Deep learning-based intelligent judgment system for onset risk and metastasis of lung cancer
Through an intelligent judgment system based on deep learning, multi-source data and deep neural network models are used to solve the accuracy of lung cancer risk and metastasis judgment, efficient and accurate lung cancer screening and preventive measures are achieved, and lung cancer screening is improved. Efficiency and cure rate.
Patent Information
- Application Number
- CN202410844555.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-27
- Publication Date
- 2024-11-19
AI Technical Summary
The existing lung cancer risk and metastasis determination system is difficult to effectively use multi-source data for accurate judgment, resulting in low efficiency in lung cancer screening and insufficient targeted prevention measures.
Using an intelligent judgment system based on deep learning, the construction of deep neural network model and the calculation of risk factor weights can be achieved to accurately determine the risk and metastasis level of lung cancer, including multi-source data integration, data preprocessing, and deep neural Network model training and risk level statistical analysis.
It improves the calculation efficiency and accuracy of lung cancer risk and metastasis determination, provides high-precision lung cancer screening tools, helps doctors to conduct targeted disease intervention and management, and reduces the incidence of lung cancer.
Smart Images

Figure CN118983082A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a deep learning-based intelligent determination system for lung cancer risk and its metastasis, and specifically to a lung cancer risk level determination system and its metastasis level determination system using a deep neural network model, belonging to the field of artificial intelligence and biomedical information processing technology. Background Art
[0002] Actively carrying out accurate research on the risk of lung cancer and its metastasis determination system will play a positive role in improving the efficiency of lung cancer screening and reducing the incidence of lung cancer. It is also of great significance to actively prevent the occurrence of lung cancer and achieve early diagnosis and treatment of lung cancer. In recent years, with the increasing level of economic development, people's living standards and living habits have changed greatly. At the same time, the living environment and meteorological and climatic conditions are constantly changing. These have brought a lot of uncertain risks to the occurrence and development of lung cancer. In the face of complex risks, determining the risk of lung cancer and its metastasis is conducive to active prevention and intervention, early diagnosis and treatment of lung cancer, and helps to reduce the incidence of lung cancer and improve the cure rate and survival rate.
[0003] Therefore, it is necessary to use biomedical information processing technology and computer-assisted diagnosis and treatment systems, and with the help of artificial intelligence, big data information technology and other means, to efficiently determine and calculate the risk of lung cancer for different individuals based on deep learning technology, in order to carry out targeted disease intervention and management and reduce the incidence of lung cancer. Summary of the invention
[0004] The purpose of the present invention is to solve technical problems such as the calculation of lung cancer risk and its metastasis determination. By using artificial intelligence, big data and other technical means and utilizing a computer-aided diagnosis and treatment system, it can better help doctors predict the risk of lung cancer and creatively propose a lung cancer risk and metastasis intelligent determination system based on deep learning.
[0005] The present invention is implemented by adopting the following technical solutions.
[0006] A deep learning-based intelligent determination system for lung cancer risk and metastasis, comprising a multi-source data preprocessing module, a deep learning model building module, a lung cancer risk determination module and a risk transfer determination module.
[0007] Among them, the multi-source data preprocessing module includes a multi-source data integration unit, a data preprocessing unit and a data grouping unit.
[0008] The multi-source data integration unit is used to integrate cross-domain data resources from multiple sources, including physical examination data, risk screening data, follow-up data, environmental pollution data, meteorological data, questionnaire survey data, etc.
[0009] The data preprocessing unit is used to remove white noise from the integrated data, fill in missing values, perform consistency processing on data with inconsistent structures or units, and unify the same metadata.
[0010] The data grouping unit is used to group the preprocessed data according to proportion, gender and age.
[0011] The deep learning model building module includes a deep neural network model training unit, a deep neural network model testing unit and a deep neural network model verification unit.
[0012] Among them, the deep neural network model training unit is used to train the deep learning model using the training set based on the deep neural network algorithm.
[0013] The deep neural network model testing unit is used to test the deep neural network model using the test set data.
[0014] The deep neural network model verification unit is used to verify the performance of the deep neural network model using the verification set data.
[0015] The lung cancer incidence risk determination module includes a risk factor weight calculation unit, a risk level determination unit and a risk level statistical analysis unit.
[0016] Among them, the risk factor weight calculation unit is used to calculate the weights of different risk factors through a deep neural network model.
[0017] The risk level determination unit is used to determine and calculate the lung cancer risk level of different individuals using the risk factor weights.
[0018] The risk level statistical analysis unit is used to perform statistical analysis on the lung cancer risk levels of different groups.
[0019] The risk transfer determination module includes a follow-up risk level determination unit, a risk transfer calculation unit and a risk transfer statistical analysis unit.
[0020] Among them, the follow-up risk level determination unit is used to determine the risk factor weights and calculate the individual's lung cancer risk level in different follow-up years.
[0021] The risk transfer calculation unit is used to calculate the risk transfer path trajectory of lung cancer in different individuals after many years of follow-up.
[0022] The risk transfer statistical analysis unit is used to perform statistical analysis on the probability of lung cancer risk transfer in different groups.
[0023] The connection relationship between the above components is as follows:
[0024] The output end of the multi-source data integration unit is connected to the input end of the data preprocessing unit to generate a data set after the multi-source data is integrated;
[0025] The output end of the data preprocessing unit is connected to the input end of the data grouping unit to generate a preprocessed input data source.
[0026] The output end of the data grouping unit is connected to the input end of the deep neural network model training unit to generate a grouped system input data source.
[0027] The output end of the deep neural network model training unit is connected to the input end of the deep neural network model testing unit to generate a trained deep neural network model.
[0028] The output end of the deep neural network model testing unit is connected to the input end of the deep neural network model verification unit to generate a tested deep neural network model.
[0029] The output end of the deep neural network model verification unit is connected to the input end of the risk factor weight calculation unit to generate a deep neural network model after verification and evaluation.
[0030] The output end of the risk factor weight calculation unit is connected to the input end of the risk level determination unit to generate weight values of different risk factors.
[0031] The output end of the risk level determination unit is connected to the input end of the risk level statistical analysis unit to generate the lung cancer risk levels of different individuals.
[0032] The output end of the risk level statistical analysis unit is connected to the input end of the follow-up risk level determination unit to generate statistical results of lung cancer risk levels in different groups.
[0033] The output end of the follow-up risk level determination unit is connected to the input end of the risk transfer calculation unit to generate the lung cancer incidence risk level of different individuals corresponding to the follow-up year.
[0034] The output end of the risk transfer calculation unit is connected to the input end of the risk transfer statistical analysis unit to generate the risk transfer path trajectory of lung cancer incidence in different individuals.
[0035] The output end of the risk transfer statistical analysis unit outputs the statistical analysis results of the risk transfer probability of lung cancer in different groups.
[0036] Furthermore, the working method of the above system comprises the following steps:
[0037] Step 1: Collect physical examination data, questionnaire data, risk screening data, environmental pollution data, meteorological data and other resources from multiple sources and across fields, integrate them, and obtain input data resources for the judgment system.
[0038] Step 2: Based on the integrated data set obtained in step 1, data preprocessing is performed to clean the multi-source data resources, including white noise removal, missing value filling, data consistency processing, metadata unification processing, etc.
[0039] Step 3: Based on the preprocessed data obtained in step 2, data grouping processing is carried out to group the preprocessed data according to proportion, gender and age.
[0040] Specifically, the following steps may be included:
[0041] Step 3.1: Divide the preprocessed dataset into training set, test set and validation set of the deep neural network model according to the ratio a:b:c.
[0042] Step 3.2: The preprocessed data set is divided into males not older than d years old, females not older than d years old, males older than d years old, and females older than d years old according to age d years old and gender.
[0043] Step 4: Based on the grouped data obtained in step 3, use the training set to carry out deep neural network model training.
[0044] Preferably, the present method provides a specific implementation method for step 4, and those skilled in the art may also use other methods to implement it.
[0045] Specifically, step 4 can be implemented by the following method:
[0046] Step 4.1: Use activation function m and learning rate n to enhance the nonlinear ability of deep neural network and improve the network fitting ability.
[0047] Step 4.2: Optimize and adjust the model parameters, derive the optimal model structure, and solve the number of model input layers e, the number of hidden layers f, the number of output layers g, the number of input nodes h, and the number of output nodes i.
[0048] Step 5: Based on the deep neural network model trained in step 4, use the test set to test the performance of the deep neural network model.
[0049] Step 6: Based on the deep neural network model obtained in step 5, use the validation set to verify and evaluate the generalization ability and model performance of the deep neural network model.
[0050] Step 7: Use the deep neural network model to calculate the weight values of different individual risk factors.
[0051] Specifically, the following steps may be included:
[0052] Step 7.1: Calculate the weight values corresponding to individual risk factors through the deep neural network model, and sort the risk factor weight values of each individual from large to small.
[0053] Step 7.2: Extract the top j important risk factors for each individual and average their weight values.
[0054] Step 8: Based on the risk factor weight values obtained in step 7, determine the individual lung cancer risk level.
[0055] Specifically, the following steps may be included:
[0056] Step 8.1: Sort the weighted mean values of the important risk factors of individuals from large to small. The risk of lung cancer for individuals ranked in the top k% is high risk, the risk of lung cancer for individuals ranked in the bottom p% is low risk, and the risk of lung cancer for individuals ranked between k% and p% is medium risk.
[0057] Step 8.2: Classify individuals according to their risk level for lung cancer.
[0058] Step 9: Based on the lung cancer risk level determined in step 8, statistical analysis of individual risk levels based on gender and age is performed.
[0059] Step 10: For multiple follow-up years corresponding to different individuals, determine their follow-up risk levels using a deep neural network model.
[0060] Specifically, the following steps may be included:
[0061] Step 10.1: Calculate the weight values of individual risk factors through the deep neural network model, and solve the weight mean for the important risk factors ranked in the top j positions.
[0062] Step 10.2: Determine the lung cancer risk level of individuals in different follow-up years according to step 8.1, and classify individuals according to risk level.
[0063] Step 11: Based on the follow-up risk transfer level determination in step 10, calculate the lung cancer risk transfer path trajectory.
[0064] Step 12: Based on the lung cancer risk transfer calculation in step 11, statistically analyze the risk transfer probabilities of different risk levels and summarize and analyze the risk transfer results.
[0065] Beneficial Effects
[0066] Compared with the classification prediction technologies such as multivariate logistic regression, support vector machine, Bayesian network, etc., which are mostly used in the prior art, the system of the present invention has the following advantages:
[0067] 1. This method is the first to build an individual lung cancer risk and metastasis determination system based on a deep neural network model. It has the advantages of high computational efficiency and high system accuracy, and can be used for large-scale population lung cancer screening;
[0068] 2. This system has the advantages of quantitative parallel calculation, fast calculation speed, high risk judgment accuracy, and easy operation. BRIEF DESCRIPTION OF THE DRAWINGS
[0069] Figure 1 Schematic diagram of the principle of the system of the present invention. DETAILED DESCRIPTION
[0070] The system of the present invention is further described in detail below with reference to the accompanying drawings and embodiments.
[0071] Example
[0072] like Figure 1 As shown, a deep learning-based intelligent determination system for lung cancer risk and metastasis includes a multi-source data preprocessing module, a deep learning model construction module, a lung cancer risk determination module and a risk transfer determination module. Among them, the multi-source data preprocessing module includes a multi-source data integration unit, a data preprocessing unit and a data grouping unit; the deep learning model construction module includes a deep neural network model training unit, a deep neural network model testing unit and a deep neural network model verification unit; the lung cancer risk determination module includes a risk factor weight calculation unit, a risk level determination unit and a risk level statistical analysis unit; the risk transfer determination module includes a follow-up risk level determination unit, a risk transfer calculation unit and a risk transfer statistical analysis unit.
[0073] Example
[0074] The working method of this system includes the following steps:
[0075] Step 1: Collect data resources from multiple sources, such as physical examination data, questionnaire survey data, risk screening data, environmental pollution data, and meteorological data, and integrate them to form the input data source Q.
[0076] Specifically, the demographic characteristics, physical health status, personal disease history, family history of disease, smoking and drinking habits, daily living habits, physical exercise habits, occupational environmental exposure conditions, and years of follow-up data of 15,000 residents in a certain area from 1992 to 2020, as well as relevant data resources such as environmental pollution conditions and climatic conditions in the place of residence are integrated to form the system input data resource Q as the input of the lung cancer risk and metastasis determination system.
[0077] Step 2: Based on the multi-source integrated data obtained in step 1, data preprocessing operations are carried out to clean the multi-source data resources. This mainly includes: white noise removal, missing value filling, data consistency processing, metadata unification processing, etc.
[0078] Step 3: Based on the pre-processed data obtained in step 2, data grouping is carried out, and the pre-processed data is grouped according to proportion, gender and age. The specific process is as follows:
[0079] Step 3.1: Divide the preprocessed dataset into training set, test set and validation set of the deep neural network model according to the ratio a:b:c.
[0080] Step 3.2: The preprocessed data set is divided into males not older than d years old, females not older than d years old, males older than d years old, and females older than d years old according to age d years old and gender.
[0081] Step 4: Based on the grouped data obtained in step 3, use the training set to carry out deep neural network model training.
[0082] Specifically, step 4 includes the following steps:
[0083] Step 4.1: Use activation function m and learning rate n to enhance the nonlinear ability of deep neural network and improve the network fitting ability.
[0084] Step 4.2: Optimize and adjust the model parameters, derive the optimal model structure, and solve the number of model input layers e, the number of hidden layers f, the number of output layers g, the number of input nodes h, and the number of output nodes i.
[0085] Step 5: Based on the deep neural network model trained in step 4, use the test set to test the performance of the deep neural network model.
[0086] Step 6: Based on the deep neural network model obtained in step 5, use the validation set to verify and evaluate the generalization ability and model performance of the deep neural network model.
[0087] Step 7: Use the deep neural network model to calculate the weights of different individual risk factors. The specific process is as follows:
[0088] Step 7.1: Calculate the weight values corresponding to individual risk factors through the deep neural network model, and sort the risk factor weight values of each individual from large to small.
[0089] Step 7.2: Extract the top j important risk factors for each individual and average their weight values.
[0090] Step 8: Based on the risk factor weight values obtained in step 7, determine the individual lung cancer risk level. The specific process is as follows:
[0091] Step 8.1: Sort the weighted mean values of the important risk factors of individuals from large to small. The risk of lung cancer for individuals ranked in the top k% is high risk, the risk of lung cancer for individuals ranked in the bottom p% is low risk, and the risk of lung cancer for individuals ranked between k% and p% is medium risk.
[0092] Step 8.2: Classify individuals according to their risk level for lung cancer.
[0093] Step 9: Based on the lung cancer risk level determined in step 8, statistical analysis of individual risk levels based on gender and age is performed.
[0094] Step 10: For multiple follow-up years corresponding to different individuals, determine their follow-up risk levels using a deep neural network model.
[0095] Specifically, the following steps may be included:
[0096] Step 10.1: Calculate the weight values of individual risk factors through the deep neural network model, and solve the weight mean for the important risk factors ranked in the top j positions.
[0097] Step 10.2: Determine the lung cancer risk level of individuals in different follow-up years according to step 8.1, and classify individuals according to risk level.
[0098] Step 11: Based on the follow-up risk transfer level determination in step 10, calculate the lung cancer risk transfer path trajectory.
[0099] Step 12: Based on the lung cancer risk transfer calculation in step 11, statistically analyze the risk transfer probabilities of different risk levels and summarize and analyze the risk transfer results.
[0100] Using the preprocessed data obtained in step 2, a support vector machine model was constructed to determine the risk of lung cancer and its metastasis, and compared with the deep neural network model to further evaluate the performance of the lung cancer risk and metastasis determination system.
[0101] The details are as follows:
[0102] The preprocessed data set was divided into the training set, test set and validation set of the deep neural network model in a ratio of 4:4:2. The deep neural network model was constructed using the follow-up data every three years starting from the follow-up baseline according to the follow-up time. The deep neural network was trained with the training set, and its performance was tested with the test set. The validation set was used to verify and evaluate it. The specific performance indicators are shown in Table 1. Furthermore, the lung cancer incidence risk and metastasis judgment model based on the support vector machine was constructed for different follow-up years. By comparison, it was found that the performance of the deep neural network model was significantly improved compared with the support vector machine model. The model accuracy in each follow-up year was increased by 8.36%, 9.19%, 8.67%, 9.25%, 10.05%, 9.82%, 9.33% and 9.08%, respectively. In addition, the deep neural network training efficiency was high and the speed was fast.
[0103] Table 1 Performance of lung cancer risk and metastasis determination model based on deep neural network
[0104]
[0105] By using the deep learning-based lung cancer risk and metastasis determination system, the lung cancer risk level and metastasis determination were carried out for different gender and age groups, and the incidence rates of high-risk, medium-risk and low-risk lung cancer were statistically analyzed, as shown in Table 2.
[0106] Table 2 The incidence of lung cancer in high-risk and medium-risk patients
[0107]
[0108] The following results were obtained by analyzing the lung cancer risk and metastasis determination system:
[0109] 1. After 10 years of follow-up, the incidence of lung cancer in men over 65 years old, who are at high risk of lung cancer, is higher than that in women over 65 years old, men under 65 years old, and women under 65 years old. The incidence of lung cancer is 1.39 times that of women over 65 years old, 3.05 times that of men under 65 years old, and 4.15 times that of women under 65 years old.
[0110] 2. Regardless of male or female, the incidence of medium-risk lung cancer is much lower than that of high-risk lung cancer. For example, for men over 65 years old, who have the highest incidence of lung cancer, the high-risk incidence was 6.84 times the medium-risk incidence after 10 years of follow-up.
[0111] 3. Through the analysis of the results of lung cancer risk level determination in the grouped population, it can be further seen that the development of a lung cancer risk and metastasis determination system will help improve the efficiency of lung cancer screening, guide targeted lung cancer intervention measures, actively prevent the occurrence and development of lung cancer, and reduce the incidence of lung cancer.
[0112] The specific examples described above are further explanations of the present invention and are not intended to limit the scope of protection of the present invention. All changes and equivalent substitutions made within the principles and spirit of the present invention should be within the scope of protection of the present invention.
Claims
1. A deep learning-based intelligent determination system for lung cancer risk and its metastasis, characterized in that: It includes a multi-source data preprocessing module, a deep learning model building module, a lung cancer risk determination module, and a risk transfer determination module; Among them, the multi-source data preprocessing module includes a multi-source data integration unit, a data preprocessing unit and a data grouping unit; The multi-source data integration unit is used to integrate cross-domain data resources from multiple sources; The data preprocessing unit is used to remove white noise from the integrated data, fill in missing values, perform consistency processing on data with inconsistent structures or units, and perform uniform processing on the same metadata; The data grouping unit is used to group the preprocessed data according to proportion, gender and age; The deep learning model building module includes a deep neural network model training unit, a deep neural network model testing unit, and a deep neural network model verification unit; Wherein, the deep neural network model training unit is used to train the deep learning model using the training set based on the deep neural network algorithm; The deep neural network model testing unit is used to test the deep neural network model using the test set data; The deep neural network model verification unit is used to verify the performance of the deep neural network model using the verification set data; The lung cancer incidence risk determination module includes a risk factor weight calculation unit, a risk level determination unit and a risk level statistical analysis unit; Among them, the risk factor weight calculation unit is used to calculate the weights of different risk factors through a deep neural network model; The risk level determination unit is used to determine and calculate the lung cancer risk level of different individuals using the risk factor weights; The risk level statistical analysis unit is used to perform statistical analysis on the lung cancer risk levels of different groups; The risk transfer determination module includes a follow-up risk level determination unit, a risk transfer calculation unit and a risk transfer statistical analysis unit; Among them, the follow-up risk level determination unit is used to determine the risk factor weights and calculate the individual's lung cancer risk level in different follow-up years; The risk transfer calculation unit is used to calculate the risk transfer path trajectory of lung cancer in different individuals after years of follow-up; The risk transfer statistical analysis unit is used to perform statistical analysis on the probability of lung cancer risk transfer in different groups; The connection relationship between the above components is as follows: The output end of the multi-source data integration unit is connected to the input end of the data preprocessing unit to generate a data set after the multi-source data is integrated; Wherein, the output end of the data preprocessing unit is connected to the input end of the data grouping unit to generate a preprocessed input data source; The output end of the data grouping unit is connected to the input end of the deep neural network model training unit to generate a grouped system input data source; The output end of the deep neural network model training unit is connected to the input end of the deep neural network model testing unit to generate a trained deep neural network model; The output end of the deep neural network model testing unit is connected to the input end of the deep neural network model verification unit to generate a tested deep neural network model; The output end of the deep neural network model verification unit is connected to the input end of the risk factor weight calculation unit to generate a deep neural network model after verification and evaluation; The output end of the risk factor weight calculation unit is connected to the input end of the risk level determination unit to generate the weight values of different risk factors; The output end of the risk level determination unit is connected to the input end of the risk level statistical analysis unit to generate the lung cancer risk level of different individuals; The output end of the risk level statistical analysis unit is connected to the input end of the follow-up risk level determination unit to generate statistical results of lung cancer risk levels in different groups; The output end of the follow-up risk level determination unit is connected to the input end of the risk transfer calculation unit to generate the lung cancer incidence risk level of different individuals corresponding to the follow-up year; The output end of the risk transfer calculation unit is connected to the input end of the risk transfer statistical analysis unit to generate the path trajectory of lung cancer risk transfer for different individuals; The output end of the risk transfer statistical analysis unit outputs the statistical analysis results of the risk transfer probability of lung cancer in different groups.
2. The deep learning-based intelligent determination system for lung cancer risk and its metastasis as claimed in claim 1, characterized in that: Multi-source data include physical examination data, risk screening data, follow-up data, environmental pollution data, meteorological data, and questionnaire survey data.
3. The deep learning-based intelligent determination system for lung cancer risk and its metastasis as claimed in claim 1, characterized in that: The working method of the system includes the following steps: Step 1: Collect resources from multiple sources and across fields, integrate them, and obtain input data resources for the judgment system; Step 2: Perform data preprocessing based on the integrated data set obtained in step 1 and perform data cleaning on multi-source data resources; Step 3: Based on the preprocessed data obtained in step 2, data grouping processing is carried out to group the preprocessed data according to proportion, gender and age; Step 4: Based on the grouped data obtained in step 3, use the training set to carry out deep neural network model training; Step 5: Based on the deep neural network model trained in step 4, the performance of the deep neural network model is tested using the test set; Step 6: Based on the deep neural network model obtained in step 5, the generalization ability and model performance of the deep neural network model are verified and evaluated using the validation set; Step 7: Use the deep neural network model to calculate the weight values of different individual risk factors; Step 8: Based on the risk factor weight values obtained in step 7, determine the individual lung cancer risk level; Step 9: Based on the lung cancer risk level determined in step 8, statistical analysis of individual risk levels based on gender and age is performed; Step 10: For multiple follow-up years corresponding to different individuals, determine their follow-up risk levels using a deep neural network model; Step 11: Based on the follow-up risk transfer level determination in step 10, calculate the lung cancer risk transfer path trajectory; Step 12: Based on the lung cancer risk transfer calculation in step 11, statistically analyze the risk transfer probabilities of different risk levels and summarize and analyze the risk transfer results.
4. The deep learning-based intelligent determination system for lung cancer risk and its metastasis as claimed in claim 3, characterized in that: Step 3 includes the following steps: Step 3.1: Divide the preprocessed data set into the training set, test set, and validation set of the deep neural network model according to the ratio a:b:c; Step 3.2: The preprocessed data set is divided into males not older than d years old, females not older than d years old, males older than d years old, and females older than d years old according to age d years old and gender.
5. The deep learning-based intelligent determination system for lung cancer risk and its metastasis as claimed in claim 3, characterized in that: Step 4 includes: Step 4.1: Use activation function m and learning rate n to enhance the nonlinear ability of deep neural network and improve the network fitting ability; Step 4.2: Optimize and adjust the model parameters, derive the optimal model structure, and solve the number of model input layers e, the number of hidden layers f, the number of output layers g, the number of input nodes h, and the number of output nodes i.
6. The deep learning-based intelligent determination system for lung cancer risk and its metastasis as claimed in claim 3, characterized in that: Step 7 includes: Step 7.1: Calculate the weight values corresponding to individual risk factors through the deep neural network model, and sort the weight values of risk factors for each individual from large to small; Step 7.2: Extract the top j important risk factors for each individual and average their weight values.
7. The deep learning-based intelligent determination system for lung cancer risk and its transfer as claimed in claim 3, wherein step 8 comprises the following steps: Step 8.1: Sort the weighted mean values of the important risk factors of individuals from large to small. The risk of lung cancer of individuals ranked in the top k% is high risk, the risk of lung cancer of individuals ranked in the bottom p% is low risk, and the risk of lung cancer of individuals ranked between k% and p% is medium risk; Step 8.2: Classify individuals according to their risk level for lung cancer.
8. The deep learning-based intelligent determination system for lung cancer risk and its metastasis as claimed in claim 3, wherein step 10 comprises the following steps: Step 10.1: Calculate the weights of individual risk factors through the deep neural network model, and solve the weight mean for the top j important risk factors; Step 10.2: Determine the lung cancer risk level of individuals in different follow-up years and classify individuals according to risk level.
Citation Information
Cited By
Lung cancer onset risk metastasis path construction system and method based on causal knowledge network
CN120932882A
