Intelligent decision-making system for clinical diagnosis decision-making analysis

Abstract

The invention is suitable for the technical field of intelligent medical treatment, and provides an intelligent decision-making system for clinical diagnosis decision analysis, which comprises a data acquisition module, a knowledge graph module, a multi-modal analysis engine, a clinical evaluation module, a clinical treatment scheme decision-making module and a feedback optimization module, the data acquisition module is used for integrating multi-source heterogeneous clinical data; the knowledge graph module is used for dynamically storing a medical entity relationship; the multi-modal analysis engine extracts features through deep learning; the clinical treatment scheme decision module is used for outputting diagnosis suggestions and providing explanations; the feedback optimization module is used for receiving a doctor correction result, automatically iterating a training model and forming closed-loop learning; according to the system, real-time association medical new discovery is supported through a dynamic knowledge graph, and hospital information system data are not effectively integrated; through cooperative work of the rule engine and the AI model, clinical medical quality and efficiency are improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical technology, and more specifically, to an intelligent decision-making system for clinical diagnosis and decision analysis. Background Art

[0002] Currently, people have an increasing demand for clinical medical resources. At the same time, human resources such as experts, doctors, nurses and other medical resources in the domestic medical field are still relatively scarce, and the quality and efficiency of clinical medicine still need to be improved.

[0003] Traditional clinical decision-making relies on physician experience and is susceptible to subjective factors, resulting in low diagnostic efficiency for complex cases. Existing CDSS systems are mostly rule-based, inflexible, and difficult to process multi-source heterogeneous data (such as imaging, genetics, and electronic medical records). They also lack dynamic learning capabilities and are unable to automatically optimize models as medical knowledge is updated. This presents the following problems:

[0004] Data silos and ineffective integration of hospital information systems (HIS, PACS, etc.) Algorithm limitations and low diagnostic accuracy of single models (such as logistic regression and decision trees) for rare diseases. Summary of the Invention

[0005] In response to the shortcomings of the existing technology, the purpose of the present invention is to provide an intelligent decision-making system for clinical diagnostic decision analysis. Through dynamic knowledge graphs, it supports real-time association of new medical discoveries (such as new pathogenic genes) and effectively integrates data from hospital information systems (HIS, PACS, etc.); through the collaborative work of rule engines (to ensure diagnosis and treatment standards) and AI models (to handle uncertainty), it improves the quality and efficiency of clinical medical care; through federated learning, it protects patient privacy and locally updates models; and through multimodal data fusion, it improves the recognition rate of rare diseases and improves the accuracy of rare disease diagnosis.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] An intelligent decision-making system for clinical diagnosis decision analysis includes a data acquisition module, a knowledge graph module, a multimodal analysis engine, a clinical evaluation module, a clinical treatment plan decision module, and a feedback optimization module; the data acquisition module is used to integrate multi-source heterogeneous clinical data; the knowledge graph module is used to dynamically store medical entity relationships; the multimodal analysis engine extracts features through deep learning; the clinical evaluation module takes the clinical manifestations of new patients as input, executes a rule reasoning mechanism to output the evaluation results of new patients; the clinical treatment plan decision module is used to output diagnostic suggestions and provide explanations; the feedback optimization module is used to receive doctor-corrected results, automatically iterate the training model, and form a closed-loop learning.

[0008] The present invention is further configured as follows: the data acquisition module is connected to the hospital HIS / LIS / PACS system, integrates structured data and unstructured data, and supports natural language processing to extract key clinical features.

[0009] The present invention is further configured such that: the data acquisition module has a built-in data cleaning unit for processing missing values, abnormal values and standardized terms.

[0010] The present invention is further configured as follows: the knowledge graph module constructs a disease-symptom-drug-gene association graph based on medical literature, guidelines, and expert consensus, and supports dynamic updates.

[0011] The present invention is further configured as follows: the knowledge graph module uses a graph neural network (GNN) to mine potential associations; the knowledge graph module uses a graph database Neo4j for storage and supports incremental updates.

[0012] The present invention is further configured as follows: the multimodal analysis engine adopts a dual-channel deep learning model, integrates deep reinforcement learning and Bayesian networks, balances exploration and utilization, combines NLP, CNN, and time series models, outputs a unified feature vector, and visualizes the decision basis through SHAP value and attention mechanism, meeting clinical audit requirements; the multimodal analysis engine integrates a reinforcement learning algorithm and optimizes the diagnostic path through a reward function.

[0013] The present invention is further configured as follows: the clinical treatment plan decision module utilizes the clinical diagnosis and treatment guidelines and case literature report knowledge in the clinical diagnosis and treatment knowledge base module, combined with the evaluation results output by the clinical evaluation module, to analyze the similarity between the clinical manifestations and clinical evaluations of new patients and historical cases, finds the diagnosis and treatment case most similar to the new patient, and provides reference for the treatment of the new patient based on the treatment plan of the most similar diagnosis and treatment case.

[0014] The present invention is further configured as follows: the clinical treatment plan decision module is used to assist manual decision-making in the treatment plan decision-making stage, and the clinical manifestations and clinical evaluations of the new patient are described as a new problem. Through similarity analysis between the new problem and existing cases, the diagnosis and treatment case most similar to the clinical manifestations of the new patient is found, and based on the treatment plan of the similar diagnosis and treatment case, a treatment plan is provided for the new patient through an auxiliary decision-making mechanism.

[0015] The advantages of the present invention are:

[0016] 1. This invention uses a dynamic knowledge graph to support real-time association of new medical discoveries (such as new pathogenic genes) and effectively integrate data from hospital information systems (HIS, PACS, etc.); it improves the quality and efficiency of clinical care through the collaboration of a rule engine (to ensure diagnosis and treatment standards) and an AI model (to handle uncertainty); and it protects patient privacy through federated learning and locally updates the model.

[0017] 2. The present invention improves the recognition rate of rare diseases through multimodal data fusion, thereby improving the accuracy of rare disease diagnosis.

[0018] 3. Through the setting of various modules, the present invention realizes the sorting, integration and standardized expression of domain knowledge and experience, realizes clinical evaluation based on rule reasoning to improve the accuracy and efficiency of evaluation, and realizes clinical treatment plan decision-making based on case reasoning to combine experiential knowledge with the efficiency of intelligent decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a framework diagram of the intelligent decision-making system for clinical diagnosis decision analysis of the present invention. DETAILED DESCRIPTION

[0020] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0021] It should be noted that, unless otherwise specified, all technical and scientific terms used in this application have the same meaning as commonly understood by ordinary technicians in the technical field to which this application belongs.

[0022] In the present invention, unless otherwise specified, directions such as "up" and "down" are generally used with respect to the directions shown in the drawings, or with respect to the vertical, perpendicular or gravity directions; similarly, for ease of understanding and description, "left" and "right" are generally used with respect to the left and right shown in the drawings; "inside" and "outside" refer to the inside and outside relative to the outline of each component itself, but the above-mentioned directions are not used to limit the present invention.

[0023] For example 1, please refer to Figure 1 , the present invention provides the following technical solutions:

[0024] The intelligent decision-making system for clinical diagnosis decision analysis specifically includes a data acquisition module, a knowledge graph module, a multimodal analysis engine, a clinical evaluation module, a clinical treatment plan decision module, and a feedback optimization module; the data acquisition module is used to integrate multi-source heterogeneous clinical data; the knowledge graph module is used to dynamically store medical entity relationships; the multimodal analysis engine extracts features through deep learning; the clinical evaluation module takes the clinical manifestations of new patients as input, executes a rule reasoning mechanism to output the evaluation results of new patients; the clinical treatment plan decision module is used to output diagnostic suggestions and provide explanations; the feedback optimization module is used to receive the doctor's correction results, automatically iterate the training model, and form a closed-loop learning.

[0025] Working principle of this embodiment 1:

[0026] Through dynamic knowledge graphs, it supports real-time association of new medical discoveries (such as new pathogenic genes) and effectively integrates data from hospital information systems (HIS, PACS, etc.); through the collaboration of rule engines (to ensure diagnosis and treatment standards) and AI models (to handle uncertainty), it improves the quality and efficiency of clinical medical care; through federated learning, it protects patient privacy and updates models locally; and through multimodal data fusion, it improves the recognition rate of rare diseases and improves the accuracy of rare disease diagnosis.

[0027] For example 2, please refer to Figure 1 This second embodiment makes the following improvements based on the first embodiment. Specifically, the data acquisition module is connected to the hospital HIS / LIS / PACS system, integrates structured data and unstructured data, and supports natural language processing to extract key clinical features; the data acquisition module has a built-in data cleaning unit for processing missing values, outliers and standardized terminology.

[0028] The knowledge graph module builds a disease-symptom-drug-gene association map based on medical literature, guidelines, and expert consensus, and supports dynamic updates; the knowledge graph module uses graph neural networks (GNNs) to mine potential associations; the knowledge graph module uses the graph database Neo4j for storage and supports incremental updates.

[0029] The multimodal analysis engine adopts a dual-channel deep learning model, integrates deep reinforcement learning and Bayesian networks, balances exploration and utilization, combines NLP, CNN, and time series models, outputs a unified feature vector, and visualizes the decision basis through SHAP values and attention mechanisms, meeting clinical audit requirements; the multimodal analysis engine integrates reinforcement learning algorithms and optimizes the diagnostic path through reward functions.

[0030] The clinical treatment plan decision module uses the clinical diagnosis and treatment guidelines and case literature report knowledge in the clinical diagnosis and treatment knowledge base module, combined with the evaluation results output by the clinical evaluation module, to analyze the similarity between the clinical manifestations and clinical evaluations of new patients and historical cases, find the diagnosis and treatment case most similar to the new patient, and provide reference for the treatment of the new patient based on the treatment plan of the most similar diagnosis and treatment case.

[0031] Working principle of the second embodiment:

[0032] HIS system (Hospital Information System)

[0033] Function:

[0034] Core management: outpatient / inpatient registration, billing, pharmacy management, electronic medical records (EMR), medical order processing, financial statistics, etc.

[0035] Coverage: Hospital-wide processes, serving as the “central nervous system” of hospital operations.

[0036] Technical features:

[0037] Adopt C / S or B / S architecture and mainstream database (Oracle, SQL Server).

[0038] Integration standards: Supports medical data protocols such as HL7 and DICOM.

[0039] Application scenarios:

[0040] Doctors issue electronic medical orders, nurses execute and record them, and financial settlements are made simultaneously.

[0041] LIS system (Laboratory Information System)

[0042] Function:

[0043] Automation of the inspection process: sample receipt, inspection equipment connection, result review, and report issuance.

[0044] Data management: Store biochemical, immunological, microbiological and other test data to support quality control analysis.

[0045] Technical features:

[0046] It can be directly connected to testing equipment (such as fully automatic biochemical analyzer) via serial port or TCP / IP, and supports ASTM protocol.

[0047] The results are automatically transmitted to HIS, reducing manual entry errors.

[0048] Application scenarios:

[0049] After the patient has blood drawn, the test results are automatically uploaded and the doctor can view them in real time in the HIS.

[0050] PACS system (Picture Archiving and Communication System)

[0051] Function:

[0052] Image management: Store CT, MRI, X-ray and other images in DICOM format, and support access, annotation, and 3D reconstruction.

[0053] Workflow optimization: Achieve "film-free" and shorten diagnosis waiting time.

[0054] Technical features:

[0055] Based on the DICOM 3.0 standard, it supports integration with RIS (Radiology Information System).

[0056] High compression ratio storage technology (such as JPEG 2000) saves server space.

[0057] Application scenarios:

[0058] Radiologists read the films in PACS, generate structured reports and send them back to HIS.

[0059] Neo4j Core Features

[0060] Native graph storage: Data is directly stored in the form of "node-relationship-property", avoiding the multi-table join overhead of relational databases.

[0061] Cypher query language: A declarative syntax designed specifically for graph data, supporting complex relational path queries (e.g., 3-hop join queries are 1,000 times more efficient than SQL).

[0062] Real-time relationship analysis: excels at processing deeply correlated data (such as disease-gene-drug interaction networks).

[0063] Natural language processing (NLP) is a branch of artificial intelligence that focuses on the interaction between computers and human language. CNN is a deep learning model specifically designed to process grid-like data (such as images). Time series models are specifically designed to process time series data or ordered sequence data. CNN can process local time series patterns (through 1D convolution). NLP models (especially Transformers) have been successfully applied to general time series problems. The modern trend is the fusion of these technologies, such as CNN-LSTM hybrid models or spatiotemporal Transformers.

[0064] The rule-based reasoning clinical assessment module is used by doctors during the clinical evaluation phase of new patients. Taking the patient's clinical presentation as input, the rule-based reasoning mechanism runs, drawing relevant diagnostic conclusions based on generative rule reasoning. When multiple rules are satisfied simultaneously, the higher-priority rule is executed first, preventing conflicts caused by simultaneous execution of multiple rules. Keyword-based search is also used within the standard library to facilitate access to relevant standards.

[0065] Through dynamic knowledge graphs, it supports real-time association of new medical discoveries (such as new pathogenic genes) and effectively integrates data from hospital information systems (HIS, PACS, etc.); through the collaboration of rule engines (to ensure diagnosis and treatment standards) and AI models (to handle uncertainty), it improves the quality and efficiency of clinical medical care; through federated learning, it protects patient privacy and updates models locally; through multimodal data fusion, it improves the recognition rate of rare diseases and the accuracy of rare disease diagnosis; through the setting of various modules, it realizes the sorting, integration and standardized expression of domain knowledge and experience, realizes clinical evaluation based on rule reasoning to improve the accuracy and efficiency of evaluation, and realizes clinical treatment plan decision-making based on case reasoning to combine experience knowledge with the efficiency of intelligent decision-making.

[0066] Obviously, the embodiments described above are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0067] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, tasks, devices, components and / or combinations thereof.

[0068] It should be noted that the terms "first," "second," and the like in the specification and claims of this application and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0069] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

[0070] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiment. All technical solutions based on the concept of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. An intelligent decision-making system for clinical diagnosis and decision analysis, characterized by: It includes data acquisition module, knowledge graph module, multimodal analysis engine, clinical evaluation module, clinical treatment plan decision module and feedback optimization module; The data acquisition module is used to integrate multi-source heterogeneous clinical data; The knowledge graph module is used to dynamically store medical entity relationships; The multimodal analysis engine extracts features through deep learning; The clinical assessment module takes the clinical manifestations of the new patient as input, executes the rule reasoning mechanism and outputs the assessment results of the new patient; The clinical treatment plan decision module is used to output diagnostic suggestions and provide explanations; The feedback optimization module is used to receive the doctor's correction results, automatically iterate the training model, and form a closed-loop learning.

2. The intelligent decision-making system for clinical diagnosis and decision analysis according to claim 1, characterized in that: The data acquisition module is connected to the hospital HIS / LIS / PACS system, integrates structured data and unstructured data, and supports natural language processing to extract key clinical features.

3. The intelligent decision-making system for clinical diagnosis and decision analysis according to claim 2, characterized in that: The data collection module has a built-in data cleaning unit for processing missing values, outliers and standardized terms.

4. The intelligent decision-making system for clinical diagnosis and decision analysis according to claim 3, characterized in that: The knowledge graph module constructs a disease-symptom-drug-gene association map based on medical literature, guidelines, and expert consensus, and supports dynamic updates.

5. The intelligent decision-making system for clinical diagnosis and decision analysis according to claim 4, characterized in that: The knowledge graph module uses a graph neural network (GNN) to mine potential associations; the knowledge graph module uses the graph database Neo4j for storage and supports incremental updates.

6. The intelligent decision-making system for clinical diagnosis and decision analysis according to claim 5, characterized in that: The multimodal analysis engine adopts a dual-channel deep learning model, integrates deep reinforcement learning and Bayesian networks, balances exploration and utilization, combines NLP, CNN, and time series models, outputs a unified feature vector, and visualizes the decision basis through SHAP value and attention mechanism, meeting clinical audit requirements; the multimodal analysis engine integrates a reinforcement learning algorithm and optimizes the diagnostic path through a reward function.

7. The intelligent decision-making system for clinical diagnosis and decision analysis according to claim 6, characterized in that: The clinical treatment plan decision module uses the clinical diagnosis and treatment guidelines and case literature report knowledge in the clinical diagnosis and treatment knowledge base module, combined with the evaluation results output by the clinical evaluation module, to analyze the similarity between the clinical manifestations and clinical evaluations of new patients and historical cases, find the diagnosis and treatment case most similar to the new patient, and provide reference for the treatment of the new patient based on the treatment plan of the most similar diagnosis and treatment case.

8. The intelligent decision-making system for clinical diagnosis and decision analysis according to claim 7, characterized in that: The clinical treatment plan decision module is used to assist manual decision-making in the treatment plan decision-making stage. The clinical manifestations and clinical evaluations of the new patient are described as a new problem. Through similarity analysis between the new problem and existing cases, the diagnosis and treatment case most similar to the clinical manifestations of the new patient is found. Based on the treatment plan of the similar diagnosis and treatment case, a treatment plan is provided to the new patient through the auxiliary decision-making mechanism.

Images