Employee pulmonary nodule screening frequency optimization method and system
Through the MDP model combined with multimodal machine learning, the frequency of employee lung nodules screening is dynamically adjusted, which solves the problem of individual differences neglected in traditional methods, realizes personalized health management, reduces health risks and costs, and improves corporate efficiency.
Patent Information
- Application Number
- CN202410389644.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-02
- Publication Date
- 2025-07-22
AI Technical Summary
Traditional fixed-frequency pulmonary nodules screening methods cannot effectively respond to individual differences in corporate employees, resulting in insufficient monitoring of high-risk employees or excessive screening of low-risk employees, and frequent screening may bring health risks.
A dynamic screening frequency adjustment model based on Markov decision-making process (MDP) is adopted, combining multimodal machine learning and data analysis, and a personalized screening plan is formulated based on individual risk factors and changes in health status.
It has achieved effective monitoring of high-risk employees, reduced the rate of missed lung cancer diagnosis, reduced unnecessary screening of low-risk employees, protected employee health, reduced health risks and management costs, and improved corporate health management efficiency and employee satisfaction.
Smart Images
Figure CN120356646A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the fields of health technology and artificial intelligence technology, and particularly relates to a method and system for optimizing the screening frequency of pulmonary nodules in employees. Background Art
[0002] When faced with the health management of enterprise employees, especially in the screening procedures for pulmonary nodules and lung cancer, the traditional fixed-frequency screening method faces significant challenges. As screening targets, the health status and lung cancer risk of enterprise employees are affected by various factors, including occupational exposure, lifestyle habits, genetic factors, etc. The fixed screening frequency ignores these individual differences, which may lead to insufficient monitoring of some high-risk employees and may also lead to over-screening of low-risk employees.
[0003] For enterprises, the health of employees directly affects work efficiency and the overall performance of the enterprise. Therefore, developing a model that can dynamically adjust the screening frequency of pulmonary nodules can not only better protect the health of employees but also optimize the enterprise's health management resources.
[0004] In addition, multiple pulmonary nodule screenings for employees also need to be carefully considered. Frequent screenings, especially those using radioactive techniques (such as low-dose computed tomography LDCT), although the radiation risk for individual examinations is relatively low, the cumulative effect may pose a risk to the long-term health of employees.
[0005] Therefore, a dynamic screening frequency adjustment model developed using the Markov decision process (MDP) can customize the screening plan according to each employee's individual risk factors, previous screening results, and changes in health status. This method can not only effectively monitor high-risk employees and reduce the missed diagnosis rate of lung cancer but also reduce unnecessary screenings for low-risk employees, thereby protecting employees from potential health risks brought by over-screening.
[0006] In summary, developing and implementing this MDP-based dynamic screening frequency adjustment model is a forward-looking health management strategy for enterprises. It not only helps improve the health level and job satisfaction of enterprise employees but also effectively controls the health management cost, which is of great significance for enhancing the social responsibility of enterprises and the long-term well-being of employees. Summary of the Invention
[0007] The present invention aims at the defects of the prior art and provides a method and system for optimizing the screening frequency of pulmonary nodules in employees.
[0008] To achieve the above invention objectives, the technical solutions adopted by the present invention are as follows:
[0009] A method for optimizing the screening frequency of pulmonary nodules in employees, comprising the following steps:
[0010] S1: Obtain data related to employees' pulmonary nodule diseases;
[0011] Data collection: Obtain the individual basic information, historical case information, and CT image data of employees from the database.
[0012] Data cleaning: Clean the collected data, including handling missing values, outliers, and duplicate values.
[0013] Data integration: Integrate the individual basic information, historical case information, and CT image data into a unified dataset.
[0014] S2: Predict the risk status of employees' pulmonary nodules based on multimodal machine learning;
[0015] S21: Data preparation: Divide the integrated dataset into a training set and a test set.
[0016] S22: Feature extraction: Extract features from the individual basic information, historical case information, and CT image data in the training set to prepare for input into the model.
[0017] S23: Based on a multimodal machine learning model, including: image CNN and structured data MLP.
[0018] S24: Model training: Use the training set to train the image CNN and structured data MLP to learn the prediction patterns of employees' pulmonary nodule risk status.
[0019] S3: Construct a transition probability matrix between the risk statuses of employees' pulmonary nodules;
[0020] S31: Define risk statuses: Determine the risk statuses of employees' pulmonary nodules, including: low risk, medium risk, and high risk.
[0021] S32: Data analysis: Analyze the historical case information to determine the transition rules and probabilities between different risk statuses.
[0022] S33: Construct the transition probability matrix: Construct the transition probability matrix between the risk statuses of employees' pulmonary nodules according to the analysis results of S32. Fill the calculated transition probabilities into a matrix. The rows and columns of the matrix correspond to different risk statuses, and each element represents the transition probability from the row status to the column status.
[0023] S4: Estimate the benefit function of the intervention strategy for employees' pulmonary nodules
[0024] S41: Determine the intervention strategies for employees in different risk statuses, including: screening frequency and treatment plans.
[0025] S42: Define the benefit function for the intervention of employees' pulmonary nodules according to the intervention strategy and the risk status of pulmonary nodules, so as to evaluate the effects and costs of different strategies.
[0026] S5: Construct a Markov decision model;
[0027] Establish a Markov decision model based on the risk status of employees' pulmonary nodules, the transition probability matrix and the benefit function.
[0028] Using the Q-table method, iteratively calculate the optimal screening frequency of employees' pulmonary nodules with the goal of maximizing the benefit function or minimizing the cost.
[0029] Furthermore, the feature extraction in S22 includes:
[0030] Individual basic information includes: age characteristics, gender characteristics, smoking history characteristics.
[0031] Historical case information includes: family medical history characteristics and past medical history characteristics.
[0032] CT image data includes: image characteristics and texture characteristics.
[0033] Furthermore, in S32, specifically:
[0034] Analysis of transfer rules: According to the historical case data, calculate the number of transfers and transfer frequencies between different risk states. Use frequency statistics or transfer rate calculation to analyze the transfer rules between different states.
[0035] Estimation of transfer probability: Based on the historical case data, estimate the transfer probability between different risk states. Use the maximum likelihood estimation or Bayesian statistical method to estimate the transfer probability.
[0036] The present invention also discloses a screening and intervention system for employees' pulmonary nodules, which can be used to implement the above-mentioned method for optimizing the screening frequency of employees' pulmonary nodules. Specifically, it includes:
[0037] Data acquisition module, with the following functions:
[0038] Data collection: Obtain the individual basic information, historical case information and CT image data of employees from the database.
[0039] Data cleaning: Clean the collected data, and process missing values, outliers and duplicate values.
[0040] Data integration: Integrate the individual basic information, historical case information and CT image data into a unified data set.
[0041] Multimodal machine learning module, with the following functions:
[0042] Data Preparation: Divide the integrated dataset into a training set and a test set.
[0043] Feature Extraction: Extract features from the individual basic information, historical case information, and CT image data in the training set.
[0044] Model Training: Based on a multi-modal machine learning model, including: image CNN and structured data MLP, train the dataset to predict the risk status of employees' lung nodules.
[0045] Transition Probability Matrix Module, with the following functions:
[0046] Define Risk Status: Determine the set of risk statuses of employees' lung nodules.
[0047] Data Analysis: Analyze the historical case information to determine the transition rules and probabilities between different risk statuses.
[0048] Construct Transition Probability Matrix: Construct the transition probability matrix between the risk statuses of employees' lung nodules based on the analysis results.
[0049] Intervention Strategy Evaluation Module, with the following functions:
[0050] Determine Intervention Strategies: Determine intervention strategies for employees in different risk statuses, including screening frequencies and treatment plans.
[0051] Define Benefit Function: According to the intervention strategies and the risk status of lung nodules, define the benefit function for the intervention of employees' lung nodules to evaluate the effects and costs of different strategies.
[0052] Markov Decision Model Module, with the following functions:
[0053] Based on the risk status of employees' lung nodules, the transition probability matrix, and the benefit function, establish a Markov decision model.
[0054] Using the Q-table method, iteratively calculate the optimal screening frequency for employees' lung nodules with the goal of maximizing the benefit function or minimizing the cost.
[0055] Result Display Module, with the following functions:
[0056] Result Visualization: Visually display the prediction results obtained from the multi-modal machine learning module.
[0057] Transition Probability Matrix Display: Display the constructed transition probability matrix between the risk statuses of employees' lung nodules to the user in the form of a table or a graph.
[0058] Intervention Strategy Evaluation Results: Visually display the effects and costs of different intervention strategies obtained from the evaluation.
[0059] Output of the Markov decision model: The output results of the Markov decision model are presented to the user, including the optimal screening frequency and the corresponding benefits or costs, so that the user can understand the best strategy recommended by the system.
[0060] Interactive interface: Provide an interactive interface for the user to adjust parameters or view results in different aspects according to their own needs, so as to achieve personalized result display.
[0061] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the above-mentioned employee lung nodule screening and intervention method is implemented.
[0062] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the above-mentioned employee lung nodule screening and intervention method is implemented.
[0063] Compared with the prior art, the advantages of the present invention are as follows:
[0064] Personalized strategy: Based on the individual's basic information, historical case information, and CT image data, the present invention can achieve personalized screening and intervention for each employee, making the strategy more accurate and effective.
[0065] Comprehensive utilization of multi-modal data: By comprehensively utilizing data of different modalities, including individual basic information, historical case information, and CT image data, the accuracy and robustness of the prediction model are improved.
[0066] Data-driven decision-making: Using machine learning and data analysis techniques, and training and optimizing the model based on a large amount of data, it can more accurately predict the risk status of employees' lung nodules and provide a scientific basis for decision-making.
[0067] Real-time monitoring and adjustment: The present invention establishes a Markov decision model, which can monitor the risk status and metastasis of employees' lung nodules in real time, and adjust the screening and intervention strategy according to the latest data to maintain the timeliness and adaptability of the strategy.
[0068] High cost-effectiveness: By evaluating and optimizing different intervention strategies, it can minimize costs and improve resource utilization efficiency on the premise of ensuring the screening effect, meeting the interests and needs of enterprises and individuals.
[0069] Improve production efficiency: By ensuring employees' health and timely intervention, it can reduce production stagnation and personnel adjustment caused by lung nodule-related diseases, and improve production efficiency and enterprise performance. Brief Description of the Drawings
[0070] Figure 1It is a flowchart of the process for optimizing the screening frequency of employees' pulmonary nodules in the implementation of the present invention. Detailed implementation manners
[0071] To make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the following further elaborates on the present invention with reference to the accompanying drawings and by way of examples.
[0072] The present invention provides a method for optimizing the screening frequency of employees' pulmonary nodules, including the following steps:
[0073] S1: Obtain data related to employees' pulmonary nodule diseases;
[0074] Data collection: Obtain the individual basic information, historical case information and CT image data of employees from the database.
[0075] Data cleaning: Clean the collected data, including handling missing values, outliers and duplicate values.
[0076] Data integration: Integrate the individual basic information, historical case information and CT image data into a unified dataset for subsequent analysis.
[0077] S2: Predict the risk status of employees' pulmonary nodules based on multimodal machine learning
[0078] S21: Data preparation: Divide the integrated dataset into a training set and a test set.
[0079] S22: Feature extraction: Extract features from the individual basic information, historical case information and CT image data in the training set to prepare for input into the model.
[0080] The features include:
[0081] Individual basic information:
[0082] Age: The age of an employee may be related to the risk of pulmonary nodules, so age can be used as a feature.
[0083] Gender: Gender may affect the incidence of pulmonary nodules, so gender can also be used as a feature.
[0084] Smoking history: There is a certain correlation between smoking and pulmonary nodules, and smoking history can be used as a feature.
[0085] Historical case information:
[0086] Family medical history: Whether an employee has a family history of pulmonary nodules or related diseases, which may affect the incidence of pulmonary nodules.
[0087] Past medical history: Whether an employee has suffered from other pulmonary or related diseases, such as lung cancer, pulmonary infection, etc.
[0088] CT image data:
[0089] Image features: Features extracted from CT images, such as the size, shape, density, and edge features of lung nodules.
[0090] Texture features: Texture features extracted from CT images using texture analysis methods, such as gray-level co-occurrence matrix (GLCM), gray-level histogram, etc.
[0091] S23: Based on a multi-modal machine learning model, including: image CNN (Convolutional Neural Network) and structured data MLP (Multi-Layer Perceptron). The image CNN can be used to process CT image data and extract features of lung nodules; while the structured data MLP can be used to process individual basic information and historical case information to predict the risk status of employees' lung nodules.
[0092] S24: Model training: Use the training set to train the image CNN and the structured data MLP to learn the prediction pattern of the risk status of employees' lung nodules.
[0093] S25: Model evaluation: Use the test set to evaluate the trained model and evaluate the performance and accuracy of the model.
[0094] S3: Construct a transition probability matrix between the risk statuses of employees' lung nodules;
[0095] S31: Define risk statuses: Determine the risk statuses of employees' lung nodules, including: low risk, medium risk, and high risk.
[0096] S32: Data analysis: Analyze historical case information to determine the transition rules and probabilities between different risk statuses, as follows:
[0097] Analysis of transition rules:
[0098] According to historical case data, calculate the number of transitions and transition frequencies between different risk statuses.
[0099] Statistical methods, such as frequency statistics or transition rate calculation, can be used to analyze the transition rules between different states.
[0100] Observe and record which transitions between states are the most common and which are less frequent.
[0101] Estimation of transition probabilities:
[0102] Based on historical case data, estimate the transition probabilities between different risk statuses.
[0103] Use maximum likelihood estimation or Bayesian statistical methods to estimate the transition probabilities.
[0104] For example, calculate the transition probability from a low-risk state to a high-risk state, and the transition probability from a high-risk state to a low-risk state.
[0105] S33: Construct a transition probability matrix: Based on the analysis results in S32, construct a transition probability matrix between the risk states of employees' pulmonary nodules. Fill the calculated transition probabilities into a matrix. The rows and columns of the matrix correspond to different risk states, and each element represents the transition probability from the row state to the column state.
[0106] S4: Estimate the benefit function of the intervention strategy for employees' pulmonary nodules
[0107] S41: Determine the intervention strategies for employees in different risk states, including: screening frequency, treatment plan, etc.
[0108] S42: Define the benefit function of the intervention for employees' pulmonary nodules according to the intervention strategy and the risk state of pulmonary nodules to evaluate the effectiveness and cost of different strategies.
[0109] The design of the benefit function considers the following aspects:
[0110] 1. Intervention effectiveness: The effectiveness of the intervention usually refers to reducing or avoiding the losses caused by pulmonary nodule-related diseases through the intervention. This can be measured by indicators such as reducing the prevalence rate, prolonging the survival time, and improving the quality of life.
[0111] 2. Cost: The cost of the intervention strategy includes direct costs (such as screening costs, treatment costs) and indirect costs (such as production losses caused by employees' absenteeism due to treatment). The cost can be fixed or related to the intensity of the intervention.
[0112] 3. Risk and uncertainty: Considering the risks and uncertainties that the intervention strategy may bring, the benefit function can include risk-related terms, such as risk-adjusted effectiveness, risk sensitivity, etc.
[0113] Based on the above considerations, the benefit function can be expressed as: u = effectiveness - cost + risk adjustment term;
[0114] Among them, effectiveness: represents the expected effectiveness of the intervention measure, which can be reducing the prevalence rate, prolonging the survival time, etc.
[0115] Cost: represents the total cost of the intervention strategy, including direct costs and indirect costs.
[0116] Risk adjustment term: Adjusted according to the risks and uncertainties that the intervention strategy may bring, which can be risk sensitivity, uncertainty adjustment, etc.
[0117] S5: Construct a Markov decision model (MDP);
[0118] A Markov decision model is established based on the risk status of employees' lung nodules, the transition probability matrix, and the benefit function.
[0119] Risk status s
[0120] Transition probability: That is, in the case of health status s t when the screening frequency a is adopted, the probability of the future health status s t+1 of the employee is denoted as R(s′|s,a) here.
[0121] Here, it is hoped to obtain the maximum benefit in the shortest time. Considering the time cost, we call this time cost the discount factor γ (0 ≤ γ ≤ 1), and the benefit function is as follows
[0122] u = r1 + γr2 + γ 2 r3 + γ 3 r4 + …
[0123] Using the Q-table method, the optimal screening frequency of employees' lung nodules is obtained through iterative calculation, aiming to maximize the benefit function or minimize the cost.
[0124] The specific steps are as follows:
[0125] S51: Definition of state space:
[0126] Determine the set of risk states of employees' lung nodules, which will constitute the state space of the Markov decision model. For example, it can be defined as three states: low risk, medium risk, and high risk.
[0127] S52: Definition of action set:
[0128] Define the set of actions that can be executed in each state. In this scenario, the actions are usually intervention strategies, such as the choice of screening frequency. For example, it can be defined as screening once a year, screening once every two years, etc.
[0129] S53: State transition probability:
[0130] Using the existing transition probability matrix, determine how the risk status of employees' lung nodules will transfer after executing a certain action. These probabilities will be used to construct the state transition probability of the Markov chain.
[0131] S54: Benefit function:
[0132] Convert the previously defined benefit function into the benefit values corresponding to each state and action. These benefit values will be used to determine the expected benefit of executing each action in each state.
[0133] S55: Construct the Markov decision:
[0134] Integrate the state space, action set, state transition probability, and reward function defined above to construct a Markov decision process. This will form a complete model that describes the transitions and rewards resulting from performing different actions in different states.
[0135] S56: Solve the MDP:
[0136] Use methods in reinforcement learning, such as the Q-table method, to solve the constructed MDP. The goal is to find an optimal policy to maximize the long-term cumulative expected reward.
[0137] In another embodiment of the present invention, an employee lung nodule screening intervention system is provided. This system can be used to implement the above-mentioned method for optimizing the screening frequency of employee lung nodules. Specifically, it includes:
[0138] Data acquisition module, with the following functions:
[0139] Data collection: Obtain the individual basic information, historical case information, and CT image data of employees from the database.
[0140] Data cleaning: Clean the collected data, dealing with missing values, outliers, and duplicate values.
[0141] Data integration: Integrate the individual basic information, historical case information, and CT image data into a unified dataset.
[0142] Multimodal machine learning module, with the following functions:
[0143] Data preparation: Divide the integrated dataset into a training set and a test set.
[0144] Feature extraction: Extract features from the individual basic information, historical case information, and CT image data in the training set.
[0145] Model training: Based on multimodal machine learning models (such as image CNN and structured data MLP), train the dataset to predict the risk status of employee lung nodules.
[0146] Transition probability matrix module, with the following functions:
[0147] Define risk status: Determine the set of risk statuses of employee lung nodules.
[0148] Data analysis: Analyze the historical case information to determine the transition rules and probabilities between different risk statuses.
[0149] Construct the transition probability matrix: Construct the transition probability matrix between the risk statuses of employee lung nodules according to the analysis results.
[0150] Intervention strategy evaluation module, with the following functions:
[0151] Determine intervention strategies: Determine intervention strategies for employees in different risk states, including screening frequencies and treatment plans.
[0152] Define the benefit function: According to the intervention strategy and the risk state of pulmonary nodules, define the benefit function for the intervention of employees' pulmonary nodules, and evaluate the effects and costs of different strategies.
[0153] The Markov decision model module has the following functions:
[0154] Establish a Markov decision model based on the risk state of employees' pulmonary nodules, the transition probability matrix, and the benefit function.
[0155] Using the Q-table method, iteratively calculate the optimal screening frequency for employees' pulmonary nodules with the goal of maximizing the benefit function or minimizing the cost.
[0156] The result display module has the following functions:
[0157] Result visualization: Visualize and display the prediction results obtained by the multi-modal machine learning module, such as presenting the prediction results of the risk state of employees' pulmonary nodules through charts, statistical data, or images.
[0158] Display the transition probability matrix: Display the constructed transition probability matrix between the risk states of employees' pulmonary nodules in the form of a table or graph to allow users to understand the transition probabilities between different risk states.
[0159] Evaluation results of intervention strategies: Visualize and display the effects and costs of different intervention strategies obtained through evaluation, such as presenting them in the form of charts or indicator data, so that users can clearly understand the advantages and disadvantages of different strategies.
[0160] Output of the Markov decision model: Display the output results of the Markov decision model to users, including the optimal screening frequency and the corresponding benefits or costs, so that users can understand the best strategy recommended by the system.
[0161] Interactive interface: Provide users with an interactive interface that allows them to adjust parameters or view results in different aspects according to their needs to achieve personalized result display.
[0162] In another embodiment of the present invention, a terminal device is provided. The terminal device includes a processor and a memory. The memory is used to store a computer program, and the computer program includes program instructions. The processor is used to execute the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions to implement the corresponding method flow or corresponding function. The processor described in the embodiment of the present invention can be used for the operation of a method for optimizing the screening frequency of employees' pulmonary nodules, including the following steps:
[0163] S1: Obtain data related to employees' pulmonary nodule diseases;
[0164] S2: Predict the risk status of employees' pulmonary nodules based on multimodal machine learning;
[0165] S3: Construct a transition probability matrix between the risk statuses of employees' pulmonary nodules;
[0166] S4: Estimate the benefit function of the intervention strategy for employees' pulmonary nodules;
[0167] S5: Construct a Markov decision model.
[0168] In another embodiment of the present invention, the present invention also provides a storage medium, specifically a computer-readable storage medium (Memory). The computer-readable storage medium is a memory device in the terminal device and is used to store programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and, of course, the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, and this storage space stores the operating system of the terminal. And, in this storage space, one or more instructions suitable for being loaded and executed by the processor are also stored. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory.
[0169] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the method for optimizing the screening frequency of employees' pulmonary nodules in the above embodiments; the one or more instructions in the computer-readable storage medium are loaded and executed by the processor to perform the following steps:
[0170] S1: Obtain data related to employees' pulmonary nodule diseases;
[0171] S2: Predict the risk status of employees' pulmonary nodules based on multimodal machine learning;
[0172] S3: Construct a transition probability matrix between the risk statuses of employees' pulmonary nodules;
[0173] S4: Estimate the benefit function of the intervention strategy for employees' pulmonary nodules;
[0174] S5: Construct a Markov decision model.
[0175] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0176] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of processes and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.
[0177] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the specified functions in Figure 1 one process or multiple processes and / or blocks Figure 1 one block or multiple blocks.
[0178] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions for implementing the steps of the function specified in one process or a plurality of processes and / or blocks Figure 1 one process or a plurality of processes and / or blocks Figure 1 or steps of the function specified in a plurality of blocks.
[0179] Those of ordinary skill in the art will appreciate that the embodiments described herein are provided to assist the reader in understanding the implementation of the present invention, and it should be understood that the scope of protection of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations that do not depart from the essence of the present invention based on the technical revelations disclosed in the present invention, and these deformations and combinations are still within the scope of protection of the present invention.
Claims
1. An optimization method for the screening frequency of employees' pulmonary nodules, characterized in that it includes the following steps: S1: Obtain data related to employees' pulmonary nodule diseases; Data collection: The database obtains the individual basic information, historical case information, and CT image data of employees; Data cleaning: Clean the collected data, including handling missing values, outliers, and duplicate values; Data integration: Integrate the individual basic information, historical case information, and CT image data into a unified dataset; S2: Predict the risk status of employees' pulmonary nodules based on multimodal machine learning; S21: Data preparation: Divide the integrated dataset into a training set and a test set; S22: Feature extraction: Extract features from the individual basic information, historical case information, and CT image data in the training set to prepare for input into the model; S23: Based on a multimodal machine learning model, including: image CNN and structured data MLP; S24: Model training: Use the training set to train the image CNN and structured data MLP to learn the prediction pattern of the risk status of employees' pulmonary nodules; S3: Construct a transition probability matrix between the risk statuses of employees' pulmonary nodules; S31: Define the risk status: Determine the risk status of employees' pulmonary nodules, including: low risk, medium risk, and high risk; S32: Data analysis: Analyze the historical case information to determine the transition rules and probabilities between different risk statuses, S33: Construct the transition probability matrix: Construct a transition probability matrix between the risk statuses of employees' pulmonary nodules according to the analysis results of S32; Fill the calculated transition probabilities into a matrix; The rows and columns of the matrix correspond to different risk statuses, and each element represents the transition probability from the row status to the column status; S4: Estimate the benefit function of the intervention strategy for employees' pulmonary nodules S41: Determine the intervention strategies for employees in different risk statuses, including: screening frequency and treatment plan; S42: Define the benefit function of the intervention for employees' pulmonary nodules according to the intervention strategy and the risk status of the pulmonary nodules to evaluate the effects and costs of different strategies; S5: Construct a Markov decision model; Establish a Markov decision model based on the risk status of employees' pulmonary nodules, the transition probability matrix, and the benefit function; Use the Q-table method to iteratively calculate the optimal screening frequency of employees' pulmonary nodules with the goal of maximizing the benefit function or minimizing the cost.
2. The method for optimizing the screening frequency of pulmonary nodules in employees according to claim 1, characterized in that: The feature extraction in S22 includes: The individual basic information includes: age feature, gender feature, smoking history feature; The historical case information includes: family history feature and past disease history feature; The CT image data includes: image feature and texture feature.
3. The method for optimizing the screening frequency of pulmonary nodules in employees according to claim 1, wherein: Specifically in S32: Transition rule analysis: According to the historical case data, calculate the number of transitions and transition frequencies between different risk statuses; Use frequency statistics or transition rate calculation to analyze the transition rules between different statuses; Transition probability estimation: Estimate the transition probabilities between different risk statuses based on the historical case data; Use the maximum likelihood estimation or Bayesian statistics method to estimate the transition probability.
4. An employee lung nodule screening and intervention system, characterized in that: This system can be used to implement an optimization method for the screening frequency of employees' pulmonary nodules described in any one of claims 1 to 3. Specifically, it includes: Data acquisition module, with the following functions: Data collection: Obtain the individual basic information, historical case information, and CT image data of employees from the database; Data cleaning: Clean the collected data, dealing with missing values, outliers, and duplicate values; Data integration: Integrate the individual basic information, historical case information, and CT image data into a unified dataset; Multimodal machine learning module, with the following functions: Data preparation: Divide the integrated dataset into a training set and a test set; Feature extraction: Extract features from the individual basic information, historical case information, and CT image data in the training set; Model training: Based on the multimodal machine learning model, including: image CNN and structured data MLP, train the dataset to predict the risk status of employees' lung nodules; Transition probability matrix module, with the following functions: Define risk status: Determine the set of risk statuses of employees' lung nodules; Data analysis: Analyze the historical case information to determine the transition rules and probabilities between different risk statuses; Construct the transition probability matrix: Construct the transition probability matrix between the risk statuses of employees' lung nodules according to the analysis results; Intervention strategy evaluation module, with the following functions: Determine intervention strategies: Determine intervention strategies for employees in different risk statuses, including screening frequencies and treatment plans; Define the benefit function: According to the intervention strategies and the lung nodule risk status, define the benefit function for the intervention of employees' lung nodules, and evaluate the effects and costs of different strategies; Markov decision model module, with the following functions: Establish a Markov decision model based on the risk status of employees' lung nodules, the transition probability matrix, and the benefit function; Using the Q-table method, iteratively calculate the optimal screening frequency for employees' lung nodules with the goal of maximizing the benefit function or minimizing the cost; Result display module, with the following functions: Result visualization: Visually display the prediction results obtained by the multimodal machine learning module; Display the transition probability matrix: Display the constructed transition probability matrix between the risk statuses of employees' lung nodules to the user in the form of a table or a graph; Display the evaluation results of intervention strategies: Visually display the effects and costs of different intervention strategies obtained by the evaluation; Output of the Markov decision model: Display the output results of the Markov decision model to the user, including the optimal screening frequency and the corresponding benefit or cost situation, so that the user can understand the best strategy recommended by the system; Interactive interface: Provide an interactive interface for the user to adjust parameters or view results in different aspects according to their own needs to achieve personalized result display.
5. A computer device, characterized in that: Including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, it implements the employee lung nodule screening and intervention method according to any one of claims 1 to 3.
6. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by the processor, it implements the employee lung nodule screening and intervention method according to any one of claims 1 to 3.