Deep learning model-based lung cancer patient screening method and device, terminal and medium

Automatically screening of lung cancer patients through deep learning models and BERT models has solved the problem of slow information collection and insufficient accuracy in the screening methods of traditional lung cancer patients, and achieved efficient and accurate lung cancer screening, reducing manual intervention and subjectivity, and improving screening efficiency and early detection rate.

CN119049731APending Publication Date: 2024-11-29NORTH CHINA DIGITAL HEALTH TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411236509.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-04
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The traditional screening method of lung cancer patients relies on manual screening, resulting in slow information collection, insufficient accuracy, labor and material resources, and diagnostic errors, and rely on the experience and subjective judgment of medical personnel, resulting in poor reusability and relatively poor accuracy.

Method used

Using the deep learning model-based screening method for lung cancer patients, lung cancer patients are automatically screened from the medical record data through the BERT model, and patient data is obtained using the medical record information system, preprocessing and feature annotation, and sample sets are constructed and model trained to realize automatic screening.

Benefits of technology

It improves the efficiency and accuracy of screening of lung cancer patients, reduces the rate of missed diagnosis, reduces manual intervention, improves the rate of early detection, avoids subjectivity and limitations, and improves screening efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119049731A_ABST
    Figure CN119049731A_ABST
Patent Text Reader

Abstract

The invention relates to the field of patient screening, and particularly discloses a lung cancer patient screening method and device based on a deep learning model, a terminal and a medium. Labeling a plurality of characteristics representing lung cancer in the illness state data; constructing a sample set by using the plurality of disease condition data of the plurality of labeled patients; training a lung cancer patient screening model by using the sample set, wherein the lung cancer patient screening model is a BERT model; and obtaining a plurality of to-be-screened disease condition data, and screening out the disease condition data of the lung cancer patient from the plurality of to-be-screened disease condition data through the trained lung cancer patient screening model. The lung cancer patients are automatically screened from the medical records through the deep learning model, and the screening efficiency and precision are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of patient screening, and in particular to a method, device, terminal and medium for screening lung cancer patients based on a deep learning model. Background Art

[0002] In the medical field, the treatment and research of specific diseases usually need to consider the individual differences and disease characteristics of patients. The demand for establishing a database platform for major diseases and screening patients with similar diseases from massive data is growing rapidly. However, the traditional screening of lung cancer patients currently relies more on follow-up, manual screening and other methods.

[0003] First, research experts put forward requirements for disease research on lung cancer patients based on research scenarios. Then engineers collect lung cancer patients through follow-up and manual SQL screening of medical records. Medical personnel review and include the population required for research. If the population fails the review, it will be filtered out. If the number of patients in the required research population is insufficient, it will be returned to the engineer to continue screening the patient population. Researchers search for lung cancer patient medical record information based on the index information of the included patients, screen out patients with complete and available patient medical record information, and include them in the research population database. Patients with incomplete medical record information or lack of necessary treatment data (clinical manifestations, medical history, examinations, tests, medical orders, and electronic medical record information) are excluded, and the exclusion continues until the available research population data meets the research requirements.

[0004] The traditional screening method has the disadvantages of slow patient information collection, little medical record data information, insufficient accuracy in patient screening, and heavy reliance on the experience and knowledge of medical personnel. There are differences in personal experience and knowledge levels of medical personnel, and the diversity of disease manifestations. This method is limited by the personal experience and subjective judgment of medical personnel, and is prone to diagnostic errors and consistency problems. At the same time, due to the lack of manual comparison of medical record content, the screened patients are unavailable and the patients have to be rescreened. Moreover, this situation has the disadvantages of high medical requirements for engineers, large workload for statistical data, high consumption of manpower and material resources, poor reusability, and relatively poor accuracy. Summary of the invention

[0005] To solve the above problems, the present invention provides a method, device, terminal and medium for screening lung cancer patients based on a deep learning model, which automatically screens lung cancer patients from medical records through a deep learning model to improve screening efficiency and accuracy.

[0006] In a first aspect, the technical solution of the present invention provides a method for screening lung cancer patients based on a deep learning model, comprising the following steps: Obtaining certain medical condition data of certain patients; Label several features that characterize lung cancer in the disease data; Construct a sample set using the labeled disease data of several patients; Using the sample set to train a lung cancer patient screening model, wherein the lung cancer patient screening model is a BERT model; A plurality of condition data to be screened are obtained, and the condition data of lung cancer patients are screened out from the plurality of condition data to be screened by using the trained lung cancer patient screening model.

[0007] In an optional embodiment, obtaining a plurality of condition data of a plurality of patients specifically includes: Enter the hospital medical record information system, enter the search criteria, filter out the medical record list that matches the search criteria, and export the medical record list into a text file; the search criteria include the date of consultation, department, and diagnosis keywords; Extract each patient ID from the medical record list; Extract patient-related data from the hospital patient management system based on each patient identifier, including admission records, imaging data and examination report information, and medical record report information; Preprocess patient-related data and patient medical record data to obtain patient condition data, including word segmentation, stop word removal, and stem extraction.

[0008] In an optional embodiment, several features characterizing lung cancer in the disease data are annotated, specifically including; Convert patient condition data into JSON format and submit it to the annotation tool; Use annotation tools to annotate several features that characterize lung cancer in the disease data.

[0009] In an optional embodiment, in the process of training a lung cancer patient screening model using a sample set, the input data is forward propagated through the model, the loss is calculated, and then the gradient is calculated through backpropagation, and finally the optimizer is used to update the model parameters.

[0010] In an optional embodiment, the lung cancer patient screening model is trained to screen the condition data of the lung cancer patient from the plurality of condition data to be screened, specifically including; Number and sort the screened disease data; According to the order of sorting from front to back, the disease data is input into the trained lung cancer patient screening model for processing to obtain the recognition result; Extract the medical condition data of patients whose identification results are lung cancer; Obtain original patient-related data and patient medical record data based on the condition data.

[0011] In a second aspect, the technical solution of the present invention provides a lung cancer patient screening device based on a deep learning model, comprising: Condition data acquisition module: obtains condition data of several patients; Feature annotation module: annotates several features that characterize lung cancer in the disease data; Sample set construction module: constructs a sample set from several annotated disease data of several patients; Model training module: using the sample set to train a lung cancer patient screening model, wherein the lung cancer patient screening model is a BERT model; Patient screening module: obtains a number of condition data to be screened, and screens out the condition data of lung cancer patients from the number of condition data to be screened through the trained lung cancer patient screening model.

[0012] In a third aspect, the technical solution of the present invention provides a terminal, including: A memory, used for storing a lung cancer patient screening program based on a deep learning model; A processor is used to implement the steps of any of the above-mentioned methods for screening lung cancer patients based on a deep learning model when executing the lung cancer patient screening program based on a deep learning model.

[0013] In a fourth aspect, the technical solution of the present invention provides a computer-readable storage medium, on which is stored a lung cancer patient screening program based on a deep learning model. When the lung cancer patient screening program based on a deep learning model is executed by a processor, the steps of the lung cancer patient screening method based on a deep learning model as described in any one of the above items are implemented.

[0014] The present invention provides a method, device, terminal and medium for screening lung cancer patients based on a deep learning model, which has the following beneficial effects compared with the prior art: the method for screening lung cancer patients based on a deep learning model can automatically extract and learn the characteristics of lung cancer from a large amount of disease data, avoiding the limitations and subjectivity of manual screening, thereby improving the efficiency and accuracy of screening. As one of the most advanced natural language processing models, the BERT model performs well in text understanding and classification tasks, and can more accurately identify lung cancer-related information in disease data. By constructing a sample set containing a large number of lung cancer patients and non-lung cancer patients, and fully training the model, the method can learn the complex patterns and laws of lung cancer patients' disease data, thereby effectively reducing the missed diagnosis rate and improving the detection rate of early lung cancer. The deep learning model can automatically process a large amount of disease data without manual intervention, greatly improving the screening efficiency. BRIEF DESCRIPTION OF THE DRAWINGS In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 paying creative work.

[0015] Figure 1 It is a flowchart of a method for screening lung cancer patients based on a deep learning model provided in an embodiment of the present invention.

[0016] Figure 2 It is a schematic block diagram of the structure of a lung cancer patient screening device based on a deep learning model provided in an embodiment of the present invention.

[0017] Figure 3 It is a schematic diagram of the structure of a terminal provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the scheme of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0020] Figure 1 : is a schematic diagram of a method for screening lung cancer patients based on a deep learning model provided by an embodiment of the present invention. Figure 1 The execution subject may be a lung cancer patient screening device based on a deep learning model. The lung cancer patient screening method based on a deep learning model provided in an embodiment of the present invention is executed by a computer device, and accordingly, the lung cancer patient screening device based on a deep learning model runs in the computer device. According to different requirements, the order of the steps in the flowchart can be changed, and some can be omitted.

[0021] like Figure 1 As shown, the method includes the following steps.

[0022] S1, obtaining some medical condition data of some patients.

[0023] The medical condition data acquired in this embodiment includes medical records, admission records, examination data, test data, etc. To achieve the acquisition of the medical condition data, the following steps are performed.

[0024] S1.1, enter the hospital medical record information system, input the search conditions, filter out the medical record list matching the search conditions, and export the medical record list into a text file; wherein the search conditions include the date of consultation, department, and diagnosis keywords.

[0025] Users search for patients and obtain the medical record data of these patients through the hospital medical record information system. They log in to the hospital medical record information system and enter the search criteria, including the date of consultation, department, diagnosis keywords, etc. The system will display a list of medical records that match the search criteria and export them to the intranet system in the form of a text file.

[0026] S1.2, extract the patient IDs in the medical record list.

[0027] The medical record list obtained in step S1.1 includes a patient ID, which refers to the patient's unique identity code, which can be an ID number, medical card number, etc. The patient ID is extracted to match the corresponding other medical condition data.

[0028] S1.3, extract patient-related data from the hospital patient management system based on each patient identifier, including admission records, imaging data and examination report information, and medical record report information.

[0029] The disease condition data of this embodiment also includes data such as admission records, imaging data and examination report information, and medical history report information. These data are extracted from the hospital management system using the patient identifier extracted in step S1.2.

[0030] Admission records: chief complaint, medical history, past medical history, family history, smoking history, occupational history and other information related to lung cancer.

[0031] Imaging data and examination report information: including X-ray, CT scan, MRI, PET-CT, lymph node examination, ultrasound examination, lung function, electrocardiogram and other imaging examination results and examination conclusion information.

[0032] Pathology report information: Marks pathological diagnosis, tumor staging and grading, and description of pathological specimens. The description of pathological specimens in the pathology report, such as size, shape, color, presence of tumors, and size, color, morphology, and texture of the tumor, are fields that are of great value for understanding the characteristics of the tumor and formulating surgical plans.

[0033] S1.4, pre-process patient-related data and patient medical record data to obtain patient condition data, including word segmentation, stop word removal, and stem extraction.

[0034] Patient-related data and patient medical record data constitute the condition data. These data are pre-processed by word segmentation, stop word removal, stem extraction, etc. to eliminate redundancy and noise and ensure that the model can accurately capture the key information in the text.

[0035] S2, labeling several features representing lung cancer in the disease data.

[0036] First, convert the patient's condition data into json format and submit it to the annotation tool. Use the annotation tool to annotate several features that characterize lung cancer in the condition data. The annotation data mainly organizes the screened medical record data into json format and submits it to the annotation tool. Users annotate the content that can represent the characteristics of lung cancer based on the diagnosis and clinical manifestations. The main content includes the annotation of medical record content that is clearly marked as lung cancer, lung malignant tumor, lung Ca, bronchial malignant tumor, bronchial cancer CA, etc. in the diagnosis. Information such as the description of confirmed lung cancer, radiotherapy or chemotherapy, description of lung malignant tumor surgery, and description of the examination conclusion confirming lung malignant tumor or lung cancer in the admission record. The annotation data also needs to annotate the features in the medical record file, such as the location, size, shape, etc. of the tumor, and the tumor that is clearly determined to be a lung malignant tumor will be annotated.

[0037] S3, constructing a sample set from the labeled disease data of several patients.

[0038] A sample library is established using the labeled sample patient condition data. Based on the medical record information in the sample library and the determined screening indicators, screening rules are formulated to achieve accurate screening of lung cancer patients.

[0039] S4, using the sample set to train a lung cancer patient screening model, wherein the lung cancer patient screening model is a BERT model.

[0040] This example uses the BERT model as a lung cancer patient screening model for training. The annotated disease data is exported to the intranet in the form of json for model training. The data needs to be divided into a training set of about 3,000 copies, a validation set, and a test set (about 300 copies). The size of the validation set can be adjusted as needed.

[0041] Initialize the model's weights and biases (parameters) for training. Initialization methods include random initialization and pre-trained model loading. Train the model on the training set. During training, the input data is forward propagated through the model, the loss is calculated, and then the gradient is calculated through backpropagation (automatic differentiation). Finally, the optimizer is used to update the model parameters. This process requires multiple iterations. The model return parameters mainly include the patient number and index number of the corresponding medical record.

[0042] Model performance and accuracy are periodically evaluated on the validation set to detect overfitting or undertraining. The accuracy and reliability of the screening model are ensured by continuously verifying and adjusting the screening model parameters and optimizing them according to new medical records and clinical data. The accuracy rate, loss and other indicators can be monitored to measure model performance and accuracy.

[0043] You can also continuously adjust hyperparameters, including learning rate, batch size, model complexity, etc., based on the performance and accuracy of the validation set to improve model performance. Finally, use an independent test set to evaluate the performance and accuracy of the final model. When the performance and accuracy requirements are met, the model can be deployed to actual application tools.

[0044] S5, obtaining a number of condition data to be screened, and screening the condition data of lung cancer patients from the number of condition data to be screened using the trained lung cancer patient screening model.

[0045] Step S4 completes the training of the lung cancer patient screening model, and then uses the trained lung cancer patient screening model to screen new lung cancer patients to obtain a large amount of lung cancer patients' condition data for scientific research.

[0046] S5.1, number and sort the screened disease data.

[0047] S5.2, according to the order of sorting from front to back, the disease condition data is input into the trained lung cancer patient screening model for processing to obtain the recognition result.

[0048] S5.3, extracting the medical condition data of the patient whose identification result is lung cancer.

[0049] S5.4, obtain original patient-related data and patient medical record data based on the condition data.

[0050] When obtaining a number of disease data to be screened, first perform operations such as word segmentation, stop word removal, and stem extraction to eliminate redundancy and noise, and retain patient identification during the preprocessing process. Then, the screened disease data are numbered and sorted, and processed in sequence through the trained lung cancer patient screening model according to the sorting. The model automatically analyzes the text information in the disease data to determine whether there are lung cancer-related diagnostic or feature words, and to determine whether the patient is a qualified lung cancer patient, and output the recognition result. The output recognition result contains the patient identification, and the patient identification is used to extract the corresponding original patient-related data and patient medical record data for the next step of scientific research to complete the accurate screening of new lung cancer patients.

[0051] An embodiment of a method for screening lung cancer patients based on a deep learning model is described in detail above. Based on the method for screening lung cancer patients based on a deep learning model described in the above embodiment, an embodiment of the present invention also provides a device for screening lung cancer patients based on a deep learning model corresponding to the method.

[0052] Figure 2 is a schematic block diagram of the structure of a lung cancer patient screening device based on a deep learning model provided by an embodiment of the present invention. The lung cancer patient screening device 200 based on a deep learning model can be divided into multiple functional modules according to the functions it performs, such as Figure 2 The functional modules may include: a disease data acquisition module 210, a feature annotation module 220, a sample set construction module 230, a model training module 240, and a patient screening module 250. The module referred to in the present invention refers to a series of computer program segments that can be executed by at least one processor and can perform fixed functions, which are stored in a memory.

[0053] The condition data acquisition module 210 is used to acquire condition data of a number of patients.

[0054] Feature labeling module 220: labels several features representing lung cancer in the disease condition data.

[0055] The sample set construction module 230 constructs a sample set from the labeled condition data of the patients.

[0056] Model training module 240: Use the sample set to train a lung cancer patient screening model, where the lung cancer patient screening model is a BERT model.

[0057] Patient screening module 250: obtains a number of condition data to be screened, and screens out condition data of lung cancer patients from the number of condition data to be screened using the trained lung cancer patient screening model.

[0058] The lung cancer patient screening device based on deep learning model of this embodiment is used to implement the aforementioned lung cancer patient screening method based on deep learning model. Therefore, the specific implementation method of the device can be seen in the embodiment part of the lung cancer patient screening method based on deep learning model in the previous text. Therefore, its specific implementation method can refer to the description of the corresponding embodiments of each part, which will not be elaborated here.

[0059] In addition, since the lung cancer patient screening device based on the deep learning model of this embodiment is used to implement the aforementioned lung cancer patient screening method based on the deep learning model, its function corresponds to that of the above method and will not be repeated here.

[0060] Figure 3A schematic diagram of the structure of a terminal 300 provided in an embodiment of the present invention includes: a processor 310, a memory 320 and a communication unit 330. The processor 310 is used to implement the following steps when implementing the lung cancer patient screening program based on the deep learning model stored in the memory 320: Obtaining certain medical condition data of certain patients; Label several features that characterize lung cancer in the disease data; Construct a sample set using the labeled disease data of several patients; Using the sample set to train a lung cancer patient screening model, wherein the lung cancer patient screening model is a BERT model; A plurality of condition data to be screened are obtained, and the condition data of lung cancer patients are screened out from the plurality of condition data to be screened by using the trained lung cancer patient screening model. The terminal 300 includes a processor 310, a memory 320 and a communication unit 330. These components communicate via one or more buses. It can be understood by those skilled in the art that the structure of the server shown in the figure does not constitute a limitation of the present invention, and it can be a bus structure or a star structure, and can also include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0061] The memory 320 can be used to store the execution instructions of the processor 310, and the memory 320 can be implemented by any type of volatile or non-volatile storage terminal or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. When the execution instructions in the memory 320 are executed by the processor 310, the terminal 300 can perform some or all of the steps in the following method embodiments.

[0062] The processor 310 is the control center of the storage terminal, and uses various interfaces and lines to connect various parts of the entire electronic terminal. It runs or executes software programs and / or modules stored in the memory 320, and calls data stored in the memory to perform various functions of the electronic terminal and / or process data. The processor can be composed of an integrated circuit (IC), for example, it can be composed of a single packaged IC, or it can be composed of multiple packaged ICs with the same or different functions. For example, the processor 310 can only include a central processing unit (CPU). In the embodiment of the present invention, the CPU can be a single computing core or multiple computing cores.

[0063] The communication unit 330 is used to establish a communication channel so that the storage terminal can communicate with other terminals, receive user data sent by other terminals or send user data to other terminals.

[0064] The present invention also provides a computer storage medium, wherein the storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM).

[0065] The computer storage medium stores a lung cancer patient screening program based on a deep learning model, and when the lung cancer patient screening program based on a deep learning model is executed by a processor, the following steps are implemented: Obtaining certain medical condition data of certain patients; Label several features that characterize lung cancer in the disease data; Construct a sample set using the labeled disease data of several patients; Using the sample set to train a lung cancer patient screening model, wherein the lung cancer patient screening model is a BERT model; A plurality of condition data to be screened are obtained, and the condition data of lung cancer patients are screened out from the plurality of condition data to be screened by using the trained lung cancer patient screening model.

[0066] Those skilled in the art can clearly understand that the technology in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solution in the embodiments of the present invention, in essence or in other words, the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, and other media that can store program codes, including several instructions for enabling a computer terminal (which can be a personal computer, a server, or a second terminal, a network terminal, etc.) to execute all or part of the steps of the methods described in each embodiment of the present invention.

[0067] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0068] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0069] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0070] The above disclosure is only a preferred embodiment of the present invention, but the present invention is not limited thereto. Any non-creative changes that can be thought of by a person skilled in the art, as well as several improvements and modifications made without departing from the principle of the present invention, should fall within the protection scope of the present invention.

Claims

1. A method for screening lung cancer patients based on a deep learning model, characterized in that: The following steps are involved: Obtaining certain medical condition data of certain patients; Label several features that characterize lung cancer in the disease data; Construct a sample set using the labeled disease data of several patients; Using the sample set to train a lung cancer patient screening model, wherein the lung cancer patient screening model is a BERT model; A plurality of condition data to be screened are obtained, and the condition data of lung cancer patients are screened out from the plurality of condition data to be screened by using the trained lung cancer patient screening model.

2. The method for screening lung cancer patients based on a deep learning model according to claim 1, characterized in that: Obtain some medical data of some patients, including: Enter the hospital medical record information system, enter the search criteria, filter out the medical record list that matches the search criteria, and export the medical record list into a text file; the search criteria include the date of consultation, department, and diagnosis keywords; Extract each patient ID from the medical record list; Extract patient-related data from the hospital patient management system based on each patient identifier, including admission records, imaging data and examination report information, and medical record report information; Preprocess patient-related data and patient medical record data to obtain patient condition data, including word segmentation, stop word removal, and stem extraction.

3. The method for screening lung cancer patients based on a deep learning model according to claim 2, characterized in that: Label several features of lung cancer in the disease data, including: Convert patient condition data into JSON format and submit it to the annotation tool; Use annotation tools to annotate several features that characterize lung cancer in the disease data.

4. The method for screening lung cancer patients based on a deep learning model according to any one of claims 1 to 3, characterized in that: When training the lung cancer patient screening model using the sample set, the input data is forward propagated through the model to calculate the loss, then the gradient is calculated through backpropagation, and finally the optimizer is used to update the model parameters.

5. The method for screening lung cancer patients based on a deep learning model according to any one of claims 1 to 3, characterized in that: The trained lung cancer patient screening model selects the condition data of lung cancer patients from a number of condition data to be screened, including: Number and sort the screened disease data; According to the order of sorting from front to back, the disease data is input into the trained lung cancer patient screening model for processing to obtain the recognition result; Extract the medical condition data of patients whose identification results are lung cancer; Obtain original patient-related data and patient medical record data based on the condition data.

6. A lung cancer patient screening device based on a deep learning model, characterized in that: include, Condition data acquisition module: obtains condition data of several patients; Feature annotation module: annotates several features that characterize lung cancer in the disease data; Sample set construction module: constructs a sample set from several annotated disease data of several patients; Model training module: using the sample set to train a lung cancer patient screening model, wherein the lung cancer patient screening model is a BERT model; Patient screening module: obtains a number of condition data to be screened, and screens out the condition data of lung cancer patients from the number of condition data to be screened through the trained lung cancer patient screening model.

7. A terminal, characterized in that: include: A memory, used for storing a lung cancer patient screening program based on a deep learning model; A processor, configured to implement the steps of the method for screening lung cancer patients based on a deep learning model as described in any one of claims 1 to 5 when executing the lung cancer patient screening program based on a deep learning model.

8. A computer-readable storage medium, characterized in that:The readable storage medium stores a lung cancer patient screening program based on a deep learning model. When the lung cancer patient screening program based on a deep learning model is executed by a processor, the steps of the lung cancer patient screening method based on a deep learning model as described in any one of claims 1 to 5 are implemented.