Prediction method for progress of early lung adenocarcinoma from MIA to IA and risk scoring model construction method thereof

Calculate risk scores and build a risk score model by calculating the risk score in the current technology, which solves the problem of difficult prediction of the early progress of lung adenocarcinoma in the existing technology, and achieves accurate prediction of the progression of early lung adenocarcinoma from MIA to IA, and improves lung adenocarcinoma The scientificity and accuracy of early detection and treatment.

CN118942540AInactive Publication Date: 2024-11-12WEST CHINA HOSPITAL SICHUAN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411436633.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-15
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The prior art is difficult to accurately predict the early risk of lung adenocarcinoma from microinvasive adenocarcinoma (MIA) to invasive adenocarcinoma (IA), resulting in difficulty in diagnosis and poor prognosis.

Method used

By obtaining mutations in specific gene sets in multiple samples, the risk score is calculated and the risk score model is constructed, and the prognosis model is constructed based on differentially mutated genes using chi-square test and machine learning algorithms (such as Logistic regression, SVM, DecisionTree, RandomForest, XGBoost) is used to construct prognostic models based on differentially mutated genes using chi-square tests and machine learning algorithms (such as Logistic regression, SVM, DecisionTree, RandomForest, XGBoost) is used to construct prognostic models based on differentially mutated genes. , predicts the development risk of early stage lung adenocarcinoma.

Benefits of technology

Accurate prediction of the progression of early lung adenocarcinoma from MIA to IA is achieved, which improves the possibility of early detection and timely treatment of lung adenocarcinoma, and improves the scientificity and accuracy of the prognostic risk score model.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118942540A_ABST
    Figure CN118942540A_ABST
Patent Text Reader

Abstract

The invention provides a progression prediction method for early lung adenocarcinoma from MIA to IA and a risk scoring model construction method thereof, and belongs to the technical field of disease progression prediction. According to the method, large panel gene detection of 1021 genes is carried out on 755 early-stage lung adenocarcinoma clinical samples, the lung adenocarcinoma prognosis risk scoring model and the corresponding lung adenocarcinoma development prediction method are constructed on the basis of the obtained data, the method has excellent applicability, various data set verification results show that the method can effectively predict the early-stage lung adenocarcinoma progress, and the method has good application prospects. Early discovery, accurate prediction and timely treatment of lung adenocarcinoma are realized clinically. According to the construction method for the prognosis risk scoring model of the lung adenocarcinoma, a specific differential mutant gene and a core algorithm model are utilized, and the differential mutant gene is closely related to an image, a pathological differentiation degree and a tumor size, so that the construction method is more scientific and accurate; and the clinical application value of the early-stage lung adenocarcinoma prognosis risk scoring model is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of disease progression prediction, and in particular to a method for predicting the progression of early lung adenocarcinoma from MIA to IA and a method for constructing a risk scoring model thereof. Background Art

[0002] Lung cancer is the world's number one cancer killer. Lung adenocarcinoma (LUAD) is its most common pathological type, and its evolution goes through multiple stages, including precursor lesions, minimally invasive adenocarcinoma (MIA) and invasive adenocarcinoma (IA). Once MIA develops into IA, the malignancy is significantly enhanced. In addition, early diagnosis of lung adenocarcinoma is difficult, which directly leads to poor prognosis for patients, with a 5-year survival rate of only 15%. However, the 5-year survival rate of patients who are diagnosed early can reach more than 92%. Therefore, early diagnosis is a key means to improve the prognosis of lung cancer patients.

[0003] For the early diagnosis of lung cancer, clinicians mainly rely on CT imaging evaluation. However, since the occurrence and development of lung cancer involves multiple genes and multiple links, the sensitivity, specificity and accuracy of the diagnosis method based solely on CT imaging evaluation are poor.

[0004] Chinese patent CN116313062A discloses a lung adenocarcinoma prognosis model, which belongs to the field of tumor molecular biology technology. The lung adenocarcinoma prognosis model is constructed based on a combination of polyamine metabolism-related genes SMS, SMOX, GPC1, SLC47A1, AZIN2 and MAOB; the construction method is: collecting training sets and validation sets for constructing a lung adenocarcinoma prognosis model of polyamine metabolism-related genes from the Cancer Genome Atlas Database and the Gene Expression Comprehensive Database; sorting polyamine metabolism-related genes from the KEGG, GO databases and literature; and constructing a prognosis model of polyamine metabolism-related genes through LASSO-COX regression.

[0005] However, this approach failed to correlate with MIA / IA transition and accurately predict the risk of early lung adenocarcinoma development. Summary of the invention

[0006] The purpose of the present invention is to provide a method for predicting the progression of early lung adenocarcinoma from MIA to IA and a method for constructing a risk scoring model thereof. Compared with the prior art, the data source is more extensive and more statistically significant, and it can accurately predict the progression of early lung adenocarcinoma from MIA to IA, thereby realizing early detection, accurate prediction and timely treatment of lung adenocarcinoma in clinical practice.

[0007] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions: The present invention provides a method for predicting the development of lung adenocarcinoma, comprising the following steps: S1. Obtain mutation status of target gene sets in multiple samples; S2. Calculate the results obtained in S1 using the following formula to obtain risk scores for multiple samples; Risk score = 1.83331403*G TP53 +1.24152921*G CDKN2A +2.02331501*G SETD2 -1.54969391*G ERBB2 -1.75892268*G BRAF -1.66240923*G TSC1 -1.59870732*G MAP2K1 -0.46874594*G CDKN1B -0.79330108*G JUN -1.24918034*G AKT1 -0.75809038*G HIST1H1D In the formula, the G TP53 is the mutation status of the TP53 gene in the sample, with mutation recorded as 1 and no mutation recorded as 0; The G CDKN2A is the mutation status of the CDKN2A gene, with mutation recorded as 1 and no mutation recorded as 0; The G SETD2 is the mutation status of the SETD2 gene, with mutation recorded as 1 and no mutation recorded as 0; The G ERBB2 is the mutation status of the ERBB2 gene, with mutation recorded as 1 and no mutation recorded as 0; The G BRAF is the mutation status of the BRAF gene, with mutation recorded as 1 and no mutation recorded as 0; The G TSC1 is the mutation status of the TSC1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G MAP2K1 is the mutation status of MAP2K1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G CDKN1B is the mutation status of the CDKN1B gene, with mutation recorded as 1 and no mutation recorded as 0; The G JUN is the mutation status of the JUN gene, with mutation recorded as 1 and no mutation recorded as 0; The G AKT1is the mutation status of the AKT1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G HIST1H1D is the mutation status of the HIST1H1D gene, with mutation recorded as 1 and no mutation recorded as 0; S3. Calculate a median based on the risk scores of the multiple samples obtained, and predict that samples with a risk score greater than the median are at high risk of developing early lung adenocarcinoma, and samples with a risk score less than the median are at low risk of developing early lung adenocarcinoma.

[0008] Preferably, the high risk of lung adenocarcinoma development refers to a higher risk of early lung adenocarcinoma developing from MIA to IA.

[0009] The present invention also provides a method for constructing a lung adenocarcinoma prognostic risk scoring model, comprising the following steps: Performing gene sequencing on the samples to obtain genome mutation information, wherein the sources of the samples include microinvasive adenocarcinoma tissue, invasive adenocarcinoma tissue and peripheral blood; Samples from peripheral blood were used as the control group to exclude germline mutation results in tissue samples and obtain somatic mutation information of the tissue sample genome; Based on the tumor pathological characteristics of the samples and the somatic mutation information of the genome of the samples, a chi-square test was performed to obtain differentially mutated genes between the groups; A risk scoring model was constructed based on differentially mutated genes between groups.

[0010] Preferably, the samples are derived from 755 patients with clinically early-stage lung adenocarcinoma.

[0011] Preferably, the chi-square test comprises the following steps: Based on the different sources of samples, the sample genes were analyzed for differential mutations between groups, and the differential mutation genes between microinvasive adenocarcinoma and invasive adenocarcinoma were obtained according to the chi-square test results (p<0.05).

[0012] Preferably, the differentially mutated genes between the microinvasive adenocarcinoma and invasive adenocarcinoma groups include AKT1, BCL2L1, BRAF, CDKN1B, CDKN2A, CDKN2C, ERBB2, HIST1H1D, JUN, MAP2K1, MED12, NF1, PRKAR1A, SETD2, SLC34A2, TERC, TP53, TSC1 and ZFP36L1.

[0013] Preferably, the methods that can be used to construct the risk scoring model include Logistic, SVM, DecisionTree, RandomForest and XGBoost.

[0014] Preferably, the constructing of the risk scoring model further comprises: A validation data set is obtained, and the risk scoring model is optimized based on the validation data set.

[0015] Preferably, optimizing the risk scoring model based on the validation data set includes: The inter-group differentially mutated genes are screened by Logistic regression analysis combined with a ten-fold cross validation method to determine the genes and the initial formula of the risk scoring model that are ultimately used to construct the risk scoring model; Substituting the gene finally used to construct the risk scoring model into the initial formula of the risk scoring model to obtain a risk scoring model; The initial formula of the risk scoring model is: Risk score = G 1 M 1 +…+G n M n , Among them, G n Indicates the mutation of the nth gene used to construct the risk scoring model, with mutation being recorded as 1 and non-mutation being recorded as 0; M n represents the coefficient of the nth gene used to construct the risk scoring model, and n represents the number of genes used to construct the risk scoring model; The median is calculated based on the risk scores of the obtained samples. Samples with risk scores greater than the median are predicted to have a high risk of developing lung adenocarcinoma, and samples with risk scores less than the median are predicted to have a low risk of developing lung adenocarcinoma.

[0016] Preferably, the genes ultimately used to construct the risk scoring model include TP53, SETD2, CDKN2A, BRAF, TSC1, MAP2K1, ERBB2, CDKN1B, JUN, AKT1 and HIST1H1D.

[0017] Preferably, the risk scoring model adopts the following formula: Risk score = 1.83331403*G TP53 +1.24152921*G CDKN2A +2.02331501*G SETD2 -1.54969391*G ERBB2 -1.75892268*G BRAF -1.66240923*G TSC1 -1.59870732*G MAP2K1 -0.46874594*G CDKN1B -0.79330108*G JUN -1.24918034*GAKT1 -0.75809038*G HIST1H1D In the formula, the G TP53 is the mutation status of the TP53 gene in the sample, with mutation recorded as 1 and no mutation recorded as 0; The G CDKN2A is the mutation status of the CDKN2A gene, with mutation recorded as 1 and no mutation recorded as 0; The G SETD2 is the mutation status of the SETD2 gene, with mutation recorded as 1 and no mutation recorded as 0; The G ERBB2 is the mutation status of the ERBB2 gene, with mutation recorded as 1 and no mutation recorded as 0; The G BRAF is the mutation status of the BRAF gene, with mutation recorded as 1 and no mutation recorded as 0; The G TSC1 is the mutation status of the TSC1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G MAP2K1 is the mutation status of MAP2K1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G CDKN1B is the mutation status of the CDKN1B gene, with mutation recorded as 1 and no mutation recorded as 0; The G JUN is the mutation status of the JUN gene, with mutation recorded as 1 and no mutation recorded as 0; The G AKT1 is the mutation status of the AKT1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G HIST1H1D is the mutation status of the HIST1H1D gene, with mutation being recorded as 1 and no mutation being recorded as 0.

[0018] Beneficial effects of the present invention: The present invention aims at clinical treatments such as lung cancer mutation targeted drugs, chemotherapy and immunity, and designs a large panel gene test of 1021 genes using 755 clinical samples of early lung adenocarcinoma, and constructs a lung adenocarcinoma prognosis risk score model and a corresponding lung adenocarcinoma development prediction method based on the obtained data. The early lung adenocarcinoma development prediction method provided by the present invention has excellent applicability, and the results of multiple data set verification show that the prediction method of the present invention can accurately predict the progression of early lung adenocarcinoma from MIA to IA, and realize the early detection and timely treatment of lung adenocarcinoma in clinical practice.

[0019] The method for constructing the early lung adenocarcinoma prognostic risk scoring model in the present invention utilizes specific differential mutation genes and a core algorithm model. The risk model is closely related to tumor mutation burden (TMB), tumor heterogeneity (MATH), imaging, pathological differentiation degree and tumor size. The differential mutation genes are also closely related to imaging, pathological differentiation degree and tumor size, making the construction method more scientific and accurate, and improving the clinical application value of the obtained early lung adenocarcinoma prognostic risk scoring model. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 The result diagram of differentially mutated genes between 19 MIA and MIA groups provided by the present invention; Figure 2 This is the AUC value result graph of the model results; Figure 3 This is the result diagram of the prognostic gene feature set; Figure 4 A gene classification result diagram for constructing the risk scoring model; Figure 5 This is a graph showing the difference in progression-free survival between the high-risk group and the low-risk group in this risk scoring model in the first validation set; Figure 6 This is a graph showing the difference in overall survival due to tumor death in the first validation set between the high-risk group and the low-risk group in this risk scoring model; Figure 7 This is a graph showing the difference in overall survival between the high-risk group and the low-risk group in this risk scoring model in the first validation set; Figure 8 This is a graph showing the difference in overall survival between the high-risk group and the low-risk group in this risk scoring model in the second validation set; Fig. 9 A result diagram showing the correlation between high-risk and low-risk groups in TMB provided by the present invention; Fig.10 A correlation result diagram of the high-risk and low-risk groups in MATH provided by the present invention; Fig.11 This is a result diagram of the correlation between the gene mutation frequencies of the high-risk and low-risk groups in the imaging results of the present invention; Fig.12 This is a result diagram of the correlation between the gene mutation frequencies of the high-risk and low-risk groups in the pathological differentiation degree results of the present invention; Fig.13 This is a result diagram of the correlation between the gene mutation frequency of the high-risk and low-risk groups in the tumor size results of the present invention; Fig.14 This is a result diagram of the correlation between TMB values ​​of high-risk and low-risk groups in imaging results of the present invention; Fig.15This is a result diagram of the correlation between the TMB values ​​of the high-risk and low-risk groups in the pathological differentiation degree results of the present invention; Fig.16 This is a result diagram of the correlation between the TMB values ​​of the high-risk and low-risk groups in the tumor size results of the present invention; Fig.17 This is a result diagram of the correlation between the MATH values ​​of the high-risk and low-risk groups in the imaging results of the present invention; Fig.18 This is a result diagram of the correlation between the MATH values ​​of the high-risk and low-risk groups in the pathological differentiation degree results of the present invention; Fig.19 This is a result diagram of the correlation between the MATH values ​​of the high-risk and low-risk groups in the tumor size results of the present invention; Fig. 20 This is a result diagram of the correlation between the VAF values ​​of the high-risk and low-risk groups in the imaging results of the present invention; Fig.21 This is a result diagram of the correlation between the VAF values ​​of the high-risk and low-risk groups in the results of the degree of pathological differentiation of the present invention; Fig. 22 The difference results of TP53 in imaging, pathological differentiation degree and tumor size are shown in the figure; Fig.23 The difference results of CDKN2A in imaging, pathological differentiation degree and tumor size are shown in the figure; Fig.24 The difference results of ERBB2 in images and tumor size are shown in the figure; Fig.25 The following are the difference results of BRAF in images and tumor size. DETAILED DESCRIPTION

[0021] The technical solutions provided by the present invention are described in detail below in conjunction with the embodiments, but they should not be construed as limiting the protection scope of the present invention.

[0022] Example Screening, sequencing and analysis of early lung adenocarcinoma samples: According to the tumor pathological characteristics of early lung adenocarcinoma, the obtained lung tissues were divided into microinvasive adenocarcinoma and invasive adenocarcinoma, and the corresponding peripheral blood of the patients was obtained as the control group; genomic DNA information was extracted from the relevant tissues and peripheral blood to generate a sequencing library, and the microinvasive adenocarcinoma, invasive adenocarcinoma and peripheral blood were subjected to DNA targeted capture sequencing using a 1021 gene panel; the 1021 gene mutation analysis process was used to obtain genomic mutation information from the 1021 gene sequencing results, and the control blood mutation information was used to exclude the germline mutation results in the tissue samples, and the final genomic somatic mutation information was obtained through database and manual composite filtering.

[0023] Get differentially mutated genes: Based on the tumor pathological characteristics (MIA and IA) and the final detected somatic mutation information, the chi-square test was used to perform differential mutation analysis between the MIA and IA groups for each gene. According to the chi-square test results p < 0.05, 19 differentially mutated genes between the MIA and IA groups were obtained (AKT1, BCL2L1, BRAF, CDKN1B, CDKN2A, CDKN2C, ERBB2, HIST1H1D, JUN, MAP2K1, MED12, NF1, PRKAR1A, SETD2, SLC34A2, TERC, TP53, TSC1, ZFP36L1) ( Figure 1 ).

[0024] Construction of prognostic model for early lung cancer: Construction and selection of risk scoring methods: Based on the 19 differentially mutated genes between MIA and IA, models were constructed using Logistic, SVM, DecisionTree, RandomForest and XGBoost. The AUC values ​​of all model results were greater than 0.7 ( Figure 2 ), the results show that the differential mutation genes provided by the present invention have excellent effectiveness and universality, and have broad application prospects. In subsequent experiments, Logistic was selected to construct the final risk assessment model.

[0025] Obtaining a training data set to construct a risk scoring model: The above differentially mutated genes are screened by Logistic regression analysis combined with a ten-fold cross validation method to determine the genes used to construct the risk scoring model and the risk scoring model, wherein the risk scoring model is expressed as: RiskScore=G 1 M 1 +…+G n M n , Among them, RiskScore represents the risk score, G n represents the nth gene mutation used to construct the risk scoring model (1: mutation, 0: no mutation), M n represents the coefficient of the nth gene used to construct the risk scoring model, and n represents the number of genes used to construct the risk scoring model.

[0026] Logistic analysis was performed to screen out 11 genes that constitute the prognostic gene signature set (gene signature) ( Figure 3The genes used to construct the risk score model include 3 mutated genes that promote tumor progression (TP53, SETD2, CDKN2A) and 8 genes that inhibit tumor progression (BRAF, TSC1, MAP2K1, ERBB2, CDKN1B, JUN, AKT1, HIST1H1D) ( Figure 4 ); The risk scoring model is expressed as follows: the risk score of each sample is calculated according to the formula, and all samples are divided into a high-risk group and a low-risk group according to the median of the risk score (samples greater than the median are divided into the high-risk group, and samples less than the median are divided into the low-risk group).

[0027] According to the classification of high-risk and low-risk groups predicted by the present invention, patients can receive different treatments, such as conservative treatment for the low-risk group and active and diverse treatments for the high-risk group. This can effectively avoid overtreatment and untimely treatment, which can not only reduce the economic burden on patients and their families and improve the quality of life of patients, but also effectively prolong the life of patients by participating in the decision-making of treatment plans, and ultimately achieve clinical cure.

[0028] The risk scoring model is expressed as: Risk score = 1.83331403*TP53+1.24152921*CDKN2A+2.02331501*SETD2-1.54969391*ERBB2-1.75892268*BRAF-1.66240923*TSC1-1.59870732*MAP2K1-0.46874594*CDKN1B-0.79330108*JUN-1.24918034*AKT1-0.75809038*HIST1H1D Each gene name in the formula represents the corresponding gene mutation status used to construct the risk scoring model (1: mutation, 0: no mutation).

[0029] Get the training data set and verify the performance of the risk scoring model: Dataset validation in the C-Bioportal cohort database: The effectiveness of the risk scoring model was verified using the somatic mutation data and clinical information of early lung adenocarcinoma in the C-Bioportal cohort database (source: https: / / www.cbioportal.org / study / summary?id=luad_tcga_pan_can_atlas_2018). The results showed that in terms of progression-free survival (PFS, Progression Free Status) Figure 5), overall survival due to tumor death (DSS, Disease-specific Survival status) ( Figure 6 ) and overall survival (OS, Overall patientsurvival status) ( Figure 7 ) There was significant biological significance between the high-risk group and the low-risk group in this risk scoring model, which proved that this risk scoring model can effectively predict the progression of early lung adenocarcinoma from MIA to IA.

[0030] American cohort data set validation: The effectiveness of the risk score model was verified using somatic mutation data and clinical information of early lung adenocarcinoma in an American cohort (source: Qian J, Zhao S, Zou Y, Rahman SMJ, Senosain MF, Stricker T, Chen H, Powell CA, Borczuk AC, Massion PP. Genomic Underpinningsof Tumor Behavior in In Situ and Early Lung Adenocarcinoma. Am J Respir CritCare Med. 2020 Mar 15;201(6):697-706. doi: 10.1164 / rccm.201902-0294OC. PMID:31747302; PMCID: PMC7068818.). The results showed that the overall survival (OS) also had significant biological significance between the high-risk group and the low-risk group in this risk score model ( Figure 8 ), proving that this risk scoring model can effectively predict the progression of early lung adenocarcinoma from MIA to IA.

[0031] Correlation between the risk score model and gene mutation frequency (Frequency), tumor mutation burden (TMB), tumor heterogeneity (MATH) and gene locus mutation frequency (VAF): The high-risk and low-risk groups showed significant differences in gene mutation frequency, TMB, Fig. 9 ) and MATH ( Fig.10 ), proving that model construction is closely related to the occurrence and development of the disease. Secondly, in terms of imaging, pathological differentiation and tumor size, the high-risk and low-risk groups have different gene mutation frequencies (Frequency) ( Fig.11 ,12,13)、TMB( Fig.14 ,15,16)、MATH( Fig.17 ,18,19) and VAF are still significantly correlated ( Fig. 20 ,21). This is closely related to the progression of MIA and IA, proving that this model can effectively predict the progression of early lung adenocarcinoma from MIA to IA, providing a solid and reliable clinical basis for clinical treatment.

[0032] Key mutation genes in the risk scoring model: Imaging, pathological differentiation and tumor size are closely related to the progression of MIA and IA. The key genes in the gene signature that promote tumor progression are TP53 and CDKN2A. TP53 ( Fig. 22 ) and CDKN2A ( Fig.23 ) showed significant differences in imaging, pathological differentiation and tumor size. The key genes related to promoting tumor progression in gene signature were ERBB2 and BRAF. ERBB2 ( Fig.24 ) and BRAF ( Fig.25 ) also had significant differences in imaging and tumor size. The above results further prove that this risk scoring model can effectively predict the progression of early lung adenocarcinoma.

[0033] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for predicting the progression of early lung adenocarcinoma from MIA to IA, characterized in that: The following steps are involved: S1. Obtain mutation status of target gene sets in multiple samples; S2. Calculate the results obtained in S1 using the following formula to obtain risk scores for multiple samples; Risk score = 1.83331403*G TP53 +1.24152921*G CDKN2A +2.02331501*G SETD2 -1.54969391*G ERBB2 -1.75892268*G BRAF -1.66240923*G TSC1 -1.59870732*G MAP2K1 -0.46874594*G CDKN1B -0.79330108*G JUN -1.24918034*G AKT1 -0.75809038*G HIST1H1D In the formula, the G TP53 is the mutation status of the TP53 gene in the sample, with mutation recorded as 1 and no mutation recorded as 0; The G CDKN2A is the mutation status of the CDKN2A gene, with mutation recorded as 1 and no mutation recorded as 0; The G SETD2 is the mutation status of the SETD2 gene, with mutation recorded as 1 and no mutation recorded as 0; The G ERBB2 is the mutation status of the ERBB2 gene, with mutation recorded as 1 and no mutation recorded as 0; The G BRAF is the mutation status of the BRAF gene, with mutation recorded as 1 and no mutation recorded as 0; The G TSC1 is the mutation status of the TSC1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G MAP2K1 is the mutation status of MAP2K1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G CDKN1B is the mutation status of the CDKN1B gene, with mutation recorded as 1 and no mutation recorded as 0; The G JUN is the mutation status of the JUN gene, with mutation recorded as 1 and no mutation recorded as 0; The G AKT1 is the mutation status of the AKT1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G HIST1H1D is the mutation status of the HIST1H1D gene, with mutation recorded as 1 and no mutation recorded as 0; S3. Calculate a median based on the risk scores of the multiple samples obtained, and predict that samples with a risk score greater than the median are at high risk of developing early lung adenocarcinoma, and samples with a risk score less than the median are at low risk of developing early lung adenocarcinoma.

2. A method for constructing a risk scoring model for early lung adenocarcinoma, characterized in that: The following steps are involved: Performing gene sequencing on the samples to obtain genome mutation information, wherein the sources of the samples include microinvasive adenocarcinoma tissue, invasive adenocarcinoma tissue and peripheral blood; Samples from peripheral blood were used as the control group to exclude germline mutation results in tissue samples and obtain somatic gene mutation information in tissue samples; Based on the tumor pathological characteristics of the samples and the somatic mutation information of the genome of the samples, a chi-square test was performed to obtain differentially mutated genes between the microinvasive adenocarcinoma and invasive adenocarcinoma groups; A risk scoring model was constructed based on differentially mutated genes between groups.

3. The construction method according to claim 2, characterized in that: The samples were derived from 755 patients with clinically early-stage lung adenocarcinoma.

4. The construction method according to claim 2, characterized in that: The chi-square test comprises the following steps: Based on the different sources of samples, the sample genes were analyzed for inter-group differential mutations, and the inter-group differential mutation genes were obtained according to the chi-square test results p<0.

05.

5. The construction method according to claim 4, characterized in that: The differentially mutated genes between the microinvasive adenocarcinoma and invasive adenocarcinoma groups include AKT1, BCL2L1, BRAF, CDKN1B, CDKN2A, CDKN2C, ERBB2, HIST1H1D, JUN, MAP2K1, MED12, NF1, PRKAR1A, SETD2, SLC34A2, TERC, TP53, TSC1 and ZFP36L1.

6. The construction method according to claim 2, characterized in that: The methods that can be used to construct the risk scoring model include Logistic, SVM, DecisionTree, RandomForest and XGBoost.

7. The construction method according to claim 2, characterized in that: The constructing of the risk scoring model further comprises: A validation data set is obtained, and the risk scoring model is optimized based on the validation data set.

8. The construction method according to claim 2, characterized in that: Optimizing the risk scoring model based on the validation data set includes: The inter-group differentially mutated genes are screened by Logistic regression analysis combined with a ten-fold cross validation method to determine the genes and the initial formula of the risk scoring model that are ultimately used to construct the risk scoring model; Substituting the gene finally used to construct the risk scoring model into the initial formula of the risk scoring model to obtain the risk scoring model; The initial formula of the risk scoring model is: Risk score = G1M1+…+G n M n , Among them, G n Indicates the mutation of the nth gene used to construct the risk scoring model, with mutation being recorded as 1 and non-mutation being recorded as 0; M n represents the coefficient of the nth gene used to construct the risk scoring model, and n represents the number of genes used to construct the risk scoring model; The median is calculated based on the risk scores of the obtained samples. Samples with risk scores greater than the median are predicted to have a high risk of developing early lung adenocarcinoma, and samples with risk scores less than the median are predicted to have a low risk of developing early lung adenocarcinoma.

9. The construction method according to claim 2, characterized in that: The genes ultimately used to construct the risk scoring model include TP53, SETD2, CDKN2A, BRAF, TSC1, MAP2K1, ERBB2, CDKN1B, JUN, AKT1 and HIST1H1D.

10. The construction method according to claim 9, characterized in that: The risk scoring model uses the following formula: Risk score = 1.83331403*G TP53 +1.24152921*G CDKN2A +2.02331501*G SETD2 -1.54969391*G ERBB2 -1.75892268*G BRAF -1.66240923*G TSC1 -1.59870732*G MAP2K1 -0.46874594*G CDKN1B -0.79330108*G JUN -1.24918034*G AKT1 -0.75809038*G HIST1H1D In the formula, the G TP53 is the mutation status of the TP53 gene in the sample, with mutation recorded as 1 and no mutation recorded as 0; The G CDKN2A is the mutation status of the CDKN2A gene, with mutation recorded as 1 and no mutation recorded as 0; The G SETD2 is the mutation status of the SETD2 gene, with mutation recorded as 1 and no mutation recorded as 0; The G ERBB2 is the mutation status of the ERBB2 gene, with mutation recorded as 1 and no mutation recorded as 0; The G BRAF is the mutation status of the BRAF gene, with mutation recorded as 1 and no mutation recorded as 0; The G TSC1 is the mutation status of the TSC1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G MAP2K1 is the mutation status of MAP2K1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G CDKN1B is the mutation status of the CDKN1B gene, with mutation recorded as 1 and no mutation recorded as 0; The G JUN is the mutation status of the JUN gene, with mutation recorded as 1 and no mutation recorded as 0; The G AKT1 is the mutation status of the AKT1 gene, with mutation recorded as 1 and no mutation recorded as 0; The G HIST1H1D is the mutation status of the HIST1H1D gene, with mutation being recorded as 1 and no mutation being recorded as 0.