Intracranial aneurysm rupture risk assessment and management system based on internet hospital

Abstract

The present application relates to the medical technical field, specifically to the intracranial aneurysm rupture risk assessment and management system based on the internet hospital, including: patient end module, the patient passes through uploading the medical history data through the system standardization questionnaire and multi-modal data acquisition mechanism to the automatic filing and label processing of basic health data, the present application realizes the consistency and comparability of health data through vectorization coding and standardization processing; the blind deconvolution technology is used to improve the brain blood vessel image definition, and the risk assessment is optimized in combination with the image and clinical characteristics; the privacy is protected through the adaptive watermark technology, the data safety and compliance are ensured in combination with the double encryption and de-identification processing; the system identifies the risk and pushes the personalized health intervention in time through the dynamic monitoring and intelligent early warning, improves the accuracy and initiative of chronic disease management. In addition, the block chain technology ensures the data traceability and anti-tampering, enhances the doctor-patient trust.

Description

Technical Field

[0001] This invention relates to the field of medical technology, specifically to an intracranial aneurysm rupture risk assessment and management system based on an internet hospital. Background Technology

[0002] With the rapid development of medical imaging technology, the early diagnosis and accurate assessment of cerebrovascular diseases increasingly rely on high-quality imaging data and comprehensive health information. However, cerebrovascular imaging data processing faces challenges such as image blurring, noise interference, and difficulties in edge recognition. Furthermore, traditional risk assessment methods often rely on limited clinical data and single imaging features, making it difficult to comprehensively and accurately assess a patient's individualized rupture risk. In addition, cerebrovascular diseases exhibit high individual variability; relying solely on traditional static assessment methods often fails to respond promptly to dynamic changes in a patient's health status, leading to incomplete individualized treatment and intervention plans.

[0003] Existing patient risk assessment systems have several shortcomings, failing to meet the actual clinical needs for data privacy protection, health trend monitoring, and accurate risk prediction. Traditional health information systems offer relatively simple encryption protection measures for uploaded data, lacking comprehensive mechanisms such as double encryption, dynamic key generation, de-identification, and blockchain log recording. This poses risks of data leakage and untraceability, failing to meet increasingly stringent data compliance requirements. Furthermore, their identification and processing of sensitive text regions in images are not intelligent enough, making it difficult to achieve automated and accurate masking, increasing the risk of patient privacy leaks. Existing systems handle time-series data such as patient blood pressure, weight, and activity levels in a crude manner, lacking smoothing filtering, trend modeling, and anomaly warning mechanisms. This makes it difficult to detect abnormal fluctuations in health status in a timely manner, affecting the effectiveness of early intervention for chronic diseases. Currently, most systems use single models or rule-based scoring systems, failing to fully integrate imaging features, clinical features, and dynamic health indicators. This results in insufficient prediction accuracy and a lack of dynamic weight adjustment and feature importance enhancement mechanisms, leading to insensitivity in identifying key high-risk factors and a tendency to miss or misdiagnose high-risk patients.

[0004] In conclusion, the development of an intracranial aneurysm rupture risk assessment and management system based on internet hospitals remains a critical issue that urgently needs to be addressed in the medical field. Summary of the Invention

[0005] The purpose of this invention is to address the problems in existing intelligent systems, such as insufficient data privacy protection, limited data processing capabilities, and a lack of dynamic monitoring and feedback mechanisms.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] This invention provides a risk assessment and management system for intracranial aneurysm rupture based on an internet hospital. The system includes a risk assessment module that extracts and analyzes patient medical history data, generates risk level assessments and visual assessment reports, including:

[0008] Acquire standardized medical history data and multimodal medical image data of patients, and preprocess the multimodal medical image data. The preprocessing includes voxel normalization using trilinear interpolation and smoothing noise and highlighting the edge features of cerebral blood vessels using the Laplacian enhancement algorithm.

[0009] Based on the preprocessed image data, the aneurysm region was extracted and multi-scale semantic segmentation was performed. The aneurysm surface complexity index H and the convexity defect index CDI were calculated.

[0010] Calculate the surface complexity H of the tumor:

[0011] ,

[0012] in, Indicates the first The area of ​​a local surface segment This represents the total surface area of ​​the aneurysm. Indicates the first The proportion of the area of ​​each segment to the total area;

[0013] Calculate the Convexity Defect Index (CDI):

[0014] ,

[0015] in, The volume of the three-dimensional convex hull of the tumor. This indicates the actual volume of the tumor itself;

[0016] The tumor surface complexity H and convexity defect index CDI are used to form image feature codes, which are then fused with clinical feature codes derived from clinical risk factors to form a unified feature matrix.

[0017] Based on the unified feature matrix, feature modeling is performed through an ensemble learning model, wherein the model optimization objective is to minimize the weighted loss function, and the importance score of each feature is calculated based on the information gain, and the feature matrix is ​​weighted to form a standardized input matrix;

[0018] Based on the standardized input matrix, the probability of aneurysm rupture is calculated using a multimodal neural network, and the final risk value is obtained by combining the risk adjustment term.

[0019] Beneficial effects

[0020] Compared with known public technologies, the technical solution provided by this invention has the following beneficial effects:

[0021] This invention utilizes multimodal data fusion technology to organically combine imaging features (such as tumor surface complexity H and convexity defect index CDI) with clinical risk factors, overcoming the limitations of traditional methods that rely on a single data source. By employing an ensemble learning model and introducing a feature importance weighting mechanism, the influence of key risk factors can be amplified, thereby constructing a more accurate and personalized risk assessment model and significantly improving the reliability of rupture risk prediction. Attached Figure Description

[0022] Figure 1 This is a flowchart of the intracranial aneurysm rupture risk assessment and management system based on an internet hospital, as described in this invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but includes other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] The present invention will now be described in further detail with reference to the accompanying drawings:

[0026] like Figure 1 As shown, this invention provides an intracranial aneurysm rupture risk assessment and management system based on an internet hospital, characterized by comprising:

[0027] The patient-side module allows patients to upload their medical history data, which is then automatically archived and tagged using the system's standardized questionnaires and multimodal data collection mechanism.

[0028] Furthermore, the operation process of the patient-side module includes: the system generates a unique ID and establishes a health record; and the data uploaded by the patient is normalized through vectorized encoding.

[0029] ,in, Indicates an age range. Indicates the patient's smoking index, This represents the hypertension weighting coefficient. Indicates symptom index,

[0030] Recovering sharp images from blurry images using blind zone convolution image restoration.

[0031] ,in, This indicates a blurred observation image. This represents a clear image that needs to be restored. The blur kernel represents the blur caused by motion blur. This represents a two-dimensional convolution operation. The gradient of an image representing edge information. Representing the overall energy of the kernel, it restores sharp details from blurred MRA images. By estimating the optimal blur kernel, the system can infer an image as close as possible to the true sharp state from the input blurred image, enhance the edge information of cerebral vascular images, and introduce image gradient information during deconvolution to focus on preserving the details of vascular contours and minute branches.

[0032] Sensitive information in patient medical images is masked using an adaptive watermarking method. The PP-OCRv3 model is used to locate text regions in the image and determine whether they contain sensitive information.

[0033]

[0034] in, A weight matrix of the same size as the image, used to control the transparency of the watermark. This represents the identified privacy region. The weight of the identified region is set to 0.9, while the weight of other regions is 0, keeping the original image unchanged.

[0035] Based on the above formula for synthesizing a protected image after covering sensitive information in the image, the formula is as follows:

[0036]

[0037] in, Represents the original image. This represents a Gaussian noise image, used as a watermark background. This indicates the final output image after it has been protected. Represents the weight matrix. Indicates the weight of the non-privacy region.

[0038] Multimodal data fusion formula:

[0039] Where b represents behavioral time-series data and h represents static structural data. Represents image data / image features. , , Indicates dynamic weights, This represents a fusion vector. By separately collecting patients' dynamic behavioral time-series data, static clinical structural data (such as medical history, family history, and genetic information), and medical imaging feature data (such as cerebral vascular morphological parameters), and combining them with dynamically adjusted weighting factors, flexible importance scores are assigned to different categories of data. Ultimately, a unified fusion vector is generated to comprehensively and accurately represent the patient's current overall health status, providing a reliable basis for subsequent risk assessment and intervention decisions.

[0040] Specifically, in the patient-side module, users complete real-name registration and health information initialization. The system automatically generates a unique ID and establishes a personal health record. The basic health information uploaded by the patient is vectorized and normalized to achieve consistency and standardization across different feature dimensions. Meanwhile, the uploaded cerebrovascular imaging data (MRA) uses blind deconvolution image restoration technology to recover clear image details from blurred images. The system uses an adaptive watermarking method to automatically locate sensitive text regions in medical images and uses a region weight matrix to cover and protect privacy information. For multi-source health data systems, a dynamic weight mechanism is used to achieve multimodal data fusion and generate a unified fusion vector.

[0041] The risk assessment module extracts and analyzes patient medical history data through integrated deep learning algorithms to generate risk level assessments and visual assessment reports.

[0042] Furthermore, the operation process of the risk assessment module includes: the system performs image preprocessing operations, using trilinear interpolation to perform voxel homogenization on the DICOM format MRA image data to ensure consistent resolution across all axes. The formula is as follows:

[0043] ,

[0044] in, This represents the grayscale value of the target point in the resampled image. Indicates the interpolation weights. The voxel value represents the original image and is used to account for interpretation errors caused by differences in the original resolution.

[0045] The image is processed using Laplacian enhancement to smooth noise while preserving blood vessel edges. ,in, Represents the two-dimensional Laplace operator. Represented as Gaussian kernel With images The convolution result, This represents the edge enhancement weight parameter, used to highlight the edge features of cerebral blood vessels. A Gaussian kernel is used for initial image denoising, followed by a Laplacian operator to emphasize the image's edges, preserving detailed structures such as cerebral blood vessels and aneurysms. By adjusting the edge enhancement weight parameter, the enhancement effect can be precisely controlled. The enhanced image makes it easier to distinguish small lesions such as aneurysms.

[0046] Based on the aforementioned prominent edge features of cerebral blood vessels, vascular structure segmentation and aneurysm region extraction are performed, followed by multi-scale semantic segmentation of the processed image. ,in, Used to measure the degree of overlap between the segmented region and the true label. Represents the binary cross-entropy loss. , Represents the weighting coefficients of the loss function.

[0047] The system constructs a morphological classification submodule, which uses shape entropy and convex hull deviation factor to calculate the surface complexity of the tumor:

[0048] ,

[0049] in, Indicates the first The area of ​​a local surface segment This represents the total surface area of ​​the aneurysm. Indicates the first The proportion of the area of ​​each segment to the total area

[0050] ,

[0051] in, The volume of the three-dimensional convex hull of the tumor. This indicates the actual volume of the tumor itself. and The greater the volume difference, the greater the surface depression of the tumor. The range of values ​​for is , A value close to 0 indicates a smooth tumor. This indicates that there are depressions on the surface of the tumor.

[0052] Furthermore, the operational process of the risk assessment module includes: the system uniformly encodes various collected clinical risk factors into vectors. The geometric parameters of the aneurysm extracted from the image are encoded into a vector. The two are combined to form a unified feature matrix: Based on fusion, feature modeling is performed using ensemble learning methods, with the core optimization objective being to minimize the weighted loss function.

[0053] ,

[0054] in, Indicates predicted value With real labels Single point loss between This represents the sum of penalty terms applied to the complexity of all weak learners.

[0055] Based on the above modeling, features Importance score Calculated using information gain:

[0056] ,in Indicates the first One characteristic, Representation of features The total number of times nodes were split during training. This represents the loss value before the split. Let represent the loss value after the split. This indicates the degree of optimization of this split, and its characteristics. Each participation in the decision split leads to a certain degree of "model improvement" (loss reduction). Summarizing all these improvements and averaging them gives its importance score. A higher score indicates that the feature is more important.

[0057] The system will weight the matrix The input matrix is ​​formed by applying the same method to the original feature matrix. ,in This represents the Hadamard product, which enhances the treatment of high-importance risk factors through feature weighting. Based on the base model, patient features are scored, and a comprehensive baseline risk value is obtained.

[0058] ,in This represents the score of the classical statistical model. This indicates the number of features involved in the statistical model. This represents the standardized value of the i-th feature variable. This indicates the weight or score of a feature in the corresponding scoring system. It strengthens high-risk features, amplifies the values ​​of highly important features, and assigns smaller weights to irrelevant or less important features to improve the model's generalization ability.

[0059] Using a multimodal neural network, image feature vectors and clinical feature vectors are jointly input according to a fusion model, and the multimodal neural network calculates the rupture probability:

[0060] ,in Indicates image feature encoding. Indicates clinical feature coding, This represents a multimodal deep fusion network. This represents the Sigmoid activation function. The final output is the probability of rupture risk.

[0061] The system introduces risk adjustment items. Final predicted risk value: Based on the final risk value, the system classifies the risk level:

[0062] ,

[0063] in, This represents the final risk value. This indicates the upper limit of low risk. The lower limit representing high risk is defined by the following rules:

[0064] Then it is judged as low risk. It is then judged as medium risk. This is considered high-risk; doctors can adjust [their procedures]. This supplements the system model's output, making the system more aligned with actual clinical judgment. For example, a doctor might discover a case with unique characteristics, requiring appropriate adjustments to the risk value. The way risk levels are categorized allows doctors to understand and interpret the system's decision-making basis, providing more reliable evidence for clinical decisions. Through doctors' fine-tuning and the setting of risk thresholds, the system output is ensured to remain within a safe range, avoiding overly extreme or inappropriate risk judgments.

[0065] Specifically, the system utilizes a comprehensive analysis of imaging data, medical history information, and clinical features to generate personalized risk assessments and visualization reports. First, it uses trilinear interpolation to voxelize MRA images, improving image quality. Then, it employs Laplacian enhancement to highlight vessel edges, aiding in the extraction of vessel and aneurysm regions. Next, the system uses multi-scale semantic segmentation algorithms and morphological analysis to quantify the surface complexity of aneurysms. Finally, through ensemble learning, the system fuses imaging features and clinical data, optimizes feature modeling, calculates risk scores, and predicts rupture risk. Based on the predicted values, the system categorizes risk levels (low, medium, and high risk) and generates personalized assessment reports to assist physicians in making precise decisions, thereby improving the effectiveness and safety of risk management.

[0066] Furthermore, the operational process of the health management module includes:

[0067] Standardize the patient's health data and generate a feature matrix:

[0068] ,in, This indicates how the patient's blood pressure curve changes over time. This represents a curve showing the patient's weight change. Indicates the patient's activity level. This indicates the patient's dietary calorie intake record. The data represents the patient's sleep duration and quality indicators, and is then normalized. ,in This represents the mean of various health data. Standard deviation,

[0069] Monitoring patient health trends using weighted moving average:

[0070] ,in, Represents the smoothing factor. This indicates real-time monitoring of health status trends. When an abnormal deviation from a set threshold is detected, the system will trigger personalized health interventions or high-risk warnings.

[0071] Specifically, the system first collects and standardizes patients' health data, including indicators such as blood pressure curves, weight changes, activity levels, dietary calorie intake, and sleep quality, and generates a unified feature matrix. After data normalization, the system ensures the consistency of the dimensions of various health data to facilitate subsequent analysis. The system uses a weighted moving average algorithm to monitor patients' health status in real time and smooths changes in health data through a smoothing factor. The system can detect trends in patients' health status and dynamically monitor them according to preset thresholds. When abnormal health trends are detected or deviations from the set thresholds are detected, the system will automatically trigger personalized health intervention measures or issue high-risk warnings, ensuring that patients can receive timely and targeted health management and risk warnings, thereby improving the accuracy and proactivity of health management.

[0072] Furthermore, the specific processes of the data security module include:

[0073] The patient's uploaded image data and health information are double-encrypted.

[0074] ,in, A symmetric encryption key generated in one go. For the patient's image data, For the patient's health information, Indicates using Encrypted images and health data content, This represents the symmetric key encrypted using the server's public key. , and This indicates the final storage, where the uploaded health dataset undergoes de-standardization processing, and the system generates immutable records for data access or operation behaviors.

[0075] Specifically, patients upload MRA images and health information through the system. The system performs double encryption on the uploaded data. It generates a one-time symmetric encryption key to encrypt the patient's image data and health information. Then, it uses the server's public key to encrypt the symmetric key. Finally, the encrypted data and key are securely stored on the server to ensure data security during transmission and storage. The system de-identifies the uploaded health dataset, removing sensitive personally identifiable information and generating an anonymous dataset to ensure effective protection of patient privacy and meet relevant data privacy compliance requirements. The system generates an immutable operation record for each data access or operation, including the user's unique identifier, operation type, operation time, and user permission set information, further enhancing the system's data security and doctor-patient trust.

[0076] A concrete example of feasibility: Let's assume Mr. Wang (45 years old, long-term smoker, with a history of hypertension) completes his first real-name registration through the patient-side module. The system automatically generates a unique ID and initializes his personal health record. He then uploads basic health information (including age, smoking index, blood pressure control level, family medical history, etc.) and cerebrovascular MRA imaging data. The system first performs vectorization encoding and normalization on the uploaded health data, standardizing each feature dimension to ensure data consistency and forming a normalized feature vector. For the uploaded MRA images, the system applies a blind zone convolution restoration method to recover clear details: And after voxel homogenization (using trilinear interpolation: Then perform Laplace enhancement: To highlight blood vessel edges and improve the visibility of small blood vessels and aneurysms, and to detect and cover sensitive areas containing sensitive information such as names and hospital numbers in the images, the system uses an adaptive watermarking mechanism. To protect privacy and prevent information leaks, sensitive information is overlaid and then combined into a protective image. The system is based on a dynamic weight fusion mechanism: Integrating Mr. Wang's multimodal health data, including static medical history, dynamic behavioral data (such as blood pressure fluctuation curves and changes in exercise volume over the past 3 months), and cerebrovascular imaging features, a unified fusion vector is formed. The fused vector is input into the ensemble learning model, where the system automatically quantifies key indicators such as aneurysm surface complexity, vessel wall irregularity, and local blood flow characteristics. By integrating imaging features and health data, it calculates a rupture risk score and classifies the risk into medium-to-high risk levels based on a set threshold.

[0077] ,

[0078] in, This represents the final risk value. This indicates the upper limit of low risk. The lower limit representing high risk is defined by the following rules: Then it is judged as low risk. It is then judged as medium risk. If it is judged as high risk, the system will generate a personalized assessment report;

[0079] Table 1

[0080]

[0081] As shown in Table 1, Mr. Wang is considered low-risk when he is in situation A, medium-risk when he is in situation B, and high-risk when he is in situation C. In situation A, the system recommends maintaining a healthy lifestyle, such as controlling diet, quitting smoking and limiting alcohol consumption, and regular exercise. In situation B, the system recommends scheduling follow-up imaging (MRA, CTA) within 6-12 months to strengthen blood pressure and blood lipid management. In situation C, the system recommends timely medical attention and completion of high-resolution vascular wall imaging to assess whether interventional treatment or surgery is needed, and early intervention to prevent rupture if necessary. The system applies a weighted moving average smoothing process to Mr. Wang's continuous health monitoring data (such as blood pressure curve, weight, and sleep quality). in, Represents the smoothing factor. This system monitors health status trends in real time. When a trend deviates abnormally from a set threshold, the system will trigger personalized health interventions or high-risk warnings. Specifically, it monitors health status trends in real time, and if blood pressure rises abnormally and exceeds a safe threshold, the system will automatically trigger a high-risk warning and push personalized health guidance, such as medication adjustment suggestions and lifestyle interventions. Regarding data security, all data uploaded by Mr. Wang is protected by a dual encryption strategy. First, a one-time symmetric key is used to encrypt the data itself. ,in, A symmetric encryption key generated in one go. For the patient's image data, For the patient's health information, Indicates using Encrypted images and health data content, This represents the symmetric key encrypted using the server's public key. , and The system indicates that the data will be stored in the final database, and then the key itself will be encrypted with the server's public key to ensure the security of the data during transmission and storage. At the same time, every data access or processing behavior is recorded by the system, and the uploaded health dataset is de-standardized to improve the data trust and privacy compliance of both doctors and patients. The system generates an immutable record of data access or operation behavior.

[0082] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A risk assessment and management system for intracranial aneurysm rupture based on an internet hospital, characterized in that, The system includes a risk assessment module: Acquire standardized medical history data and multimodal medical image data of patients, and preprocess the multimodal medical image data. The preprocessing includes voxel normalization using trilinear interpolation and smoothing noise and highlighting the edge features of cerebral blood vessels using the Laplacian enhancement algorithm. Based on the preprocessed image data, the aneurysm region was extracted and multi-scale semantic segmentation was performed. The aneurysm surface complexity index H and the convexity defect index CDI were calculated. ， in, The volume of the three-dimensional convex hull of the tumor. This indicates the actual volume of the tumor itself; ， in, Indicates the first The area of ​​a local surface segment This represents the total surface area of ​​the aneurysm. Indicates the first The proportion of the area of ​​each segment to the total area; The tumor surface complexity H and convexity defect index CDI are used to form image feature codes, which are then fused with clinical feature codes derived from clinical risk factors to form a unified feature matrix. Based on the unified feature matrix, feature modeling is performed through an ensemble learning model, wherein the model optimization objective is to minimize the weighted loss function, and the importance score of each feature is calculated based on the information gain, and the feature matrix is ​​weighted to form a standardized input matrix; Based on the standardized input matrix, the probability of aneurysm rupture is calculated using a multimodal neural network, and the final risk value is obtained by combining the risk adjustment term. The system also includes a patient-side module: The system generates a unique ID and establishes a health record, and normalizes the data uploaded by patients through vectorized encoding. Recovering sharp images from blurry images using blind zone convolution image restoration. ,in, This indicates a blurred observation image. This represents a clear image that needs to be restored. The blur kernel represents the blur caused by motion blur. This represents a two-dimensional convolution operation. The gradient of an image representing edge information. Representing the overall energy of the kernel, it restores sharp details from blurred MRA images. By estimating the optimal blur kernel, the system can infer an image as close as possible to the true sharp state from the input blurred image, enhance the edge information of cerebral vascular images, and introduce image gradient information during deconvolution to focus on preserving the details of vascular contours and minute branches. Sensitive information in patient medical images is masked using an adaptive watermarking method. The PP-OCRv3 model is used to locate text regions in the image and determine whether they contain sensitive information. in, A weight matrix of the same size as the image, used to control the transparency of the watermark. This represents the identified privacy region. The weight of the identified region is set to 0.9, while the weight of other regions is 0, keeping the original image unchanged. Based on the above formula for synthesizing a protected image after covering sensitive information in the image, the formula is as follows: in, Represents the original image. This represents a Gaussian noise image, used as a watermark background. This indicates the final output image after it has been protected. Represents the weight matrix. Indicates the weight of the non-privacy region. Multimodal data fusion formula: Where b represents behavioral time-series data, h represents static structural data, and I represents image data / image features. , , Indicates dynamic weights, This represents the fusion vector.

2. The intracranial aneurysm rupture risk assessment and management system based on an internet hospital as described in claim 1, characterized in that, Calculate the probability of rupture The formula is: ,in Indicates image feature encoding. Indicates clinical feature coding, This represents a multimodal deep fusion network. This represents the Sigmoid activation function. This represents the final output probability of rupture risk.

3. The intracranial aneurysm rupture risk assessment and management system based on an internet hospital as described in claim 2, characterized in that, Final predicted risk value The formula is: ； in, This is a risk adjustment item.

4. The intracranial aneurysm rupture risk assessment and management system based on an internet hospital as described in claim 3, characterized in that, Based on the final risk value The system classifies risk levels as follows: ， in, This represents the final risk value. This indicates the upper limit of low risk. The lower limit representing high risk is defined by the following rules: Then it is judged as low risk. It is then judged as medium risk. This is then judged as high risk.

5. The intracranial aneurysm rupture risk assessment and management system based on an internet hospital as described in claim 1, characterized in that, The system also includes a health management module, which continuously uploads patients' daily health data and pushes health management suggestions and follow-up reminders through an intelligent reminder system. Standardize the patient's health data and generate a feature matrix: ,in, This indicates how the patient's blood pressure curve changes over time. This represents a curve showing the patient's weight change. Indicates the patient's activity level. This indicates the patient's dietary calorie intake record. The data represents the patient's sleep duration and quality indicators, and is then normalized. ,in This represents the mean of various health data. The standard deviation is used to monitor patients' health trends using a weighted moving average.

6. The intracranial aneurysm rupture risk assessment and management system based on an internet hospital as described in claim 1, characterized in that, The system also includes a data security module that uses blockchain and tokenization technology to protect data privacy and security. The patient's uploaded image data and health information are double-encrypted. ,in, A symmetric encryption key generated in one go. For the patient's image data, For the patient's health information, Indicates using Encrypted images and health data content, This represents the symmetric key encrypted using the server's public key. , and This indicates the final storage, where the uploaded health dataset undergoes de-standardization processing, and the system generates immutable records for data access or operation behaviors.

Images