Intelligent diagnosis auxiliary platform for cardiovascular disease complicated with cognitive impairment

Abstract

The invention relates to the technical field of biomedical engineering, and discloses an intelligent diagnosis auxiliary platform for cardiovascular diseases complicated with cognitive impairment. Comprising a multi-modal data collection module, a multi-modal data analysis module, a multi-modal model construction module, an intelligent diagnosis model training module, a decision support system design module, a clinical trial module and a maintenance and update module. The defects that a traditional cardiovascular disease combined cognitive impairment diagnosis model is low in screening rate, high in diagnosis and treatment technical requirement for doctors and obvious in limitation of diagnosis and treatment modes are overcome, and clinical medical records, digital clock testing, biological markers and electrocardiogram multi-modal data characteristics are fused through the multi-modal technology; a set of objective, accurate and high-operability intelligent diagnosis auxiliary system is developed, and the intelligent auxiliary system finds potential and controllable etiology and risk factors through early screening of cardiovascular disease patients, so that early intervention is performed, and the progress of cognitive impairment is effectively delayed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of biomedical engineering, in particular to an intelligent diagnosis auxiliary platform for cardiovascular disease combined with cognitive impairment. BACKGROUND

[0002] Cardiovascular disease and cognitive impairment are major diseases that lead to increasing global disease burden and are the main causes of death in urban and rural residents. Their prevention and treatment are of great significance to public health. Cardiovascular disease significantly increases the risk of cognitive impairment, and cognitive impairment increases the disease burden of patients and affects disease prognosis, forming a vicious cycle of interaction.

[0003] In view of the current situation that there is no specific evidence-based strategy for screening of cardiovascular disease combined with cognitive impairment at home and abroad, and the vast majority of cardiology clinics in China do not conduct systematic cognitive function assessment on patients, resulting in significant constraints on early identification and accurate diagnosis. Existing diagnostic screening tools, such as the Mini-Mental State Examination and the Montreal Cognitive Assessment, have obvious limitations in sensitivity and specificity in cognitive impairment screening. Screening methods that rely on patient self-reporting or family feedback may miss patients with severe cognitive impairment who are not effectively detected. In addition, domestic and foreign studies on the prevalence of patients with cardiovascular disease combined with cognitive impairment are 11% to 85%, with large differences in results. Therefore, there is an urgent need to establish an objective, accurate, and operable cardiovascular disease combined with cognitive impairment screening and diagnosis auxiliary system with strong generalizability, so as to intervene as soon as possible.

[0004] In recent years, intelligent diagnosis auxiliary systems have rapidly developed in the medical field, relying on artificial intelligence, machine learning, and big data analysis technologies, which can objectively diagnose and identify potential health risks. Currently, commonly used intelligent cognitive assessment tools abroad include serious games, iVitality online research platform, medical devices for cognitive screening and monitoring, digital clock test (DCTclock), etc. In China, there are XX version of vigilance and memory test, electronic cognitive screen, and cognitive impairment risk rapid screening tool, etc. Among them, DCTclock is an internationally recognized neuropsychological screening test, which has the advantages of convenience, accuracy, and efficiency, and is an effective screening tool for distinguishing between normal elderly people and patients with cognitive impairment. However, these intelligent assessment tools mainly target cognitive impairment in the elderly and focus on distinguishing between mild cognitive impairment and dementia. There is no intelligent diagnosis auxiliary system specifically for cardiovascular disease combined with cognitive impairment.

[0005] The present application develops an intelligent diagnosis auxiliary system for cardiovascular disease combined with cognitive impairment, which helps medical personnel to make more accurate diagnoses and assist in decision-making through multi-modal data fusion technology, and improves early screening and diagnosis capabilities. Thus, the early screening and diagnosis capabilities for cardiovascular disease combined with cognitive impairment are comprehensively improved. SUMMARY

[0006] (1) Technical problems solved

[0007] In response to the shortcomings of existing technologies, the present invention provides an intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment, which has the advantages of screening for combined cognitive impairment, as well as early diagnosis and intervention, and solves the problem of significantly limited sensitivity and specificity of traditional scale diagnostic methods.

[0008] (2) Technical solution

[0009] To achieve the above objectives, the present invention provides the following technical solutions: an intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment, comprising a multimodal data collection module, a multimodal data analysis module, a multimodal model construction module, an intelligent diagnostic model training module, a decision support system design module, a clinical trial module, and a maintenance and update module;

[0010] The multimodal data collection module is responsible for collecting and preprocessing multimodal data, where the multimodal data sources include medical records, vital signs, electrocardiograms, biological markers, digital clock drawing tests, and patient-reported symptoms;

[0011] The multimodal data analysis module is responsible for providing a calculation formula to calculate the data collected in the multimodal data collection module, and at the same time inputting the digital clock image and the electrocardiogram image into the pre-trained deep convolutional neural network model, using the pre-trained deep convolutional neural network model as the basic framework of the multimodal model construction module, and the calculation formula as the calculation basis for the construction of the multimodal model;

[0012] The multimodal model building module uses a deep convolutional neural network to extract image features of digital clock images and electrocardiograms, uses a language model to extract key information from doctors' medical records and patients' self-reports, and uses a multimodal interaction network to fuse information from different modalities, and jointly generates a unified multimodal feature vector representation with a calculation formula;

[0013] The intelligent diagnosis model training module fine-tunes the pre-trained large model to adapt to the multimodal data analysis in this scenario, performs model selection and training, and performs model verification, and constructs an intelligent diagnosis model based on the calculation results;

[0014] The decision support system design module is responsible for designing the data input module, diagnosis engine, personalized decision suggestion module, user interface and visual construction decision support system architecture;

[0015] The clinical trial module is integrated into the hospital information system, system evaluation and optimization, and physician training and trial;

[0016] The maintenance and update module is responsible for model updates, user feedback and iteration.

[0017] Preferably, the multi-modal data collection module comprises a medical record information collection unit, a physiological data acquisition unit and a biological marker acquisition unit, the medical record information collection unit acquires medical record data by structured processing and information extraction of medical record texts, the physiological data acquisition unit provides physiological data by acquiring high-quality electrocardiogram waveforms through professional equipment, and the biological marker acquisition unit acquires biological data by testing biological marker samples of the patient, and the medical record information collection unit, the physiological data acquisition unit and the biological marker acquisition unit are connected with the multi-modal data analysis module through the network after numbering the internal data of the units.

[0018] Preferably, the medical record information collection unit numbers the probability of occurrence of cardiovascular diseases in the patient's family, the probability of occurrence of cardiovascular diseases in the patient's family, the probability of occurrence of diabetes in the patient's family and the probability of occurrence of hypertension in the patient's family according to the characteristics of the medical record data, and the probability of occurrence of cardiovascular diseases in the patient's family, the probability of occurrence of cardiovascular diseases in the patient's family, the probability of occurrence of diabetes in the patient's family and the probability of occurrence of hypertension in the patient's family are numbered as X1, H1, J1 and L1, respectively.

[0019] Preferably, the physiological data acquisition unit numbers the probability of abnormal diagnosis of electrocardiogram waveforms of the patient, the digital clock image features and the image features of electrocardiogram according to the characteristics of the physiological data, and the probability of abnormal diagnosis of electrocardiogram waveforms of the patient, the digital clock image features and the image features of electrocardiogram are numbered as X2, F t , F d .

[0020] Preferably, the biological marker acquisition unit numbers the probability of exceeding the concentration of low-density lipoprotein cholesterol in the blood of the patient according to the characteristics of the biological data, and the probability of exceeding the concentration of low-density lipoprotein cholesterol in the blood of the patient is numbered as X3.

[0021] Preferably, the multi-modal data analysis module comprises a diagnosis probability calculation unit, a multi-modal feature fusion unit and a risk assessment unit.

[0022] Preferably, the diagnosis probability calculation unit calculates the probability Gv of the patient suffering from cardiovascular diseases combined with cognitive impairment according to the medical record data, the physiological data and the biological data, and the calculation formula is:

[0023]

[0024] In the formula, Gv represents the probability that a patient suffers from cardiovascular disease combined with cognitive impairment, X1 represents the probability of cardiovascular disease occurring in the patient's family, X2 represents the probability of the patient's electrocardiogram waveform being diagnosed abnormally, X3 represents the probability that the patient's low-density lipoprotein cholesterol concentration in the blood exceeds the standard, w1, w2, and w3 represent the proportion of the probability of cardiovascular disease occurring in the patient's family, the probability of the patient's electrocardiogram waveform being diagnosed abnormally, and the probability of the patient's low-density lipoprotein cholesterol concentration in the blood exceeding the standard in the probability of the patient suffering from cardiovascular disease combined with cognitive impairment, respectively. b represents a constant, which serves as the intercept term of the model.

[0025] Preferably, the multimodal feature fusion unit calculates the multimodal feature fusion index Sz based on the physiological data, and the calculation formula is:

[0026] Sz=F t *y t +F d *y d

[0027] In the formula, Sz represents the multimodal feature fusion index, F t represents the characteristics of the digital clock image, F d Represents the image features of the electrocardiogram, y t 、y d They respectively represent the weights of the digital clock image features and the electrocardiogram image features in the multimodal feature fusion index.

[0028] Preferably, the risk assessment unit calculates the patient's disease risk score Qr based on the medical record data, and the calculation formula is:

[0029] Q r={X1*Xa1+H1*Ha1+J1*Ja1+L1*La1}*100

[0030] In the formula, Qr represents the risk score of the patient's disease, X1 represents the probability of cardiovascular disease in the patient's family, H1 represents the probability of cardiovascular disease in the patient's family, J1 represents the probability of diabetes in the patient's family, L1 represents the probability of hypertension in the patient's family, Xa1, Ha1, Ja1, and La1 represent the proportion of the probability of cardiovascular disease in the patient's family, the probability of cardiovascular disease in the patient's family, the probability of diabetes in the patient's family, and the probability of hypertension in the patient's family in the patient's disease risk score, respectively.

[0031] Preferably, the probability Gv of the patient suffering from cardiovascular disease combined with cognitive impairment is used to evaluate the risk of the individual suffering from the disease, to provide a quantitative basis for clinical decision-making, the multi-modal feature fusion index Sz is used to comprehensively evaluate the physiological and pathological state of the patient, to provide a reference for personalized treatment, and the risk score Qr of the disease of the patient is used to identify high-risk patients, to help doctors take targeted prevention and intervention measures.

[0032] Compared with the prior art, the intelligent diagnosis auxiliary platform for cardiovascular disease combined with cognitive impairment provided by the application has the following beneficial effects:

[0033] 1. The application breaks through the shortcomings of the traditional cardiovascular disease combined with cognitive impairment diagnosis model, such as low screening rate, high requirement for doctor's diagnosis and treatment technology, and obvious limitation of diagnosis and treatment method, through the mutual cooperation of the multi-modal data collection module, the multi-modal model construction module, the intelligent diagnosis model training module, the decision support system design module, the clinical trial module, and the maintenance and update module, and develops an objective, accurate, operable, and highly generalizable intelligent diagnosis auxiliary system by fusing the multi-modal data features of clinical medical records, digital clock test, biological markers, and electrocardiogram. The intelligent auxiliary system can effectively delay the progression of cognitive impairment by early screening of cardiovascular disease patients, discovering potential and controllable causes and risk factors, and then early intervention.

[0034] 2. The probability Gv of the patient suffering from cardiovascular disease combined with cognitive impairment calculated by the application can be used to evaluate the risk of the individual suffering from the disease, to provide a quantitative basis for clinical decision-making, and the multi-modal feature fusion index Sz can help to more comprehensively understand the physiological and pathological state of the patient, to provide a reference for personalized treatment. According to the risk score Qr of the disease of the patient calculated by the risk assessment unit, doctors can identify high-risk patients and take targeted prevention and intervention measures, thereby reducing the risk of disease occurrence and development. Through the cooperative work of the above calculation formula, the intelligent diagnosis auxiliary platform can comprehensively utilize multi-modal data to improve the diagnostic accuracy of cardiovascular disease associated cognitive impairment.

[0035] 3. The application improves the diagnosis and treatment ability of clinical doctors for cardiovascular disease patients combined with cognitive impairment by evaluating the objectivity and accuracy of the cardiovascular disease combined with cognitive impairment diagnosis model, thereby improving the clinical diagnosis and treatment efficiency, saving manpower, time and training, reducing the consumption of social medical resources, preventing the progression of cognitive impairment of cardiovascular disease patients, reducing the burden of major diseases, and achieving multiple benefits of economy, society and ecology. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 The accompanying drawings illustrate the structure of the application. DETAILED DESCRIPTION

[0037] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0038] Please refer to Figure 1 , the intelligent diagnosis auxiliary platform for cardiovascular disease combined with cognitive impairment includes a multi-modal data collection module, a multi-modal data analysis module, a multi-modal model construction module, an intelligent diagnosis model training module, a decision support system design module, a clinical trial module, a maintenance and update module;

[0039] The multi-modal data collection module is responsible for collecting and preprocessing multi-modal data. The multi-modal data sources include medical record information, vital signs, electrocardiogram, biological markers, digital clock test and patient-reported symptoms.

[0040] The multi-modal data analysis module is responsible for providing calculation formulas to calculate the data collected in the multi-modal data collection module. Meanwhile, the digital clock image and electrocardiogram image are input into the pre-trained deep convolutional neural network model. The pre-trained deep convolutional neural network model is used as the basic framework of the multi-modal model construction module, and the calculation formulas are used as the calculation basis of the multi-modal model construction. After multi-layer convolution and pooling operation, the image feature vector is extracted. After preprocessing (word segmentation, part-of-speech tagging, etc.) of the text data, the BERT model is input to obtain the text semantic representation. Finally, the image feature vector and the text feature vector are fused through the multi-modal fusion network.

[0041] The multi-modal model construction module uses deep convolutional neural network to extract the image features of digital clock image and electrocardiogram. The language model is used to extract the key information in the doctor's medical record and the patient's self-report. The multi-modal interaction network is used to fuse the information of different modalities, and the unified multi-modal feature vector representation is generated together with the calculation formula.

[0042] The intelligent diagnosis model training module fine-tunes the pre-trained large model to adapt to the multi-modal data analysis in this scenario, selects and trains the model, and verifies the model. According to the calculation result, the intelligent diagnosis model is constructed.

[0043] The decision support system design module is responsible for designing the data input module, the diagnosis engine, the personalized decision suggestion module, the user interface and the visual construction of the decision support system architecture.

[0044] The clinical trial module is integrated into the hospital information system, system evaluation and optimization, and doctor training and trial.

[0045] The maintenance and update module is responsible for model updating, user feedback and iteration.

[0046] Advantages: the application breaks through the shortcomings of low screening rate of traditional cardiovascular disease combined with cognitive impairment diagnosis model, high requirement for doctor's diagnosis and treatment technology, and obvious limitation of diagnosis and treatment mode, through the mutual cooperation of the multi-modal data collection module, the multi-modal model construction module, the intelligent diagnosis model training module, the decision support system design module, the clinical trial module and the maintenance and update module, a set of objective, accurate, operable and strong generalizable intelligent diagnosis auxiliary system is developed by fusing the clinical medical record, the digital clock test, the biological marker and the electrocardiogram multi-modal data characteristics, the intelligent auxiliary system can effectively delay the progression of cognitive impairment through early screening of cardiovascular disease patients, finding potential and controllable causes and risk factors, and then early intervention.

[0047] The multi-modal data collection module includes a medical record information collection unit, a physiological data acquisition unit and a biological marker acquisition unit, the medical record information collection unit obtains medical record data through structured processing and information extraction of medical record texts, the physiological data acquisition unit obtains high-quality electrocardiogram waveform graphs through professional equipment to provide physiological data, which provides important clues for the diagnosis of heart disease, and the biological marker acquisition unit obtains biological data by testing the biological marker samples of the patient, the medical record information collection unit, the physiological data acquisition unit and the biological marker acquisition unit number the internal data and connect with the multi-modal data analysis module through the network.

[0048] The medical record information collection unit numbers the probability of occurrence of cardiovascular disease in the patient's family, the probability of occurrence of cardiovascular disease in the patient's family, the probability of occurrence of diabetes in the patient's family and the probability of occurrence of hypertension in the patient's family according to the characteristics of the medical record data, and the probability of occurrence of cardiovascular disease in the patient's family, the probability of occurrence of cardiovascular disease in the patient's family, the probability of occurrence of diabetes in the patient's family and the probability of occurrence of hypertension in the patient's family are numbered as X1, H1, J1 and L1 respectively.

[0049] The physiological data acquisition unit numbers the probability of abnormal diagnosis of the patient's electrocardiogram waveform graph, the digital clock image feature and the image feature of the electrocardiogram according to the characteristics of the physiological data, and the probability of abnormal diagnosis of the patient's electrocardiogram waveform graph, the digital clock image feature and the image feature of the electrocardiogram are numbered as X2, F t , F d .

[0050] The biological marker acquisition unit numbers the probability of over-standard concentration of low-density lipoprotein (LDL) cholesterol in the patient's blood according to the characteristics of the biological data, and the probability of over-standard concentration of low-density lipoprotein (LDL) cholesterol in the patient's blood is numbered as X3.

[0051] The multi-modal data analysis module comprises a diagnosis probability calculation unit, a multi-modal feature fusion unit and a risk assessment unit.

[0052] The diagnosis probability calculation unit calculates the probability Gv of the patient suffering from cardiovascular disease combined with cognitive impairment according to the medical record data, physiological data and biological data, and the calculation formula is:

[0053]

[0054] In the formula, Gv represents the probability of the patient suffering from cardiovascular disease combined with cognitive impairment, X1 represents the probability of the patient's family suffering from cardiovascular disease, X2 represents the probability of the patient's electrocardiogram waveform abnormal diagnosis, X3 represents the probability of the patient's blood low-density lipoprotein (LDL) cholesterol concentration exceeding the standard, w1, w2 and w3 respectively represent the proportion of the probability of the patient's family suffering from cardiovascular disease, the probability of the patient's electrocardiogram waveform abnormal diagnosis and the probability of the patient's blood low-density lipoprotein (LDL) cholesterol concentration exceeding the standard in the probability of the patient suffering from cardiovascular disease combined with cognitive impairment, and b represents a constant as an intercept term of the model.

[0055] The diagnosis probability calculation unit sets a probability threshold according to the medical record data, physiological data and biological data. When the probability Gv of the patient suffering from cardiovascular disease combined with cognitive impairment is higher than the probability threshold, the family members of the patient should be given detailed physical examination, preventive treatment of cardiovascular disease combined with cognitive impairment should be carried out, and the monitoring frequency of the patient's physical physiological data should be increased.

[0056] The multi-modal feature fusion unit calculates a multi-modal feature fusion index Sz according to the physiological data, and the calculation formula is:

[0057] Sz=F t *y t +F d *y d

[0058] In the formula, Sz represents the multi-modal feature fusion index, F t represents the image feature of the digital clock image, F d represents the image feature of the electrocardiogram, y t and y d respectively represent the weight of the digital clock image feature and the image feature of the electrocardiogram in the multi-modal feature fusion index.

[0059] Low risk: when the multi-modal feature fusion index Sz<0.5, it is considered that the risk of the patient is low, and special measures do not need to be taken immediately;

[0060] Medium risk: when 0.5≤ multimodal feature fusion index Sz<0.75, the risk of the patient is considered to be medium, and regular monitoring and follow-up are recommended;

[0061] High risk: when multimodal feature fusion index Sz≥0.75, the risk of the patient is considered to be high, and more active medical intervention measures need to be taken.

[0062] Advantages: through the analysis and evaluation of the range of calculation results, early screening of cardiovascular disease patients is realized, potential and controllable causes and risk factors are found, and early intervention is realized, which can effectively delay the progression of cognitive impairment.

[0063] The risk assessment unit calculates the risk score Qr of the patient's disease according to the medical record data, and the calculation formula is:

[0064] Qr={X1*Xa1+H1*Ha1+J1*Ja1+L1*La1}*100

[0065] In the formula, Qr represents the risk score of the patient's disease, X1 represents the probability of cardiovascular disease in the patient's family, H1 represents the probability of cardiovascular disease in the patient's family, J1 represents the probability of diabetes in the patient's family, L1 represents the probability of hypertension in the patient's family, Xa1, Ha1, Ja1, and La1 represent the proportion of the probability of cardiovascular disease in the patient's family, the probability of cardiovascular disease in the patient's family, the probability of diabetes in the patient's family, and the proportion of the probability of hypertension in the patient's family in the risk score of the patient's disease.

[0066] When the risk score Qr of the patient's disease is higher than 60, the patient needs to take corresponding prevention or treatment measures, including adjusting the lifestyle, regular medical monitoring, and starting drug treatment to reduce the risk of cardiovascular disease.

[0067] The probability Gv of the patient suffering from cardiovascular disease combined with cognitive impairment is used to assess the individual risk of the disease and provide quantitative basis for clinical decision-making, the multimodal feature fusion index Sz is used to comprehensively evaluate the physiological and pathological state of the patient and provide reference for personalized treatment, and the risk score Qr of the patient's disease is used to identify high-risk patients and help doctors take targeted prevention and intervention measures, thereby reducing the risk of disease occurrence and development.

[0068] Advantages are: the probability Gv of the patient suffering from cardiovascular disease combined with cognitive impairment obtained by calculation can be used to evaluate the risk of individual disease, provide quantitative basis for clinical decision, according to the multi-modal feature fusion index Sz, it is helpful to more comprehensively understand the physiological and pathological state of the patient, and provide reference for personalized treatment, according to the risk score Qr of the disease of the patient calculated by the risk assessment unit, it can help doctors to identify high-risk patients and take targeted prevention and intervention measures, so as to reduce the risk of occurrence and development of the disease, through the cooperative work of the above calculation formula, the intelligent diagnosis auxiliary platform can comprehensively utilize multi-modal data, improve the diagnosis accuracy of cardiovascular disease associated cognitive impairment, the platform integrates a variety of technologies and methods, aiming at providing more accurate and comprehensive support for the diagnosis and treatment of cardiovascular disease associated cognitive impairment, which is helpful to improve the prognosis and quality of life of patients.

[0069] Advantages are: the present application improves the diagnosis and treatment ability of clinicians for cardiovascular disease patients combined with cognitive impairment by evaluating the objectivity and accuracy of the cardiovascular disease combined with cognitive impairment diagnosis model, thereby improving the clinical diagnosis and treatment efficiency, saving manpower, time and training, reducing the consumption of social medical resources, preventing and treating the progress of cognitive impairment of cardiovascular disease patients, reducing the burden of major diseases, thereby realizing the multiple benefits of economy, society and ecology.

[0070] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. Intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment, characterized by: It includes multimodal data collection module, multimodal data analysis module, multimodal model construction module, intelligent diagnosis model training module, decision support system design module, clinical trial module, and maintenance and update module; The multimodal data collection module is responsible for collecting and preprocessing multimodal data, where the multimodal data sources include medical records, vital signs, electrocardiograms, biological markers, digital clock drawing tests, and patient-reported symptoms; The multimodal data analysis module is responsible for providing a calculation formula to calculate the data collected in the multimodal data collection module, and at the same time inputting the digital clock image and the electrocardiogram image into the pre-trained deep convolutional neural network model, using the pre-trained deep convolutional neural network model as the basic framework of the multimodal model construction module, and the calculation formula as the calculation basis for the construction of the multimodal model; The multimodal model building module uses a deep convolutional neural network to extract image features of digital clock images and electrocardiograms, uses a language model to extract key information from doctors' medical records and patients' self-reports, and uses a multimodal interaction network to fuse information from different modalities, and jointly generates a unified multimodal feature vector representation with a calculation formula; The intelligent diagnosis model training module fine-tunes the pre-trained large model to adapt to the multimodal data analysis in this scenario, performs model selection and training, and performs model verification, and constructs an intelligent diagnosis model based on the calculation results; The decision support system design module is responsible for designing the data input module, diagnosis engine, personalized decision suggestion module, user interface and visual construction decision support system architecture; The clinical trial module is integrated into the hospital information system, system evaluation and optimization, and physician training and trial; The maintenance and update module is responsible for model updates, user feedback and iteration.

2. The intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment according to claim 1, characterized in that: The multimodal data collection module includes a medical record information collection unit, a physiological data collection unit, and a biological marker collection unit. The medical record information collection unit obtains medical record data by structured processing and information extraction of medical record texts. The physiological data collection unit obtains high-quality electrocardiogram waveforms using professional equipment to provide physiological data. The biological marker collection unit obtains biological data by testing the patient's biological marker samples. After the medical record information collection unit, physiological data collection unit, and biological marker collection unit number the data within the unit, they are connected to the multimodal data analysis module via a network.

3. The intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment according to claim 2, characterized in that: The medical record information collection unit numbers the probability of cardiovascular disease occurring in the patient's family, the probability of cardiovascular disease occurring in the patient's family, the probability of diabetes occurring in the patient's family, and the probability of hypertension occurring in the patient's family according to the characteristics of the medical record data. The probability of cardiovascular disease occurring in the patient's family, the probability of cardiovascular disease occurring in the patient's family, the probability of diabetes occurring in the patient's family, and the probability of hypertension occurring in the patient's family are numbered X1, H1, J1, and L1, respectively.

4. The intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment according to claim 2, characterized in that: The physiological data acquisition unit numbers the probability of abnormal diagnosis of the patient's electrocardiogram waveform, the digital clock image feature, and the electrocardiogram image feature according to the physiological data features. The probability of abnormal diagnosis of the patient's electrocardiogram waveform, the digital clock image feature, and the electrocardiogram image feature are numbered as X2, F t 、F d .

5. The intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment according to claim 2, characterized in that: The biological marker collection unit numbers the probability of the patient's blood low-density lipoprotein cholesterol concentration exceeding the standard according to the biological data characteristics, and the probability of the patient's blood low-density lipoprotein cholesterol concentration exceeding the standard is numbered X3.

6. The intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment according to claim 1, characterized in that: The multimodal data analysis module includes a diagnosis probability calculation unit, a multimodal feature fusion unit and a risk assessment unit.

7. The intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment according to claim 6, characterized in that: The diagnostic probability calculation unit calculates the probability Gv that the patient suffers from cardiovascular disease combined with cognitive impairment based on the medical record data, physiological data and biological data. The calculation formula is: In the formula, Gv represents the probability that a patient suffers from cardiovascular disease combined with cognitive impairment, X1 represents the probability of cardiovascular disease occurring in the patient's family, X2 represents the probability of the patient's electrocardiogram waveform being diagnosed abnormally, X3 represents the probability that the patient's low-density lipoprotein cholesterol concentration in the blood exceeds the standard, w1, w2, and w3 represent the proportion of the probability of cardiovascular disease occurring in the patient's family, the probability of the patient's electrocardiogram waveform being diagnosed abnormally, and the probability of the patient's low-density lipoprotein cholesterol concentration in the blood exceeding the standard in the probability of the patient suffering from cardiovascular disease combined with cognitive impairment, respectively. b represents a constant, which serves as the intercept term of the model.

8. The intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment according to claim 6, characterized in that: The multimodal feature fusion unit calculates the multimodal feature fusion index Sz based on the physiological data, and the calculation formula is: Sz=F t *y t +F d *y d In the formula, Sz represents the multimodal feature fusion index, F t represents the characteristics of the digital clock image, F d Represents the image features of the electrocardiogram, y t 、y d They respectively represent the weights of the digital clock image features and the electrocardiogram image features in the multimodal feature fusion index.

9. The intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment according to claim 6, characterized in that: The risk assessment unit calculates the patient's disease risk score Qr based on the medical record data, and the calculation formula is: Q r={X1*Xa1+H1*Ha1+J1*Ja1+L1*La1}*100 In the formula, Qr represents the risk score of the patient's disease, X1 represents the probability of cardiovascular disease in the patient's family, H1 represents the probability of cardiovascular disease in the patient's family, J1 represents the probability of diabetes in the patient's family, L1 represents the probability of hypertension in the patient's family, Xa1, Ha1, Ja1, and La1 represent the proportion of the probability of cardiovascular disease in the patient's family, the probability of cardiovascular disease in the patient's family, the probability of diabetes in the patient's family, and the probability of hypertension in the patient's family in the patient's disease risk score, respectively.

10. The intelligent diagnostic assistance platform for cardiovascular disease combined with cognitive impairment according to claim 9, characterized in that: The probability Gv that a patient suffers from cardiovascular disease combined with cognitive impairment is used to assess individual disease risk and provide a quantitative basis for clinical decision-making. The multimodal feature fusion index Sz is used to comprehensively assess the patient's physiological and pathological status and provide a reference for personalized treatment. The patient's disease risk score Qr is used to identify high-risk patients and help doctors take targeted prevention and intervention measures.

Images