Medical linear accelerator quality control data analysis system

Abstract

The invention belongs to the technical field of medical equipment quality control, and particularly relates to a medical linear accelerator quality control data analysis system which comprises a data acquisition module, a data analysis module, a deep learning module, a data storage module and a user interaction module. The data acquisition module is used for acquiring dosimetry data, mechanical performance data and image data of the medical linear accelerator; the data analysis module is connected with the data acquisition module and is used for counting data records in a database, extracting feature information, establishing and analyzing personalized data contents and intelligently screening data through a B / S framework and an advanced computer technology and a deep learning module, so that quick response of analysis data is realized, and the efficiency and the accuracy of quality control work are improved; the change trend of quality control data is predicted by using an LSTM (Long Short-Term Memory Network), potential unstable factors are identified, and the period and frequency of quality control are adjusted.

Description

Technical Field

[0001] The invention relates to the technical field of medical equipment quality control, and in particular to a medical linear accelerator quality control data analysis system. Background Art

[0002] Medical linear accelerators are key equipment for tumor radiotherapy, and their performance stability and the accuracy of quality control data are directly related to the treatment effect and patient safety. However, traditional quality control data management methods mostly rely on paper records or local computer storage, which has problems such as low data processing efficiency, inconvenient query, and difficulty in achieving real-time monitoring. In addition, the quality control analysis of the imaging system mostly relies on manual visual recognition, which is easily affected by subjective factors and it is difficult to ensure the accuracy and consistency of the analysis results; it is impossible to predict the changing trend of quality control data, identify potential unstable factors, and adjust the cycle and frequency of quality control. For this reason, this application proposes a medical linear accelerator quality control data analysis system. Summary of the invention

[0003] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0004] In view of the above-mentioned and / or existing problems in quality control of existing medical devices, the present invention is proposed.

[0005] Therefore, the purpose of the present invention is to provide a medical linear accelerator quality control data analysis system, which realizes comprehensive automated management and analysis of quality control data through B / S architecture and advanced computer technology, improves the efficiency and accuracy of quality control work, and uses a deep learning module to count data records in the database, extract feature information, establish and analyze personalized data content, and intelligently screen the data to achieve rapid response to analysis data; LSTM (long short-term memory network) is used to predict the changing trend of quality control data, identify potential unstable factors, and adjust the cycle and frequency of quality control. User information is fed back to the personalized learning module to establish a user's personal feature database, realize the rapid execution of data instructions, quickly obtain analysis results, and ensure the timeliness of quality control data.

[0006] To solve the above technical problems, according to one aspect of the present invention, the present invention provides the following technical solutions:

[0007] A medical linear accelerator quality control data analysis system, comprising:

[0008] Data acquisition module: used to acquire medical linear accelerator dosimetry data, mechanical properties data and imaging data;

[0009] Data analysis module: connected to the data acquisition module, using Python Django framework and MySQL database, combined with open source toolkits to perform data processing and analysis, to achieve automated analysis of the collected quality control data, including dose stability analysis, mechanical property stability analysis, and image quality analysis;

[0010] Deep learning module: connected with the data analysis module, counts data records in the database, extracts feature information, establishes personalized data content for analysis, performs intelligent screening of data, and realizes rapid response of analysis data;

[0011] Data storage module: connected to the data storage module, storing quality control data and analysis results on the server, supporting long-term storage and rapid retrieval;

[0012] User interaction module: connected to the data storage module, providing a web-based user interface and supporting multi-role access;

[0013] The user interaction module is connected to the data analysis module, and the user information is fed back to the personalized learning module to establish a user's personal feature database to achieve rapid execution of data instructions;

[0014] The system is based on a B / S architecture, provides user interaction through a Web interface, and uses the Python Django framework and MySQL database to achieve automated collection, analysis, storage, and management of quality control data.

[0015] As a preferred solution of the medical linear accelerator quality control data analysis system described in the present invention, the data acquisition module supports the reception and processing of multiple data formats, such as image files and log files in DICOM format.

[0016] As a preferred solution of the medical linear accelerator quality control data analysis system described in the present invention, the data analysis module provides a visualization tool to help users intuitively understand the changing trend and distribution status of quality control data.

[0017] As a preferred solution of the medical linear accelerator quality control data analysis system described in the present invention, the data storage module adopts a three-layer MTV (Model-Template-View) architecture to ensure data security and maintainability.

[0018] As a preferred solution of the medical linear accelerator quality control data analysis system described in the present invention, in which: in the user interaction module, the user can perform interactive operations on quality control data through the interface, support data export in formats such as Excel, facilitate subsequent analysis and research, provide complete user authority management functions, and ensure data security and compliance.

[0019] As a preferred solution of the medical linear accelerator quality control data analysis system described in the present invention, the data acquisition module collects dosimetry data, mechanical property data and imaging data in real time through an interface connection with the medical linear accelerator and its supporting equipment (such as the OBI system), and supports the reception and analysis of multiple data formats.

[0020] As a preferred solution of the medical linear accelerator quality control data analysis system described in the present invention, the data storage module adopts MySQL database as the data storage center, designs a reasonable database table structure, and supports classified storage and rapid retrieval of quality control data.

[0021] As a preferred solution of the medical linear accelerator quality control data analysis system described in the present invention, the user interaction module provides a Web-based user interface and supports multi-role access. Users can query, review, export and other operations of quality control data through this module, and support data export in Excel format. At the same time, the module also includes a permission management function to ensure the security and compliance of data access.

[0022] As a preferred solution of the medical linear accelerator quality control data analysis system described in the present invention, it also includes a security module for protecting the system's data security and user privacy, including functions of data encryption transmission, access control, and log auditing.

[0023] As a preferred solution of the medical linear accelerator quality control data analysis system described in the present invention, when the deep learning module works, the following method steps are adopted:

[0024] Step 1: Statistical analysis of data, collecting and collating data for automated analysis of quality control data;

[0025] Step 2: Screen the collected data, extract and analyze the data information, establish a data feature model, use LSTM (Long Short-Term Memory Network) to predict the change trend of quality control data, identify potential unstable factors, and adjust the cycle and frequency of quality control;

[0026] Step 3: According to the data feature model, bring in the subsequent calculation data and retrieve relevant data to achieve rapid response of data processing;

[0027] Step 4: Establish a connection between the user's feature database and the data feature model. When the user executes an operation instruction, the data is quickly identified and retrieved to achieve rapid execution of the data instruction.

[0028] Compared with the prior art: the present invention proposes a medical linear accelerator quality control data analysis system, which realizes comprehensive automatic management and analysis of quality control data through B / S architecture and advanced computer technology, improves the efficiency and accuracy of quality control work, and uses a deep learning module to count data records in the database, extract feature information, establish personalized data content for analysis, and intelligently screen the data to achieve rapid response to analysis data; LSTM (long short-term memory network) is used to predict the change trend of quality control data, identify potential unstable factors, and adjust the cycle and frequency of quality control. User information is fed back to the personalized learning module to establish a user's personal feature database, realize rapid execution of data instructions, quickly obtain analysis results, and ensure the timeliness of quality control data;

[0029] Improve quality control efficiency: Through automated data analysis, reduce the workload of manual analysis and improve quality control efficiency;

[0030] Ensure data accuracy: Use advanced algorithms for data analysis to ensure the accuracy and consistency of analysis results;

[0031] Realize information management: The Web-based architecture realizes online storage and management of quality control data, facilitating real-time query and statistical analysis of data;

[0032] Multi-role support: Provide multi-role access control to meet the needs of different users and improve the flexibility and scalability of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be described in detail below in combination with the accompanying drawings and detailed embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0034] Figure 1 It is a schematic diagram of the system architecture of the present invention;

[0035] Figure 2 It is a schematic diagram of the overall structure of the present invention;

[0036] Figure 3 This is a schematic diagram of deep learning of the present invention. DETAILED DESCRIPTION

[0037] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0038] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0039] Secondly, the present invention is described in detail with reference to schematic diagrams. When describing the embodiments of the present invention in detail, for the sake of convenience, the cross-sectional diagrams showing the device structure will not be partially enlarged according to the general scale, and the schematic diagrams are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0040] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0041] The present invention provides a medical linear accelerator quality control data analysis system, which realizes comprehensive automatic management and analysis of quality control data through B / S architecture and advanced computer technology, and improves the efficiency and accuracy of quality control work. Figure 1-Figure 3 , including: data acquisition module, data analysis module, data storage module and user interaction module;

[0042] The system is based on a B / S architecture, provides user interaction through a Web interface, and uses the Python Django framework and MySQL database to achieve automated collection, analysis, storage, and management of quality control data.

[0043] Data acquisition module:

[0044] By collecting medical linear accelerator dosimetry data, mechanical properties data and imaging data.

[0045] Supports receiving and processing of multiple data formats, such as DICOM image files, log files, etc.

[0046] The data acquisition module collects dosimetry data, mechanical property data and imaging data in real time by interfacing with the medical linear accelerator and its supporting equipment (such as the OBI system), and supports the reception and analysis of multiple data formats.

[0047] Data Analysis Module:

[0048] The Python Django framework and MySQL database are used in combination with open source toolkits (such as Pydicom and Pylinac) for data processing and analysis.

[0049] Realize automated analysis of the collected quality control data, including dose stability analysis, mechanical property stability analysis, and image quality analysis (such as spatial resolution, low contrast resolution, geometric distortion, etc.).

[0050] Provides a variety of visualization tools such as trend charts and bar charts to help users intuitively understand the changing trends and distribution status of quality control data.

[0051] The data analysis module has built-in multiple professional algorithms for automatic processing and analysis of the collected quality control data, including but not limited to dose stability analysis, mechanical positioning accuracy analysis, image quality analysis (such as spatial resolution, low contrast resolution, etc.), and provides data visualization tools to display the analysis results.

[0052] Deep learning module: connected with the data analysis module, counts data records in the database, extracts feature information, establishes personalized data content for analysis, intelligently filters data, and realizes rapid response to analysis data; and, in conjunction with the trend charts and bar charts provided by the data analysis module, predicts the changing trend of quality control data based on feature information and identifies potential unstable factors; predicts the changing trend of quality control data and can make adaptive adjustments to the cycle and frequency of quality control.

[0053] Data storage module:

[0054] Store quality control data and analysis results on the server, supporting long-term preservation and rapid retrieval.

[0055] The three-layer MTV (Model-Template-View) architecture is adopted to ensure the security and maintainability of data.

[0056] The data storage module uses MySQL database as the data storage center, designs a reasonable database table structure, and supports classified storage and rapid retrieval of quality control data.

[0057] User interaction module:

[0058] Provides a web-based user interface that supports multi-role access (such as technician, physicist, technical supervisor, and system administrator).

[0059] Users can interact with quality control data through the interface, which supports data export in formats such as Excel to facilitate subsequent analysis and research.

[0060] Provides comprehensive user rights management functions to ensure data security and compliance.

[0061] The user interaction module provides a web-based user interface that supports multi-role access (such as technicians, physicists, technical supervisors, etc.). Users can query, review, and export quality control data through this module, and support data export in Excel format. At the same time, this module also includes permission management functions to ensure the security and compliance of data access.

[0062] The user interaction module is connected to the data analysis module, and the user information is fed back to the personalized learning module to establish a user's personal feature database to achieve rapid execution of data instructions.

[0063] This system adopts a three-tier B / S (browser / server) architecture, namely the front-end user interaction layer, the middle business logic layer, and the back-end data persistence layer. This architecture model enables the system to efficiently and securely process requests and data from different users.

[0064] Front-end user interaction layer: Use HTML, CSS and JavaScript to develop a user-friendly web interface, providing intuitive operation experience and rich interactive functions. Users can access the system through a web browser to complete operations such as query, analysis, review and management of quality control data.

[0065] The middle business logic layer is built on the Python Django framework, which is responsible for processing requests from the front end, calling the corresponding data processing logic and algorithms, interacting with the back-end database, and returning the processing results to the front end for display. The Django framework provides powerful URL routing, template engine, and ORM (object-relational mapping) functions, which simplifies the development process of Web applications.

[0066] Backend data persistence layer: Use MySQL database to store quality control data, analysis results and system configuration information. The database design adopts standardized table structure and index optimization to ensure data integrity, consistency and query efficiency. Through the ORM framework and Django integration, automatic data management and operation are realized.

[0067] Data acquisition and interface to achieve real-time data acquisition: through the interface with the medical linear accelerator and its imaging system (such as OBI system), the real-time acquisition of quality control data is realized. The system supports the reception of multiple data formats, including but not limited to DICOM format image files, log files, etc. Adopt event-driven or polling mechanism to obtain the latest quality control data from the accelerator system regularly or on demand.

[0068] Data standardization: pre-process and standardize the collected raw data to ensure data consistency and comparability. The processing process includes steps such as data cleaning, format conversion, and unit unification.

[0069] Data analysis module built-in analysis algorithm: The system has built-in multiple quality control data analysis algorithms, including dosimetry analysis (such as dose stability, uniformity, etc.), mechanical performance analysis (such as isocenter accuracy, gantry rotation stability, etc.) and image quality analysis (such as spatial resolution, low contrast resolution, geometric distortion, etc.). These algorithms are designed based on medical physics principles and statistical methods and can accurately reflect the performance status of the accelerator.

[0070] Automated analysis process: Users select the data set and analysis parameters to be analyzed through the web interface, and the system automatically calls the corresponding analysis algorithm for processing and presents the analysis results to users in the form of charts, reports, etc. The analysis process supports batch processing and multi-task concurrent execution to improve analysis efficiency.

[0071] Data storage and management

[0072] Database design: tree-structured database table design is adopted to realize the classified storage and management of quality control data. The database tables include quality control record table, analysis result table, equipment information table, user information table, etc. The tables are associated with each other through foreign keys to form a complete quality control data system.

[0073] Data backup and recovery: Back up the database regularly to ensure data security and recoverability. The system supports manual and automatic data backup, and users can set backup strategies as needed. At the same time, a data recovery mechanism is provided to enable rapid recovery when data is lost or damaged.

[0074] User interaction and permission management

[0075] Multi-role access control: The system supports multi-role user access, including technicians, physicists, technical supervisors, and system administrators. Different roles have different operating permissions and view interfaces to ensure the security and compliance of the system.

[0076] User interface design: The front-end user interface design is concise and clear, providing intuitive operation prompts and help documents. Users can quickly locate the required data and analysis results through menu navigation, search filtering and other functions. At the same time, it supports the data export function in Excel format, which is convenient for users to conduct further data processing and analysis.

[0077] Review and approval process: The system has built-in quality control result review and approval process to ensure the standardization and effectiveness of quality control work. Reviewers can review the analysis results through the system and give approval opinions. The approval results will directly affect the status of quality control records and subsequent processing procedures.

[0078] The system architecture diagram is usually divided into three main layers from top to bottom: user interface layer, business logic layer and data access layer (persistence layer). Each layer communicates through a clear interface.

[0079] User Interface Layer

[0080] Located at the top layer, it is the interface for users to interact with the system.

[0081] In this system, the presentation layer is presented through a web browser, using HTML, CSS, and JavaScript to build dynamic web pages.

[0082] Users access the system through a browser to perform operations such as query, analysis, and export, and view the analysis results and visual charts of quality control data.

[0083] Business logic layer

[0084] Located between the presentation layer and the data access layer, it is responsible for processing user requests and business rules.

[0085] In this system, the business logic layer is implemented by the Python Django framework.

[0086] Views in the Django framework process user requests, call models for data operations, and use templates to render HTML pages and return them to users.

[0087] This layer also contains various quality control data analysis algorithms for processing and analyzing data collected from the accelerator and its imaging system.

[0088] Data access layer (persistence layer)

[0089] Located at the bottom layer, responsible for data storage and retrieval.

[0090] In this system, the data access layer uses the MySQL database to store quality control data, analysis results, and user information.

[0091] The Django framework's ORM (object-relational mapping) tool is used to map database tables to Python classes, simplifying database operations.

[0092] Inter-layer communication:

[0093] Users send HTTP requests to the web server through their browsers.

[0094] The web server forwards the request to the Django framework for processing.

[0095] Django view functions receive requests, call business logic for processing, and may need to query or update data from the database.

[0096] After processing, the view function renders the result as an HTML page and returns it to the user's browser through the web server for display.

[0097] The system also includes a security module to protect the system's data security and user privacy, including data encryption transmission, access control, log auditing and other functions.

[0098] The system supports customized development and can expand functions and customize interfaces according to the specific needs of different medical institutions.

[0099] When the deep learning module works, the following method steps are adopted:

[0100] Step 1: Statistical analysis of data, collecting and collating data for automated analysis of quality control data, mainly focusing on image data;

[0101] Step 2: Screen the collected data, extract and analyze the data information, establish a data feature model, use LSTM (Long Short-Term Memory Network) to predict the change trend of quality control data, identify potential unstable factors, and adjust the cycle and frequency of quality control;

[0102] After analyzing the image data, record the feature information of the image, and select a suitable deep learning model architecture based on the nature of the problem and the characteristics of the data. Common models include convolutional neural networks (CNN), recurrent neural networks (RNN), long short-term memory networks (LSTM), etc. The choice of model depends on the type of data (such as images, text, sequences) and the requirements of the task (such as classification, regression, generation); use the preprocessed data to train the model; during the training process, the model learns the patterns and features in the data by adjusting parameters; adjust the model's hyperparameters, including learning rate, number of layers, number of nodes, etc. By adjusting the hyperparameters, the performance of the model can be further optimized.

[0103] Step 3: According to the data feature model, bring in the subsequent calculation data and retrieve relevant data to achieve rapid response of data processing;

[0104] Step 4: Establish a connection between the user's feature database and the data feature model. When the user executes an operation instruction, the data is quickly identified and retrieved to achieve rapid execution of the data instruction.

[0105] According to the above method steps, a quick comparison and identification is performed after image acquisition, and after user interaction, the control instructions are quickly identified and executed and brought into the data feature model. The more data is processed, the more feature data there is. Subsequent data retrieval does not need to compare image data one by one. Key data can be quickly retrieved directly according to the features, and the changing trend of quality control data can be predicted based on the feature information to identify potential unstable factors. By predicting the changing trend of quality control data, adaptive adjustments can be made to the cycle and frequency of quality control, achieving rapid response of data processing, improving the computing speed of data analysis, quickly obtaining analysis results, and ensuring the timeliness of quality control data.

[0106] Specifically, LSTM (Long Short-Term Memory Network) is used to predict the changing trend of quality control data, identify potential unstable factors, and adjust the cycle and frequency of quality control. When working, LSTM (Long Short-Term Memory Network) adopts the following sequential process:

[0107] S1: Data preparation and processing of LSTM model, collecting quality control data, the data includes timestamps and indicators related to quality control, cleaning and preprocessing the data, data cleaning uses the following operations:

[0108] Remove outliers: Detect and remove obviously abnormal measurements through statistical analysis (such as box plots, Z-score, etc.).

[0109] Fill missing values: Use methods such as mean interpolation, linear interpolation, or previous interpolation to fill missing items in the data.

[0110] Smoothing: For short-term fluctuations in data, methods such as moving average can be used to smooth the data and reduce the interference of noise on the prediction.

[0111] Data normalization: Use MinMaxScaler to normalize the data:

[0112] MinMaxScaler: Scales the data to the range of 0 to 1. The formula is:

[0113]

[0114] in:

[0115] X is the original data.

[0116] X min is the minimum value in the original data.

[0117] X max is the maximum value in the original data.

[0118] max and min are the maximum and minimum values ​​of the scaled target range, which are 1 and 0 by default (i.e. scaled to the [0, 1] range).

[0119] Creation of time window data: By using the sliding window method, the time series data can be divided into multiple small training samples and the data can be converted into a suitable format:

[0120] Given a time window size, assuming the window size is window_size, the model will use the data of the past window_size time steps to predict the data of the next time step.

[0121] For example, if window_size=30, each training sample contains data from the past 30 days, and the goal is to predict the value on the 31st day.

[0122] S2: Build an LSTM model, which contains the following layers:

[0123] Input layer: The input of LSTM is a three-dimensional array of shape (samples, timesteps, features), where samples represents the number of samples, timesteps represents the number of time steps (i.e., window size), and features represents the number of features in each time step (for a single indicator, features=1);

[0124] LSTM layer: contains one or more LSTM layers. Each LSTM layer consists of multiple LSTM units, which are responsible for capturing dependencies in the time series. You can set return_sequences=True to let each LSTM layer output a sequence of time steps instead of just the output of the final time step;

[0125] Dropout layer: To prevent overfitting, a Dropout layer is usually added between LSTM layers;

[0126] Output layer: If the task is a regression problem (for example, predicting the value of a quality control indicator), the output layer is a single neuron used to predict continuous values.

[0127] Hyperparameter adjustment: Adjust the hyperparameters of the LSTM model, including:

[0128] Number of LSTM units: The number of units in each LSTM layer. Generally, more units can capture more features, but it is also easy to cause overfitting.

[0129] Learning rate: A hyperparameter that controls the speed at which the model is updated. An appropriate learning rate helps improve convergence speed and accuracy.

[0130] Batch size: The number of training examples used in each update. Smaller batch sizes help the model train in smaller steps, but may require more computing resources.

[0131] S3: Model training and evaluation. Divide the data into training set, validation set and test set. In time series prediction, data division needs to be done in chronological order to avoid using future data for training (to avoid data leakage). Usually, the training and test set division ratio can be 80%-20% or 70%-30%.

[0132] Use the training set to train the LSTM model, and use the validation set to monitor overfitting during training. During training, early stopping can be used to prevent overfitting.

[0133] After the model is trained, it is evaluated on the test set. The evaluation indicators include:

[0134] Mean square error (MSE) is a common method to measure the difference between the model prediction value and the true value. Its formula is as follows:

[0135]

[0136] in:

[0137] n is the sample size.

[0138] is the true value of the i-th sample.

[0139] is the predicted value of the ith sample.

[0140] It represents the square of the difference between the true value and the predicted value of the i-th sample.

[0141] MSE measures the overall prediction performance of the model by calculating the square of the prediction error of each sample and then averaging it. The smaller the MSE value, the better the prediction performance of the model, that is, the closer the predicted value is to the true value. MSE is sensitive to outliers because outliers can cause the square of the error to increase significantly.

[0142] R² (coefficient of determination) is an important statistical indicator to measure the degree to which the model explains the variation of the target variable. The following is the expression of R²:

[0143] r 2

[0144] in:

[0145] SSR (Sum of Squares due to Regression) is the sum of squares of regression, which represents the part explained by the independent variables in the model.

[0146] SST (Total Sum of Squares) is the total sum of squares, which represents the total variation of the dependent variable.

[0147] R² Value Range: The value of R² lies between 0 and 1.

[0148] R² value close to 1: indicates that the model can well explain the variation of the dependent variable, that is, the model fit is good. This means that the values ​​predicted by the model are very close to the actual observed values.

[0149] R² value is close to 0: It means that the model has poor ability to explain the variation of the dependent variable, that is, the model fit is poor. This may be because the model is too simple to capture the complex relationship in the data, or there is no significant linear relationship between the independent variable and the dependent variable.

[0150] S4: Identify potential unstable factors, use the LSTM model to predict the future trend of quality control data, and identify potential unstable factors in combination with actual business rules: By comparing the predicted future data trend with historical data, abnormal fluctuations or trend changes can be identified. For example, a sudden and large fluctuation in the predicted value of a quality indicator may indicate potential unstable factors; threshold setting: set an abnormal value threshold. When the predicted value exceeds the threshold, the system can alarm and remind the operator to take further measures.

[0151] S5: Adjust the quality control cycle and frequency. Adjust the quality control cycle and frequency according to the prediction results of the LSTM model and the identified potential unstable factors, including increasing the frequency during high-risk periods and reducing the frequency during low-risk periods. The details are as follows:

[0152] Increase frequency during high-risk periods: If the predicted trend shows that the quality fluctuations are large during a certain period of time, the frequency of quality control testing during that period can be increased.

[0153] Reduce frequency during low-risk periods: When the predicted quality indicator trend is relatively stable, the frequency of quality control testing can be reduced to save resources.

[0154] Although the present invention has been described above with reference to the embodiments, various modifications may be made thereto and parts thereof may be replaced by equivalents without departing from the scope of the present invention. In particular, as long as there is no structural conflict, the various features in the embodiments disclosed in the present invention may be used in combination with each other in any manner, and the fact that these combinations are not exhaustively described in this specification is only for the sake of omitting space and saving resources. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A medical linear accelerator quality control data analysis system, characterized in that: include: Data acquisition module: used to acquire medical linear accelerator dosimetry data, mechanical properties data and imaging data; Data analysis module: connected to the data acquisition module, using Python Django framework and MySQL database, combined with open source toolkits to perform data processing and analysis, to achieve automated analysis of the collected quality control data, including dose stability analysis, mechanical property stability analysis, and image quality analysis; Deep learning module: connected with the data analysis module, counts data records in the database, extracts feature information, establishes personalized data content for analysis, performs intelligent screening of data, and realizes rapid response of analysis data; Data storage module: connected to the data storage module, storing quality control data and analysis results on the server, supporting long-term storage and rapid retrieval; User interaction module: connected to the data storage module, providing a web-based user interface and supporting multi-role access; The user interaction module is connected to the data analysis module, and the user information is fed back to the personalized learning module to establish a user's personal feature database to achieve rapid execution of data instructions; The system is based on a B / S architecture, provides user interaction through a Web interface, and uses the Python Django framework and MySQL database to achieve automated collection, analysis, storage, and management of quality control data.

2. A medical linear accelerator quality control data analysis system according to claim 1, characterized in that: The data acquisition module supports receiving and processing multiple data formats, such as image files and log files in DICOM format.

3. A medical linear accelerator quality control data analysis system according to claim 1, characterized in that: The data analysis module provides visualization tools to help users intuitively understand the changing trends and distribution status of quality control data.

4. A medical linear accelerator quality control data analysis system according to claim 1, characterized in that: The data storage module adopts a three-layer MTV (Model-Template-View) architecture to ensure data security and maintainability.

5. The medical linear accelerator quality control data analysis system according to claim 1, characterized in that: In the user interaction module, users can perform interactive operations on quality control data through the interface, support data export in multiple formats to facilitate subsequent analysis and research, provide complete user authority management functions, and ensure data security and compliance.

6. A medical linear accelerator quality control data analysis system according to claim 1, characterized in that: The data acquisition module collects dosimetry data, mechanical property data and imaging data in real time by interfacing with the medical linear accelerator and its supporting equipment (such as the OBI system), and supports the reception and analysis of multiple data formats.

7. The medical linear accelerator quality control data analysis system according to claim 1, characterized in that: The data storage module uses MySQL database as the data storage center, designs a reasonable database table structure, and supports classified storage and rapid retrieval of quality control data.

8. The medical linear accelerator quality control data analysis system according to claim 1, characterized in that: The user interaction module provides a web-based user interface that supports multi-role access. Users can use this module to interact with quality control data and support data export in Excel format. At the same time, the module also includes a permission management function to ensure the security and compliance of data access.

9. The medical linear accelerator quality control data analysis system according to claim 1, characterized in that: It also includes a security module to protect the system's data security and user privacy, including functions such as data encryption transmission, access control, and log auditing.

10. The medical linear accelerator quality control data analysis system according to claim 1, characterized in that: When the deep learning module works, the following method steps are adopted: Step 1: Statistical analysis of data, collecting and collating data for automated analysis of quality control data; Step 2: Screen the collected data, extract and analyze the data information, establish a data feature model, use LSTM (Long Short-Term Memory Network) to predict the change trend of quality control data, identify potential unstable factors, and adjust the cycle and frequency of quality control; Step 3: According to the data feature model, bring in the subsequent calculation data and retrieve relevant data to achieve rapid response of data processing; Step 4: Establish a connection between the user's feature database and the data feature model. When the user executes an operation instruction, the data is quickly identified and retrieved to achieve rapid execution of the data instruction.

Images