Tumor information question and answer processing method, tumor information recognition model training method and device, computer equipment, storage medium and computer program product

In the tumor information question and answer processing method, the tumor information recognition model generated by the model fine-tuning training technology is solved, and efficient and accurate tumor information recognition and personalized output results are achieved.

CN119028562APending Publication Date: 2024-11-26SHANGHAI ORIGIMED CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410821114.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-24
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

In the prior art, the pathological recognition model obtained by using pathological information training is low in accuracy when used for the identification of tumor-related information, and it is difficult to achieve efficient and accurate tumor information recognition.

Method used

Provide a tumor information question and answer processing method. By obtaining and analyzing the tumor information question and answer request sent by the terminal, inputting the to-process tumor type data, tumor stage data, gene data and treated data into the tumor information identification model, and using the model fine-tuning training technology Generate pre-generated question and answer models to improve identification accuracy.

Benefits of technology

Targeted output results are achieved through individualized pending data, which significantly improves the accuracy and adaptability of tumor information identification and improves the ability to identify the association between tumor type, stage, gene and treatment data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119028562A_ABST
    Figure CN119028562A_ABST
Patent Text Reader

Abstract

The invention relates to a tumor information question and answer processing method, a tumor information recognition model training method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring a tumor information question and answer request sent by a terminal, analyzing a tumor information inquiry statement carried in the tumor information question and answer request, extracting to-be-processed tumor type data, to-be-processed tumor stage data, to-be-processed gene data and to-be-processed treated data included in the tumor information inquiry statement, the to-be-processed tumor type data, the to-be-processed tumor stage data, the to-be-processed gene data and the to-be-processed treated data are input into a tumor information recognition model, result data output by the tumor information recognition model are received, a display instruction is sent to a terminal, and the display instruction carries the result data. By adopting the method, the tumor information processing accuracy can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method for processing tumor information questions and answers, a method, device, computer equipment, storage medium and computer program product for training a tumor information recognition model. Background Art

[0002] With the development of computer technology in the medical field, more and more computer technology is used to generate pathology recognition models to realize the recognition and confirmation of pathology information. Computer technology can also be used to generate treatment information for confirmed pathology information.

[0003] In the related art, the pathology recognition model is mainly obtained by training using the collected pathology information. However, the pathology recognition model obtained in this way is mainly used to determine various pathologies, but has low accuracy when used to recognize tumor-related information. Summary of the invention

[0004] Based on this, it is necessary to provide a tumor information question and answer processing method, a tumor information recognition model training method, device, computer equipment, computer-readable storage medium and computer program product that can improve the accuracy of information determination in response to the above technical problems.

[0005] In a first aspect, the present application provides a tumor information question-answering processing method, comprising:

[0006] Obtaining a tumor information question and answer request sent by the terminal, wherein the tumor information question and answer request data carries a tumor information query statement;

[0007] Parsing the tumor information query statement, extracting the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed included in the tumor information query statement;

[0008] Inputting the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed into the tumor information recognition model, and receiving the result data corresponding to the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed output by the tumor information recognition model; wherein the tumor information recognition model is obtained by training the pre-generated question-answering model using the model fine-tuning training technology according to the sample data; the sample data includes tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and sample data of the association relationship between any two of the tumor type sample data, tumor stage sample data, gene sample data, and treatment sample data;

[0009] A display instruction is sent to the terminal device, the display instruction carries the result data, and the display instruction is used to instruct the terminal to display the result data.

[0010] In one embodiment, before the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed are input into the tumor information recognition model, the method also includes: obtaining literature report data; identifying sample data included in the literature report data, the sample data including tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data, and treatment sample data; obtaining a pre-generated question-answering model, using model fine-tuning training technology to train the sample data, and generating a tumor information recognition model.

[0011] In one embodiment, identifying sample data included in literature report data includes: identifying sample data included in literature report data to obtain initial tumor type data, initial tumor stage data, initial gene data, initial treatment data, and an initial association relationship between any two of the initial tumor type data, initial tumor stage data, initial gene data, and initial treatment data; clustering the initial tumor type data, initial tumor stage data, initial gene data, and initial treatment data to obtain tumor type sample data, tumor stage sample data, gene sample data, and treatment data sample data; generating association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data, and treatment data sample data according to the initial association relationship; generating sample data according to the tumor type sample data, tumor stage sample data, gene sample data, treatment data sample data, and association relationship sample data.

[0012] In one embodiment, the initial treatment data includes initial medication data, and generating sample data based on tumor type sample data, tumor stage sample data, gene sample data, treatment data sample data, and associated relationship sample data includes: obtaining data source information corresponding to the initial medication data, level data of the data source information, and effect data of the initial medication data; determining priority information of the initial medication data based on the data source information, the level data of the data source information, and the effect data of the initial medication data; generating sample data based on the tumor type sample data, tumor stage sample data, gene sample data, treatment data sample data, priority information of the initial medication data, and associated relationship sample data.

[0013] In a second aspect, the present application provides a tumor information recognition model training method, comprising:

[0014] Obtain literature report data;

[0015] Identify sample data included in the literature report data, the sample data including tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and sample data of association relationships between any two of the tumor type sample data, tumor stage sample data, gene sample data, and treatment sample data;

[0016] Obtain a pre-generated question-answering model, use model fine-tuning training technology, train sample data, and generate a tumor information recognition model.

[0017] In a third aspect, the present application provides a tumor information question-answering processing device, comprising:

[0018] A request acquisition module is used to acquire a tumor information question and answer request sent by a terminal, wherein the tumor information question and answer request data carries a tumor information query statement;

[0019] A parsing module, used for parsing tumor information query statements, and extracting tumor type data to be processed, tumor stage data to be processed, gene data to be processed, and treated data to be processed included in the tumor information query statements;

[0020] A result acquisition module is used to input the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed into the tumor information recognition model, and receive the result data output by the tumor information recognition model corresponding to the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed; wherein the tumor information recognition model is obtained by training the pre-generated question-answering model using the model fine-tuning training technology according to the sample data; the sample data includes tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and sample data of the association relationship between any two of the tumor type sample data, tumor stage sample data, gene sample data, and treatment sample data;

[0021] The instruction generation module is used to send a display instruction to the terminal, the display instruction carries result data, and the display instruction is used to instruct the terminal to display the result data.

[0022] In a fourth aspect, the present application provides a tumor information recognition model training, including:

[0023] Literature acquisition module, used to obtain literature report data;

[0024] A sample data acquisition module is used to identify sample data included in the literature report data, the sample data including tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data and treatment sample data;

[0025] The training module is used to obtain the pre-generated question-answering model, use the model fine-tuning training technology, train the sample data, and generate a tumor information recognition model.

[0026] In a fifth aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0027] In a sixth aspect, the present application further provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0028] In a seventh aspect, the present application also provides a computer program product, including a computer program, which implements the steps in the above-mentioned method embodiments when executed by a processor.

[0029] The above-mentioned tumor information question and answer processing method, tumor information recognition model training method, device, computer equipment, storage medium and computer program product use individualized tumor type data to be processed, tumor stage data to be processed, gene data to be processed and treated data to be processed to achieve recognition, which can improve the accuracy of tumor information recognition, and can achieve targeted output result data based on individualized tumor type data to be processed, tumor stage data to be processed, gene data to be processed and treated data to be processed, thereby effectively improving adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0031] FIG1 is a diagram showing an application environment of a tumor information question-answering processing method and a tumor information recognition model training method in one embodiment;

[0032] FIG2 is a schematic diagram of a flow chart of a tumor information question-answering processing method in one embodiment;

[0033] 3(a)-3(d) are schematic diagrams of a tumor information question-answering processing method in one embodiment;

[0034] Figure 4 is a flowchart of a tumor information recognition model training method in one embodiment;

[0035] Figure 5 is a structural block diagram of a tumor information question-answering processing device in one embodiment;

[0036] Figure 6 is a structural block diagram of a tumor information recognition model training device in one embodiment;

[0037] FIG. 7 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0039] The tumor information question and answer processing method provided in the embodiment of the present application can be applied to the application environment shown in Figure 1. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The server 104 obtains the tumor information question and answer request sent by the terminal 102, and the server 104 parses the tumor information query statement carried in the tumor information question and answer request, extracts the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed included in the tumor information query statement, and inputs the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed into the tumor information recognition model, receives the result data corresponding to the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed output by the tumor information recognition model, and the server 104 sends a display instruction to the terminal 102, and the display instruction carries the result data, and the display instruction is used to instruct the terminal 102 to display the result data. The terminal 102 may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers.

[0040] In an exemplary embodiment, as shown in FIG2 , a tumor information question-answering processing method is provided, which is described by taking the method applied to the server in FIG1 as an example, and includes the following steps 202 to 208. Among them:

[0041] Step 202: Obtain a tumor information question and answer request sent by a terminal.

[0042] Optionally, the tumor information question and answer request carries a tumor information query statement.

[0043] In one embodiment, the user can use the terminal to input a tumor information query statement. After the tumor information query statement is input, the terminal generates a tumor information question and answer request based on the acquired tumor information query statement.

[0044] In one embodiment, a tumor information inquiry software may be formed on the terminal, and the user may use the tumor information inquiry software to input a tumor information inquiry statement.

[0045] In one embodiment, the input content can be limited so that the tumor information query statement that the terminal can obtain meets the requirements. For example, it is limited to input tumor type, tumor stage data, gene data, and treatment data. Optionally, the terminal can detect the input content to detect whether the input content meets the requirements. If it does not meet the requirements, a prompt message can be output so that the user can enter a statement that meets the requirements according to the prompt message.

[0046] In an optional embodiment, the tumor information query software can be formed in the form of a question and answer, as shown in Figure 3 (a) to Figure 3 (d), the right side is the input tumor information query statement to be queried, and the left side is the result data returned by the server for the tumor information query statement. Optionally, the tumor information query statement may include the input personalized tumor type, tumor stage data, gene data and treatment data, and may also include specific query information. For example, as shown in Figure 3 (a), the tumor information query statement includes "colorectal tumor, colorectal adenocarcinoma, T3N1M0, HER2 positive. Radical colorectal surgery plus lymph node dissection, adjuvant chemotherapy has been completed. Gene mutations detected after surgery: JAK1 c2404-2_2405del (splicing site change) system mutation, CTNNA1 Q176* system mutation, TP3 R213* system mutation, ERBB2 exon16del systemic mutation. Detected Chinese medicine indicators: low HED, MSS, TMB-H. The doctor said that I don’t need to take any more medicine after completing chemotherapy, so what’s the point of doing this genetic test? ", which includes the inquiry information "What’s the point of doing this genetic test", and the sentence information "Colorectal tumor, colorectal adenocarcinoma, T3N1M0, HER2 positive. Radical colorectal surgery plus lymph node dissection, adjuvant chemotherapy has been completed. Gene mutations detected after surgery: JAK1 c2404-2_2405del (splicing site change) systemic mutation, CTNNA1 Q176* systemic mutation, TP3 R213* systemic mutation, ERBB2 exon16del systemic mutation. Detected Chinese medicine indicators: low HED, MSS, TMB-H. ", the input tumor type, tumor stage data, gene data and treated data can be obtained from the sentence information.

[0047] The tumor type data is used to characterize the type of the tumor. Optionally, the tumor type data may include benign tumors or malignant tumors. Optionally, the benign tumor may be a hemangioma, a lymphangioma, a fibroma, a lipoma, a neurogenic tumor, a teratoma, or a carotid body tumor; the malignant tumor may be gastric cancer, liver cancer, osteosarcoma, or breast cancer, etc.

[0048] Tumor stage data is used to characterize the stage information of the tumor, such as the early stage of gastric cancer, the middle stage of gastric cancer, or the late stage of gastric cancer.

[0049] The gene data is used to characterize the gene information related to the tumor. Optionally, the gene data includes but is not limited to population gene data, hereditary gene data, etc.

[0050] The treated data is used to characterize the treatments that have been performed on the tumor, such as surgical treatment, drug treatment, etc.

[0051] Step 204 , parse the tumor information query statement, and extract the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed included in the tumor information query statement.

[0052] Optionally, after the server obtains the tumor information query request, it obtains the tumor information query statement from the tumor information query request, and parses the tumor information query statement to extract the tumor type data, tumor stage data, gene data and treated data included in the tumor information query statement, and uses the extracted tumor type data, tumor stage data, gene data and treated data as tumor type data to be processed, tumor stage data to be processed, gene data to be processed and treated data to be processed.

[0053] In one embodiment, a text recognition model can be pre-set, and the acquired tumor information query request can be input into the text recognition model to identify the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed included in the tumor information query request. The specific recognition model is not limited, as long as text recognition can be achieved. For example, sample data including tumor type, tumor stage data, gene data, treated data, and query information can be pre-set, and the sample data can be trained using a neural network algorithm to obtain a text recognition model.

[0054] Step 206, input the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed into the tumor information recognition model, and receive the result data output by the tumor information recognition model corresponding to the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed.

[0055] Step 208: Send a display instruction to the terminal device, where the display instruction carries the result data.

[0056] Among them, the tumor information recognition model is obtained by training the pre-generated question-answering model based on sample data using model fine-tuning training technology.

[0057] Optionally, the sample data includes tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data and treatment sample data.

[0058] Optionally, the pre-generated question-answering model is chatGLM-6B. Optionally, the model fine-tuning training technology can be LoRA technology, wherein the LoRA technology allows the behavior of the model to be adjusted without significantly increasing the number of parameters by introducing a low-rank matrix into the weight matrix of the model.

[0059] Optionally, fine-tuning the pre-generated question and answer model based on LoRA technology can include: obtaining the pre-generated question and answer model; importing sample data into the pre-generated question and answer model for pre-training; based on the LoRA method, freezing the pre-trained model weight parameters and generating a new network layer; training the new network layer based on the sample data and updating the corresponding parameters; importing the new network layer into the pre-generated question and answer model.

[0060] In an exemplary embodiment, before inputting the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed into the tumor information recognition model, the tumor information recognition model can include: obtaining literature report data, identifying sample data included in the literature report data; obtaining a pre-generated question-answering model, using model fine-tuning training technology to train the sample data, and generating a tumor information recognition model.

[0061] Optionally, the literature report data may include, but are not limited to, tumor research literature, clinical guidelines, pathology reports, and case analysis reports.

[0062] Optionally, the literature report data can be supplemented regularly. In one embodiment, tumor knowledge can be obtained from the website regularly. Exemplarily, information is obtained by visiting relevant knowledge websites, and the drugs and research content are filled. Exemplarily, the latest research content is regularly obtained from the relevant meeting materials released after the end of tumor-related conferences. Optionally, the conferences include but are not limited to ASCO, AACR, CSCO, etc. Exemplarily, the content of the PubMed website can be supplemented, where the PubMed website is the world's largest medical and biological website. The website will provide an automatic download function for the abstract part of the relevant article.

[0063] In one embodiment, after obtaining a literature report, the literature report is parsed to obtain tumor type data, tumor stage data, gene data, treatment data, and association relationship data between any two of the tumor type data, tumor stage data, gene data, and treatment data from the literature report; the tumor type data, tumor stage data, gene data, treatment data, and association relationship data between any two of the tumor type data, tumor stage data, gene data, and treatment data obtained from the literature report are used as samples to form sample data, that is, the sample data includes tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data, and treatment sample data.

[0064] Among them, the association relationship sample data is used to characterize the association between any two of the tumor type sample data, tumor stage sample data, gene sample data, and treatment sample data. Optionally, through the association relationship, when any one of the tumor type sample data, tumor stage sample data, gene sample data, and treatment sample data is obtained, the corresponding data can be determined according to the association relationship, for example: for stage III lung cancer, the corresponding treatment data can be a combination of surgical treatment and immunotherapy.

[0065] Optionally, the correlation data between any two of the tumor type data, tumor stage data, gene data, and treatment data obtained from the literature report can be subjected to correlation analysis, and the closeness of the correlation data between any two of the tumor type data, tumor stage data, gene data, and treatment data can be determined through correlation analysis. The correlation data between any two of the tumor type data, tumor stage data, gene data, and treatment data can be annotated according to the closeness, so as to realize the training of the tumor information recognition model, so that the tumor information recognition model can give accurate result data based on the correlation data. For example, according to the correlation data, it is determined that the stage III lung cancer, the corresponding treatment data can be a combination of surgical treatment and immunotherapy, and the "treatment data can be a combination of surgical treatment and immunotherapy" is output as the result data. Among them, the closeness is used to characterize the correlation program between the tumor type data, tumor stage data, gene data, and treatment data. The closeness can clearly determine the correlation program between any two of the tumor type data, tumor stage data, gene data, and treatment data. When realizing the generation of the tumor information recognition model, the correlation between the tumor type data, tumor stage data, gene data, and treatment data is considered, and the result data can be output in a targeted manner to improve the accuracy of the result data output.

[0066] Optionally, identifying sample data included in literature report data includes: identifying sample data included in literature report data to obtain initial tumor type data, initial tumor stage data, initial gene data, initial treatment data, and initial association relationship between any two of the initial tumor type data, initial tumor stage data, initial gene data, and initial treatment data; clustering the initial tumor type data, initial tumor stage data, initial gene data, and initial treatment data to obtain tumor type sample data, tumor stage sample data, gene sample data, and treatment data sample data; generating association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data, and treatment data sample data according to the initial association relationship; generating sample data according to the tumor type sample data, tumor stage sample data, gene sample data, treatment data sample data, and association relationship sample data. Optionally, the clustering method can be hierarchical clustering, density clustering, or fuzzy C-means clustering, and the specific selected clustering method is not limited, as long as clustering can be achieved.

[0067] In one embodiment, the initial treatment data includes initial medication data, and generating sample data based on tumor type sample data, tumor stage sample data, gene sample data, treatment data sample data, and associated relationship sample data includes: obtaining data source information corresponding to the initial medication data, level data of the data source information, and effect data of the initial medication data; determining priority information of the initial medication data based on the data source information, the level data of the data source information, and the effect data of the initial medication data; generating sample data based on the tumor type sample data, tumor stage sample data, gene sample data, treatment data sample data, priority information of the initial medication data, and associated relationship sample data.

[0068] The initial medication data is used to characterize the medication data used to treat a certain tumor. The initial medication data includes, but is not limited to, medication information, tumors that can be treated, treatment effect information, etc.

[0069] Among them, the data source information is used to determine which literature report the initial medication data comes from.

[0070] Among them, the level data of the data source information is used to determine the level information of the literature report. Optionally, different levels can be set for different literature reports, and different levels can be set according to the type of literature report. For example, tumor research literature is the highest level, clinical guidelines are the second level, pathology reports are the third level, and case analysis reports are the lowest level. Different levels can also be set according to the way the literature report is obtained. For example, the literature report obtained from the knowledge website is the highest level, the literature report obtained from the PubMed website is the middle level, and the literature report obtained from the conference materials released by the conference is the lowest level.

[0071] Among them, the priority information of the initial medication data is used to characterize the order in which the drugs are recommended. Optionally, after obtaining the initial medication data, the server obtains the data source information corresponding to the initial medication data, and the data source information determines the level data corresponding to the data source data. The effect data corresponding to the initial medication data can also be obtained, and then the level information of the initial medication data is determined based on the obtained data source information, the level data of the quantity source information, and the effect data of the initial medication data. By comprehensively determining the priority information of the initial medication data through the data source information, the level data of the data source information, and the effect data of the initial medication data, the accuracy of determining the priority information of the initial medication data can be improved, thereby improving the accuracy of the output information of the tumor information recognition model.

[0072] Optionally, the priority information of the initial medication data can also be determined based on the efficacy of the drug, whether the drug is included in medical insurance, and / or the number of times the drug appears in the literature report. Exemplarily, when the number of times a drug appears in a literature report is the same, the priority information of the medication data is determined based on the efficacy of the drug and / or whether the drug is included in medical insurance. Alternatively, when it is determined based on the efficacy of the drugs that the efficacy of related drugs is not much different, and they are all included in medical insurance, the priority of the medication data can be determined based on the number of times the drug appears in the literature report, and the priority of the drug with more appearances is higher. The priority of the drug can also be determined based on the number of times the drug appears in different literature reports. When the drug appears in different literature reports, it is considered that the priority information of the medication data is higher.

[0073] Optionally, the priority information of the initial medication data can be comprehensively determined based on the data source information, the level data of the data source information, the effect data of the initial medication data, the efficacy of the drug, whether the drug is included in the medical insurance and / or the number of times the drug appears in the literature report, etc. In one embodiment, after the priority information of the initial medication data is determined, the sample data can be generated based on the priority information of the initial medication data, the tumor type sample data, the tumor stage sample data, and the gene sample data.

[0074] In one embodiment, if the treatment data sample data includes other data in addition to the initial medication data, such as surgical treatment methods, sample data can be generated based on the treatment data sample data, priority information of the initial medication data, tumor type sample data, tumor stage sample data, and gene sample data.

[0075] The display instruction is used to instruct the terminal to display the result data.

[0076] In one embodiment, the server parses the tumor information query statement to obtain the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed included in the tumor information query statement, and inputs them into a pre-generated tumor information recognition model. The tumor information recognition model is used to output result data corresponding to the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed.

[0077] Optionally, after acquiring the result data output by the tumor information recognition model, the server encapsulates the result data into a display instruction, and sends the display instruction to the terminal, so that the terminal displays the result data carried in the display instruction under the instruction of the display instruction.

[0078] Optionally, when the terminal displays the carried result data according to the received display instruction, the result data can be input into the answer of the dialog box to form a question and answer format, as shown in Figures 3(a) to 3(d).

[0079] In one embodiment, the result data may be a targeted response to the content included in the tumor information query statement. For example, the gene data obtained from the tumor information query statement is parsed, and the query information in the tumor information query statement can be answered.

[0080] For example, the input is:

[0081] "Pathological diagnosis: colorectal tumor, colorectal adenocarcinoma, T3N1M0, HER2 positive. Radical colorectal surgery plus lymph node dissection, adjuvant chemotherapy has been completed. Gene mutations detected after surgery: JAK1 c2404-2_2405del (splicing site change) system mutation, CTNNA1 Q176* system mutation, TP3 R213* system mutation, ERBB2 exon16del system mutation. Traditional Chinese medicine indicators detected: low HED, MSS, TMB-H. The doctor said that I don’t need to take any more medicine after completing chemotherapy, so what’s the point of doing this genetic test?"

[0082] The corresponding output content can be:

[0083] "Regarding the gene variation you mentioned, we refer to the gravity population frequency database, especially the Chinese population:

[0084] Colorectal tumors, CTNNA1 mutations account for 3.1%

[0085] In colorectal tumors, ERBB2 mutations accounted for 7.9%

[0086] Colorectal tumors, JAK1 mutations account for 1.5%

[0087] In colorectal tumors, TP53 point mutations / indels were 55.9%

[0088] Colorectal cancer, TP53 truncation is 19.6%

[0089] According to your query information, the results are as follows:

[0090] Treatment for colorectal cancer usually includes surgery, chemotherapy, radiotherapy, and therapy directed at specific molecular targets. In your case, you have completed radical colorectal surgery and adjuvant chemotherapy, which is the standard treatment procedure.

[0091] Genetic testing plays an important role in the treatment of colorectal cancer, especially in the following aspects:

[0092] Prognosis assessment: Certain gene variations, such as TP53 mutations, may be associated with cancer prognosis. Knowing these variations can help doctors better assess the risk of disease progression.

[0093] Treatment decisions: Genetic testing can reveal certain targetable molecular alterations, such as mutations in ERBB2 (HER2), which may qualify patients for HER2-targeted therapy.

[0094] Drug sensitivity: Genetic variations can affect a patient’s response to certain drugs. For example, patients with MSS (microsatellite stable) status may not benefit from certain immunotherapies, while patients with TMB-H (high tumor mutation burden) may respond better to immunotherapy.

[0095] Risk assessment: Certain hereditary cancer syndromes, such as Lynch syndrome, are associated with an increased risk of colorectal cancer. Genetic testing can help determine which genetic disturbance is present.

[0096] In your case, your doctor may decide that the current treatment is sufficient and that no further drug therapy is needed immediately after completing chemotherapy. However, the results of genetic testing provide important information about the characteristics of the disease, which may be useful for future monitoring and treatment decisions. For example, if the disease recurs or progresses in the future, this gene variation information may help select a more personalized treatment plan.

[0097] In summary, the results of genetic testing help understand the biology of the disease, provide a basis for possible future treatment options, and may influence monitoring strategies and risk assessment for family members. It is important to communicate with your doctor and understand the specific implications of these genetic variants for your personal treatment plan. "

[0098] In the above Q&A, personalized genetic variation information is provided for the patient's specific situation (colorectal adenocarcinoma, T3N1M0), HER2 positivity, and completion of surgery and adjuvant chemotherapy, and the frequency of these variations in colorectal cancer is explained, which helps patients understand the position of their condition in the population. Molecular interpretation: The potential significance of the genetic variation detected in the patient (JAK1, CTNNA1, TP53, ERBB2) and other important indicators (HED, MSS, TMB-H) is explained in detail, which helps patients and doctors understand the molecular characteristics of the disease and provide a basis for future treatment decisions. Future treatment recommendations: Although the patient has completed standard treatment, the answer still proposes the potential impact of genetic test results on future treatment options, such as ERBB2 (HER2) mutations may make patients suitable for HER2 targeted therapy, and TMB-H status may have a better response to immunotherapy. Prognosis assessment: The answer mentions that certain genetic variations (such as TP53 mutations) may be related to disease prognosis, which helps patients and doctors assess the risk of future disease progression.

[0099] In one embodiment, the result data may also be a selective result. As shown in FIG3 (b), it provides a comprehensive analysis covering the impact of gene variation on treatment and prognosis, genetic risk, and potential impact on family members. The following is an evaluation of the answer: Personalized treatment recommendations: The answer provides personalized treatment recommendations for the patient's genetic variation, especially the BRCA2 germline mutation, including the sensitivity of PARP inhibitors, which is very important for the patient's treatment selection. Genetic risk interpretation: The answer clearly explains the inheritance pattern of germline mutations and the risk of the patient's children inheriting the mutation, which helps the patient understand the risks that family members may face. Family member management: Provides recommendations for genetic counseling and genetic testing for family members (especially the patient's children), which is an important step in managing genetic risks and early prevention. Prognostic assessment: It points out that BRCA2 mutations may be associated with a poor prognosis, which helps patients and doctors make more cautious decisions when considering future monitoring and treatment strategies. Postoperative adjuvant therapy guidance: According to current clinical guidelines, recommendations for postoperative adjuvant therapy are proposed, including whole breast radiation therapy, chemotherapy adjuvant, and PARP inhibitor intensive adjuvant, which helps patients understand the standard process of postoperative treatment. Applicability of information: The answer emphasizes that the information provided is for reference only, and encourages patients to consult their attending physicians for specific medical behaviors, which reflects respect for patient safety and medical practice. In the case in the figure above, the model focuses on the BRCA2 gene germline mutations with a high level of evidence in the molecular interpretation part, which have important guiding significance for treatment prognosis and genetic counseling. However, the description of other genes is relatively brief. The model omits some mutations with a lower level of evidence but with certain guiding significance for treatment, such as: NF1 gene mutations are associated with sensitivity to the drug selumetinib (level C evidence, clinical guidelines recommend the use of selumetinib to treat patients with recurrent or progressive pilocytic astrocytoma with NF1 mutations); and tips related to immunotherapy, such as the association of MDM4 gene amplification with immune hyperprogression [PMID: 28351930], TERT gene changes are associated with enhanced tumor immunogenicity and anti-tumor immune inflammation, which may lead to prolonged overall survival (OS) of cancer patients treated with immune checkpoint inhibitors [PMID: 32564494].

[0100] In one of the embodiments, the result data can also be used to interpret comprehensive genetic variants and provide reasonable treatment recommendations, as shown in Figure 3 (c). Personalized treatment recommendations: The answer provides personalized treatment recommendations based on the patient's genetic test results, considering targeted treatment options for MET exon14del mutations and NTRK3 rearrangements, which helps patients understand specific treatment options for their condition. Molecular interpretation: The answer provides a detailed molecular interpretation of the patient's genetic variants, explaining the potential therapeutic significance of each variant and the corresponding targeted drugs, such as MET kinase inhibitors and TRK inhibitors. The importance of genetic counseling: The answer emphasizes the risk that TP53 germline mutations may be associated with Li-Fraumeni syndrome and recommends that the patient undergo genetic counseling, which helps the patient understand the genetic risks of himself and his family members. Explanation of treatment options: The answer explains that based on the patient's TMB-L and low HED indicators, immunotherapy may not be the first choice, which helps the patient understand why chemotherapy may be selected as adjuvant therapy. Comprehensive treatment considerations: The answer provides comprehensive treatment considerations, including chemotherapy and targeted therapy, and recommends that the patient discuss with the attending physician to develop a personalized treatment plan. In this case, the model did not show some treatment suggestions with relatively low evidence levels, such as: patients with CDKN2A mutations may be sensitive to palbociclib (Sensitive-C, TAPUR study) [NCT02693535, PMID: 20522552] and abemaciclib (Sensitive-D), are also predictors of immunotherapy resistance and poor prognosis in non-small cell lung cancer, and Ba / F3 cells carrying MET exon 14 skipping and MET D1228Y mutations are resistant to crizotinib, sirotinib and cabozantinib [PMID: 31279006].

[0101] In one of the embodiments, the query information may not exist in the tumor information query statement, and the result data may be a comprehensive data result. As shown in Figure 3 (d), when answering questions about the selection of neoadjuvant and adjuvant treatment options for stage III lung cancer, the answers mainly focus on immunotherapy and EGFR-TKI treatment, which have now entered the perioperative stage. Special emphasis is placed on biomarkers related to immunotherapy, such as MSS and HED, and EGFR targets that are sensitive to EGFR-TKI. MET amplification and BRCA2 gene mutations are not specifically mentioned in the answers. According to the current standard treatment guidelines, MET inhibitors are mainly used for systemic treatment of patients with advanced lung cancer, while PARP inhibitors have not yet been included in the lung cancer treatment guidelines due to less evidence in lung cancer. Similarly, in the adjuvant treatment stage, although there are a variety of EGFR-TKI drugs to choose from, such as icotinib, gefitinib, or erlotinib, given that osimertinib has the highest level of evidence and has been included in medical insurance, we chose to recommend osimertinib and did not discuss other drugs in detail. This choice reflects our preference for the treatment options with the most robust evidence and the clearest clinical benefit when providing treatment recommendations, while also taking into account drug availability and economic factors. We encourage physicians to consider the patient's specific circumstances and the latest clinical research data when considering treatment options to develop the most appropriate personalized treatment plan.

[0102] In the above-mentioned tumor information question and answer processing method, a tumor information query statement is obtained from the tumor information answer request, and the tumor information query statement can be parsed to obtain the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed and the treated data to be processed included in the tumor information query statement, and the obtained tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed and the treated data to be processed are input into a pre-generated tumor information recognition model, and the result data output by the tumor recognition model is received. When performing recognition, the above-mentioned embodiment uses individualized tumor type data to be processed, tumor stage data to be processed, gene data to be processed and treated data to be processed to achieve recognition, which can improve the accuracy of tumor information recognition, and can achieve targeted output result data based on the individualized tumor type data to be processed, tumor stage data to be processed, gene data to be processed and treated data to be processed, thereby effectively improving adaptability.

[0103] In an exemplary embodiment, Figure 4 As shown, a tumor information recognition model training method is provided, which is applied to the server in FIG1 , and includes steps 302 to 306, wherein:

[0104] Step 302, obtaining literature report data.

[0105] Optionally, the literature report data may include, but are not limited to, tumor research literature, clinical guidelines, case reports, and case analysis reports.

[0106] Step 304, identifying sample data included in the literature report data, the sample data including tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data, and treatment sample data.

[0107] Step 306, obtain a pre-generated question-answering model, use model fine-tuning training technology, train sample data, and generate a tumor information recognition model.

[0108] In the above embodiment, when realizing the generation of the tumor information recognition model, the tumor type, tumor stage, genetic data, treatment data, and the correlation between any two of the tumor type, tumor stage, genetic data, and treatment data are taken into consideration, and the tumor-related data information is comprehensively considered, so that the generated tumor information recognition model is accurate, and the result data obtained by using the tumor information recognition model for information recognition is accurate.

[0109] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0110] Based on the same inventive concept, the embodiment of the present application also provides a tumor information question and answer processing device for implementing the tumor information question and answer processing method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more tumor information question and answer processing device embodiments provided below can refer to the limitations of the tumor information question and answer processing method above, and will not be repeated here.

[0111] In an exemplary embodiment, Figure 5 As shown, a tumor information question-answering processing device 400 is provided, comprising: a request acquisition module 402, a parsing module 404, a result acquisition module 406 and an instruction sending module 408, wherein:

[0112] The request acquisition module 402 is used to acquire a tumor information question and answer request sent by the terminal, wherein the tumor information question and answer request data carries a tumor information query statement;

[0113] The parsing module 404 is used to parse the tumor information query statement and extract the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed and the treated data to be processed included in the tumor information query statement;

[0114] The result acquisition module 406 is used to input the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed into the tumor information recognition model, and receive the result data output by the tumor information recognition model corresponding to the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed; wherein the tumor information recognition model is obtained by training the pre-generated question-answering model using the model fine-tuning training technology according to the sample data; the sample data includes tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data, and treatment sample data;

[0115] The instruction generation module 408 is used to send a display instruction to the terminal. The display instruction carries the result data, and the display instruction is used to instruct the terminal to display the result data.

[0116] In one embodiment, the tumor information question and answer processing module also includes a model generation module, which is used to obtain literature report data; identify sample data included in the literature report data, the sample data including tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data and treatment sample data; obtain a pre-generated question and answer model, use model fine-tuning training technology to train the sample data, and generate a tumor information recognition model.

[0117] In one embodiment, the module generation module is also used to identify sample data included in the literature report data, obtain initial tumor type data, initial tumor stage data, initial gene data, initial treatment data, and an initial association relationship between any two of the initial tumor type data, initial tumor stage data, initial gene data and initial treatment data; cluster the initial tumor type data, initial tumor stage data, initial gene data and initial treatment data to obtain tumor type sample data, tumor stage sample data, gene sample data and treatment data sample data; generate association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data and treatment data sample data according to the initial association relationship; generate sample data according to the tumor type sample data, tumor stage sample data, gene sample data, treatment data sample data and association relationship sample data.

[0118] In one of the embodiments, the model generation module is also used to obtain data source information, level data of the data source information, and effect data of the initial medication data corresponding to the initial medication data when the initial treatment data includes initial medication data; determine priority information of the initial medication data based on the data source information, level data of the data source information, and effect data of the initial medication data; and generate sample data based on tumor type sample data, tumor stage sample data, gene sample data, treatment data sample data, priority information of the initial medication data, and association relationship sample data.

[0119] Based on the same inventive concept, the embodiment of the present application also provides a tumor information recognition model training device for implementing the above-mentioned tumor information recognition model training method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in one or more tumor information recognition model training device embodiments provided below can refer to the limitations of the tumor information model training method above, and will not be repeated here.

[0120] In an exemplary embodiment, Figure 6 As shown, a tumor information recognition model training device 500 is provided, comprising: a literature acquisition module 502, a sample data acquisition module 504 and a training module 506, wherein:

[0121] The document acquisition module 502 is used to acquire document report data.

[0122] The sample data acquisition module 504 is used to identify the sample data included in the literature report data, the sample data including tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, tumor stage sample data, gene sample data and treatment sample data.

[0123] The training module 506 is used to obtain a pre-generated question-answering model, use model fine-tuning training technology, train sample data, and generate a tumor information recognition model.

[0124] Each module in the above-mentioned tumor information question-answering processing device and tumor information identification model training device can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0125] In an exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be shown in FIG7. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a tumor information question and answer processing method and a tumor information recognition model training method are implemented.

[0126] Those skilled in the art will understand that the structure shown in FIG. 7 is merely a block diagram of a partial structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different arrangement of components.

[0127] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0128] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0129] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0130] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0131] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0132] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0133] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A tumor information question-answering processing method, characterized in that: The method comprises: Acquire a tumor information question and answer request sent by a terminal, wherein the tumor information question and answer request carries a tumor information query statement; Parsing the tumor information query statement, extracting the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed included in the tumor information query statement; Inputting the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed into a tumor information recognition model, and receiving result data output by the tumor information recognition model corresponding to the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed; wherein the tumor information recognition model is obtained by training a pre-generated question-answering model using a model fine-tuning training technology according to sample data; the sample data includes tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, the tumor stage sample data, the gene sample data, and the treatment sample data; A display instruction is sent to the terminal, wherein the display instruction carries the result data, and the display instruction is used to instruct the terminal to display the result data.

2. The method according to claim 1, characterized in that Before inputting the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed into the tumor information recognition model, the method further includes: Obtain literature report data; Identifying sample data included in the literature report data, the sample data including tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, the tumor stage sample data, the gene sample data, and the treatment sample data; A pre-generated question-answering model is obtained, and the sample data is trained using model fine-tuning training technology to generate the tumor information recognition model.

3. The method according to claim 2, characterized in that The identifying of the sample data included in the literature report data comprises: Identify the sample data included in the literature report data, and obtain initial tumor type data, initial tumor stage data, initial gene data, initial treatment data, and an initial association relationship between any two of the initial tumor type data, the initial tumor stage data, the initial gene data, and the initial treatment data; Clustering the initial tumor type data, the initial tumor stage data, the initial gene data, and the initial treatment data to obtain the tumor type sample data, the tumor stage sample data, the gene sample data, and the treatment data sample data; Generate association relationship sample data between any two of the tumor type sample data, the tumor stage sample data, the gene sample data and the treatment data sample data according to the initial association relationship; The sample data is generated according to the tumor type sample data, the tumor stage sample data, the gene sample data, the treatment data sample data and the association relationship sample data.

4. The method according to claim 3, characterized in that The initial treatment data includes initial medication data, and the generating of the sample data according to the tumor type sample data, the tumor stage sample data, the gene sample data, the treatment data sample data, and the association relationship sample data includes: Obtaining data source information corresponding to the initial medication data, level data of the data source information, and effect data of the initial medication data; Determining priority information of the initial medication data according to the data source information, the level data of the data source information, and the effect data of the initial medication data; The sample data is generated according to the tumor type sample data, the tumor stage sample data, the gene sample data, the treatment data sample data, the priority information of the initial medication data, and the association relationship sample data.

5. A tumor information recognition model training method, characterized in that: The method comprises: Obtain literature report data; Identifying sample data included in the literature report data, the sample data including tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, the tumor stage sample data, the gene sample data, and the treatment sample data; A pre-generated question-answering model is obtained, and the sample data is trained using model fine-tuning training technology to generate the tumor information recognition model.

6. A tumor information question-answering processing device, characterized in that: The device comprises: A request acquisition module, used to acquire a tumor information question and answer request sent by a terminal, wherein the tumor information question and answer request data carries a tumor information query statement; A parsing module, used for parsing the tumor information query statement, and extracting the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed included in the tumor information query statement; A result acquisition module is used to input the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed into a tumor information recognition model, and receive result data output by the tumor information recognition model corresponding to the tumor type data to be processed, the tumor stage data to be processed, the gene data to be processed, and the treated data to be processed; wherein the tumor information recognition model is obtained by training a pre-generated question-answering model using a model fine-tuning training technology according to sample data; the sample data includes tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, the tumor stage sample data, the gene sample data, and the treatment sample data; The instruction sending module is used to send a display instruction to the terminal, wherein the display instruction carries the result data, and the display instruction is used to instruct the terminal to display the result data.

7. A tumor information recognition model training device, characterized in that: The device comprises: Literature acquisition module, used to obtain literature report data; a sample data acquisition module, configured to identify sample data included in the literature report data, wherein the sample data includes tumor type sample data, tumor stage sample data, gene sample data, treatment sample data, and association relationship sample data between any two of the tumor type sample data, the tumor stage sample data, the gene sample data, and the treatment sample data; The training module is used to obtain the pre-generated question-answering model, use the model fine-tuning training technology to train the sample data, and generate the tumor information recognition model.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 or 5 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 or 5 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 or 5 are implemented.