scoring on the key features to obtain the key features including the scores.

[0030] As an implementation mode, the medical decision agent performs diagnosis and treatment reasoning according to the key features and decision task instructions issued by the task planning agent, generates and displays a diagnosis report, including:

[0031] The medical decision-making agent searches for a number of most similar case records according to the key features;

[0032] The medical decision-making agent performs diagnosis and treatment reasoning based on the key features, the preset axSpA diagnostic criteria and the several most similar case records to obtain the diagnosis report.

[0033] Compared with traditional technologies, the multi-agent medical diagnosis and treatment system described in this application has the following beneficial effects:

[0034] This system optimizes the entire diagnostic process through the collaborative cooperation of multiple agents. Each agent is responsible for specific tasks, such as data management, tool use, and medical decision-making. They can work together efficiently, avoiding the inefficiency and work overlap that may be caused by a single agent, and ensuring the smooth connection and execution of each link. It improves diagnostic efficiency and reduces the workload of doctors. Automated data processing and preliminary analysis allow doctors to focus on more complex clinical decisions without spending a lot of time on data collation and preliminary analysis. This efficiency improvement not only helps doctors reduce work pressure, but also improves the quality of overall medical services. Among them, the tool-using agent can automatically process the patient's imaging data and clinical data, extract key information, and avoid errors and time-consuming problems that may be caused by manual intervention. The medical decision-making agent integrates various data sources, including medical records, imaging data, and laboratory test results, and uses deep learning algorithms for comprehensive analysis to provide highly accurate diagnostic results. Compared with traditional methods, the system can process more complex data and conduct more detailed analysis to help doctors make more accurate diagnosis and treatment decisions.

[0035] For better understanding and implementation, the present application is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A schematic diagram of a multi-agent medical diagnosis and treatment system according to an embodiment of the present application.

[0037] Figure 2 This is a schematic diagram of the steps in the application of a multi-agent medical diagnosis and treatment system according to one embodiment of the present application.

[0038] 101. Task planning agent; 102. Data management agent; 103. Tool usage agent; 104. Medical decision-making agent. DETAILED DESCRIPTION

[0039] In order to make the objectives, technical solutions and advantages of the present application more clear, the embodiments of the present application will be further described in detail below with reference to the accompanying drawings.

[0040] It should be clear that the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the embodiments of the present application.

[0041] When the following description relates to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. In the description of the present application, it should be understood that the terms "first", "second", "third", etc. are only used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence, nor can they be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances. The singular forms of "a", "said" and "the" used in the present application and the appended claims are also intended to include the majority form, unless the context clearly indicates other meanings. The words "if" / "if" used herein can be interpreted as "at the time of" or "when" or "in response to determination".

[0042] In addition, in the description of this application, unless otherwise specified, "plurality" means two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects are in an "or" relationship.

[0043] Among them, the multi-agent medical diagnosis and treatment system of the present application is a diagnosis and treatment system for axial spondyloarthritis (axSpA). AxSpA is a chronic inflammatory joint disease with axial spondyloarthritis as the main manifestation. The disease often affects the sacroiliac joints and may cause bamboo-like changes in the spine, seriously affecting the patient's quality of life. Patients mainly present with chronic back pain and stiffness in the pelvis and lower back, but any part of the spine may be affected. However, the sensitivity and specificity of any of these criteria for diagnosing axSpA are no more than about 80%, which means that not every axSpA patient has this back pain, and chronic back pain caused by other causes needs to be identified. MRI can observe active sacroiliitis early and is the main source of imaging information for axSpA diagnosis and treatment decisions. Understanding the extraction of key information with diagnostic value from imaging images and combining comprehensive analysis of lesion sites are very important links in the diagnosis and treatment of axSpA. Therefore, the diagnosis of axSpA depends on the recognition of the clinical pattern of the disease based on clinical, laboratory and imaging features.

[0044] See also Figure 1, which is a schematic diagram of a multi-agent medical diagnosis and treatment system according to an embodiment of the present application, including multiple agents; the multiple agents include: a task planning agent 101, a data management agent 102, a tool usage agent 103, and a medical decision-making agent 104. Among them, the core of the agent is a large language model, which can be, for example, Qwen or DeepSeek, etc. Each agent of the agent sets a prompt and will convert the input text into an instruction and send it to the large language model.

[0045] The task planning agent 101 analyzes the input content and generates a corresponding task process, and the task planning agent 101 sequentially issues task instructions to the data management agent 102, the tool usage agent 103, and the medical decision-making agent 104 according to the task process; the task instructions include data task instructions, tool task instructions, and decision task instructions.

[0046] Among them, the main function of the task planning agent 101 is to analyze the user's needs and design a reasonable task process according to the needs. It coordinates the interaction between each agent according to the input instructions to ensure that each agent can work efficiently and cooperatively to complete the diagnosis task. The task planning agent 101 takes the large language model as the core and has strong natural language understanding and reasoning capabilities, and can quickly generate detailed task instructions according to the medical problems or symptom descriptions input by the user.

[0047] The data management agent 102 extracts patient data from the database according to the data task instructions issued by the task planning agent 101 and sends it to the task planning agent 101.

[0048] Among them, the data management agent 102 is mainly responsible for extracting relevant patient data from the hospital or clinic database, including medical record information, historical examination records, imaging data, etc. It can directly access the hospital's electronic medical record system (EMR) to query the patient's diagnosis history, treatment records, and related biomarker data, providing basic data support for subsequent analysis.

[0049] The tool usage agent 103 performs feature extraction processing on the patient data according to the patient data and tool task instructions issued by the task planning agent 101 to obtain key features and sends them to the task planning agent 101.

[0050] Among them, the tool usage agent 103 is responsible for performing data preprocessing and calling deep learning models. According to the instructions of the task planning agent 101, it will select appropriate software tools to perform tasks such as image processing, data cleaning, and feature analysis according to different disease analysis needs, and use corresponding deep learning algorithms to extract features from the patient's imaging data.

[0051] The medical decision-making agent 104 performs diagnosis and treatment reasoning based on the key features and decision task instructions issued by the task planning agent 101, and generates and displays a diagnostic report.

[0052] Among them, the medical decision-making agent 104 is used to comprehensively analyze and judge the patient data and the corresponding key features, and finally generate a diagnosis result of axSpA. The medical decision-making agent 104 will calculate the weights of different symptoms according to the algorithm model, so as to obtain the most likely diagnosis conclusion. At the same time, the medical decision-making agent 104 will also give personalized treatment suggestions based on the diagnosis result, including options such as drug treatment, physical therapy, and surgical intervention, to help doctors formulate personalized treatment plans for patients.

[0053] In a feasible embodiment, the task planning agent 101 analyzes the input content and generates a corresponding task process, including:

[0054] The task planning agent 101 integrates the user input and the information of each agent according to a preset text template to obtain the input content;

[0055] The task planning agent 101 analyzes the input content through a large language model to obtain the task process.

[0056] In a feasible embodiment, the task planning agent 101 analyzes the input content through a large language model to obtain the steps of the task process, including:

[0057] The task planning agent 101 analyzes the input content through a large language model to obtain the text task process output by the large language model;

[0058] The task planning agent 101 converts the text task process into a list task process, and sends task instructions to the corresponding agent for execution according to the list task process.

[0059] In a feasible embodiment, the data management agent 102 extracts patient data from the database according to the data task instructions issued by the task planning agent 101 and sends it to the task planning agent 101, including:

[0060] The data management agent 102 generates a data query code according to the data task instructions;

[0061] The data management agent 102 queries the corresponding patient data from the database with the data query code.

[0062] In a feasible embodiment, the patient data includes patient electronic medical record information, imaging data, item examination data, and clinical data.

[0063] In a feasible embodiment, the tool uses agent 103 to perform feature extraction processing on the patient data and tool task instructions issued by the task planning agent 101 to obtain key features and send them to the task planning agent 101, including:

[0064] The tool uses agent 103 to perform data preprocessing on the patient data issued by the task planning agent 101 to obtain the preprocessed patient data;

[0065] The tool uses agent 103 to perform feature extraction processing on the preprocessed patient data according to the tool task instructions to obtain the key features.

[0066] In a feasible embodiment, the patient data includes imaging data; the tool uses agent 103 to perform data preprocessing on the patient data issued by the task planning agent 101 to obtain the preprocessed patient data, including:

[0067] The tool uses agent 103 to perform image enhancement, normalization, and Resize processing on the patient data to obtain the preprocessed patient data.

[0068] In a feasible embodiment, the tool uses agent 103 to perform feature extraction processing on the preprocessed patient data according to the tool task instructions to obtain the key features, including:

[0069] The tool uses agent 103 to extract the target region corresponding to the tool task instructions from the imaging data through a segmentation network, and then performs feature extraction on the target region according to the scoring model corresponding to the tool task instructions to obtain the key features;

[0070] Or the tool uses agent 103 to perform statistical analysis on the clinical data to identify the key features related to axSpA.

[0071] In a feasible embodiment, after performing feature extraction on the target region according to the scoring model corresponding to the tool task instructions to obtain the key features, it includes:

[0072] The tool uses agent 103 to perform SPARCC scoring and SSS scoring on the key features to obtain the key features including the scores.

[0073] The SPARCC score is obtained through the SPARCC scoring tool, which is the core imaging tool for evaluating active inflammation in axial spondyloarthritis. It quantitatively analyzes the range, depth, and signal intensity of bone marrow edema in the sacroiliac joint through the MRI STIR sequence.

[0074] The SSS score is obtained through the SSS scoring tool, which is used to quantify the degree of structural damage in the sacroiliac joint and is graded based on the characteristics of bone erosion, sclerosis, and ankylosis on X-ray / CT.

[0075] Among them, the tool is trained using various software tools corresponding to the intelligent agent 103. For example, the training content of the SPARCC scoring tool includes image normalization operation, ROI region segmentation operation (Unet), and edema recognition operation (three different ResNets are used for edema recognition, brightness recognition, and depth recognition respectively); the training content of the SSS scoring tool includes image normalization operation, ROI region segmentation operation (Unet), and structured scoring operation (four different ResNets are used for the classification of ankylosis, fatty metaplasia, erosion, and backfill respectively).

[0076] In a feasible embodiment, the medical decision-making intelligent agent 104 performs diagnosis and treatment reasoning according to the key features and decision task instructions issued by the task planning intelligent agent 101, and generates and displays a diagnostic report, including:

[0077] The medical decision-making intelligent agent 104 searches for several most similar case records according to the key features;

[0078] The medical decision-making intelligent agent 104 performs diagnosis and treatment reasoning according to the key features, the preset axSpA diagnostic criteria, and the several most similar case records to obtain the diagnostic report.

[0079] Among them, the diagnostic report includes three parts: diagnosis result, diagnosis reason, and treatment suggestion. The diagnostic report presents the diagnosis result and treatment suggestion in a concise and easy-to-understand form by simplifying the form of the displayed content. The diagnostic report can be displayed through the system interface to show the analysis result of patient data, the final diagnosis, the recommended treatment plan, and the next suggestion. At the same time, the patient data and diagnosis result will be stored in the database for subsequent tracking and review.

[0080] Please refer to Figure 2 , for the application of the multi-intelligent-agent medical diagnosis and treatment system of the present application, it can be realized through the following steps:

[0081] S1. Analyze user input: The user interacts through the visual interface provided by the system and inputs their needs, including information such as symptom descriptions, medical histories, and imaging data. After receiving the user input, the task planning agent first uses large language models (such as GPT, Qwen, etc.) to parse the input information, understand the user's specific intentions, and based on this, formulate a suitable task process. This process is not only a semantic analysis of the user input, but also a reasonable derivation according to the diagnostic logic, so as to ensure that the system can accurately identify the user's needs. The task planning agent is also responsible for coordinating the working order and interaction logic of each agent to ensure that each step can be executed smoothly, avoiding unnecessary delays or duplicate tasks. If any doubts or uncertainties occur during the analysis process, the task planning agent will request further input or confirmation from the user to ensure the accuracy of the task process.

[0082] S2. Obtain diagnostic data: After determining the task process, the data management agent starts to work to obtain relevant data of the patient. These data may include electronic medical record information (such as patient chief complaints, medical histories, family histories, etc.), imaging data (such as MRI, CT scans, etc.), laboratory test results (such as blood test data, HLA - B27 test results), etc. The data management agent automatically obtains this data by querying the hospital database and sends it back to the task planning agent. If there are problems during the data acquisition process (such as some data not being correctly extracted), the data management agent will re - send a query request to the hospital database until the data is completely acquired or the maximum number of query attempts is reached. During the whole process, the task planning agent will track the progress of data acquisition in real - time to ensure the correct execution of the task.

[0083] S3. Analyze patient data: After receiving the patient data, the tool - using agent first performs preliminary pre - processing on the data. This may include operations such as data cleaning, missing value filling (filling with mean or mode), standardization (transforming features into a normal distribution), image enhancement (histogram equalization), normalization, Resize, etc. Then, the tool - using agent calls the corresponding deep - learning model to perform feature extraction and analysis of the imaging data. For example, based on the segmentation network (Unet), extract the sacroiliac joint area in the MRI image and perform SPARCC scoring and SSS scoring on this area, or perform statistical analysis on the clinical data to identify key features related to axSpA. The tool - using agent will send the analysis results back to the task planning agent, and the task planning agent will determine whether the task is completed. If not, it will regenerate instructions and send them to the tool - using agent until the task is completed or the preset maximum number of iterations is reached.

[0084] S4. Obtain and display the diagnostic report corresponding to the medical decision: The medical decision intelligent body will comprehensively analyze the extracted features, and combine the patient's historical medical records, imaging features, laboratory data, etc. for comprehensive judgment. This intelligent body will first search its long-term memory to find the top 3 records similar to this case. Then, according to the preset gold standard for axSpA diagnosis and the case references found, it will judge whether the patient meets the diagnostic criteria for axSpA. In addition, the medical decision intelligent body will also provide personalized treatment suggestions according to the patient's specific condition. These suggestions may include drug treatment plans, rehabilitation training suggestions, or surgical plans, etc., to assist doctors in making decisions. In addition, the medical decision intelligent body will simulate the process of a real doctor's progress in diagnosis, store each diagnosed case in its own long-term memory library for reference in future similar cases and to improve the diagnostic quality. The medical decision intelligent body generates a diagnostic report based on the specific condition and treatment suggestions, and presents the diagnostic results and treatment suggestions to the user in a concise and easy-to-understand form by simplifying the form of the displayed content.

[0085] The device embodiments described above are merely illustrative. The components described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this application. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0086] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0087] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, and the combination of processes and / or blocks in the flowchart and / or block diagram can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate for realizing in the process Figure 1 one process or multiple processes and / or blocks Figure 1means for selected functions in one or more blocks. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured article including an instruction means, and the instruction means implements in the process Figure 1 one process or more processes and / or blocks Figure 1 selected functions in one block or more blocks.

[0088] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide for implementing in the process Figure 1 one process or more processes and / or blocks Figure 1 steps of selected functions in one block or more blocks.

[0089] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0090] The memory may include non-permanent memory in the computer-readable medium, in the form of random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable medium.

[0091] Computer-readable media include permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information accessible by a computing device. As defined herein, computer-readable media do not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0092] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, commodity or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, commodity or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, commodity or device comprising the element.

[0093] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A multi-agent medical diagnosis and treatment system, characterized in that: The method comprises a plurality of intelligent agents, wherein the plurality of intelligent agents comprises a task planning intelligent agent, a data management intelligent agent, a tool use intelligent agent and a medical decision-making intelligent agent; The task planning agent analyzes the input content and generates a corresponding task flow, and the task planning agent issues task instructions to the data management agent, the tool use agent and the medical decision agent in sequence according to the task flow; the task instructions include data task instructions, tool task instructions and decision task instructions; The data management agent extracts patient data from the database and sends it to the task planning agent according to the data task instruction issued by the task planning agent; The tool-using agent performs feature extraction processing on the patient data according to the patient data and tool task instructions issued by the task-planning agent to obtain key features and send them to the task-planning agent; The medical decision-making agent performs diagnosis and treatment reasoning according to the key features and decision-making task instructions issued by the task planning agent, generates and displays a diagnosis report.

2. The multi-agent medical diagnosis and treatment system according to claim 1, characterized in that: The task planning agent analyzes the input content and generates a corresponding task process, including: The task planning agent integrates user input and information of each agent according to a preset text template to obtain input content; The task planning agent analyzes the input content through a large language model to obtain the task process.

3. The multi-agent medical diagnosis and treatment system according to claim 2, characterized in that: The task planning agent analyzes the input content through a large language model to obtain the steps of the task flow, including: The task planning agent analyzes the input content through the large language model to obtain the text task flow output by the large language model; The task planning agent converts the text task flow into a list task flow, and sends task instructions to the corresponding agent for execution according to the list task flow.

4. The multi-agent medical diagnosis and treatment system according to claim 1, characterized in that: The data management agent extracts patient data from a database and sends the patient data to the task planning agent according to the data task instruction issued by the task planning agent, including: The data management agent generates a data query code according to the data task instruction; The data management agent uses the data query code to query the corresponding patient data from the database.

5. The multi-agent medical diagnosis and treatment system according to claim 1, characterized in that: The patient data includes the patient's electronic medical record information, imaging data, project examination data and clinical data.

6. The multi-agent medical diagnosis and treatment system according to claim 5, characterized in that: The tool using agent performs feature extraction processing on the patient data according to the patient data and the tool task instruction issued by the task planning agent to obtain key features and send them to the task planning agent, including: The tool uses an agent to preprocess the patient data sent by the task planning agent to obtain the preprocessed patient data; The tool using agent performs feature extraction processing on the pre-processed patient data according to the tool task instruction to obtain the key features.

7. The multi-agent medical diagnosis and treatment system according to claim 6, characterized in that: The patient data includes image data; the tool uses an agent to preprocess the patient data sent by the task planning agent to obtain the preprocessed patient data, including: The tool uses an intelligent agent to perform image enhancement, normalization and resize processing on the patient data to obtain the pre-processed patient data.

8. The multi-agent medical diagnosis and treatment system according to claim 7, characterized in that: The tool using agent performs feature extraction processing on the pre-processed patient data according to the tool task instruction to obtain the key features, including: The tool using agent extracts the target area corresponding to the tool task instruction from the image data through a segmentation network, and then extracts features of the target area according to the scoring model corresponding to the tool task instruction to obtain the key features; Or the tool uses an intelligent agent to perform statistical analysis on the clinical data to identify the key features associated with axSpA.

9. The multi-agent medical diagnosis and treatment system according to claim 8, characterized in that: After extracting features of the target area according to the scoring model corresponding to the tool task instruction to obtain the key features, the method includes: The tool uses an intelligent agent to perform SPARCC scoring and SSS scoring on the key features to obtain the key features including the scores.

10. The multi-agent medical diagnosis and treatment system according to claim 1, characterized in that: The medical decision agent performs diagnosis and treatment reasoning according to the key features and decision task instructions issued by the task planning agent, generates and displays a diagnosis report, including: The medical decision-making agent searches for a number of most similar case records according to the key features; The medical decision-making agent performs diagnosis and treatment reasoning based on the key features, the preset axSpA diagnostic criteria and the several most similar case records to obtain the diagnosis report.

Images