Intelligent management method and system for clinical data of medical examination

Abstract

The invention relates to a medical examination clinical data intelligent management method and system, and belongs to the technical field of medical examination data management.The system comprises a data acquisition module, a dynamic graph analysis module and construction data, and the construction data comprises examination data, unstructured texts, time sequence data and an external knowledge base; data acquired by the data acquisition module comprises inspection equipment output data, case report text data, wearable equipment output data and digital pathological section data, and the dynamic map analysis module is constructed based on a ClinicalBERT model according to input construction data. According to the invention, multi-modal diagnosis can be carried out on patient information based on the constructed dynamic atlas analysis module according to actual clinical data, rapid diagnosis can be carried out on diseases more comprehensively through a disease diagnosis mode driven by cross-modal comparative learning, and probabilities of different causes can be displayed and analyzed according to multi-modal cause reasoning. Therefore, medical staff can be effectively assisted in disease determination.

Description

Technical Field

[0001] The present invention relates to the technical field of medical test data management, and in particular to a method and system for intelligent management of medical test clinical data. Background Art

[0002] Medical laboratory clinical data plays a vital role in the diagnosis and treatment of patients. Through intelligent management of clinical data, clinical data can be better utilized to provide patients with personalized medical services.

[0003] Existing medical laboratory clinical data management systems have the following limitations: 1. Traditional medical laboratory clinical data management systems rely on fixed rules for data analysis and cannot dynamically adapt to changes in individualized patient data (such as fluctuations in chronic disease indicators), resulting in insufficient static analysis capabilities; 2. Test data, imaging data, and genomic data are scattered across different systems and lack deep associations (such as the inability to associate elevated tumor markers with CT imaging features); 3. Treatment recommendations are simplistic, and existing diagnostic systems mostly generate recommendations based on single-modality data, ignoring the synergistic effects of multi-dimensional evidence.

[0004] To facilitate the intelligent statistics, cross-modal comparative analysis, multimodal diagnosis of diseases, and personalized treatment recommendations of clinical data, a method and system for intelligent management of medical laboratory clinical data is proposed. Based on the constructed dynamic graph analysis module, multimodal diagnosis of patient information can be performed according to actual clinical data. The disease diagnosis model driven by cross-modal comparative learning can quickly diagnose diseases more comprehensively. Based on multimodal etiological reasoning, the probabilities of different causes can be displayed and analyzed, which can effectively assist medical staff in disease identification and provide treatment plans based on past treatment plans, thereby further improving the work efficiency of medical staff. By monitoring the treatment effect, the treatment plan can be replaced and optimized in a timely manner, thereby improving the accuracy of multimodal diagnosis and the adoption rate of clinicians. Summary of the Invention

[0005] The present invention provides a method and system for intelligent management of medical laboratory clinical data, which solves the problems raised by the above-mentioned existing background technologies, realizes intelligent statistics of clinical data, cross-modal comparative analysis, multimodal diagnosis of diseases and personalized treatment recommendations, and is suitable for hospitals, regional medical platforms and remote diagnosis and treatment scenarios.

[0006] The present invention solves the above-mentioned technical problems with the following solution: a method and system for intelligent management of clinical data of medical tests, comprising a data acquisition module and a dynamic graph analysis module, and constructed data, wherein the constructed data comprises test data, unstructured text, time series data, and an external knowledge base; the data collected by the data acquisition module comprises test equipment output data, case report text data, wearable device output data, and digital pathology slide data; the dynamic graph analysis module is constructed based on the ClinicalBERT model according to the input constructed data;

[0007] The management method includes the following steps:

[0008] S1: Knowledge graph construction: Test data, unstructured text, time series data, and external knowledge base are used as construction data input into the ClinicalBERT model to build a dynamic graph analysis module. The ClinicalBERT model extracts entity test indicators, disease information, treatment plans, and imaging features from medical records, defines the association between test indicators and diseases, such as "elevated CRP → inflammatory response", and establishes drug-test indicator relationships, such as "statins → may cause elevated creatine kinase", thereby building a dynamic graph analysis module;

[0009] S2: Clinical data acquisition: The output data of the testing equipment, case report text data, wearable device output data, and digital pathology slide data are pre-processed by the data acquisition module and then input into the dynamic graph analysis module;

[0010] S3: Dynamic Reasoning and Etiology Analysis: 1. Real-time Anomaly Detection: The streaming processing engine Flink monitors clinical data in real time, triggering anomaly thresholds to issue early warning reports, such as blood creatinine > 442 μmol / L → risk of acute kidney injury. 2. Multimodal Etiology Reasoning: Graph path reasoning is performed. The symptom "fever + lymphocytopenia" is input. The dynamic graph analysis module analyzes related diseases, such as viral infections and blood diseases, and combines test results, such as HIV antibody positivity, to narrow the scope and generate suspected causes, such as HIV infection combined with opportunistic infections. Probabilistic reasoning is performed on suspected causes, using a Bayesian network to evaluate the probabilities of different causes, such as P(leukemia | abnormal white blood cells + positive bone marrow biopsy) = 85%. Uncertainty handling: If genetic testing shows a BRCA1 mutation, the breast cancer risk assessment weight is increased. 3. Conflict Resolution and Interpretation: Multi-source data consistency verification is performed. When test results conflict with imaging reports, such as elevated tumor markers but no mass on CT scan, the conflict is flagged and a reexamination prompt is prompted. High-confidence conclusions, such as the cancer staging standards in the NCCN guidelines, are prioritized based on authoritative guidelines.

[0011] S4: Personalized treatment recommendations: Determine the disease type and output the prediction accuracy based on abnormal test data and multimodal causal reasoning results, provide treatment plans, sort by efficacy-side effect balance, monitor the treatment effect, and dynamically track test data such as liver function ALT / AST. If abnormally elevated, trigger a recommendation to change the treatment plan, and warn of drug interactions such as abnormal INR caused by the combination of warfarin and antibiotics.

[0012] On the basis of the above technical solution, the present invention can also be improved as follows.

[0013] Furthermore, in S1: the test data includes structured data such as blood routine, biochemical indicators, and microbial culture results; the unstructured text includes doctor's medical record descriptions, scientific research literature, and clinical guidelines; the time series data includes dynamic changes in patients' multiple test results such as blood sugar and tumor marker trends; the external knowledge base includes medical ontologies SNOMEDCT, UMLS, drug knowledge base DrugBank, and gene-disease association OMIM.

[0014] Furthermore, in S2: the data preprocessing includes outlier correction and missing value filling. Based on the patient's historical data and population distribution, if the Z-score>3 is considered abnormal, GAN is used to generate reasonable alternative values, the data acquisition module is corrected for outliers and marked as corrected data for reference only, and the missing values of the data are filled and supplemented through multimodal cross-validation and marked as filled data for reference only.

[0015] Furthermore, in said S4: based on the patient's genotype, such as the CYP2C19 metabolizer type, contraindicated drugs are excluded, for example, clopidogrel is ineffective for slow metabolizers, and the dosage is adjusted in combination with the test indicators, for example, the dosage of vancomycin needs to be reduced in patients with renal insufficiency.

[0016] Furthermore, in said S4: dynamic efficacy prediction of the treatment regimen is performed, the treatment regimen such as the chemotherapy regimen FOLFOX is input, and the engine simulates the effect of the drug on the test index such as predicting the risk of neutropenia.

[0017] The beneficial effects of the present invention are as follows: the present invention provides a method and system for intelligent management of medical test clinical data, which has the following advantages:

[0018] 1. By integrating test equipment output data, case report text data, wearable device output data, and digital pathology slide data, a patient-centric dynamic knowledge graph is constructed, overcoming the limitations of traditional systems that rely on a single data source;

[0019] 2. A causal engine based on reinforcement learning and Bayesian modeling can identify potential causal relationships between diseases and test indicators and provide treatment plans with evidence level annotations, avoiding decision-making bias driven by a single rule;

[0020] 3. Based on the constructed dynamic graph analysis module, multimodal diagnosis of patient information can be performed according to actual clinical data. The disease diagnosis model driven by cross-modal comparative learning can quickly diagnose the disease more comprehensively. The probabilities of different causes can be displayed and analyzed based on multimodal etiology reasoning, which can effectively assist medical staff in determining the disease and provide treatment plans based on past treatment plans, thereby further improving the work efficiency of medical staff. By monitoring the treatment effect, the treatment plan can be replaced and optimized in time, thereby improving the accuracy of multimodal diagnosis and the adoption rate of clinicians. It is suitable for hospitals, regional medical platforms and remote diagnosis and treatment scenarios.

[0021] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and to implement it according to the contents of the description, the following preferred embodiments of the present invention are described in detail with reference to the accompanying drawings. The specific implementation methods of the present invention are given in detail by the following embodiments and the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0023] Figure 1 This is a system flow chart of a medical test clinical data intelligent management method and system provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0024] The following is combined with Figure 1 The principles and features of the present invention are described, and the examples given are only for the purpose of explaining the present invention and are not intended to limit the scope of the present invention. The following paragraphs describe the present invention in more detail by way of example with reference to the accompanying drawings. The advantages and features of the present invention will become more apparent from the following description and claims. It should be noted that the drawings are in a very simplified form and are not in exact proportions, and are only used for the purpose of conveniently and clearly assisting in illustrating the embodiments of the present invention.

[0025] It should be noted that when a component is referred to as being "fixed to" another component, it may be directly on the other component or there may also be a central component. When a component is considered to be "connected to" another component, it may be directly connected to the other component or there may also be a central component. When a component is considered to be "set on" another component, it may be directly set on the other component or there may also be a central component. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0026] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. The terms used in this specification of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0027] like Figure 1 As shown, the present invention provides a method and system for intelligent management of medical test clinical data, including a data acquisition module, a dynamic graph analysis module, and construction data, characterized in that: the construction data includes test data, unstructured text, time series data, and an external knowledge base; the data collected by the data acquisition module includes test equipment output data, case report text data, wearable device output data, and digital pathology slice data; the dynamic graph analysis module is constructed based on the ClinicalBERT model according to the input construction data.

[0028] The specific working principle and method of use of the present invention are as follows:

[0029] S1: Knowledge graph construction: Test data, unstructured text, time series data, and external knowledge bases are used as construction data to input into the ClinicalBERT model to build a dynamic graph analysis module. Test data includes structured data such as blood routine, biochemical indicators, and microbial culture results. Unstructured text includes doctor's medical record descriptions, scientific research literature, and clinical guidelines. Time series data includes the dynamic changes of patients' multiple test results, such as blood sugar and tumor marker trends. External knowledge bases include medical ontologies SNOMED CT, UMLS, drug knowledge base DrugBank, and gene-disease association OMIM. The ClinicalBERT model extracts entity test indicators, disease information, treatment plans, and imaging features from medical records, defines the association between test indicators and diseases, such as "elevated CRP → inflammatory response", and establishes drug-test indicator relationships, such as "statins → may cause elevated creatine kinase", thereby building a dynamic graph analysis module;

[0030] S2: Clinical data collection: The output data of the testing equipment, case report text data, wearable device output data, and digital pathology slide data are preprocessed by the data collection module and then input into the dynamic graph analysis module. The data preprocessing includes outlier correction and missing value filling. Based on the patient's historical data and population distribution, if the Z-score>3 is considered abnormal, GAN is used to generate reasonable replacement values. The data collection module is corrected for outliers and marked as corrected data for reference only. The missing values of the data are filled and supplemented through multimodal cross-validation and marked as filled data for reference only.

[0031] S3: Dynamic Reasoning and Etiology Analysis: 1. Real-time Anomaly Detection: The streaming processing engine Flink monitors clinical data in real time, triggering anomaly thresholds to issue early warning reports, such as blood creatinine > 442 μmol / L → risk of acute kidney injury. 2. Multimodal Etiology Reasoning: Graph path reasoning is performed. The dynamic graph analysis module analyzes related diseases, such as viral infections and blood diseases, based on the input symptoms "fever + lymphocytopenia." It then narrows the scope by combining test results, such as HIV antibody positivity, to generate suspected causes, such as HIV infection combined with opportunistic infections. Probabilistic reasoning is performed on suspected causes, using a Bayesian network to evaluate the probabilities of different causes, such as P(leukemia | abnormal white blood cells + positive bone marrow biopsy) = 85%. Uncertainty handling: If genetic testing reveals a BRCA1 mutation, the breast cancer risk assessment weight is increased. 3. Conflict Resolution and Interpretation: Multi-source data consistency verification is performed. When test results conflict with imaging reports, such as elevated tumor markers but no mass on CT scan, the conflict is flagged and a reexamination prompt is prompted. High-confidence conclusions, such as the cancer staging criteria in the NCCN guidelines, are prioritized based on authoritative guidelines.

[0032] S4: Personalized treatment recommendations: Based on abnormal test data and multimodal etiological reasoning results, the disease type is determined and the prediction accuracy is output. Treatment plans are provided, ranked by efficacy-side effect balance, and treatment effects are monitored. Contraindicated drugs are excluded based on the patient's genotype, such as CYP2C19 metabolic type. For example, clopidogrel is ineffective for slow metabolizers. The dosage is adjusted based on test indicators, such as the need to reduce the dosage of vancomycin for patients with renal insufficiency. Test data such as liver function ALT / AST are tracked dynamically. If abnormally elevated, a recommendation to change the treatment plan is triggered. Dynamic efficacy prediction of the treatment plan is performed. Treatment plans are input, such as the chemotherapy plan FOLFOX. The engine simulates the impact of drugs on test indicators, such as predicting the risk of neutropenia, and warns of drug interactions, such as abnormal INR caused by the combination of warfarin and antibiotics.

[0033] It should be noted that, in this document, relational terms such as first and second are used solely to distinguish one entity or operation from another, and do not necessarily require or imply any actual relationship or order between these entities or operations. Matters not described in detail in this specification are well known to those skilled in the art.

[0034] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any ordinary technician in this industry can smoothly implement the present invention as shown in the drawings and described above. However, any equivalent changes, modifications and evolutions made by technicians familiar with this profession without departing from the scope of the technical solution of the present invention using the technical content disclosed above are all equivalent embodiments of the present invention. At the same time, any equivalent changes, modifications and evolutions made to the above embodiments based on the essential technology of the present invention are still within the scope of protection of the technical solution of the present invention.

Claims

1. An intelligent management system for medical laboratory clinical data, comprising a data acquisition module and a dynamic graph analysis module, and constructing data, characterized in that: The constructed data includes test data, unstructured text, time series data, and external knowledge bases. The collected data of the data acquisition module includes test equipment output data, case report text data, wearable device output data, and digital pathology slide data. The dynamic graph analysis module is constructed based on the ClinicalBERT model according to the input constructed data. The management method includes the following steps: S1: Knowledge graph construction: Test data, unstructured text, time series data, and external knowledge base are used as construction data input into the ClinicalBERT model to build a dynamic graph analysis module. The ClinicalBERT model extracts the test indicators, disease information, treatment plans, and imaging features of entities from medical records, defines the association between test indicators and diseases, and establishes the drug-test indicator relationship, thereby building a dynamic graph analysis module; S2: Clinical data acquisition: The output data of the testing equipment, case report text data, wearable device output data, and digital pathology slide data are pre-processed by the data acquisition module and then input into the dynamic graph analysis module; S3: Dynamic Reasoning and Cause Analysis:

1. Real-time anomaly detection: The streaming processing engine Flink monitors clinical data in real time and triggers anomaly thresholds for early warning reports; 2. Multimodal etiology reasoning: Perform graph path reasoning. Input symptoms. The dynamic graph analysis module analyzes related diseases, combines test results to narrow the scope, generates suspected causes, and performs probabilistic reasoning on the suspected causes. Use Bayesian networks to evaluate the probabilities of different causes and handle uncertainty.

3. Conflict resolution and interpretation: Perform multi-source data consistency verification. When the test results conflict with the imaging report, mark the conflict and prompt for re-examination. Based on authoritative guidelines, prioritize high-confidence conclusions. S4: Personalized treatment recommendations: Determine the disease type and output the prediction accuracy based on abnormal test data and multimodal causal reasoning results, provide treatment plans, sort by efficacy-side effect balance, monitor the treatment effect, and dynamically track test data. If abnormalities increase, trigger a recommendation to change the treatment plan and warn of drug interactions.

2. A medical test clinical data intelligent management method according to claim 1, characterized in that: In S1: the test data includes structured data such as blood routine, biochemical indicators, and microbial culture results; the unstructured text includes doctor's medical record descriptions, scientific research literature, and clinical guidelines; the time series data includes the dynamic changes of patients' multiple test results; and the external knowledge base includes the medical ontology SNOMED CT, UMLS, the drug knowledge base DrugBank, and the gene-disease association OMIM.

3. The method for intelligent management of medical laboratory clinical data according to claim 1, characterized in that: In S2: the data preprocessing includes outlier correction and missing value filling. Based on the patient's historical data and population distribution, GAN is used to generate reasonable alternative values. The data acquisition module is corrected for outliers and marked as corrected data for reference only. The missing values of the data are filled and supplemented through multimodal cross-validation and marked as filled data for reference only.

4. A medical test clinical data intelligent management method according to claim 1, characterized in that: In said S4: contraindicated drugs are excluded according to the patient's genotype, and the dosage is adjusted in combination with the test indicators.

5. The method for intelligent management of medical laboratory clinical data according to claim 1, characterized in that: In said S4: dynamic efficacy prediction of the treatment plan is performed, the treatment plan is input, and the engine simulates the influence of the drug on the test index.

Images