Gastric cancer risk detection device based on big data analysis technology

Abstract

The invention provides a gastric cancer risk detection device based on a big data analysis technology, and relates to the field of medical treatment. The gastric cancer disease risk detection device based on the big data analysis technology comprises a data acquisition module which comprises multi-modal data acquisition including genome data, metabolome data, iconography data, environmental data and electronic health records, and is used for acquiring genome data of a user, especially gene mutation related to gastric cancer, and detecting the risk of the gastric cancer disease according to the genome data, the metabolome data, the iconography data, the environmental data and the electronic health records. The method comprises the following steps of: acquiring a plurality of health records of a patient, performing gene data acquisition by using gene sequencing equipment, transmitting metabolome data into a system, associating the metabolome data with other health data of the user, uploading the metabolome data to the system through a network, and acquiring medical record data of the patient from medical records of a hospital, including past medical histories, family medical histories, medication records and the like, by using the electronic health records. By arranging the data acquisition module, the accuracy of gastric cancer risk assessment is improved, and by arranging the data storage module, safe sharing and traceability of data are realized, and the authenticity and reliability of the data are ensured.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a gastric cancer disease risk detection device based on big data analysis technology. Background Art

[0002] With the continuous development of artificial intelligence, artificial intelligence technology has also been continuously introduced into the field of medical imaging. Artificial intelligence technology can automatically screen for existing diseased tissues based on medical images, thereby assisting doctors in diagnosis. This can not only save doctors' labor and increase detection speed, but also effectively reduce the missed diagnosis rate.

[0003] Although the existing technology detects the potential risk of gastric cancer, the vast majority of gastric cancers are adenocarcinomas, which have no obvious symptoms in the early stages, or present with non-specific symptoms such as upper abdominal discomfort and belching, which are often similar to the symptoms of chronic gastric diseases such as gastritis and gastric ulcers. This problem is easily overlooked, resulting in a low early diagnosis and detection rate for gastric cancer. Therefore, a person skilled in the art provides a gastric cancer disease risk detection device based on big data analysis technology to solve the problems raised in the above background technology. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the shortcomings of the prior art, the present invention provides a gastric cancer disease risk detection device based on big data analysis technology, which solves the problem of analyzing gastric cancer disease by integrating multiple data and using big data for comprehensive analysis during the diagnosis process.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a gastric cancer disease risk detection device based on big data analysis technology, including a data acquisition module, including multimodal data acquisition, the multimodal data acquisition includes genomic data, metabolomics data, imaging data, environmental data and electronic health records, the genomic data is obtained by cooperating with a gene testing agency to obtain the user's genomic data, especially gene mutations related to gastric cancer, and gene sequencing equipment is used to collect gene data, and the data is transmitted to the system, the metabolomics data is collected by collecting biological samples such as blood and urine of the user, and metabolites are analyzed by a mass spectrometer, and the metabolomics data is transmitted to the system and associated with other health data of the user, the imaging data is obtained by using CT and MRI to obtain the user's stomach imaging data, and the data is uploaded to the system, the environmental data is obtained by entering environmental data such as air quality and dietary intake collected at the patient's home or workplace, and uploading the data to the system through the network, and the electronic health record obtains the patient's medical record data from the hospital's medical record, including past medical history, family medical history, medication records, etc.;

[0008] A data preprocessing module, including data cleaning and standardization and multimodal data fusion, wherein the data cleaning and standardization include missing value processing, outlier detection and data standardization;

[0009] Data storage module, including data encryption and storage and data sharing;

[0010] Feature processing module, including feature extraction and feature selection;

[0011] Risk assessment and prediction module, including personalized risk prediction model and real-time risk assessment;

[0012] Diagnosis and decision-making module, including intelligent diagnostic assistance and treatment plan recommendations.

[0013] Preferably, the data preprocessing module processes missing data by using the mean filling method, and for missing key data, after marking, reminds the user to supplement it in time, and then uses the Z-score method to detect outliers, and corrects or deletes them, and standardizes the obtained different types of data. Finally, the multimodal fusion technology in deep learning is used to fuse genomic data, metabolomics data, imaging data and electronic health records, and through feature alignment and feature selection, the correlation and consistency between different modal data are ensured. The following is the formula used in the data preprocessing process:

[0014] Mean filling method:

[0015]

[0016] Among them, x i is a missing value, x j is the number of observations with non-missing values, and n is the number of non-missing values;

[0017] Z-score method:

[0018]

[0019] Where x is the observed value, μ is the mean, σ is the standard deviation, and |Z|>3 is usually considered an outlier.

[0020] Preferably, the data storage module uses blockchain technology to encrypt, store and transmit the patient's data, and utilizes smart contract technology to ensure the security and integrity of the data. At the same time, a data sharing platform is established to integrate data resources from different sources, and secure data sharing is achieved through blockchain technology, allowing different medical institutions and scientific research institutions to share data while protecting the privacy of patients.

[0021] Preferably, the risk assessment and prediction module trains a personalized risk prediction model by using XGBoost and deep neural networks, and uses transfer learning to apply the knowledge learned from large-scale data sets to specific individuals to generate personalized risk predictions, evaluate the accuracy, precision, recall and F1-score indicators of the obtained model data, optimize the model according to the obtained results, and finally process and analyze the user's data in real time according to the improved model, and perform risk assessment on the patient. When abnormal data or high risk is detected, the system triggers an early warning to remind the user to conduct further inspections or take preventive measures. The following are the formulas needed in the model building process:

[0022] XGBoost:

[0023]

[0024] Among them, l is the loss function, is the predicted value, Ω(f k ) is the regularization term;

[0025] Deep Neural Networks:

[0026]

[0027] Among them, W is the weight matrix, b is the bias, and σ is the activation function;

[0028] Accuracy:

[0029]

[0030] Accuracy:

[0031]

[0032] Recall:

[0033]

[0034] F1-score:

[0035]

[0036] Among them, TP is true positive, TN is true negative, FP is false positive, and FN is false negative.

[0037] Preferably, the feature processing module collects gene mutations and genetic markers related to gastric cancer in patients, extracts imaging features such as lesion size, shape, density, and extracts features such as past medical history, family medical history, and medication records, and uses the LASSO regression method to screen the above data to screen out features that have an important impact on gastric cancer risk prediction.

[0038] Preferably, the diagnosis and decision-making module uses a deep learning algorithm to automatically analyze imaging data, assist doctors in identifying and diagnosing gastric cancer, and provide diagnostic recommendations and lesion quantitative analysis to help doctors make more accurate diagnoses. It also recommends personalized treatment plans based on the patient's medical records and the latest medical research. Finally, it uses natural language processing technology to analyze the doctor's diagnostic records and the patient's medical records to automatically generate diagnostic reports and treatment recommendations.

[0039] Preferably, the feature processing module needs to use the following formula in the process of data screening:

[0040] LASSO regression:

[0041]

[0042] Among them, y i is the target variable, x ij is a feature, β j is the regression coefficient and λ is the regularization parameter.

[0043] (III) Beneficial effects

[0044] The present invention provides a gastric cancer disease risk detection device based on big data analysis technology. It has the following beneficial effects:

[0045] 1. By setting up a data acquisition module and integrating genomic data, metabolomic data, imaging data and electronic health records, multimodal data fusion can provide a more comprehensive analysis of the patient's health status, thereby improving the accuracy of gastric cancer risk assessment. At the same time, multimodal data can help identify individualized risk factors and potential treatment targets, and support personalized treatment plans.

[0046] 2. By setting up a data storage module and using blockchain technology through data encryption and smart contracts, the security and privacy of patient data can be guaranteed and data leakage can be prevented. Blockchain technology can realize the secure sharing and traceability of data, ensure the authenticity and reliability of data, and promote the effective use of medical data.

[0047] 3. By setting up a risk assessment and prediction module and training a personalized risk prediction model using the XGBoost method, customized risk assessment can be provided based on the individual's specific risk factors. This personalized prediction helps to formulate more effective prevention and treatment strategies. At the same time, the model can be used to conduct real-time risk assessment of user data and provide early warning information in a timely manner, which is crucial for early detection and intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of the structure of the present invention;

[0049] Figure 2 It is a structural schematic diagram of the data acquisition module of the present invention;

[0050] Figure 3 It is a structural schematic diagram of the data preprocessing module of the present invention;

[0051] Figure 4 It is a schematic diagram of the structure of the risk assessment and prediction module of the present invention;

[0052] Figure 5 It is a structural schematic diagram of the data storage module of the present invention. DETAILED DESCRIPTION

[0053] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] Embodiment 1:

[0055] like Figure 1-5 As shown, an embodiment of the present invention provides a gastric cancer disease risk detection device based on big data analysis technology, including a data acquisition module, including multimodal data acquisition, the multimodal data acquisition includes genomic data, metabolomics data, imaging data, environmental data and electronic health records, the genomic data is obtained by cooperating with a genetic testing agency to obtain the user's genomic data, especially gene mutations related to gastric cancer, and using a gene sequencing device to collect gene data, and transmit the data to the system, the metabolomics data is collected by collecting biological samples such as blood and urine of the user, using a mass spectrometer to analyze metabolites, and the metabolomics data is transmitted to the system and associated with other health data of the user, and the imaging data is obtained by using CT and MRI obtain the user's gastric imaging data and upload the data to the system. Environmental data is obtained by entering environmental data such as air quality and dietary intake collected at the patient's home or workplace and uploading it to the system through the Internet. Electronic health records obtain the patient's medical record data from the hospital's medical records, including past medical history, family medical history, medication records, etc. By setting up a data collection module and integrating genomic data, metabolomics data, imaging data and electronic health records, multimodal data fusion can provide a more comprehensive analysis of the patient's health status, thereby improving the accuracy of gastric cancer risk assessment. At the same time, multimodal data can help identify individualized risk factors and potential treatment targets, and support personalized treatment plans.

[0056] The data preprocessing module includes data cleaning and standardization and multimodal data fusion, wherein data cleaning and standardization include missing value processing, outlier detection and data standardization.

[0057] Data storage module, including data encryption and storage and data sharing.

[0058] Feature processing module, including feature extraction and feature selection.

[0059] Risk assessment and prediction module, including personalized risk prediction model and real-time risk assessment.

[0060] Diagnosis and decision-making module, including intelligent diagnostic assistance and treatment plan recommendations.

[0061] The data preprocessing module handles missing data by using the mean filling method, and for missing key data, it marks and reminds users to supplement them in time. Secondly, it uses the Z-score method to detect outliers, and corrects or deletes them. The different types of data obtained are standardized. Finally, the multimodal fusion technology in deep learning is used to fuse genomic data, metabolomics data, imaging data and electronic health records, and through feature alignment and feature selection, the correlation and consistency between different modal data are ensured. The following are the formulas used in the data preprocessing process:

[0062] Mean filling method:

[0063]

[0064] Among them, x i is a missing value, x j is the number of observations with non-missing values ​​and n is the number of non-missing values.

[0065] Z-score method:

[0066]

[0067] Where x is the observed value, μ is the mean, σ is the standard deviation, and |Z|>3 is usually considered an outlier.

[0068] The data storage module uses blockchain technology to encrypt, store and transmit patient data, and uses smart contract technology to ensure the security and integrity of the data. At the same time, it establishes a data sharing platform to integrate data resources from different sources, and realizes secure data sharing through blockchain technology, so that different medical institutions and scientific research institutions can share data under the premise of protecting the privacy of patients. By setting up a data storage module and using blockchain technology through data encryption and smart contracts, the security and privacy of patient data can be guaranteed and data leakage can be prevented. Blockchain technology realizes secure data sharing and traceability, ensures the authenticity and reliability of data, and promotes the effective use of medical data.

[0069] The risk assessment and prediction module uses XGBoost and deep neural networks to train personalized risk prediction models, and uses transfer learning to apply the knowledge learned from large-scale data sets to specific individuals to generate personalized risk predictions. The obtained model data is evaluated for accuracy, precision, recall, and F1-score indicators, and the model is optimized based on the results. Finally, the user's data is processed and analyzed in real time according to the improved model, and the patient's risk is assessed. When abnormal data or high risk is detected, the system triggers an early warning to remind the user to conduct further inspections or take preventive measures. By setting up the risk assessment and prediction module and training the personalized risk prediction model through the XGBoost method, customized risk assessments can be provided based on the individual's specific risk factors. This personalized prediction helps to formulate more effective prevention and treatment strategies. At the same time, the model can be used to perform real-time risk assessments on the user's data and provide early warning information in a timely manner, which is crucial for early detection and intervention. The following are the formulas needed in the model building process:

[0070] XGBoost:

[0071]

[0072] Among them, l is the loss function, is the predicted value, Ω(f k ) is the regularization term.

[0073] Deep Neural Networks:

[0074]

[0075] Among them, W is the weight matrix, b is the bias, and σ is the activation function.

[0076] Accuracy:

[0077]

[0078] Accuracy:

[0079]

[0080] Recall:

[0081]

[0082] F1-score:

[0083]

[0084] Among them, TP is true positive, TN is true negative, FP is false positive, and FN is false negative.

[0085] The feature processing module collects the patient's gastric cancer-related gene mutations and genetic markers, extracts the size, shape, density, etc. of the lesions in the imaging features, extracts the features such as the previous medical history, family medical history, and medication records, and uses the LASSO regression method to screen the above data to screen out the features that have an important impact on the risk prediction of gastric cancer. The diagnosis and decision-making module uses a deep learning algorithm to automatically analyze the imaging data, assist doctors in identifying and diagnosing gastric cancer, and provide diagnostic suggestions and lesion quantitative analysis to help doctors make more accurate diagnoses. At the same time, based on the patient's medical records and the latest medical research, it recommends personalized treatment plans. Finally, it uses natural language processing technology to analyze the doctor's diagnostic records and the patient's medical records to automatically generate diagnostic reports and treatment recommendations.

[0086] The feature processing module needs to use the following formula in the process of data screening;

[0087] LASSO regression:

[0088]

[0089] Among them, y i is the target variable, x ij is a feature, β j is the regression coefficient, and λ is the regularization parameter

[0090] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A gastric cancer disease risk detection device based on big data analysis technology, characterized in that: It includes a data acquisition module, including multimodal data acquisition, and the multimodal data acquisition includes genomic data, metabolomic data, imaging data, environmental data and electronic health records. The genomic data is obtained by cooperating with a genetic testing agency to obtain the user's genomic data, especially gene mutations related to gastric cancer, and using a gene sequencing device to collect gene data, and the data is transmitted to the system. The metabolomic data is obtained by collecting biological samples such as blood and urine from the user, using a mass spectrometer to analyze metabolites, and the metabolomic data is transmitted to the system and associated with other health data of the user. The imaging data is obtained by using CT and MRI to obtain the user's stomach imaging data, and the data is uploaded to the system. The environmental data is obtained by entering environmental data such as air quality and dietary intake collected at the patient's home or workplace, and uploading the data to the system through the network. The electronic health record obtains the patient's medical record data from the hospital's medical record, including past medical history, family medical history, medication records, etc.; A data preprocessing module, including data cleaning and standardization and multimodal data fusion, wherein the data cleaning and standardization include missing value processing, outlier detection and data standardization; Data storage module, including data encryption and storage and data sharing; Feature processing module, including feature extraction and feature selection; Risk assessment and prediction module, including personalized risk prediction model and real-time risk assessment; Diagnosis and decision-making module, including intelligent diagnostic assistance and treatment plan recommendations.

2. The gastric cancer disease risk detection device based on big data analysis technology according to claim 1, characterized in that: The data preprocessing module processes missing data by using the mean filling method, and for missing key data, it marks and reminds users to supplement them in time. Secondly, it uses the Z-score method to detect outliers, and corrects or deletes them. The different types of data obtained are standardized. Finally, the multimodal fusion technology in deep learning is used to fuse genomic data, metabolomics data, imaging data and electronic health records, and through feature alignment and feature selection, the correlation and consistency between different modal data are ensured. The following are the formulas used in the data preprocessing process: Mean filling method: Among them, x i is a missing value, x j is the number of observations with non-missing values, and n is the number of non-missing values; Z-score method: Where x is the observed value, μ is the mean, σ is the standard deviation, and |Z|>3 is usually considered an outlier.

3. The gastric cancer disease risk detection device based on big data analysis technology according to claim 1, characterized in that: The data storage module uses blockchain technology to encrypt, store and transmit patient data, and utilizes smart contract technology to ensure the security and integrity of the data. At the same time, it establishes a data sharing platform to integrate data resources from different sources, and realizes secure data sharing through blockchain technology, allowing different medical institutions and scientific research institutions to share data while protecting the privacy of patients.

4. The gastric cancer disease risk detection device based on big data analysis technology according to claim 1, characterized in that: The risk assessment and prediction module uses XGBoost and deep neural networks to train personalized risk prediction models, and uses transfer learning to apply the knowledge learned from large-scale data sets to specific individuals to generate personalized risk predictions. The obtained model data is evaluated for accuracy, precision, recall, and F1-score indicators, and the model is optimized based on the obtained results. Finally, the user's data is processed and analyzed in real time based on the improved model, and the patient's risk is assessed. When abnormal data or high risk is detected, the system triggers an early warning to remind the user to conduct further inspections or take preventive measures. The following are the formulas needed in the model building process: XGBoost: Among them, l is the loss function, is the predicted value, Ω(f k ) is the regularization term; Deep Neural Networks: Among them, W is the weight matrix, b is the bias, and σ is the activation function; Accuracy: Accuracy: Recall: F1-score: Among them, TP is true positive, TN is true negative, FP is false positive, and FN is false negative.

5. A gastric cancer disease risk detection device based on big data analysis technology according to claims 1 to 4, characterized in that: The feature processing module collects the patient's gastric cancer-related gene mutations and genetic markers, extracts the lesion size, shape, density, etc. in the imaging features, extracts the features such as the previous medical history, family medical history, and medication records, and uses the LASSO regression method to screen the above data to screen out the features that have an important impact on the prediction of gastric cancer risk.

6. A gastric cancer disease risk detection device based on big data analysis technology according to claims 1 to 4, characterized in that: The diagnosis and decision-making module uses deep learning algorithms to automatically analyze imaging data, assist doctors in identifying and diagnosing gastric cancer, and provide diagnostic recommendations and lesion quantitative analysis to help doctors make more accurate diagnoses. It also recommends personalized treatment plans based on the patient's medical records and the latest medical research. Finally, it uses natural language processing technology to analyze the doctor's diagnostic records and the patient's medical records to automatically generate diagnostic reports and treatment recommendations.

7. The gastric cancer disease risk detection device based on big data analysis technology according to claim 5, characterized in that: The feature processing module needs to use the following formula in the process of data screening: LASSO regression: Among them, y i is the target variable, x ij is a feature, β j is the regression coefficient and λ is the regularization parameter.

Images