Heart failure early warning biomarker conjoint analysis system and method thereof

Abstract

The invention discloses a biomarker conjoint analysis system and method for early warning of heart failure, and the system comprises a data collection module which is used for collecting various biomarker data, clinical data and patient individualized data in real time; the multi-modal data fusion module is used for preprocessing and fusing biomarker data, iconography data, clinical records and individual difference data from different sources, and the deep learning analysis module is used for carrying out feature extraction and analysis by utilizing a deep learning model based on fused data and outputting a heart failure early-stage risk assessment result. By integrating various biomarkers, iconography data and individual difference data, the heart failure risk of a patient can be comprehensively evaluated, the limitation of only depending on a single marker is avoided, the accuracy and sensitivity of early warning are remarkably improved, multi-dimensional data are comprehensively analyzed by using a deep learning model, and the accuracy and sensitivity of early warning are improved. Potential modes and complex relations in the data can be automatically found, and the prediction accuracy is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medical diagnosis and biomedical engineering technology, in particular to a biomarker joint analysis system for early warning of heart failure and a method thereof. BACKGROUND

[0002] The biomarker joint analysis system for early warning of heart failure and the method thereof integrate multiple biomarker data and combine advanced analysis techniques to achieve efficient identification and prediction of the early stage of heart failure. The system uses joint analysis of biomarkers (such as BNP, cTnI, myocardial enzymes, etc.) and clinical data, combined with machine learning algorithms, to comprehensively evaluate the risk of heart failure in patients, predict changes in the disease in advance, and provide accurate early warning information for early intervention and personalized treatment, significantly improving the survival rate and quality of life of patients.

[0003] In the prior art, the biomarker joint analysis system and method for early warning of heart failure have the following shortcomings: limited marker selection: existing systems usually rely on a few traditional biomarkers such as BNP, cTnI, etc., but these markers may not fully reflect the complex mechanisms of heart failure. Therefore, single or limited marker analysis may result in insufficient accuracy and sensitivity of early warning; insufficient data integration and analysis capability: in practical applications, the early warning of heart failure involves multiple biomarkers, clinical data (such as imaging, medical history, etc.) and patient individual characteristics. Existing technologies often cannot effectively integrate and analyze these multi-dimensional heterogeneous data, lack strong data fusion and multi-modal analysis capabilities, and the performance of the early warning system is limited when dealing with complex information; limitations of algorithms and models: many existing systems rely on traditional statistical analysis or simple machine learning methods, lacking the application of more complex deep learning algorithms. This makes it difficult for the model to extract deep patterns from a large amount of high-dimensional data, thereby affecting the prediction accuracy, especially in the case of large individual differences, which may not provide accurate early warning; lack of individualization: existing technologies usually use standardized biomarkers and uniform prediction models, lacking individualized analysis for different patients. For example, factors such as patient age, gender, lifestyle, genetic background, etc. have different effects on heart failure, but existing methods often fail to fully consider these individual differences, resulting in large differences in early warning performance among different patients; sensitivity and specificity of early warning: although existing systems can detect some early signs of heart failure, there are still deficiencies in sensitivity and specificity. Some systems may lack sufficient predictive ability in the initial stages of the disease, or be easily affected by external factors (such as drug interference, infection, etc.), leading to false positives or false negatives; insufficient real-time and operability: some existing systems have poor real-time performance in clinical applications, usually relying on laboratory tests, and cannot achieve dynamic monitoring and immediate early warning. In addition, clinicians may have difficulty directly operating complex analysis tools, affecting the actual application effect of the system.

[0004] Therefore, we propose a biomarker joint analysis system and method for early warning of heart failure. SUMMARY

[0005] To achieve the above purpose, the present application provides the following technical solution: a biomarker joint analysis system for early warning of heart failure, comprising the following steps:

[0006] S1: Data acquisition module: for real-time acquisition of multiple biomarker data, clinical data and patient individual data;

[0007] S2: Multi-modal data fusion module: for preprocessing and fusion of biomarker data, imaging data, clinical records and individual difference data from different sources;

[0008] S3: Deep learning analysis module: based on the fusion data, using deep learning model for feature extraction and analysis, output heart failure early risk assessment results;

[0009] S4: Personalized early warning module: according to the deep learning analysis results and individual differences of patients, provide personalized heart failure risk warning for each patient;

[0010] S5: Real-time monitoring module: for real-time monitoring of patient biomarker data, and dynamically adjusting the warning results according to the monitoring data;

[0011] S6: Report output module: generate dynamic warning report according to real-time analysis results, provide warning and treatment suggestions for clinicians.

[0012] Preferably, wherein the data acquisition module further comprises: a plurality of biomarker data acquisition unit, including BNP, cTnI, myocardial enzyme, renal function index, etc.; imaging data acquisition unit, including heart ultrasound, CT scan, MRI data acquisition; individual difference data acquisition unit, including patient age, gender, weight, genetic background, life habit information acquisition.

[0013] Preferably, wherein the multi-modal data fusion module further comprises: data preprocessing unit, for removing noise data, filling missing data, and standardizing the data; feature selection unit, through statistical method or machine learning algorithm to select the most relevant features, and carry out dimension reduction processing.

[0014] Preferably, wherein the deep learning analysis module adopts deep learning model: convolutional neural network (CNN), used to extract important features in imaging data; recurrent neural network (RNN), used to process time series data such as patient's biomarker dynamic change; multilayer perceptron (MLP), used to integrate the analysis results of multi-modal data and output heart failure early risk assessment value.

[0015] Preferably, wherein the personalized early warning module realizes personalized early warning by the following ways: individual data weighting, based on patient individual characteristics (such as genetic background, life habit, etc.) to adjust the weight of biomarker; adaptive learning mechanism, according to the patient's historical data and real-time data to automatically adjust the early warning model, in order to improve the accuracy of early warning.

[0016] A biomarker combined analysis method for early warning of heart failure, characterized by comprising the following steps:

[0017] S1: Real-time acquisition of a plurality of biomarker data, imaging data, clinical data and patient individual difference data;

[0018] S2: Preprocessing the collected data, including noise removal, missing data imputation, and data standardization;

[0019] S3: Feature extraction and fusion of the preprocessed data through deep learning algorithms, outputting early risk assessment results of heart failure;

[0020] S4: Providing personalized early warning information for patients according to the early risk assessment results of heart failure, and generating reports;

[0021] S5: Real-time monitoring of patient data, real-time adjustment of early warning information and treatment recommendations according to dynamic changes.

[0022] Preferably, the deep learning analysis algorithm in step S3 includes:

[0023] Convolutional Neural Network (CNN): for processing imaging data;

[0024] Recurrent Neural Network (RNN): for processing time series data of patient biomarkers;

[0025] Multilayer Perceptron (MLP): for integrating multi-modal data and making final assessment.

[0026] Compared with the prior art, the present application provides a biomarker joint analysis system and method for early warning of heart failure, which has the following beneficial effects:

[0027] 1. The biomarker joint analysis system and method for early warning of heart failure can comprehensively assess the risk of heart failure in patients by integrating various biomarkers, imaging data, and individual difference data, avoiding the limitations of relying on a single marker, significantly improving the accuracy and sensitivity of early warning, and using a deep learning model to comprehensively analyze multi-dimensional data can automatically discover potential patterns and complex relationships in the data, improving the accuracy of prediction, especially when dealing with a large amount of dynamic data, it can effectively adapt to complex clinical scenarios, dynamically adjust the early warning model according to the individual differences of patients (such as genetic background, age, gender, etc.), avoiding the problem of poor applicability of a unified standard to different patients, thereby realizing more accurate personalized risk assessment and treatment recommendations.

[0028] 2. The biomarker joint analysis system and method for early warning of heart failure can monitor the biomarker data and health status of patients in real time, dynamically adjust the early warning results according to real-time data, ensure timely intervention measures during disease progression, and maximize the reduction of the risk of heart failure and deterioration, through a simplified operation process and intuitive report output, clinicians can easily interpret the early warning information provided by the system, assist in decision-making, and improve the actual effect and work efficiency of clinical application. DETAILED DESCRIPTION

[0029] The technical solutions in the embodiments of the present application will be described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work are within the protection scope of the present application.

[0030] EMBODIMENT

[0031] Embodiments of a biomarker joint analysis system for early warning of heart failure and a method thereof

[0032] A biomarker joint analysis system for early warning of heart failure comprises the following steps:

[0033] S1: a data acquisition module for acquiring real-time multiple biomarker data, clinical data and patient individualized data;

[0034] S2: a multi-modal data fusion module for preprocessing and fusing biomarker data, imaging data, clinical records and individual difference data from different sources;

[0035] S3: a deep learning analysis module for feature extraction and analysis based on the fusion data using a deep learning model, and outputting early risk assessment results of heart failure;

[0036] S4: a personalized early warning module for providing personalized heart failure risk warning for each patient according to the deep learning analysis results and patient individual differences;

[0037] S5: a real-time monitoring module for real-time monitoring of the patient's biomarker data and dynamically adjusting the early warning results according to the monitoring data;

[0038] S6: a report output module for generating a dynamic early warning report according to the real-time analysis results, and providing early warning and treatment suggestions for clinicians.

[0039] Specifically, the data acquisition module further comprises: a multiple biomarker data acquisition unit including BNP, cTnI, myocardial enzyme, renal function indicators, etc.; an imaging data acquisition unit including heart ultrasound, CT scan, MRI data acquisition, etc.; and an individual difference data acquisition unit including patient age, gender, weight, genetic background, lifestyle information acquisition, etc.

[0040] Specifically, the multi-modal data fusion module further includes a data preprocessing unit for removing noise data, filling missing data, and standardizing the data; and a feature selection unit for selecting the most relevant features based on statistical methods or machine learning algorithms and performing dimensionality reduction processing.

[0041] Specifically, the deep learning analysis module employs a deep learning model that includes a convolutional neural network (CNN) for extracting important features from imaging data, a recurrent neural network (RNN) for processing time series data such as patient biomarker dynamics, and a multi-layer perceptron (MLP) for integrating the analysis results of multi-modal data and outputting an early risk assessment value for heart failure.

[0042] Specifically, the individualized early warning module achieves individualized early warning through the following methods: individualized data weighting, adjusting the weight of biomarkers based on patient individual characteristics (such as genetic background, lifestyle habits, etc.); adaptive learning mechanism, automatically adjusting the early warning model based on patient historical data and real-time data to improve the accuracy of early warning.

[0043] A biomarker joint analysis method for early warning of heart failure, characterized by the following steps:

[0044] S1: Real-time collection of multiple biomarker data, imaging data, clinical data, and patient individual difference data;

[0045] S2: Preprocessing of collected data, including noise removal, missing data filling, and data standardization;

[0046] S3: Feature extraction and fusion of preprocessed data through deep learning algorithms, outputting early risk assessment results for heart failure;

[0047] S4: Providing individualized early warning information for patients based on the early risk assessment results for heart failure, and generating a report;

[0048] S5: Real-time monitoring of patient data, real-time adjustment of early warning information based on dynamic changes, and provision of treatment recommendations.

[0049] Specifically, the deep learning analysis algorithm in step S3 includes:

[0050] Convolutional Neural Network (CNN): for processing imaging data;

[0051] Recurrent Neural Network (RNN): for processing patient biomarker time series data;

[0052] Multi-layer Perceptron (MLP): for integrating multi-modal data and performing final evaluation.

[0053] Through the technical scheme, the heart failure risk of the patient can be comprehensively evaluated by integrating various biomarkers, imaging data and individual difference data, the limitation of relying on a single marker is avoided, the accuracy and sensitivity of early warning are significantly improved, the deep learning model is used for comprehensive analysis of multi-dimensional data, potential patterns and complex relationships in the data can be automatically found, the prediction accuracy is improved, especially when dealing with a large amount of dynamic data, the complex clinical scene can be effectively adapted, the early warning model is dynamically adjusted according to the individual differences (such as genetic background, age, gender, etc.) of the patient, the problem that the uniform standard is not suitable for different patients is avoided, so that more accurate personalized risk assessment and treatment suggestions are realized, the system can monitor the biomarker data and health status of the patient in real time, and dynamically adjust the early warning result according to the real-time data, so that timely intervention measures are provided in the disease progression process, the occurrence and aggravation risk of heart failure is minimized, through the simplified operation process and intuitive report output, the clinician can easily interpret the early warning information provided by the system, assist in decision-making, and improve the actual effect and work efficiency of clinical application.

[0054] Data acquisition module: real-time acquisition of multi-modal data of patients using biosensors, intelligent medical devices and hospital information systems (HIS). Biomarker data is collected by laboratory equipment (such as blood testing instruments, electrocardiographs, etc.), imaging data is obtained by cardiac imaging equipment (such as CT, MRI, ultrasound, etc.), and individual difference data is recorded by questionnaire survey, electronic medical record, etc.

[0055] Data fusion and preprocessing: first, the collected data is cleaned, standardized and noise removed. For missing data, appropriate imputation algorithms (such as KNN imputation, mean imputation, etc.) are used for processing. Then, the most relevant features are selected by using feature selection methods (such as L1 regularization, principal component analysis), and redundant information is reduced.

[0056] Deep learning analysis module: important features in imaging data are extracted using convolutional neural network (CNN), time series changes of biomarkers are analyzed using recurrent neural network (RNN), and multi-modal data are comprehensively analyzed by combining multi-layer perceptron (MLP) to finally assess the early risk of heart failure.

[0057] Personalized early warning and real-time monitoring: according to the individual differences (such as gender, age, etc.) of the patient, the weighting value of the data is adjusted to form a personalized early warning model. When the patient's data changes, the early warning result is updated in real time, and treatment suggestions are provided.

[0058] Report output module: the system generates an intuitive and easy-to-understand report according to the analysis result, displays the risk assessment score of heart failure, early warning prompt and further treatment suggestions for reference by clinicians.

[0059] While embodiments of the application have been shown and described, it is to be understood that the application is not limited to the details of the embodiments described, since numerous changes, modifications, substitutions and variations can be made thereto without departing from the spirit and scope of the application as defined by the appended claims and their equivalents.

Claims

1. A biomarker combined analysis system for early warning of heart failure, characterized by: The following steps are involved: S1: Data acquisition module: used to collect multiple biomarker data, clinical data and patient individualized data in real time; S2: Multimodal data fusion module: used to preprocess and fuse biomarker data, imaging data, clinical records, and individual difference data from different sources; S3: Deep learning analysis module: Based on the fusion data, it uses the deep learning model to extract and analyze features and output the early risk assessment results of heart failure; S4: Personalized early warning module: Provides each patient with personalized heart failure risk warning based on deep learning analysis results and individual patient differences; S5: Real-time monitoring module: used to monitor the patient's biomarker data in real time and dynamically adjust the early warning results based on the monitoring data; S6: Report output module: Generates dynamic warning reports based on real-time analysis results to provide warnings and treatment recommendations to clinicians.

2. The biomarker combined analysis system for early warning of heart failure according to claim 1, characterized in that: in, The data acquisition module further includes: multiple biomarker data acquisition units, including BNP, cTnI, myocardial enzymes, renal function indicators, etc.; imaging data acquisition units, including cardiac ultrasound, CT scan, MRI and other data acquisition; individual difference data acquisition units, including patient age, gender, weight, genetic background, living habits and other information collection.

3. The biomarker combined analysis system for early warning of heart failure according to claim 1, characterized in that: in, The multimodal data fusion module further includes: a data preprocessing unit for removing noise data, filling missing data, and standardizing the data; a feature selection unit for selecting the most relevant features through statistical methods or machine learning algorithms and performing dimensionality reduction.

4. The biomarker combined analysis system for early warning of heart failure according to claim 1, characterized in that: in, The deep learning analysis module uses the following deep learning models: convolutional neural network (CNN), which is used to extract important features from imaging data; recurrent neural network (RNN), which is used to process time series data such as the dynamic changes of patients' biomarkers; and multi-layer perceptron (MLP), which is used to integrate the analysis results of multimodal data and output early risk assessment values ​​for heart failure.

5. The biomarker combined analysis system for early warning of heart failure according to claim 1, characterized in that: in, The personalized early warning module achieves personalized early warning through the following methods: individualized data weighting, adjusting the weight of biomarkers based on individual patient characteristics (such as genetic background, lifestyle habits, etc.); Adaptive learning mechanism automatically adjusts the early warning model based on patient historical data and real-time data to improve early warning accuracy.

6. A combined biomarker analysis method for early warning of heart failure, characterized by: The following steps are involved: S1: Real-time collection of multiple biomarker data, imaging data, clinical data, and patient individual difference data; S2: Preprocess the collected data, including noise removal, missing data filling and data standardization; S3: Use deep learning algorithms to extract and fuse features from pre-processed data and output early risk assessment results for heart failure; S4: Provide patients with personalized early warning information and generate reports based on the results of early heart failure risk assessment; S5: By monitoring patient data in real time, early warning information can be adjusted in real time according to dynamic changes and treatment recommendations can be provided.

7. The method for combined analysis of biomarkers for early warning of heart failure according to claim 6, characterized in that: The deep learning analysis algorithm in step S3 includes: Convolutional Neural Network (CNN): used to process imaging data; Recurrent Neural Network (RNN): used to process patient biomarker time series data; Multilayer Perceptron (MLP): used to integrate multimodal data and perform final evaluation.