Abnormity recognition system for trend characteristics of chronic disease monitoring data

Abstract

The invention relates to the technical field of medical information, in particular to an anomaly recognition system for trend characteristics of chronic disease monitoring data. The system comprises a chronic disease monitoring data integration module, a chronic disease feature change analysis module, a chronic disease composite feature analysis module and a chronic disease trend feature anomaly recognition module, and can perform chronic disease classification and associated chronic disease integration according to real-time chronic disease monitoring data. Generating integrated associated chronic disease data; performing tissue structure mapping reconstruction of the chronic disease patient and disease mutation characteristic factor analysis of the chronic disease patient according to the integrated and associated chronic disease data, and generating disease mutation characteristic factor data of the chronic disease patient; according to the disease mutation characteristic factor data of the chronic disease patient, performing disease composite characteristic analysis on the chronic disease patient, and performing chronic disease trend abnormal characteristic analysis to generate chronic disease trend abnormal characteristic data. According to the invention, accurate identification of chronic disease trend abnormity is realized.

Description

Technical Field

[0001] The present invention relates to the field of medical information technology, and in particular to an abnormality recognition system for trend characteristics of chronic disease monitoring data. Background Art

[0002] With the aging of the population and the continued rise in the incidence of chronic diseases, chronic diseases have become one of the main factors affecting public health. Common chronic diseases include cardiovascular disease, diabetes, and chronic respiratory diseases. They have a long course, slow progression, and are prone to recurrence. They are often accompanied by complex physiological changes and multi-system interactions, requiring long-term monitoring and intervention management. With the rapid development of digital medicine, with the help of powerful computer data processing capabilities and machine learning algorithms, data trends can be analyzed in real time and accurately, and abnormal changes in data fluctuations can be quickly captured. However, existing chronic disease anomaly recognition technologies have difficulty in capturing key abnormal changes in the course of the disease, especially in the early stages of disease evolution. It is impossible to predict trends for chronic diseases, which can easily lead to delays in preventive interventions for chronic diseases, thereby affecting the effectiveness and timeliness of clinical interventions. Summary of the Invention

[0003] Based on this, the present invention provides an anomaly identification system for trend characteristics of chronic disease monitoring data to solve at least one of the above technical problems.

[0004] To achieve the above objectives, a chronic disease monitoring data trend feature anomaly identification system is proposed, which includes the following modules:

[0005] The chronic disease monitoring data integration module is used to obtain real-time chronic disease monitoring data, classify chronic diseases based on the real-time chronic disease monitoring data, and generate classified chronic disease data; and integrate related chronic diseases based on the classified chronic disease data to generate integrated related chronic disease data;

[0006] A chronic disease characteristic change analysis module is used to reconstruct the tissue structure mapping of chronic disease patients based on the integrated related chronic disease data to generate tissue structure mapping data of chronic disease patients; to analyze the disease mutagenic characteristic factors of chronic disease patients based on the tissue structure mapping data of chronic disease patients to generate disease mutagenic characteristic factor data of chronic disease patients;

[0007] The chronic disease composite characteristic analysis module is used to analyze the comorbidity of chronic disease patients based on the disease mutagenic characteristic factor data of chronic disease patients and generate chronic disease comorbidity data; and to analyze the disease composite characteristic of chronic disease patients based on the chronic disease comorbidity data and generate chronic disease composite characteristic data;

[0008] The chronic disease trend feature abnormality identification module is used to analyze the chronic disease trend abnormality features based on the disease composite feature data of chronic disease patients, generate chronic disease trend abnormality feature data, and feed back the chronic disease trend abnormality feature data to the terminal.

[0009] The present invention has the beneficial effects of providing timely, accurate, and comprehensive basic information for subsequent analysis by acquiring chronic disease monitoring data in real time, including a chronic disease feature change analysis module, a chronic disease composite feature analysis module, and a chronic disease trend feature anomaly identification module. This system can capture disease dynamics immediately and avoid biased assessments of disease conditions due to data lags. Chronic disease classification based on real-time data allows complex and diverse chronic diseases to be categorized according to unified standards, helping medical staff and researchers gain a clearer and more systematic understanding of the characteristics of different types of chronic diseases, improving disease identification and management efficiency. By integrating the classified chronic disease data with related chronic diseases, potential correlations between different chronic diseases can be uncovered. This correlation integration not only helps to comprehensively understand the pathogenesis and interaction mechanisms of chronic diseases, but also provides a basis for developing more comprehensive and accurate disease prevention and treatment strategies. By integrating related chronic disease data, patients' potential complications can be more accurately predicted, preventive measures can be taken in advance, and the risk of disease and treatment costs can be reduced. In the public health sector, this system can help relevant departments formulate targeted disease prevention and control policies, rationally allocate medical resources, improve resource utilization efficiency, and more effectively prevent, control, and manage chronic diseases. Reconstructing the organizational structure of chronic disease patients based on integrated, interrelated chronic disease data can transcend the traditional single-dimensional understanding of a patient's condition. From an organizational perspective, it can deeply analyze the complex connections and changes between various systems and organs caused by chronic diseases. The resulting organizational structure mapping data for chronic disease patients provides an intuitive understanding of the extent and scope of chronic disease impact on various tissues and organs, as well as the interactions between them. Analyzing the disease-causing mutation signatures of chronic disease patients can pinpoint key factors that contribute to the onset and progression of chronic diseases and the various symptoms they trigger, such as genetics, lifestyle, environmental factors, and underlying conditions. This allows for more accurate assessment of disease progression and prediction of potential disease progression, enabling the development of more targeted, personalized treatment plans. Comorbidity analysis of chronic disease patients based on disease-causing mutation signature data can comprehensively and systematically identify the mechanisms and patterns of interaction among multiple chronic diseases. The resulting chronic disease comorbidity data provides physicians with a holistic understanding of a patient's complex condition, avoiding a single-disease focus during treatment while ignoring the connections and impacts of other comorbidities. When formulating treatment plans, we comprehensively consider the priority and treatment sequence of various diseases to improve treatment effectiveness and reduce the occurrence of complications. Further analysis of the complex disease characteristics of chronic disease patients based on chronic disease comorbidity data can further explore the unique characteristics and manifestations presented by the coexistence of multiple chronic diseases.Composite disease feature data can help doctors diagnose conditions more accurately and differentiate between patients, enabling more refined medical services. Chronic disease trend anomaly analysis based on composite disease feature data from patients with chronic diseases can be used to assist in diagnosis, more accurately assess the condition, develop appropriate treatment plans, and improve diagnostic accuracy and treatment effectiveness.

[0010] Therefore, the anomaly recognition system of trend characteristics of chronic disease monitoring data of the present invention can capture key abnormal changes in the course of the disease, and by monitoring the early stages of the evolution of chronic diseases, make trend predictions for chronic diseases to ensure timely preventive intervention for chronic diseases, thereby achieving the effectiveness and timeliness of clinical intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 This is a module diagram of a system for identifying abnormalities in trend characteristics of chronic disease monitoring data according to the present invention;

[0012] Figure 2 for Figure 1 Functional diagram of the chronic disease monitoring data integration module;

[0013] Figure 3 for Figure 1 Functional diagram of the chronic disease characteristic change analysis module. DETAILED DESCRIPTION

[0014] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the embodiments described are 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 those skilled in the art without making any creative work are within the scope of protection of the present invention.

[0015] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0016] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0017] To achieve this, please refer to Figures 1 to 3 The present invention provides a chronic disease monitoring data trend feature anomaly recognition system, comprising the following modules:

[0018] The chronic disease monitoring data integration module is used to obtain real-time chronic disease monitoring data, classify chronic diseases based on the real-time chronic disease monitoring data, and generate classified chronic disease data; and integrate related chronic diseases based on the classified chronic disease data to generate integrated related chronic disease data;

[0019] A chronic disease characteristic change analysis module is used to reconstruct the tissue structure mapping of chronic disease patients based on the integrated related chronic disease data to generate tissue structure mapping data of chronic disease patients; to analyze the disease mutagenic characteristic factors of chronic disease patients based on the tissue structure mapping data of chronic disease patients to generate disease mutagenic characteristic factor data of chronic disease patients;

[0020] The chronic disease composite characteristic analysis module is used to analyze the comorbidity of chronic disease patients based on the disease mutagenic characteristic factor data of chronic disease patients and generate chronic disease comorbidity data; and to analyze the disease composite characteristic of chronic disease patients based on the chronic disease comorbidity data and generate chronic disease composite characteristic data;

[0021] The chronic disease trend feature abnormality identification module is used to analyze the chronic disease trend abnormality features based on the disease composite feature data of chronic disease patients, generate chronic disease trend abnormality feature data, and feed back the chronic disease trend abnormality feature data to the terminal.

[0022] In the embodiment of the present invention, please refer to Figure 1 FIG. 1 is a schematic diagram of a module of a system for identifying anomalies of trend characteristics of chronic disease monitoring data according to the present invention. In this example, the system for identifying anomalies of trend characteristics of chronic disease monitoring data includes the following modules:

[0023] S1: Chronic disease monitoring data integration module, used to obtain real-time chronic disease monitoring data, classify chronic diseases based on the real-time chronic disease monitoring data, and generate classified chronic disease data; integrate related chronic diseases based on the classified chronic disease data, and generate integrated related chronic disease data;

[0024] In an embodiment of the present invention, medical IoT devices deployed in medical institutions and community health service centers, such as smart blood glucose meters, ambulatory blood pressure monitors, and wearable electrocardiogram (ECG) monitoring devices, collect chronic disease monitoring data such as patients' blood glucose levels, blood pressure fluctuations, and electrocardiogram (ECG) waveforms in real time. For example, the data collection frequency is set to every 15 minutes, and the data is transmitted to a server in a regional medical data center via Bluetooth Low Energy (BLE) technology. The server uses the Hadoop Distributed File System (HDFS) to store raw data and the Spark Streaming framework for real-time data processing. The International Classification of Diseases (ICD-10) standard is used to classify real-time data, and a classification model based on a decision tree algorithm is established. For example, if the fasting blood glucose value in the monitoring data exceeds 7.0 mmol / L for three consecutive times and the glycated hemoglobin (HbA1c) is ≥6.5%, the patient is diagnosed with diabetes and the relevant data is classified into the diabetes data set. When classifying hypertension data, if the systolic blood pressure is ≥140 mmHg or the diastolic blood pressure is ≥90 mmHg, and these criteria are met on at least three different days, the patient is classified as having hypertension. In the integration stage of associated chronic diseases, the Apriori algorithm is used to mine association rules between different disease data sets.

[0025] S2: Chronic disease characteristic change analysis module, used to reconstruct the tissue structure mapping of chronic disease patients based on the integrated related chronic disease data to generate tissue structure mapping data of chronic disease patients; perform disease mutagenic characteristic factor analysis on chronic disease patients based on the tissue structure mapping data of chronic disease patients to generate disease mutagenic characteristic factor data of chronic disease patients;

[0026] In an embodiment of the present invention, based on the integrated and correlated chronic disease data generated by the chronic disease monitoring data integration module, 3D medical image reconstruction technology is used to process the patient's CT and MRI image data. The Marching Cubes algorithm is used to construct 2D tomographic images into 3D models, enabling reconstruction of tissue structure mapping for chronic disease patients. Taking patients with both hypertension and diabetes as an example, by analyzing brain MRI images, the spatial distribution relationship between hypertension-induced microvascular lesions and diabetes-induced neuropathy in brain tissue can be observed, generating tissue structure mapping data containing information such as organ morphology, lesion location, and vascular status. In the analysis of disease-induced characteristic factors, principal component analysis (PCA) is used to reduce the dimensionality of the tissue structure mapping data and extract key features. Genetic factors such as the ACE gene I / D polymorphism associated with hypertension and the TCF7L2 gene locus associated with diabetes are screened from the patient's genetic test data. A dataset of disease-induced characteristic factors is constructed by combining lifestyle data such as daily salt intake, exercise, and smoking history. The importance of each characteristic factor is calculated using a random forest algorithm, identifying age, BMI, and family medical history as the main disease-induced characteristic factors.

[0027] S3: Chronic disease composite feature analysis module, used to perform comorbidity analysis on chronic disease patients based on their disease mutagenic feature factor data, and generate chronic disease comorbidity data; perform disease composite feature analysis on chronic disease patients based on their chronic disease comorbidity data, and generate disease composite feature data on chronic disease patients;

[0028] In an embodiment of the present invention, a chronic disease comorbidity analysis model is constructed using a Bayesian network based on the disease mutagenic characteristic factor data generated by the chronic disease characteristic change analysis module. Taking chronic obstructive pulmonary disease (COPD) and cardiovascular disease as an example, the patient's lung function test data, electrocardiogram data, and inflammatory factor (such as C-reactive protein) levels in the blood are input, and the probability of simultaneous occurrence of different diseases is calculated through a Bayesian network. If the model predicts that the probability of COPD patients with concurrent cardiovascular disease exceeds a certain value, the relevant data is marked as comorbidity data, and a chronic disease comorbidity data set is generated. In the disease composite feature analysis stage, the DBSCAN algorithm in cluster analysis is used to process the chronic disease comorbidity data. Taking the data of patients suffering from diabetes, hypertension and coronary heart disease as an example, patients with similar disease characteristics are clustered according to indicators such as the patient's blood sugar control level, blood pressure fluctuation amplitude, and degree of coronary artery stenosis.

[0029] S4: Chronic disease trend feature abnormality identification module, used to analyze chronic disease trend abnormality features based on the disease composite feature data of chronic disease patients, generate chronic disease trend abnormality feature data, and feed back the chronic disease trend abnormality feature data to the terminal.

[0030] In an embodiment of the present invention, the ARIMA model in time series analysis is used to analyze the abnormal trend characteristics of chronic diseases for the disease composite feature data of the chronic disease composite feature analysis module. Taking the blood glucose data of diabetic patients as an example, the daily fasting blood glucose values ​​of the past several months are modeled. If the error between the blood glucose value predicted by the model for the next three days and the actual monitoring value exceeds a certain ratio, it is determined that the blood glucose trend is abnormal, and the abnormal time point, fluctuation amplitude and other information are recorded to generate chronic disease trend abnormal feature data. The ECharts visualization library is used to convert the chronic disease trend abnormal feature data into a visual chart. The abnormal blood glucose trend changes are displayed with a line chart, and the abnormal data segments are highlighted with a red line. The correlation strength between multiple chronic disease composite features is presented through a heat map, and the darker the color, the closer the correlation. The generated visualization data is pushed to clinical terminal devices in real time via the WebSocket protocol, such as the hospital's electronic medical record system, the doctor's mobile ward round terminal, and the patient's smartphone APP. Doctors can use the electronic medical record system to quickly check abnormal disease trends in patients and formulate treatment plans such as adjusting insulin dosage and optimizing antihypertensive drug combinations; patients can receive health management tips on smartphone apps, such as dietary recommendations and exercise reminders, to effectively manage their chronic diseases.

[0031] Furthermore, the chronic disease monitoring data integration module includes the following functions:

[0032] Access real-time chronic disease monitoring data;

[0033] Perform time-series synchronization processing on real-time chronic disease monitoring data to generate time-series synchronized chronic disease data;

[0034] Based on the preset chronic disease clustering algorithm, the time series synchronized chronic disease data is processed for chronic disease classification to generate classified chronic disease data;

[0035] Perform chronic disease feature analysis based on classified chronic disease data to generate chronic disease feature data;

[0036] Perform chronic disease cross-feature analysis based on chronic disease feature data to generate chronic disease cross-feature data;

[0037] Based on the cross-feature data of chronic diseases, the classified chronic disease data are associated with chronic diseases to generate integrated associated chronic disease data.

[0038] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Schematic diagram of the functional flow of the chronic disease monitoring data integration module. In this embodiment, the chronic disease monitoring data integration module includes the following functions:

[0039] S11: Obtain real-time chronic disease monitoring data;

[0040] In an embodiment of the present invention, multi-source medical Internet of Things devices are deployed in hospitals, community health service centers, and patient homes, including smart blood glucose meters, dynamic blood pressure monitors, wearable electrocardiogram recorders, and blood oxygen saturation monitors. For example, the smart blood glucose meter collects fingertip blood glucose concentration at regular intervals and transmits the data to the local gateway via near-field communication (NFC) technology. The dynamic blood pressure monitor uses the oscillometric method to automatically measure systolic blood pressure, diastolic blood pressure, and mean arterial pressure every hour and sends the data via the Bluetooth 4.0 protocol. The wearable electrocardiogram recorder collects electrocardiogram signals at a sampling frequency of 250Hz and transmits them to the regional medical data center server via Wi-Fi. The blood oxygen saturation monitor uses the principle of photoelectric blood oxygen detection to record blood oxygen saturation data every 5 minutes and aggregates it to the server via the ZigBee network. The server uses the distributed storage architecture Ceph to store real-time chronic disease monitoring data collected by different devices in an object storage pool in timestamp order to ensure data integrity and traceability.

[0041] S12: Perform time series synchronization processing on the real-time chronic disease monitoring data to generate time series synchronized chronic disease data;

[0042] In this embodiment of the present invention, due to differences in sampling frequency and data transmission time among various monitoring devices, time series synchronization processing is required. An interpolation method based on timestamp alignment is employed, for example, using the hourly measurement time of an ambulatory blood pressure monitor as the reference time axis. For non-hourly data collected by a smart blood glucose meter, a linear interpolation formula is used. This processing is performed on all monitoring data, unifying data from different sources into the same time series to generate time-series synchronized chronic disease data.

[0043] S13: performing chronic disease classification processing on the time series synchronized chronic disease data based on a preset chronic disease clustering algorithm to generate classified chronic disease data;

[0044] In an embodiment of the present invention, the DBSCAN density clustering algorithm is used to classify chronic diseases. A density threshold MinPts and a neighborhood radius ε are set (determined based on statistical analysis of clinical data. Taking the blood glucose fluctuation range of diabetic patients as an example, 0.3mmol / L can effectively distinguish different blood glucose states). Taking the three types of disease data of diabetes, hypertension, and coronary heart disease as examples, the blood glucose value, blood pressure value, and electrocardiogram ST segment offset in the time-series synchronized chronic disease data are used as feature vectors. For a certain set of data points (for example, blood glucose 6.2mmol / L, systolic blood pressure 130mmHg, ST segment offset 0.1mV), the number of data points within the ε neighborhood is calculated. If it exceeds MinPts, it is divided into a cluster. All data points are traversed and data with similar features are clustered into one category. For example, clusters with blood glucose continuously higher than 7.0mmol / L are marked as diabetes, clusters with systolic blood pressure long-term greater than 140mmHg are marked as hypertension, and clusters with abnormal ST segment offset and accompanied by chest tightness symptoms are marked as coronary heart disease. Finally, classified chronic disease data is generated.

[0045] S14: performing chronic disease feature analysis based on the classified chronic disease data to generate chronic disease feature data;

[0046] In the embodiment of the present invention, principal component analysis (PCA) is used to extract chronic disease features for classified chronic disease data. Taking diabetes data as an example, the original data contains multiple feature dimensions such as fasting blood glucose, postprandial blood glucose, glycosylated hemoglobin, insulin resistance index, etc. First, the covariance matrix is ​​calculated. where x i is the i-th sample data, is the sample mean, and n is the number of samples. Then the covariance matrix is ​​decomposed to obtain the eigenvalues ​​λ1≥λ2≥…≥λ 10 And the corresponding eigenvectors. Select several principal components with large cumulative contributions, and perform weighted combination of the corresponding eigenvectors to form new chronic disease characteristic data.

[0047] S15: performing chronic disease cross-feature analysis based on the chronic disease feature data to generate chronic disease cross-feature data;

[0048] In an embodiment of the present invention, the association rule mining algorithm Apriori is used. The minimum support threshold and the minimum confidence threshold are set. Taking the two types of data, diabetes and hypertension, as an example, the proportion of patients suffering from both diabetes and hypertension in the total patients is calculated. If the proportion reaches a certain proportion (for example, greater than the minimum support of 0.15), the support condition is met. The probability of diabetic patients suffering from hypertension is further calculated. If it is 85% (for example, greater than the minimum confidence of 0.8), the association rule "diabetes → hypertension" is generated. By traversing all classified chronic disease data, cross-feature relationships such as "coronary heart disease → hyperlipidemia" and "diabetes + hypertension → renal damage" are mined.

[0049] S16: Based on the cross-feature data of chronic diseases, the classified chronic disease data is integrated with associated chronic diseases to generate integrated associated chronic disease data.

[0050] In an embodiment of the present invention, the classified chronic disease data are associated and integrated based on the cross-feature data of chronic diseases. The graph database Neo4j is used to construct a chronic disease relationship network, with each chronic disease as a node and the cross-feature relationship as an edge. Taking the cross-feature of "diabetes → hypertension" as an example, a "diabetes" node and a "hypertension" node are created in the graph database, and the attribute graph model is used to add node attributes (such as disease diagnosis criteria, typical symptoms), and a directed edge is established between the two nodes. The attributes of the edge include the association strength (determined according to the confidence calculated by the Apriori algorithm) and the association evidence (such as the number of patients with comorbidities). All chronic disease cross-feature data are traversed, and each case in the classified chronic disease data is added to the corresponding relationship network according to the cross-feature relationship to form a complete integrated and associated chronic disease data. For example, when there is new patient data (suffering from diabetes and hypertension at the same time), it is automatically matched to the constructed relationship network node, and the relevant statistical information of the nodes and edges is updated to realize the dynamic integration and association analysis of chronic disease data.

[0051] Furthermore, the chronic disease characteristic change analysis module includes the following functions:

[0052] Conduct regulatory signal molecule analysis on patients with chronic diseases based on the integration of related chronic disease data to generate regulatory signal molecule data;

[0053] Performing signal molecule transduction abnormality analysis on regulatory signal molecule data to generate signal molecule transduction abnormality data;

[0054] Based on the abnormal signal molecule transduction data, the perturbation regulatory factors of chronic disease patients are analyzed, including the perturbation regulatory factors of chronic disease patients in terms of changes in the number of signal transduction proteins, changes in the function of signal transduction proteins, excessive signal transduction that promotes cell proliferation, insufficient signal transduction that inhibits cell proliferation, signal transduction pathways, and abnormally activated and inhibited immune pathways, to generate perturbation regulatory factor data;

[0055] Conduct perturbation pathogenesis analysis on patients with chronic diseases based on perturbation regulator data to generate perturbation pathogenesis data;

[0056] Reconstructing the tissue structure mapping of chronic disease patients based on the disturbance pathogenesis data and the disturbance regulatory factor data to generate tissue structure mapping data of chronic disease patients;

[0057] The disease mutagenic characteristic factor analysis of chronic disease patients is performed based on the tissue structure mapping data of chronic disease patients to generate disease mutagenic characteristic factor data of chronic disease patients.

[0058] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Schematic diagram of the functional flow of the chronic disease characteristic change analysis module. In this embodiment, the chronic disease characteristic change analysis module includes the following functions:

[0059] S21: Analyze regulatory signal molecules in patients with chronic diseases based on the integrated related chronic disease data to generate regulatory signal molecule data;

[0060] In an embodiment of the present invention, a liquid chromatography-mass spectrometry (LC-MS) instrument is used to analyze blood and urine samples from patients with chronic diseases to obtain concentration data for various regulatory signaling molecules in the samples. For example, a 5 ml blood sample and a 10 ml urine sample are collected from a diabetic patient. After the samples enter the LC-MS instrument, the liquid chromatography section uses a C18 column with acetonitrile and 0.1% formic acid aqueous solution as the mobile phase, separating the molecules in the sample via gradient elution at a flow rate of 0.3 ml / min. The separated molecules enter the mass spectrometer and are ionized using an electrospray ionization (ESI) source in positive ion mode with a spray voltage of 3.5 kilovolts and a drying gas temperature of 250°C. The mass spectrometer scan range is set to m / z 50-1500, with a scan rate of 1 scan per second. This allows for accurate detection of the mass-to-charge ratio and abundance of regulatory signaling molecules such as insulin, glucagon, and tumor necrosis factor-α (TNF-α). By comparing the mass-to-charge ratio, retention time and other information of known signal molecules with the standard substance database, the specific type of regulatory signal molecules in the sample is determined, and its concentration is calculated using the external standard method based on the peak area to generate regulatory signal molecule data containing information such as molecule name, concentration, and sample collection time.

[0061] S22: performing signal molecule transduction abnormality analysis on the regulatory signal molecule data to generate signal molecule transduction abnormality data;

[0062] In an embodiment of the present invention, a combination of protein immunoblotting (Western Blot) technology and real-time fluorescence quantitative PCR (qPCR) technology is used to analyze abnormal signal molecule transduction of regulatory signal molecule data. Taking the study of the insulin signal transduction pathway as an example, total protein and total RNA are extracted from patient tissue samples (such as muscle tissue, 50 mg each time). In the Western Blot experiment, the extracted total protein is subjected to SDS-PAGE gel electrophoresis with a gel concentration of 12%. The electrophoresis voltage is set to 80 volts (stack gel stage) and 120 volts (separation gel stage), and the electrophoresis time is about 2 hours. After the electrophoresis is completed, the protein is transferred to a polyvinylidene fluoride (PVDF) membrane by semi-dry transfer at 15 volts for 90 minutes. After blocking the PVDF membrane with 5% skim milk for 1 hour, the primary antibody against insulin receptor substrate-1 (IRS-1) is added (dilution ratio 1:1000) and incubated overnight at 4°C. The next day, horseradish peroxidase (HRP)-conjugated secondary antibody (1:5000 dilution) was added and incubated at room temperature for 1 hour. IRS-1 protein expression and phosphorylation levels were analyzed using a chemiluminescent substrate for color development and gel imaging. In addition, total RNA was reverse transcribed into cDNA for qPCR. β-actin was used as an internal reference gene. IRS-1-specific primers (upstream primer: 5'-ATGCTGCTGCTGCTGCTG-3'; downstream primer: 5'-CTGCTGCTGCTGCTGCTG-3') were designed. The reaction system was prepared according to the SYBR Green PCR Master Mix instructions. Amplification was performed on a real-time fluorescence quantitative PCR instrument using the following reaction conditions: 95°C denaturation for 30 seconds, followed by 40 cycles of 95°C denaturation for 5 seconds and 60°C annealing for 30 seconds. The relative expression of IRS-1 was calculated using the 2^(-ΔΔCt) formula. Compared with the reference data of normal people, if the IRS-1 protein expression level decreases by more than a certain proportion, or the phosphorylation level decreases by more than a certain proportion, or the gene expression level decreases by more than a certain proportion, it is judged as abnormal insulin signaling molecule transduction, and signaling molecule transduction abnormality data containing information such as abnormal molecules, abnormal indicators, and abnormality degree are generated.

[0063] S23: Analyze the perturbation regulatory factors of chronic disease patients based on abnormal signal molecule transduction data, including perturbation regulatory factors in chronic disease patients with changes in the number of signal transduction proteins, changes in the function of signal transduction proteins, excessive signal transduction that promotes cell proliferation, insufficient signal transduction that inhibits cell proliferation, signal transduction pathways, and abnormally activated and inhibited immune pathways, to generate perturbation regulatory factor data;

[0064] In an embodiment of the present invention, immunohistochemical staining technology and flow cytometry are used to analyze the perturbation regulatory factors of chronic disease patients using abnormal signal molecule transduction data. Taking lung cancer patients as an example, tumor tissue sections (thickness 4 microns) were taken for immunohistochemical staining using the streptavidin-biotin-peroxidase complex (SABC) method. After the sections were dewaxed to water, antigen retrieval was performed (in a pH 6.0 citrate buffer, 95°C water bath for 20 minutes), endogenous peroxidase was blocked with 3% hydrogen peroxide solution for 10 minutes, and an anti-epidermal growth factor receptor (EGFR) primary antibody was added (dilution ratio 1:200) and incubated at 4°C overnight. The next day, biotin-labeled secondary antibody and SABC complex were added in sequence, DAB was used for color development, hematoxylin was used for counterstaining, and the sections were dehydrated and transparentized before sealing. Under an optical microscope, Image-Pro Plus software was used to count and analyze the optical density of EGFR-positive cells to calculate the positive cell rate and protein expression intensity. At the same time, 2 ml of peripheral blood was collected from the patient, and flow cytometry was used to detect T cell subsets and related signal transduction proteins. After the blood sample was treated with red blood cell lysis buffer, fluorescently labeled anti-CD3, anti-CD4, anti-CD8, and anti-phospho-AKT (p-AKT) antibodies were added (dilution ratio 1:100). The sample was incubated at room temperature in the dark for 30 minutes. After washing with PBS, the sample was detected on a flow cytometer using excitation wavelengths of 488 nm and 633 nm. FlowJo software was used to analyze the cell population ratio and the expression levels of signal transduction proteins. If the EGFR protein expression intensity is detected to be increased by more than a certain percentage compared with normal tissue, or the proportion of p-AKT-positive T cells is decreased by a certain percentage, and the ERK1 / 2 phosphorylation level in the MAPK signal transduction pathway that promotes cell proliferation is increased, the expression of PTEN protein that inhibits cell proliferation is decreased, and the NF-κB signal transduction pathway is abnormally activated (such as a 70% increase in nuclear p65 protein expression), and the TGF-β immune pathway is inhibited (such as a decrease in TGF-β1 secretion), then the influence coefficient of each perturbation regulatory factor is calculated based on these data (such as influence coefficient = (abnormal value - normal value) / normal value), generating perturbation regulatory factor data that includes information such as changes in the number and function of signal transduction proteins, perturbations in various signaling pathways and immune pathways, and influence coefficients.

[0065] S24: Analyze the perturbation pathogenesis of patients with chronic diseases based on the perturbation regulator data to generate perturbation pathogenesis data;

[0066] In the embodiment of the present invention, based on the disturbance regulation factor data, the system biology modeling method is used to analyze the disturbance pathogenesis of chronic disease patients. Taking atherosclerosis as an example, a mathematical model including multiple modules such as lipid metabolism, inflammatory response, and vascular endothelial cell function is constructed. In the lipid metabolism module, based on the concentration data of low-density lipoprotein (LDL) and high-density lipoprotein (HDL) and the activity data of related enzymes (such as lipoprotein lipase), the Michaelis-Menten equation ( Where V is the reaction rate, V max is the maximum reaction rate, [S] is the substrate concentration, K m A kinetic model for LDL oxidation and clearance is established (where TNF-α is the Michaelis constant). In the inflammatory response module, a signal transduction network model based on the law of mass action is established based on the concentration data of cytokines such as TNF-α and interleukin-6 (IL-6). In the vascular endothelial cell function module, a cell function regulation model is established based on the concentration data of nitric oxide (NO) and endothelin-1 (ET-1). The perturbation regulatory factor data is used as the input parameter of the model. For example, when it is detected that the inflammatory response is enhanced due to an increase in TNF-α concentration, the rate constant of the relevant reaction in the inflammatory response module is adjusted. The model is solved using numerical simulation methods (such as the Runge-Kutta method) to analyze the interaction between the modules and the dynamic changes of the system. If the simulation results show accelerated LDL oxidation, increased inflammatory cell infiltration, and aggravated vascular endothelial cell damage, and these changes are consistent with clinical symptoms and pathological manifestations, the perturbation pathogenesis of the chronic disease is determined, and perturbation pathogenesis data containing information such as key links in the pathogenesis, influencing factors, and action pathways are generated.

[0067] S25: reconstructing the tissue structure mapping of chronic disease patients based on the perturbation pathogenesis data and the perturbation regulatory factor data to generate tissue structure mapping data of chronic disease patients;

[0068] In an embodiment of the present invention, with the help of high-resolution microscopy imaging technology and computed tomography (CT) and magnetic resonance imaging (MRI) technology, the tissue structure mapping reconstruction of chronic disease patients is performed based on the perturbation pathogenesis data and the perturbation regulatory factor data. Taking patients with chronic kidney disease as an example, the kidney tissue is first ultrathinly sliced ​​(thickness 50 nanometers), and the ultrastructures such as glomeruli and renal tubules are observed using a transmission electron microscope (TEM). The accelerating voltage is set to 120 kV and the magnification is 10,000-100,000 times to obtain detailed images such as the mitochondrial morphology and endoplasmic reticulum structure in the cell. At the same time, the patient is subjected to an enhanced CT scan, and the scanning parameters are set to: tube voltage 120 kV, tube current 250 mA, layer thickness 1 mm, pitch 1.0, and iohexol contrast agent (350 mgI / ml, injection rate 3 ml / s) is injected intravenously to obtain a three-dimensional structural image of the kidney. MRI technology was used, employing T1WI and T2WI sequences with scan parameters of TR / TE (repetition time / echo time) = 500 / 15ms (T1WI) and 3000 / 80ms (T2WI), a 5 mm slice thickness, and a 256×256 matrix to acquire soft tissue contrast images of the kidney. TEM, CT, and MRI images were imported into the 3D image reconstruction software Mimics. Image processing algorithms such as threshold segmentation and region growing were used to extract the renal cortex, medulla, glomeruli, and tubules. Combined with information on tubular epithelial cell damage and glomerular sclerosis from perturbation pathogenesis data and the effects of relevant cytokines on tissue structure from perturbation regulatory factor data, the reconstructed 3D models were modified and annotated, such as highlighting damaged tubular areas and sclerotic glomeruli in the models. This generated tissue structure mapping data for patients with chronic diseases, including information on tissue morphology, location of pathological changes, and severity of lesions.

[0069] S26: performing a disease mutagenic characteristic factor analysis on the chronic disease patient based on the tissue structure mapping data of the chronic disease patient to generate disease mutagenic characteristic factor data on the chronic disease patient.

[0070] In an embodiment of the present invention, the random forest algorithm in machine learning is used to perform disease mutagenic characteristic factor analysis on the tissue structure mapping data of patients with chronic diseases. Taking patients with Alzheimer's disease as an example, the hippocampal volume, cerebral cortex thickness, β-amyloid protein (Aβ) plaque density, tau protein phosphorylation degree and other data in the tissue structure mapping data are used as feature variables, and the patient's clinical diagnosis results (whether suffering from Alzheimer's disease) are used as label variables. The number of decision trees of the random forest model is set to 100, the maximum depth of each tree is 8, and the minimum number of sample splits is 5. Training samples are extracted from the original data set with replacement by the bootstrap method to construct each decision tree. In the process of node splitting of the decision tree, the Gini impurity ( Where k is the number of categories, p i The splitting criterion (where ( is the proportion of samples in category i)) is used, and the feature that results in the greatest reduction in Gini impurity is selected for splitting. After training, disease-induced mutagenic signatures are determined based on the average importance of each feature variable across all decision trees (measured by the average reduction in node impurity after the feature variable is used to split the node). If Aβ plaque density has the highest average importance score, followed by hippocampal volume and tau protein phosphorylation, these features are identified as key disease-induced mutagenic signatures. This generates disease-induced mutagenic signature data for patients with chronic diseases, including feature name, importance score, and degree of association with the disease.

[0071] Furthermore, the tissue structure mapping reconstruction of chronic disease patients based on the disturbance pathogenesis data and the disturbance regulatory factor data includes:

[0072] Analyze the tissue interaction characteristics of patients with chronic diseases based on the perturbation pathogenesis data to generate tissue interaction characteristics data of patients with chronic diseases;

[0073] In an embodiment of the present invention, a multimodal imaging technology is combined with proteomic analysis to analyze tissue interaction characteristics. Taking patients with diabetic nephropathy as an example, a two-photon microscope is first used to image kidney tissue sections. The excitation light wavelength is set to 800nm ​​and the scanning speed is 1000Hz to obtain the cell morphology of glomeruli and renal tubules and the distribution images of fluorescent marker proteins between cells, so as to observe the direct contact between cells. At the same time, liquid chromatography-tandem mass spectrometry (LC-MS / MS) technology is used to perform proteomic detection on the patient's kidney tissue samples. 100mg of tissue sample is taken each time, and after lysis and enzymatic hydrolysis, it is separated by C18 reverse phase chromatography column. Mobile phase A is 0.1% formic acid aqueous solution, and mobile phase B is 0.1% formic acid acetonitrile solution. The separation is completed within 60 minutes with a gradient of 0-90% B, and the flow rate is set to 300nL / min. The separated peptides were then transferred to a mass spectrometer using data-dependent acquisition mode, with a primary mass spectrometer scanning range of m / z 300-1800 and a resolution of 120,000. The top 20 most intense ions were then selected for secondary mass spectrometry analysis, with a collision energy of 30 eV. The detected protein data were compared with the Uniprot database to determine the protein types and content. Based on information on diabetes-induced glomerular basement membrane thickening and renal tubular epithelial cell damage in the perturbation pathogenesis data, combined with changes in the expression of extracellular matrix proteins (such as collagen IV and laminin) in the proteomics data, and changes in the distribution of cell junction proteins (such as tight junction protein ZO-1) observed by two-photon microscopy, the material exchange and signal transmission relationships between different tissues were analyzed. If the expression of collagen IV in the glomerular basement membrane is found to increase by a certain proportion compared to normal, and the fluorescence intensity of the ZO-1 protein in the renal tubular epithelial cells is decreased by a certain proportion, it indicates that the interactive function of material filtration and reabsorption between the glomerulus and the renal tubule is impaired, and ultimately tissue interaction characteristic data of patients with chronic diseases is generated, which includes information such as changes in intertissue material flow, information flow, and cell connection status.

[0074] Preferably, the disturbance deviation amount of the metabolism of the chronic disease patient is calculated based on the disturbance adjustment factor data to generate metabolic disturbance deviation value data of the chronic disease patient;

[0075] In an embodiment of the present invention, the perturbation deviation of the metabolism of patients with chronic diseases is calculated by metabolomics analysis technology. Taking obese type 2 diabetic patients as an example, 5 mL of fasting blood sample and 10 mL of urine sample were collected from the patients, and metabolite detection was performed using gas chromatography-mass spectrometry (GC-MS) and ultra-performance liquid chromatography-mass spectrometry (UPLC-MS). In the GC-MS test, the blood sample was protein precipitated and derivatized, and 1 μL was injected. The chromatographic column used was DB-5MS (30m×0.25mm×0.25μm), and the heating program was an initial temperature of 80°C for 1 min, then increased to 320°C at 10°C / min and maintained for 5 min. The carrier gas was helium with a flow rate of 1 mL / min. The mass spectrometer used an electron bombardment ion source (EI), an electron energy of 70eV, and a scanning range of m / z50-600. When UPLC-MS was used to detect urine samples, an ACQUITY UPLC HSS T3 column (100mm×2.1mm, 1.8μm) was used. The mobile phase A was 0.1% formic acid in water, and the mobile phase B was acetonitrile. The separation was completed within 15 minutes with a gradient of 0-95% B, a flow rate of 0.3mL / min, and an electrospray ionization source (ESI) was used for mass spectrometry. The scan range was m / z100-1500 in positive ion mode. The detected metabolite data were compared with the Human Metabolome Database (HMDB) to determine the type and concentration of the metabolites. The mean value x and standard deviation σ of the metabolite concentration in the normal population were set as the reference standard. For a specific metabolite (such as glucose), the calculation formula for the perturbation deviation D was Where D represents the perturbation deviation of the metabolite in the patient, x is the detected concentration of the metabolite in the patient, represents the mean concentration of the metabolite in the normal population.

[0076] Preferably, the tissue function loss analysis of the chronic disease patients is performed on the tissue interaction characteristic data of the chronic disease patients based on the metabolic disturbance deviation value data of the chronic disease patients to generate the tissue function loss data of the chronic disease patients;

[0077] In an embodiment of the present invention, a tissue function quantitative assessment model is used to perform tissue function loss analysis based on metabolic disturbance deviation data. Taking liver tissue as an example, an assessment model is constructed based on the deviation of metabolites such as triglycerides, alanine aminotransferase (ALT), and aspartate aminotransferase (AST) in the metabolic disturbance deviation data, combined with the functional characteristics of the liver in lipid metabolism, detoxification, etc. For example, the weight coefficient of triglycerides is set to w1=0.3, the weight coefficient of ALT is set to w2=0.4, and the weight coefficient of AST is set to w3=0.3. The calculation formula for the liver tissue function loss index L is L=w1D 甘油三酯 +w2D ALT +w3D AST , where D 甘油三酯Indicates the deviation value of triglyceride metabolism level, D ALT Indicates the deviation value of alanine aminotransferase (ALT), a sensitive indicator of the degree of liver cell damage, D AST Indicates the deviation value of aspartate aminotransferase (AST). Combined with the abnormal information on material exchange between hepatocytes and liver sinusoidal endothelial cells in the tissue interaction characteristic data, a comprehensive assessment of tissue function loss is made. If the calculated liver tissue function loss index exceeds the normal threshold (for example, set to 3), and the tissue interaction characteristic data shows a decrease in the secretion of metabolites by hepatocytes into the liver sinusoids, it is determined that the liver tissue has functional loss, and tissue function loss data for chronic disease patients is generated, including information such as tissue name, functional loss index, and functional impairment manifestations.

[0078] Preferably, the tissue structure mapping of the chronic disease patient is reconstructed based on the tissue function loss data of the chronic disease patient and the tissue interaction characteristic data of the chronic disease patient to generate the tissue structure mapping data of the chronic disease patient.

[0079] In an embodiment of the present invention, three-dimensional modeling and data fusion techniques are used to reconstruct tissue structure mapping. Taking cardiac tissue as an example, a three-dimensional image of the patient's heart is first acquired using cardiac magnetic resonance imaging (CMR). Scanning parameters are: repetition time (TR) = 800ms, echo time (TE) = 30ms, slice thickness 8mm, and a 256×256 matrix. Simultaneously, the CMR image is corrected by combining information on reduced intercellular gap junction protein expression in the tissue interaction data center and results of muscle contractile function loss in the tissue function loss data center. Using the medical image processing software MIMICS, threshold segmentation is performed to separate the left ventricle, right ventricle, and myocardium. A physical model-based deformation algorithm is then used to adjust the model based on tissue function loss and changes in interaction characteristics. For example, due to decreased myocardial contractile function, the degree of myocardial tissue deformation during systole is reduced, and due to a decrease in intercellular gap junction protein, the electrical conduction characteristic parameters between myocardial cells are adjusted. Ultimately, three-dimensional visualization of tissue structure mapping data for patients with chronic diseases is generated, containing information on cardiac tissue morphology, functional changes, and abnormal intercellular interactions, intuitively demonstrating the impact of chronic diseases on cardiac tissue structure.

[0080] Furthermore, the analysis of the mutagenic characteristic factors of chronic disease patients based on the tissue structure mapping data of chronic disease patients includes:

[0081] Performing regulatory stress analysis on chronic disease patients based on tissue structure mapping data of chronic disease patients to generate regulatory stress data on chronic disease patients;

[0082] In an embodiment of the present invention, microfluidic chip technology is combined with a single-cell sequencing method to perform stress regulation analysis in patients with chronic diseases. Taking patients with chronic obstructive pulmonary disease (COPD) as an example, cell samples are extracted from the patient's bronchoalveolar lavage fluid, and the samples are injected into the cell capture channel of the microfluidic chip. The chip channel size is 50 microns × 50 microns, and single cells are allowed to enter independent reaction chambers in sequence through the principle of fluid dynamics. In each reaction chamber, a reaction solution containing a variety of stress marker detection reagents, such as heat shock protein 70 (HSP70) detection antibodies and reactive oxygen species (ROS) fluorescent probes, is added and incubated at a constant temperature of 37°C for 30 minutes. The cells in the reaction chamber are imaged using a fluorescence microscope, the excitation light wavelength is set to 488nm to detect ROS fluorescence signals, 561nm to detect HSP70 antibody fluorescence signals, and the imaging resolution is 1024×1024 pixels. Simultaneously, single-cell RNA sequencing was performed on each cell using the 10xGenomics Chromium platform. Cells were lysed, reverse transcribed, and amplified by PCR to construct a cDNA library. High-throughput sequencing was performed on an Illumina NovaSeq sequencer, with a sequencing depth of 50,000 reads per cell. Fluorescence imaging data were combined with single-cell sequencing data for analysis. Cells were considered to be in a stress state if their ROS fluorescence intensity exceeded twice the average intensity of normal cells, their HSP70 gene expression was upregulated by 1.5-fold, and their inflammatory response-related IL-6 gene expression increased by 3-fold. By calculating the proportion of cells in a stress state and the average fold-over expression of stress-related genes, a stress-regulating data set for chronic disease patients was generated, including information on the percentage of stressed cells, stress marker concentrations, and changes in stress-related gene expression. For example, if 30% of cells were detected to be in a stress state, the average ROS fluorescence intensity was 2.5 times that of normal cells, and the average HSP70 gene expression was 1.8 times that of normal cells, these data would be recorded in the stress-regulating data set.

[0083] Preferably, feedback function attenuation analysis of chronic disease patients is performed based on the regulatory stress data of chronic disease patients to generate feedback function attenuation data of chronic disease patients;

[0084] In an embodiment of the present invention, a dynamic metabolic monitoring system and bioelectrical signal detection technology are used to analyze the feedback function attenuation of patients with chronic diseases. Taking diabetic patients as an example, a continuous glucose monitor (CGM) is used to collect interstitial fluid glucose concentration every 5 minutes for 24 hours to obtain a blood glucose fluctuation curve. Simultaneously, an implantable glucose sensor is used to detect changes in extracellular glucose concentration in real time. The sensor uses a glucose oxidase electrode and the detection principle is based on the current signal generated by glucose oxidation, with a sensitivity of 1 μA / (mmol / L). During an oral glucose tolerance test (OGTT), patients take 75 grams of glucose orally. Blood samples are collected at 0, 30, 60, 90, and 120 minutes after taking 75 grams of glucose. High-performance liquid chromatography (HPLC) is used to measure the concentrations of hormones such as insulin, glucagon, and C-peptide. The HPLC column is a C18 column (250 mm × 4.6 mm, 5 μm), the mobile phase is acetonitrile-0.1% formic acid aqueous solution (20:80, v / v), the flow rate is 1 mL / min, and the detection wavelength is 214 nm. Based on the blood glucose fluctuation curve and hormone concentration changes, a feedback function attenuation index is calculated. Taking insulin secretion feedback as an example, the formula The response coefficient of insulin secretion to changes in blood glucose was calculated, where ΔI is the change in insulin concentration and ΔG is the change in blood glucose concentration. Similar calculations were performed for other feedback regulatory systems, such as glucagon, to generate feedback function attenuation data for chronic disease patients, including information on the attenuation degree and response coefficient of each feedback regulatory system.

[0085] Preferably, the disease mutagenic characteristic factor analysis of chronic disease patients is performed based on the feedback function attenuation data and the regulatory stress data of chronic disease patients to generate disease mutagenic characteristic factor data of chronic disease patients.

[0086] In an embodiment of the present invention, a multivariate linear regression model is combined with a principal component analysis method to perform a characteristic factor analysis of the disease mutation of patients with chronic diseases based on feedback function attenuation data and regulatory stress data. Taking hypertensive patients as an example, the degree of feedback attenuation of the renin-angiotensin-aldosterone system (RAAS) and the degree of attenuation of the baroreflex in the feedback function attenuation data, as well as the concentration of vascular smooth muscle cell stress markers and the expression multiple of inflammatory factors in the regulatory stress data are used as independent variables, and the degree of increase in the patient's blood pressure is used as the dependent variable. First, all independent variables are standardized, and the formula is: Where x is the original data, is the mean, s is the standard deviation. Then use principal component analysis to reduce the dimension of the independent variable, calculate the eigenvalues ​​and eigenvectors of the correlation coefficient matrix, and select the principal components whose cumulative contribution rate exceeds a certain proportion. Assuming that three principal components are extracted, PC1, PC2, and PC3, their expressions are where a ijis the eigenvector coefficient, x j ′ A multivariate linear regression model was established, and the regression coefficients of the principal components were calculated using the least squares method. If the absolute value of the regression coefficient of the principal component PC1 was the largest and it was primarily composed of the concentration of vascular smooth muscle cell stress markers and the degree of RAAS feedback attenuation, these two factors were identified as the main disease-inducing characteristic factors. Ultimately, disease-inducing characteristic factor data for chronic disease patients were generated, including characteristic factor names, influence weights, and degree of association with the disease, providing a basis for the diagnosis and treatment of chronic diseases.

[0087] Furthermore, the stress regulation analysis of chronic disease patients based on the tissue structure mapping data of chronic disease patients includes:

[0088] Performing physiological electrical signal analysis on chronic disease patients based on their tissue structure mapping data, performing spectrum analysis and wavelet energy transform on their tissue structure mapping data to identify the distribution of low-frequency / medium-frequency / high-frequency electrical signals in different organ tissue structures of chronic disease patients, and generating physiological electrical signal data for chronic disease patients;

[0089] In an embodiment of the present invention, a multi-channel bioelectric signal acquisition system is used to collect physiological electrical signals from patients with chronic diseases. Taking heart disease patients as an example, 12-lead ECG electrodes are pasted on the patient's chest, limbs and other parts. The electrodes are made of Ag / AgCl material, with an impedance of less than 5kΩ, and the sampling frequency is set to 1000Hz; at the same time, gastrointestinal electrodes are placed on the abdomen, and a pressure of 0.05N is applied to each square centimeter of the electrode surface, with a sampling frequency of 500Hz. The acquisition system transmits the electrical signal to the data acquisition card through a shielded cable. The A / D conversion accuracy of the data acquisition card is 16 bits. The analog electrical signal is converted into a digital signal and stored in the server. The collected digital signal is subjected to spectral analysis and wavelet energy transformation. The spectral analysis uses the fast Fourier transform (FFT) algorithm to convert the time domain signal into a frequency domain signal. The formula is Where x(n) is the time domain signal sequence, N is the number of sampling points, and k is the frequency index. Taking the ECG signal as an example, 0-0.5Hz is divided into the low frequency band, 0.5-4Hz is the medium frequency band, and 4-40Hz is the high frequency band. The signal energy proportion of each frequency band is calculated. The wavelet energy transform uses the db4 wavelet basis function to decompose the signal into three layers. The calculation formula is where ψ j,k (n) is the wavelet basis function, j is the number of decomposition levels, and k is the translation parameter. By calculating the energy values ​​of different frequency bands and decomposition levels, the distribution of low-frequency, medium-frequency, and high-frequency electrical signals in the tissue structures of organs such as the heart and gastrointestinal tract is identified, generating physiological electrical signal data for patients with chronic diseases that includes information such as signal frequency, energy value, and corresponding organs.

[0090] Preferably, the coordinated response physiological electrical signal data of the chronic disease patient is analyzed based on the physiological electrical signal data of the chronic disease patient to generate the coordinated response physiological electrical signal data of the chronic disease patient;

[0091] In the embodiment of the present invention, the cross-correlation analysis method is used to analyze the coordinated response physiological electrical signals. Taking diabetic patients as an example, pancreatic electrical signals and gastrointestinal electrical signals related to blood sugar regulation are selected for analysis. The cross-correlation function calculation formula is: Where x(t) is the pancreatic electrical signal, y(t) is the gastrointestinal electrical signal, and τ is the time delay. Set the time delay range from -10s to 10s and calculate the cross-correlation coefficient under different delay times. When the cross-correlation coefficient R corresponding to a certain delay time τ0 is xy When (τ0) reaches its maximum value and exceeds the normal threshold, it indicates that the pancreas and gastrointestinal tract have a synergistic response relationship at time τ0. Pairwise cross-correlation analysis is performed on all collected organ electrical signals to construct an organ electrical signal synergistic response matrix.

[0092] Preferably, the pathological trigger factor analysis of the chronic disease patient is performed based on the coordinated response physiological electrical signal data of the chronic disease patient to generate the pathological trigger factor data;

[0093] In an embodiment of the present invention, a decision tree algorithm is used to analyze pathological trigger factors based on the collaborative response physiological electrical signal data. Taking patients with chronic kidney disease as an example, the high-frequency energy ratio of kidney electrical signals, the correlation coefficient between kidney and heart electrical signals, and the creatinine concentration in urine are used as feature variables, and whether the patient's renal function deteriorates (with the increase in blood creatinine exceeding the baseline value as the judgment standard) is used as the label variable. Gini impurity is used as the node splitting criterion, and the calculation formula is: Where k is the number of categories, p i is the proportion of samples in the i-th class. In the process of building a decision tree, the feature that reduces the Gini impurity the most is selected for splitting.

[0094] Preferably, tissue imbalance analysis of chronic disease patients is performed based on the pathological trigger factor data to generate tissue imbalance data of chronic disease patients;

[0095] In the embodiment of the present invention, tissue imbalance analysis is performed based on pathological trigger factor data. Taking patients with hypertension as an example, the abnormal degree of vascular smooth muscle cell electrical signals, the synergy of vascular endothelial cells and smooth muscle cells electrical signals, angiotensin concentration and other indicators in the pathological trigger factor data are used as original feature variables. First, the original feature variables are standardized, and then the covariance matrix is ​​used to calculate the original feature variables. Perform eigendecomposition on the covariance matrix and obtain the eigenvalues ​​λ1≥λ2≥…≥λ m and the corresponding eigenvector, where n is the total number of samples, xi represents the normalized feature vector of the i-th sample, Represents the mean vector of all samples. Select the first three principal components and project the original feature variables into the principal component space. Use the K-means clustering algorithm and use the Euclidean distance as the similarity measure between samples. The formula is: Where x and y represent the principal component vectors of any two samples, and d(x, y) represents the Euclidean distance between the two samples. Through multiple iterative calculations, patients are divided into different tissue imbalance categories. If a patient category has a significantly higher score on principal component 1 than other categories, and this principal component is primarily determined by the degree of abnormality in vascular smooth muscle cell electrical signals, this indicates that these patients have vascular smooth muscle tissue imbalance. This generates tissue imbalance data for patients with chronic diseases, including information such as tissue imbalance category, imbalance characteristics, and patient grouping.

[0096] Preferably, the physiological electrical signal feedback deviation value of the chronic disease patient is calculated based on the tissue imbalance data of the chronic disease patient and the coordinated response physiological electrical signal data of the chronic disease patient to generate the physiological electrical signal feedback deviation value data of the chronic disease patient;

[0097] In an embodiment of the present invention, a regression analysis method is used to calculate the feedback deviation value of the physiological electrical signal. Taking patients with chronic liver disease as an example, the low-frequency energy of the liver electrical signal and the synergistic response strength of the liver and spleen electrical signals are selected as analysis indicators. x1 is the low-frequency energy of the patient's liver electrical signal, and x2 is the synergistic response strength of the patient's liver and spleen electrical signals, wherein the low-frequency energy of the liver electrical signal reflects the slow-changing characteristics of the organ's functional activity, and the synergistic response strength reflects the degree of functional coupling between organs. A linear regression model y=b0+b1x1+b2x2 is established, wherein y is the electrical signal feedback index under normal physiological conditions, and b0, b1, and b2 are regression coefficients to be calculated. By the least squares method Calculated, where n is the number of samples, x i1 and x i2 is the eigenvalue of the i-th sample, y i After the model training is completed, the measured values ​​of the target patients are substituted into the model to obtain the predicted feedback value. Then, the physiological electrical signal feedback deviation value is calculated based on the reference value of the normal population. Where E represents the deviation degree of the patient's physiological feedback, and y is the feedback value predicted by the regression model. It is the average value of physiological feedback of normal people.

[0098] Preferably, based on the physiological electrical signal feedback deviation value data of the chronic disease patient, the chronic disease patient's tissue structure mapping data is subjected to a regulation stress analysis to generate the chronic disease patient's regulation stress data.

[0099] In an embodiment of the present invention, stress analysis is performed on the tissue structure mapping data based on the physiological electrical signal feedback deviation value data. Taking Alzheimer's patients as an example, the indicators such as the cerebral cortex electrical signal feedback deviation, the hippocampus and prefrontal cortex electrical signal coordination deviation in the physiological electrical signal feedback deviation value data are fused with the indicators such as the hippocampus atrophy degree and β-amyloid protein plaque density in the tissue structure mapping data to form a feature vector. The radial basis function (RBF) is used as the kernel function of the SVM, and the formula is: Where γ is the kernel function parameter, x i and x j Represents two sample points in the input feature vector space, ||x i -x j || 2 The square of the Euclidean distance between two sample points is used to measure the similarity between samples. The optimal value is determined by cross-validation. The penalty parameter C is set to 10, and the data is divided into a training set and a test set, with the training set accounting for 70% and the test set accounting for 30%. The SVM model is trained using the training set data to minimize the structural risk. Where w is the weight vector, ξ i The optimal classification hyperplane is obtained by using the slack variable as the test set data. The trained model is then fed with the test data. If the model predicts that a patient is in a state of high regulatory stress, and the deviation value of their cerebral cortical electrical signal feedback exceeds the normal threshold and the degree of hippocampal atrophy reaches moderate, then the patient is determined to have a significant regulatory stress response.

[0100] Furthermore, the feedback function attenuation analysis of chronic disease patients based on the stress regulation data of chronic disease patients includes:

[0101] The feedback chain impact analysis of chronic disease patients is conducted on the regulatory stress data of chronic disease patients. The tissue interaction characteristic data of chronic disease patients is used as the core node, the physiological electrical signal data of chronic disease patients is used as the edge path, and the pathological trigger factor data is used as the impact mechanism of the feedback of the corresponding physiological system of the chronic disease patient to generate the feedback chain impact network data of chronic disease patients.

[0102] In an embodiment of the present invention, the graph database Neo4j is used to construct a feedback chain influence network for patients with chronic diseases. Taking patients with diabetes and cardiovascular diseases as an example, the material exchange relationship between the glomerulus and the renal tubule, the signal transmission relationship between the myocardial cells and the vascular endothelial cells, etc. in the tissue interaction characteristic data are used as core nodes, and each node contains attribute information such as the tissue name, interaction status, and the degree of pathological changes. For example, the glomerular node records data such as the degree of basement membrane thickening and changes in filtration function. The physiological electrical signal data is used as the edge path, and the connection relationship and weight of the edge are determined based on the calculated mutual correlation coefficient of the electrical signals between organs. If the mutual correlation coefficient between the cardiac electrical signal and the kidney electrical signal is 0.8, a directed edge is established between the cardiac tissue node and the kidney tissue node, and the edge weight is set to 0.8. The attributes of the edge include information such as the signal conduction direction and the frequency range of the electrical signal. The pathological trigger factor data is integrated into the network as a feedback influence mechanism. For example, in diabetic patients, the pathological trigger of endothelial cell damage caused by hyperglycemia is identified by updating the attributes of the relevant node (endothelial cell node), adding a "hyperglycemia impact degree" field, and establishing new directed edges between other affected tissue nodes (such as the cardiomyocyte node), annotating the edge attributes with "hyperglycemia-endothelial damage-myocardial function changes." By traversing all tissue interaction characteristic data, physiological electrical signal data, and pathological trigger factor data, a complete feedback chain impact network is constructed, ultimately generating feedback chain impact network data for chronic disease patients, including node attributes, edge connectivity, weights, and impact mechanisms.

[0103] Preferably, physiological conduction time delay analysis of chronic disease patients is performed based on feedback chain impact network data of chronic disease patients to generate physiological conduction time delay data of chronic disease patients;

[0104] In an embodiment of the present invention, a signal transmission delay calculation model is used to perform physiological conduction time lag analysis. Taking the interactive feedback between the nervous system and the endocrine system as an example, in the feedback chain influence network, the signal transmission path between the neuron node and the endocrine gland cell node is found. The time lag is calculated based on the propagation speed and path length of the electrical signal in the physiological electrical signal data. For example, the length of the electrical signal transmission path between the neuron and the endocrine gland cell is L (obtained by measuring the three-dimensional spatial distance in the tissue structure mapping data), and the propagation speed v of the electrical signal in the nerve fiber is 100m / s (the normal neural electrical signal propagation speed is known), then the simple time lag calculation formula is: For more complex situations, consider the signal delay in synaptic transmission, body fluid transport, etc. In the synaptic transmission link, introduce the synaptic transmission delay k1 (according to experimental data statistics, the general chemical synaptic transmission delay is about 0.5-2ms, here we take 1ms). In the body fluid transport link, according to the diffusion speed v of hormones in the blood, h and L hCalculating delay time Among them L h It represents the length of the transport path of hormones or signal molecules from the secretion point to the target site in body fluids (such as blood). h The effective transmission speed of hormones in the blood depends on the blood flow velocity and hormone distribution characteristics. The total delay T is the delay of each link, such as the delay time of nerve conduction, synaptic transmission and body fluid signal transmission. total =T+k1+T h , where T is the time it takes for the electrical signal to conduct along the neuron, k1 is the signal conversion delay at the synapse, and T h This calculation is performed for all signal transmission pathways within a feedback chain-influencing network, such as the signal transmission delay between hypothalamic neurons and pancreatic islet cells. This calculation considers processes such as neural electrical signal transmission, neurotransmitter release and diffusion, and hormone secretion and blood transport. This generates physiological transmission delay data for chronic disease patients, including information such as the starting and ending points of the pathway, the delay time of each link, and the total delay.

[0105] Preferably, feedback function attenuation analysis of chronic disease patients is performed on the regulatory stress data of chronic disease patients based on the physiological conduction time delay data of chronic disease patients to generate feedback function attenuation data of chronic disease patients.

[0106] In an embodiment of the present invention, a feedback function attenuation evaluation model is used to perform feedback function attenuation analysis on the regulatory stress data. Taking patients with hypothyroidism as an example, data related to the thyroid-pituitary feedback axis are selected for analysis. First, a normal feedback regulation model is established based on the physiological conduction delay data and feedback regulation parameters of the normal population. Assuming that under normal circumstances, the total time delay for the change in thyroid hormone (T3, T4) concentration to be transmitted to the pituitary gland and trigger the secretion regulation of thyroid stimulating hormone (TSH) is T normal , the regulatory coefficient of TSH on thyroid hormone secretion is r normal For patients, their physiological conduction delay data T patient The changes in thyroid hormone concentration and TSH secretion in the stress data are substituted into the model. The feedback function attenuation index F is calculated using the formula: where r patient is the patient's TSH regulation coefficient. Similar calculations are performed for all feedback regulation systems in the patient's body, such as the insulin-blood glucose feedback system and the renin-angiotensin-aldosterone feedback system, to generate feedback function attenuation data for chronic disease patients, including information such as the feedback system name, attenuation index, and key influencing factors.

[0107] Furthermore, the analysis of characteristic factors of disease mutagenesis in chronic disease patients based on feedback function attenuation data and regulatory stress data of chronic disease patients includes:

[0108] Conduct disease conversion factor analysis on chronic disease patients based on their feedback function attenuation data to generate disease conversion factor data;

[0109] In an embodiment of the present invention, a Logistic regression model is used to analyze the disease conversion factor of the feedback function attenuation data of patients with chronic diseases. Taking the possibility that patients with hypertension may be converted to coronary heart disease as an example, the feedback attenuation degree of the renin-angiotensin-aldosterone system (RAAS), the baroreflex attenuation degree, the vascular endothelial cell function feedback attenuation value, etc. in the feedback function attenuation data are used as independent variables, and whether the patient is converted from hypertension to coronary heart disease (converted to 1, not converted to 0) is used as the dependent variable. The independent variables are standardized. Then a Logistic regression model is constructed. where β1,β2,…,β n is the regression coefficient, which indicates the direction and intensity of the influence of each feedback function attenuation variable on the probability of disease transformation, x1, x2,…, x n is the standardized independent variable. The regression coefficient is solved by the maximum likelihood estimation method, so that the likelihood function Reach the maximum value, where m is the number of samples, y i Is the i-th patient converted, x i is the standardized feedback function attenuation vector of the i-th patient. If the absolute value of the regression coefficient β of a certain independent variable (such as the degree of RAAS system feedback attenuation) is large and the significance level p is less than a certain value, then the factor is determined to be a disease conversion factor. Calculate the odds ratio of each factor OR = e β , which indicates the multiple of the probability of disease transformation occurring when the factor changes by one unit. Ultimately, disease transformation factor data are generated, including the name of the disease transformation factor, regression coefficient, odds ratio, significance level, and other information.

[0110] Preferably, disease crosstalk analysis is performed on patients with chronic diseases based on the disease conversion factor data to generate disease crosstalk data;

[0111] In this embodiment of the present invention, the association rule mining algorithm Apriori is used to perform disease crosstalk analysis on patients with chronic diseases. Taking the patient group with both diabetes and chronic kidney disease as an example, factors such as insulin resistance and hyperglycemia toxicity related to diabetes in the disease conversion factor data, and factors such as decreased glomerular filtration rate and proteinuria related to chronic kidney disease are used as analysis objects. The minimum support threshold is set to 0.1, and the minimum confidence threshold is set to 0.8. The support calculation formula is: Where X represents the antecedent factor set, such as diabetes-related factors, Y represents the posterior factor set, such as chronic kidney disease-related factors, |X∪Y| represents the number of samples containing both factors X and Y, N is the total number of samples, and the confidence calculation formula is It indicates the proportion of samples containing factor X that also contain factor Y.

[0112] Preferably, a chronic disease coupled disturbance analysis is performed on the regulatory stress data of chronic disease patients based on the disease crosstalk data to generate chronic disease coupled disturbance data;

[0113] In an embodiment of the present invention, a system dynamics model is used to perform chronic disease coupled perturbation analysis on the regulated stress data of patients with chronic diseases. Taking patients with coupled hypertension and coronary heart disease as an example, the correlation between vascular endothelial damage and atherosclerosis in the disease crosstalk data is combined with the information such as the concentration of vascular smooth muscle cell stress markers and the expression multiple of inflammatory factors in the regulated stress data to construct a system dynamics model. The model includes material flow (such as the flow of cholesterol and inflammatory factors in the blood), information flow (such as intercellular signal transduction) and feedback loops (such as blood pressure regulation feedback). Set state variables (such as the amount of cholesterol deposited on the blood vessel wall C, the number of inflammatory cells I, and the rate variables (such as the cholesterol deposition rate R) C , inflammatory cell proliferation rate R I and auxiliary variables (such as the degree of endothelial cell damage D). Equations are established based on biological principles, such as the cholesterol deposition rate R C =k1×blood cholesterol concentration×(1-D), where k1 is a constant and the inflammatory cell proliferation rate R I =k2×D×inflammatory factor concentration, where k2 is a constant. Solve the model using numerical simulation methods (such as the Runge-Kutta method). Set the time step to 1 day and the simulation period to 365 days. Analyze the model output results. If the simulation shows that the amount of cholesterol deposits on the vascular wall increases by a certain percentage compared to the initial value after 180 days, and the number of inflammatory cells increases by a certain percentage at the same time, it indicates that there is a coupling disturbance between hypertension and coronary heart disease. Calculate the coupling disturbance intensity index Where ΔC is the rate of change of cholesterol deposition, and ΔI is the rate of change of the number of inflammatory cells.

[0114] Preferably, the tissue structure mapping data of chronic disease patients are analyzed for disease mutagenic characteristic factors of chronic disease patients based on the chronic disease coupled perturbation data to generate disease mutagenic characteristic factor data of chronic disease patients.

[0115] In an embodiment of the present invention, a method combining random forest and feature importance ranking is used to perform disease mutagenic characteristic factor analysis on the tissue structure mapping data of chronic disease patients based on the coupled perturbation data of chronic diseases. Taking patients with chronic obstructive pulmonary disease (COPD) and cardiovascular disease as an example, the concentration changes of inflammatory factors (such as TNF-α, IL-6) and oxidative stress indicators (such as ROS levels) in the coupled perturbation data are integrated with the degree of lung tissue fibrosis, heart ventricular wall thickness and other information in the tissue structure mapping data as feature variables. For example, the number of decision trees of the random forest model is set to 200, the maximum depth of each tree is 10, and the minimum number of sample splits is 5. Training samples are extracted from the original data set with replacement by the bootstrap method to construct each decision tree. In the process of node splitting of the decision tree, the Gini impurity is used Where k is the number of categories, p i The splitting criterion (where ( is the proportion of samples in category i)) was used, and the feature that resulted in the greatest decrease in Gini impurity was selected for splitting. After training, each feature variable was ranked according to its average importance across all decision trees (measured by the average reduction in node impurity after the feature variable was used to split the node). If changes in TNF-α concentration had the highest average importance score, followed by the degree of lung fibrosis and ROS levels, these three factors were identified as key disease-inducing features.

[0116] Furthermore, the chronic disease composite feature analysis module includes the following functions:

[0117] Conducting correlation tissue change analysis on chronic disease patients based on disease mutagenic characteristic factor data of chronic disease patients to generate correlation tissue change data on chronic disease patients;

[0118] In an embodiment of the present invention, spatial mapping and image analysis techniques are used to analyze the association of disease mutagenic characteristic factor data of chronic disease patients with tissue changes. Taking diabetic patients as an example, the disease mutagenic characteristic factor data includes information such as hyperglycemic toxicity and insulin resistance. High-resolution microscopy technology is used to obtain images of tissue sections of the patient's pancreas, liver, kidney, etc. The thickness of the pancreatic sections is set to 5 microns, and HE staining and immunohistochemistry staining (for insulin receptors and GLUT4 transporters) are used. The images are processed using ImageJ software, and the tissue cell structure is extracted through threshold segmentation and morphological operations. For pancreatic tissue, the number density of pancreatic beta cells is calculated using the formula: cell number density = total number of pancreatic beta cells / pancreatic tissue area. For example, in a pancreatic tissue image of a diabetic patient, the pancreatic tissue area is measured to be 0.1 square millimeters. Through manual counting and software-assisted identification, the total number of pancreatic beta cells is 500. Therefore, the cell number density = 500 / 0.1 = 5000 / square millimeter. By comparing the pancreatic beta cell density with that of the normal population (assuming 8,000 cells / square millimeter), the degree of reduction in the number of pancreatic beta cells is obtained. At the same time, the liver tissue images are analyzed for fatty degeneration, and the degree of fatty degeneration is assessed by calculating the ratio of the area of ​​fat vacuoles to the total area of ​​liver cells. If the area of ​​fat vacuoles in a patient's liver tissue accounts for a high proportion, while the ratio is low in the normal population, it is determined that the liver has obvious fatty degeneration. By combining the analysis results of various tissues, the associated tissue change data of patients with chronic diseases is generated, which includes information such as tissue name, cell structure change indicators, and degree of functional damage.

[0119] Preferably, a comorbidity analysis of chronic disease patients is performed on the associated tissue change data of chronic disease patients based on the chronic disease coupling disturbance data, so as to generate chronic disease comorbidity data by analyzing the co-occurrence frequency, temporal coupling degree and symptom synergy index corresponding to the chronic disease coupling disturbance data, and performing comorbidity screening on the associated tissue change data of chronic disease patients based on the co-occurrence frequency, temporal coupling degree and symptom synergy index;

[0120] In an embodiment of the present invention, a comorbidity analysis is performed on the associated tissue change data using a quantitative analysis method. Taking patients with both hypertension and coronary heart disease as an example, the chronic disease coupling disturbance data contains information such as vascular endothelial damage and elevated inflammatory factors. The co-occurrence frequency is calculated using the formula: co-occurrence frequency = number of patients with both diseases / total number of patients. The degree of temporal coupling is calculated using a cross-correlation analysis method. Taking blood pressure fluctuation data and coronary artery stenosis degree change data as an example, the time window is set to 30 days and the sampling interval is 1 day. The cross-correlation function formula is: Where x(n) is the blood pressure value sequence, y(n) is the coronary artery stenosis degree sequence, and where τ is the mean of the blood pressure and stenosis series, N is the number of data points, and τ is the time delay. The cross-correlation coefficient for different values ​​of τ is calculated, and the maximum value is taken as the indicator of the degree of temporal coupling. The symptom synergy index is calculated and a symptom scoring system is established. For example, separate scoring criteria are set for hypertension symptoms (headache, dizziness) and coronary heart disease symptoms (chest pain, palpitations). The symptom synergy index formula is: where s 1i is the symptom score of hypertension at time point i, s 2i is the symptom score of CHD at time point i, and n is the number of time points. If the calculated symptom synergy index is greater than a certain value, it indicates that the symptoms of the two diseases have significant synergy.

[0121] Preferably, the disease composite characteristic data of chronic disease patients are analyzed based on the chronic disease comorbidity data to generate the disease composite characteristic data of chronic disease patients.

[0122] In an embodiment of the present invention, principal component analysis (PCA) and cluster analysis methods are used to analyze the composite disease characteristics of chronic disease comorbidity data. Taking data from patients with diabetes, hypertension, and coronary heart disease as an example, the comorbidity data includes multiple parameters, such as blood glucose levels, blood pressure, degree of coronary artery stenosis, pancreatic beta cell count, and endothelial cell function indicators. The data is first normalized, and then the covariance matrix is ​​calculated. Eigendecomposition is performed on the covariance matrix to obtain eigenvalues ​​and corresponding eigenvectors. The first three principal components are selected, and the original data is projected into the principal component space. For example, the first principal component primarily reflects the combined effects of blood glucose and blood pressure, while the second principal component primarily reflects the relationship between coronary artery stenosis and pancreatic beta cell count. The projected data is clustered using the K-means clustering algorithm, using Euclidean distance as the inter-sample similarity metric. Through multiple iterative calculations, patients are divided into different categories. The data characteristics of each category are analyzed. If a patient in a certain category exhibits hyperglycemia and hypertension accompanied by severe coronary artery stenosis and a significant decrease in pancreatic beta cells, these characteristics are summarized as the composite disease characteristics of that category.

[0123] Furthermore, the chronic disease trend characteristic abnormality identification module includes the following functions:

[0124] Conduct disease regional analysis of chronic disease patients based on their composite disease characteristic data to generate chronic disease regional data;

[0125] In an embodiment of the present invention, a method of fusion of medical image analysis and histopathological data is used to perform disease regional analysis on the disease composite feature data of patients with chronic diseases. Taking patients with diabetic nephropathy as an example, the disease composite feature data covers information such as renal function indicators (such as blood creatinine, glomerular filtration rate), urine test data (degree of proteinuria), and microstructural change indicators of renal tissue (proportion of tubular atrophy, proportion of glomerular sclerosis area). Renal tissue sections were imaged using an electron microscope, with the slice thickness set to 50 nanometers and the accelerating voltage set to 120 kilovolts to obtain ultrastructural images of the renal tubules and glomeruli. At the same time, renal ultrasound contrast imaging technology was used, using SonoVue contrast agent (bolus injection via elbow vein at a dose of 0.02 mL / kg), scanning was performed in low mechanical index (MI=0.08) mode to obtain renal blood perfusion images, with parameters set to a frame rate of 10 frames / second and a depth of 8 cm. Electron microscopy images were imported into ImageJ software, and through threshold segmentation and morphological operations, renal tubular and glomerular regions were extracted. The proportion of tubular atrophy to total tubular area was calculated using the formula: tubular atrophy ratio = atrophic tubular area / total tubular area × 100%. Blood perfusion parameters such as time to peak perfusion (TTP) and peak perfusion intensity (PI) were calculated for different renal regions using contrast-enhanced ultrasound images using specialized analysis software (such as Qontrast). Based on these calculations, the kidneys were classified into different diseased areas. If the proportion of tubular atrophy in a given area exceeded a certain value, and the time to peak perfusion was prolonged by more than a certain percentage of the normal mean (the normal mean was obtained from statistical data of healthy subjects), the area was marked as severely damaged.

[0126] Preferably, the chronic disease regional data is used to monitor the onset cycle fluctuations of chronic disease patients to generate onset cycle fluctuation data;

[0127] In an embodiment of the present invention, time series analysis and periodicity detection algorithms are used to monitor the periodic fluctuations of chronic disease regional data. Taking patients with coronary heart disease as an example, multiple data are collected from different areas of the patient's heart (such as the anterior wall and inferior wall of the left ventricle), including electrocardiogram ST segment offset (measured by 12-lead electrocardiogram at 60ms after the J point, with an accuracy of 0.01mV), cardiac magnetic resonance imaging (MRI) myocardial delayed enhancement (using contrast agent gadopentetate dimeglumine, dose 0.1mmol / kg, through T1WI sequence imaging, TR / TE=500 / 15ms, layer thickness 8mm), etc. The Fourier transform algorithm is used to perform frequency domain analysis on the time series data, and the electrocardiogram ST segment offset data is converted from the time domain to the frequency domain. The formula is: Where x(n) is the time domain signal sequence, N is the number of sampling points, and k is the frequency index. By analyzing the spectrum, the periodic components in the data are identified. If there is a significant peak at a certain frequency f0 (the peak intensity exceeds 3 times the average intensity), the period corresponding to this frequency is determined to be the onset period. The onset period fluctuation index is calculated in combination with the autocorrelation function. The autocorrelation function formula is: in is the mean, and τ is the time delay. If the peak position of the autocorrelation function shifts significantly within a certain time period (e.g., the shift exceeds 20% of the normal period), and the degree of delayed myocardial enhancement detected by MRI fluctuates by more than a certain amplitude within the same time period, then the cardiac region is judged to have periodic fluctuations.

[0128] Preferably, the abnormal trend characteristics of chronic diseases are analyzed on the composite characteristic data of chronic disease patients based on the onset cycle fluctuation data to generate abnormal trend characteristics data of chronic diseases. The abnormal trend characteristics data of chronic diseases are visualized and fed back to the clinical terminal equipment for doctor-assisted diagnosis and individual health management prompts.

[0129] In an embodiment of the present invention, a method combining statistical models and machine learning is used to analyze the abnormal characteristics of chronic disease trends in the disease composite characteristic data of patients with chronic diseases. Taking patients with chronic obstructive pulmonary disease (COPD) as an example, the disease composite characteristic data includes lung function indicators (FEV1, FEV1 / FVC), blood gas analysis data (arterial oxygen partial pressure PaO2, carbon dioxide partial pressure PaCO2), lung tissue density values ​​of chest CT, etc. First, a generalized linear model (GLM) is used to fit the data. Taking FEV1 data as an example, a model is established: E(Y) = β0+β1X1+β2X2+…+β n X n , where Y is the FEV1 measurement value, X1, X2,…, X n are influencing factors (such as disease course, smoking amount), β0, β1,…, β n is the regression coefficient, which is solved by the maximum likelihood estimation method. Calculate the residual between the model prediction value and the actual value. If the standard deviation of the residual exceeds the normal range (the normal range is determined by training with data from a healthy population), it is preliminarily determined that there is a trend abnormality. Then use the random forest model for further analysis. For example, set the number of decision trees to 150, the maximum depth of each tree to 8, and the minimum number of sample splits to 5. Use the disease characteristic data of the past 12 months as input to predict the values ​​for the next 3 months. Calculate the mean absolute error (MAE) between the predicted value and the actual value: where y i is the actual value, is the predicted value, and n is the number of samples. If MAE exceeds the normal threshold (determined by training with normal patient data), it is finally determined that there is a trend abnormality. The abnormal characteristic data of chronic disease trends are processed through the D3.js visualization library. The Sankey diagram is used to show the characteristic evolution relationship between different disease areas, and the bar chart is used to compare the indicator differences between abnormal areas and normal areas. The visualization data is pushed to clinical terminal devices through the RESTful API interface, such as the hospital's electronic medical record system, the doctor's mobile ward round terminal, and the patient's smartphone APP. Doctors can quickly locate abnormal areas and abnormal indicators through the electronic medical record system and formulate targeted treatment plans; patients can view their own disease trends on the APP and receive personalized health management tips.

[0130] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced within the present invention.

[0131] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A chronic disease monitoring data trend feature anomaly recognition system, characterized by: The following steps are involved: The chronic disease monitoring data integration module is used to obtain real-time chronic disease monitoring data, classify chronic diseases based on the real-time chronic disease monitoring data, and generate classified chronic disease data; and integrate related chronic diseases based on the classified chronic disease data to generate integrated related chronic disease data; A chronic disease characteristic change analysis module is used to reconstruct the tissue structure mapping of chronic disease patients based on the integrated related chronic disease data to generate tissue structure mapping data of chronic disease patients; to analyze the disease mutagenic characteristic factors of chronic disease patients based on the tissue structure mapping data of chronic disease patients to generate disease mutagenic characteristic factor data of chronic disease patients; The chronic disease composite characteristic analysis module is used to analyze the comorbidity of chronic disease patients based on the disease mutagenic characteristic factor data of chronic disease patients and generate chronic disease comorbidity data; and to analyze the disease composite characteristic of chronic disease patients based on the chronic disease comorbidity data and generate chronic disease composite characteristic data; The chronic disease trend feature abnormality identification module is used to analyze the chronic disease trend abnormality features based on the disease composite feature data of chronic disease patients, generate chronic disease trend abnormality feature data, and feed back the chronic disease trend abnormality feature data to the terminal.

2. The abnormality identification system for trend characteristics of chronic disease monitoring data according to claim 1 is characterized in that: The chronic disease monitoring data integration module includes the following functions: Access real-time chronic disease monitoring data; Perform time-series synchronization processing on real-time chronic disease monitoring data to generate time-series synchronized chronic disease data; Based on the preset chronic disease clustering algorithm, the time series synchronized chronic disease data is processed for chronic disease classification to generate classified chronic disease data; Perform chronic disease feature analysis based on classified chronic disease data to generate chronic disease feature data; Perform chronic disease cross-feature analysis based on chronic disease feature data to generate chronic disease cross-feature data; Based on the cross-feature data of chronic diseases, the classified chronic disease data are associated with chronic diseases to generate integrated associated chronic disease data.

3. The abnormality identification system for trend characteristics of chronic disease monitoring data according to claim 1 is characterized in that: The chronic disease characteristic change analysis module includes the following functions: Conduct regulatory signal molecule analysis on patients with chronic diseases based on the integration of related chronic disease data to generate regulatory signal molecule data; Performing signal molecule transduction abnormality analysis on regulatory signal molecule data to generate signal molecule transduction abnormality data; Based on the abnormal signal molecule transduction data, the perturbation regulatory factors of chronic disease patients are analyzed, including the perturbation regulatory factors of chronic disease patients in terms of changes in the number of signal transduction proteins, changes in the function of signal transduction proteins, excessive signal transduction that promotes cell proliferation, insufficient signal transduction that inhibits cell proliferation, signal transduction pathways, and abnormally activated and inhibited immune pathways, to generate perturbation regulatory factor data; Conduct perturbation pathogenesis analysis on patients with chronic diseases based on perturbation regulator data to generate perturbation pathogenesis data; Reconstructing the tissue structure mapping of chronic disease patients based on the disturbance pathogenesis data and the disturbance regulatory factor data to generate tissue structure mapping data of chronic disease patients; The disease mutagenic characteristic factor analysis of chronic disease patients is performed based on the tissue structure mapping data of chronic disease patients to generate disease mutagenic characteristic factor data of chronic disease patients.

4. The abnormality identification system for trend characteristics of chronic disease monitoring data according to claim 3 is characterized in that: The tissue structure mapping reconstruction of chronic disease patients based on the disturbance pathogenesis data and the disturbance regulatory factor data includes: Analyze the tissue interaction characteristics of patients with chronic diseases based on the perturbation pathogenesis data to generate tissue interaction characteristics data of patients with chronic diseases; Calculating the metabolic disturbance deviation of chronic disease patients based on the disturbance adjustment factor data to generate metabolic disturbance deviation value data of chronic disease patients; Based on the metabolic disturbance deviation value data of chronic disease patients, tissue interaction characteristic data of chronic disease patients are used to analyze tissue function loss of chronic disease patients, thereby generating tissue function loss data of chronic disease patients; Based on the tissue function loss data of chronic disease patients, the tissue interaction characteristic data of chronic disease patients are reconstructed to map the tissue structure of chronic disease patients and generate tissue structure mapping data of chronic disease patients.

5. The abnormality identification system for trend characteristics of chronic disease monitoring data according to claim 4 is characterized in that: The analysis of the disease mutagenic characteristic factors of chronic disease patients based on the tissue structure mapping data of chronic disease patients includes: Performing regulatory stress analysis on chronic disease patients based on tissue structure mapping data of chronic disease patients to generate regulatory stress data on chronic disease patients; Conduct feedback function attenuation analysis on chronic disease patients based on their regulatory stress data, and generate feedback function attenuation data on chronic disease patients; Based on the feedback function attenuation data and regulatory stress data of chronic disease patients, the disease mutagenic characteristic factor analysis of chronic disease patients is performed to generate disease mutagenic characteristic factor data of chronic disease patients.

6. The abnormality identification system for trend characteristics of chronic disease monitoring data according to claim 5 is characterized in that: The method of performing regulatory stress analysis on chronic disease patients based on tissue structure mapping data of chronic disease patients includes: Performing physiological electrical signal analysis on chronic disease patients based on their tissue structure mapping data, performing spectrum analysis and wavelet energy transform on their tissue structure mapping data to identify the distribution of low-frequency / medium-frequency / high-frequency electrical signals in different organ tissue structures of chronic disease patients, and generating physiological electrical signal data for chronic disease patients; performing a coordinated response physiological electrical signal analysis on the chronic disease patients based on the physiological electrical signal data of the chronic disease patients, and generating coordinated response physiological electrical signal data of the chronic disease patients; Analyze the pathological trigger factors of chronic disease patients based on the coordinated response physiological electrical signal data of chronic disease patients to generate pathological trigger factor data; Analyze tissue imbalance in patients with chronic diseases based on pathological trigger factor data to generate tissue imbalance data for patients with chronic diseases; calculating the physiological electrical signal feedback deviation value of the chronic disease patient based on the tissue imbalance data of the chronic disease patient and the coordinated response physiological electrical signal data of the chronic disease patient, thereby generating the physiological electrical signal feedback deviation value data of the chronic disease patient; Based on the physiological electrical signal feedback deviation value data of chronic disease patients, the tissue structure mapping data of chronic disease patients are analyzed to generate the regulation stress data of chronic disease patients.

7. The abnormality identification system for trend characteristics of chronic disease monitoring data according to claim 6 is characterized in that: The feedback function attenuation analysis of chronic disease patients based on the stress regulation data of chronic disease patients includes: The feedback chain impact analysis of chronic disease patients is conducted on the regulatory stress data of chronic disease patients. The tissue interaction characteristic data of chronic disease patients is used as the core node, the physiological electrical signal data of chronic disease patients is used as the edge path, and the pathological trigger factor data is used as the impact mechanism of the feedback of the corresponding physiological system of the chronic disease patient to generate the feedback chain impact network data of chronic disease patients. Conduct physiological conduction delay analysis on chronic disease patients based on feedback chain impact network data of chronic disease patients, and generate physiological conduction delay data on chronic disease patients; Based on the physiological conduction time delay data of chronic disease patients and the regulatory stress data of chronic disease patients, feedback function attenuation analysis of chronic disease patients is performed to generate feedback function attenuation data of chronic disease patients.

8. The abnormality identification system for trend characteristics of chronic disease monitoring data according to claim 5 is characterized in that: The analysis of the characteristic factors of disease mutagenesis in chronic disease patients based on the feedback function attenuation data and the regulatory stress data of chronic disease patients includes: Conduct disease conversion factor analysis on chronic disease patients based on their feedback function attenuation data to generate disease conversion factor data; Conduct disease crosstalk analysis on patients with chronic diseases based on disease conversion factor data to generate disease crosstalk data; Based on the disease crosstalk data, chronic disease coupled disturbance analysis is performed on the regulatory stress data of chronic disease patients to generate chronic disease coupled disturbance data; Based on the chronic disease coupled perturbation data, the tissue structure mapping data of chronic disease patients are analyzed for the disease mutagenic characteristic factors of chronic disease patients to generate the disease mutagenic characteristic factor data of chronic disease patients.

9. The abnormality identification system for trend characteristics of chronic disease monitoring data according to claim 8, characterized in that: The chronic disease composite feature analysis module includes the following functions: Conducting correlation tissue change analysis on chronic disease patients based on disease mutagenic characteristic factor data of chronic disease patients to generate correlation tissue change data on chronic disease patients; Performing comorbidity analysis on the associated tissue change data of chronic disease patients based on the chronic disease coupling disturbance data, analyzing the co-occurrence frequency, temporal coupling degree, and symptom synergy index corresponding to the chronic disease coupling disturbance data, and performing comorbidity screening on the associated tissue change data of chronic disease patients based on the co-occurrence frequency, temporal coupling degree, and symptom synergy index to generate chronic disease comorbidity data; The composite disease characteristic data of chronic disease patients are analyzed based on the chronic disease comorbidity data to generate the composite disease characteristic data of chronic disease patients.

10. The abnormality identification system for trend characteristics of chronic disease monitoring data according to claim 1, characterized in that: The chronic disease trend feature abnormality recognition module includes the following functions: Conduct disease regional analysis of chronic disease patients based on their composite disease characteristic data to generate chronic disease regional data; Monitor the onset cycle fluctuations of chronic disease patients in chronic disease area data and generate onset cycle fluctuation data; Based on the onset cycle fluctuation data, the chronic disease trend abnormality feature data of the chronic disease patients is analyzed on the composite disease characteristic data to generate chronic disease trend abnormality feature data. The chronic disease trend abnormality feature data is visualized and fed back to the clinical terminal equipment for doctor-assisted diagnosis and individual health management prompts.

Images