Method for Evaluating Prognosis of Survival Period of Patients with Lung Adenocarcinoma, Computer Device and Storage Medium

Abstract

The present invention discloses a method for prognostic evaluation of the survival period of lung adenocarcinoma patients, a computer device, and a storage medium. The prognostic index is calculated by the following formula: Prognostic index = β1 * TNM stage + β2 * ERSS; where β1 and β2 are weight coefficients, and ERSS is the endoplasmic reticulum stress score. The present invention evaluates the correlation between endoplasmic reticulum stress-related genes and the prognosis of lung adenocarcinoma patients according to a machine learning method, establishes an endoplasmic reticulum stress score, and combines the endoplasmic reticulum stress score with the TNM stage to establish a survival prediction model. Experiments show that the survival prediction model of the present invention has a good predictive effect on the overall survival period, and solves the problem that the TNM stage alone cannot accurately evaluate the survival period of patients.

Description

Technical Field

[0001] The present invention relates to the field of biomedical technologies, and particularly to a method for prognostic evaluation of the survival period of lung adenocarcinoma patients, a computer device, and a storage medium. Background Art

[0002] Survival risk prediction is a very important part of the medical safety management system. Analyzing and summarizing the risk prediction method based on the accumulated disease condition data is beneficial to guiding individualized treatment of patients.

[0003] In China, the incidence and fatality rate of lung cancer rank first among various malignant tumors, bringing a huge health and economic burden to the people of our country. [1] According to histological type, lung cancer can be divided into small cell carcinoma, adenocarcinoma, squamous cell carcinoma, large cell carcinoma, etc. Among them, lung adenocarcinoma accounts for more than 40% of lung cancer and is the most numerous type of lung cancer. [2] The eighth edition of the TNM staging of the International Union Against Cancer classifies patients into stages I, II, III, and IV according to clinical and pathological features such as tumor size, vascular and nerve invasion, lymph node, and distant tissue metastasis. [3] Currently, the TNM staging is the main basis for clinicians to predict the survival of lung cancer patients. However, its prediction efficiency is still not satisfactory. [4] Finding a new prognostic prediction scheme is an urgent problem to be solved currently. Discovering prognostic-related molecular targets and establishing a molecular-TNM staging combined prediction model is one of the most effective solutions currently.

[0004] Under the condition of protein stability in normal cells, endoplasmic reticulum stress sensors, including activating transcription factor 6 (ATF6), inositol-requiring enzyme 1α (IRE1α), and PRKR-like endoplasmic reticulum kinase (PERK), are in an inactivated state. In the tumor microenvironment, multiple factors, such as hypoxia, abnormal nutrient supply, accumulation of intracellular reactive oxygen species, and low pH, can interfere with protein folding in the endoplasmic reticulum. [5-8] The accumulation of misfolded proteins disrupts protein stability and activates the sensors, thereby promoting strong endoplasmic reticulum stress in cancer cells. The activation of the sensors can promote the unfolded protein response (UPR), which can restore endoplasmic reticulum homeostasis and promote cell adaptation to stress and enhance survival ability. [9] Interestingly, endoplasmic reticulum stress only plays a carcinogenic role when it is moderate. Uncontrolled endoplasmic reticulum stress-induced excessive UPR will lead to cell death.

[10] .

[0005] The role of endoplasmic reticulum stress in lung adenocarcinoma is still controversial. A research report shows that the overexpression of POU4F3 upregulates endoplasmic reticulum stress and thus inhibits tumor progression in lung adenocarcinoma.

[11] Reactive oxygen species-mediated endoplasmic reticulum stress inhibits tumors in lung cancer cells.

[12] The endoplasmic reticulum stress pathway may not be regulated by neutrophil arginase-1, is released by activated human neutrophils or dead cells, and induces apoptosis in cancer cells.

[13] In addition, endoplasmic reticulum stress has also been reported to be associated with cisplatin resistance in lung cells.

[14] However, some studies have also reported that endoplasmic reticulum stress may play a pro-cancer role in lung cancer. The study by Yamashita et al. found that endoplasmic reticulum stress promoted epithelial-mesenchymal transition and cell invasion in lung adenocarcinoma.

[15] In summary, these findings suggest that endoplasmic reticulum stress may be a promising therapeutic target in lung adenocarcinoma, and it is of great value to comprehensively explore the relationship between endoplasmic reticulum stress and lung adenocarcinoma. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method for evaluating the prognosis of the survival period of lung adenocarcinoma patients, a computer device and a storage medium, aiming at the deficiencies of the prior art, so as to solve the problem that the TNM staging cannot accurately evaluate the survival period of patients.

[0007] To solve the above technical problems, the technical solution adopted by the present invention is: a method for evaluating the prognosis of the survival period of lung adenocarcinoma patients, which calculates a prognosis index by the following formula and uses the prognosis index to evaluate the death risk of lung adenocarcinoma patients:

[0008] Prognosis index = β1 * TNM staging + β2 * ERSS;

[0009] Wherein, β1 and β2 are weight coefficients, 0 < β1 < 5, 0 < β2 < 5; when the TNM staging is I, it is assigned a value of 1, when the staging is II, it is assigned a value of 2, when the staging is III, it is assigned a value of 3, and when the staging is IV, it is assigned a value of 4; ERSS is the endoplasmic reticulum stress score; ERSS = λ1 * SERPINH1 expression level + λ2 * DSG2 expression level + λ3 * GPR37 expression level + λ4 * PCSK9 expression level + λ5 * TRPA1 expression level + λ6 * F2 expression level + λ7 * CDKN3 expression level - λ8 * DMD expression level - λ9 * NR3C2 expression level - λ10 * CFTR expression level - λ11 * CYP1A2 expression level - λ12 * MAPT expression level - λ13 * SYT2 expression level - λ14 * CYP2D6 expression level - λ15 * SCN4A expression level - λ16 * NUPR1 expression level - λ17 * PIK3CG expression level - λ18 * DERL3 expression level; the value ranges of λ1 to λ18 are all 0 to 1.

[0010] In the present invention, the Youden index for predicting the 1-year, 3-year, and 5-year overall survival rates of patients is calculated using the survivalROC package, and the optimal cut-off points for the prognostic index to predict the 1-year, 3-year, and 5-year survival of patients are 7.477722, 7.672388, and 7.437397, respectively.

[16] To predict the survival of patients 1 year after diagnosis, using 7.477722 as the cut-off point of the prognostic index has the best predictive value. When the prognostic index > 7.477722, the predicted patient status is death; when the prognostic index ≤ 7.477722, the predicted patient status is survival. Similarly, to predict the survival of patients 3 years after diagnosis, using 7.672388 as the cut-off point of the prognostic index has the best predictive value. When the prognostic index > 7.672388, the predicted patient status is death; when the prognostic index ≤ 7.672388, the patient status is survival. To predict the survival of patients 5 years after diagnosis, using 7.437397 as the cut-off point of the prognostic index has the best predictive value. When ERSS > 7.437397, the predicted patient status is death; when the prognostic index ≤ 7.437397, the predicted patient status is survival.

[0011] The present invention calculates the prognostic index by combining TNM staging and ERSS, and uses this prognostic index to evaluate the survival period of patients with lung adenocarcinoma. The ROC curve shows that combining TNM staging and ERSS has a better predictive effect on the survival of patients with lung adenocarcinoma, solving the problem that the application of TNM staging cannot accurately evaluate the survival period of patients.

[0012] Cox regression analysis shows that when β1 is 1.58167 and β2 is 3.41433, the survival prediction effect of the method of the present invention is the best.

[0013] LASOO regression shows that when λ1 is 0.13882291, λ2 is 0.09216343, λ3 is 0.07294183, λ4 is 0.06978185, λ5 is 0.06473978, λ6 is 0.03490076, λ7 is 0.02606376, λ8 is 0.01049731, λ9 is 0.01071941, λ10 is 0.01443196, λ11 is 0.01872414, λ12 is 0.01915136, λ13 is 0.03551274, λ14 is 0.03613574, λ15 is 0.0400695, λ16 is 0.0432351, λ17 is 0.07008271, λ18 is 0.0722234, the survival prediction effect of ERSS is the best.

[0014] In the present invention, the calculation process of the endoplasmic reticulum stress score ERSS includes:

[0015] 1) Screen N1 endoplasmic reticulum stress-related genes with a correlation coefficient greater than 7 points;

[0016] 2) Perform differential expression analysis on the N1 endoplasmic reticulum stress-related genes through the limma package in R language to obtain N2 endoplasmic reticulum stress-related genes with significant differential expression in tumor tissues and adjacent tissues;

[0017] 3) Perform univariate cox regression screening on the N2 endoplasmic reticulum stress-related genes through the survival package in R language to obtain N3 genes significantly related to the overall survival of patients;

[0018] 4) Take the intersection of the N2 endoplasmic reticulum stress-related genes and the N3 genes significantly related to the overall survival of patients through a Venn diagram to obtain N4 endoplasmic reticulum stress genes that are both significantly related to the overall survival and significantly differentially expressed in tumor tissues;

[0019] 5) Perform LASSO regression through the glmnet package in R language to further screen the N4 genes, obtain 18 key genes of endoplasmic reticulum stress, and assign a weight coefficient to each gene to obtain the endoplasmic reticulum stress score ERSS.

[0020] The main technical advantages of the gene screening process of the present invention are as follows:

[0021] 1) The present invention screens endoplasmic reticulum stress-related genes, and while establishing a prediction model (ERSS), for the first time comprehensively analyzes the relationship between endoplasmic reticulum stress and the survival period of lung adenocarcinoma patients;

[0022] 2) The present invention screens out key genes closely related to the occurrence, development and survival of lung adenocarcinoma through a progressive method of differential expression analysis of tumor tissues and normal tissues, univariate Cox survival analysis, LASSO regression analysis, and multivariate Cox regression analysis. Compared with the TNM stage, the evaluation process of the present invention is more objective and accurate.

[0023] As an inventive concept, the present invention also provides a method for prognostic evaluation of the survival period of lung adenocarcinoma patients based on the endoplasmic reticulum stress score. Calculate the endoplasmic reticulum stress score ERSS using the following formula, and use the endoplasmic reticulum stress score ERSS to evaluate the death risk of lung adenocarcinoma patients:

[0024] ERSS = λ1 * SERPINH1 expression level + λ2 * DSG2 expression level + λ3 * GPR37 expression level + λ4 * PCSK9 expression level + λ5 * TRPA1 expression level + λ6 * F2 expression level + λ7 * CDKN3 expression level - λ8 * DMD expression level - λ9 * NR3C2 expression level - λ10 * CFTR expression level - λ11 * CYP1A2 expression level - λ12 * MAPT expression level - λ13 * SYT2 expression level - λ14 * CYP2D6 expression level - λ15 * SCN4A expression level - λ16 * NUPR1 expression level - λ17 * PIK3CG expression level - λ18 * DERL3 expression level; where the value ranges of λ1 to λ18 are all 0 to 1.

[0025] In the present invention, the Youden index for predicting the 1-year, 3-year, and 5-year overall survival rates of patients by ERSS is calculated using the survivalROC package, and the optimal cut-off points for ERSS to predict the 1-year, 3-year, and 5-year survival of patients are 1.520982, 1.516202, and 1.828536, respectively.

[16] . If you want to predict the survival status of patients 1 year after diagnosis, using 1.520982 as the ERSS cut-off point has the best predictive value. When ERSS > 1.520982, the predicted patient status is death; when ERSS ≤ 1.520982, the predicted patient status is survival. Similarly, if you want to predict the survival status of patients 3 years after diagnosis, using 1.516202 as the ERSS cut-off point has the best predictive value. When ERSS > 1.516202, the predicted patient status is death; when ERSS ≤ 1.516202, the patient status is survival. If you want to predict the survival status of patients 5 years after diagnosis, using 1.828536 as the ERSS cut-off point has the best predictive value. When ERSS > 1.828536, the predicted patient status is death; when ERSS ≤ 1.828536, the predicted patient status is survival.

[0026] Multivariate Cox analysis shows that ERSS is a prognostic risk factor independent of TNM staging (Figure 2(B), Figure 3(E), Figure 3(H)), that is, for patients with the same TNM stage, ERSS can still well predict the therapeutic effect of patients. For example, in the GSE30219 cohort, since 94.5% of the patients are stage I (Table 1), TNM staging can no longer predict patient survival, and at this time ERSS can be used as the main basis for predicting patient survival (Figure 3(B)).

[0027] The present invention also provides a prognostic system for the survival period of lung adenocarcinoma patients, which includes a nomogram. The first row of the nomogram is a score scale, and the score scale value range is 0 to 100;

[0028] The second line is the endoplasmic reticulum stress score (ERSS) value scale. The range of ERSS values is 0 to 3.5, and the length of the ERSS value scale is the same as that of the score scale. ERSS value 0 corresponds to score scale value 0, and ERSS value 3.5 corresponds to score scale value 100;

[0029] The third line is the TNM staging scale. The range of TNM staging values is 1 to 4. TNM staging value 1 corresponds to score scale value 0, and TNM staging value 4 corresponds to score scale value 32;

[0030] The fourth line is the total score scale. The range of values of the total score scale is 0 to 140, and the length of the total score scale is the same as that of the score scale;

[0031] The left endpoints of the score scale, ERSS value scale, TNM staging scale, and total score scale are aligned; the right endpoint of the score scale, the right endpoint of the ERSS value scale, and the left endpoint of the total score scale are aligned;

[0032] The fifth line is the probability of survival greater than 1 year. The range of probability values is 0.95 to 0.05. Probability value 0.95 corresponds to total score scale value 38, and 0.05 corresponds to total score scale value 132;

[0033] The sixth line is the probability of survival greater than 3 years. The range of probability values is 0.95 to 0.05. Probability value 0.95 corresponds to total score scale value 18, and 0.05 corresponds to total score scale value 112;

[0034] The seventh line is the probability of survival greater than 5 years. The range of probability values is 0.9 to 0.05. Probability value 0.95 corresponds to total score scale value 22, and 0.05 corresponds to total score scale value 99;

[0035] Add the value of the score scale corresponding to the ERSS value and the value of the score scale corresponding to the TNM stage to obtain the total score value. The probability values in the fifth to seventh lines corresponding to the total score value are the survival probabilities of patients with lung adenocarcinoma for more than 1 year, more than 3 years, and more than 5 years, respectively;

[0036] The endoplasmic reticulum stress score (ERSS) is calculated by the above method.

[0037] The technical advantages of the nomogram in the present invention are mainly as follows: 1) Compared with simply applying TNM staging, the nomogram in the present invention adds the ERSS index. Combining ERSS and TNM staging can more accurately predict the survival period of patients than simply applying TNM staging; 2) The nomogram can intuitively display the prediction model constructed in the present invention, and the model can also be conveniently applied in clinical practice.

[0038] As an inventive concept, the present invention also provides a computer device, comprising a memory, a processor, and a computer program stored on the memory; the processor executes the computer program to implement the steps of the method of the present invention.

[0039] As an inventive concept, the present invention also provides a computer program product, comprising a computer program / instructions; when the computer program / instructions are executed by a processor, the steps of the method of the present invention are implemented.

[0040] As an inventive concept, the present invention also provides a computer-readable storage medium, on which a computer program / instructions are stored; when the computer program / instructions are executed by a processor, the steps of the method of the present invention are implemented.

[0041] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention evaluates the correlation between endoplasmic reticulum stress-related genes and the prognosis of lung adenocarcinoma patients according to a machine learning method, establishes an endoplasmic reticulum stress score, and combines the endoplasmic reticulum stress score and the TNM stage to establish a survival prediction model. Experiments show that the survival prediction model of the present invention has a good prediction effect on the overall survival period, and solves the problem that the TNM stage alone cannot accurately evaluate the survival period of patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] FIG. 1 is a process of constructing an ERSS in the TCGA cohort according to an embodiment of the present invention: FIG. 1(A) analyzes the differential expression of endoplasmic reticulum stress-related genes in tumor tissues and normal tissues, and finds that 54 genes are significantly up-regulated in tumor tissues, 103 genes are significantly down-regulated, and 607 genes have no obvious difference; FIG. 1(B) takes the intersection of differentially expressed genes and survival-significantly related genes in a Venn diagram to obtain 45 differentially expressed genes significantly related to survival; FIGS. 1(C)-1(D) further screen the 45 genes by LASSO regression and find 18 key genes; FIG. 1(E) is an expression heat map of the 18 key genes; FIG. 1(F) is the coefficients of the 18 genes constituting the ERSS.

[0043] Figure 2 shows the process of evaluating the prognostic significance of ERSS in the TCGA cohort and establishing a combined model of ERSS and TNM staging for the embodiments of the present invention: In Figure 2(A), the overall survival of lung adenocarcinoma patients with high ERSS was significantly shorter (P<0.0001); In Figure 2(B), multivariate Cox regression indicated that TNM staging and ERSS were independent risk factors affecting patient survival; In Figure 2(C), the combined prediction model of TNM staging and ERSS was visualized according to the nomogram for convenient clinical application; In Figure 2(D), the predictive efficacy of ERSS for the 1-year, 3-year, and 5-year death risks of lung adenocarcinoma patients; In Figure 2(E), the predictive efficacy of TNM staging for the 1-year, 3-year, and 5-year death risks of lung adenocarcinoma patients; In Figure 2(F), the predictive efficacy of the combined model of TNM staging and ERSS for the 1-year, 3-year, and 5-year death risks of lung adenocarcinoma patients.

[0044] Figure 3 shows the process of externally validating the combined model in the GEO cohort for the embodiments of the present invention: In the GSE30219 cohort, in Figure 3(A), the Kaplan-Meier survival curve showed that the overall survival of patients with high ERSS expression was significantly shorter (P = 0.005); In Figure 3(B), Cox regression indicated that ERSS was the only independent risk factor; In Figure 3(C), the predictive efficacy of the combined model for the 1-year, 3-year, and 5-year death risks of lung adenocarcinoma patients was 0.775, 0.675, and 0.658, respectively. In the GSE31210 cohort, in Figure 3(D), the Kaplan-Meier survival curve showed that the overall survival of patients with high ERSS expression was shorter (P = 0.059); In Figure 3(E), Cox regression indicated that both TNM staging and ERSS were independent risk factors; (F) The predictive efficacy of the combined model for the 1-year, 3-year, and 5-year death risks of lung adenocarcinoma patients was 0.919, 0.799, and 0.717, respectively. In the GSE72094 cohort, in Figure 3(G), the Kaplan-Meier survival curve showed that the overall survival of patients with high ERSS expression was shorter (P = 0.0035); In Figure 3(H), Cox regression indicated that both TNM staging and ERSS were independent risk factors; In Figure 3(I), the predictive efficacy of the combined model for the 1-year, 3-year, and 5-year death risks of lung adenocarcinoma patients was 0.695, 0.710, and 0.739, respectively.

[0045] Figure 4 This is an example of using the prognostic risk model of the present invention. Detailed implementation manners

[0046] The specific construction method of the combined survival prediction model in the embodiments of the present invention is as follows:

[0047] (1) Through the UCSC Xena website ( https: / / xenabrowser.net / ), gene expression and clinical data of 517 lung adenocarcinoma patients were obtained from the Cancer Genome Atlas (TCGA) database, and from the GEO database ( http: / / www.ncbi.nlm.nih.gov / geo / Gene expression and clinical data of lung adenocarcinoma patients in three cohorts, GSE30219, GSE31210, and GSE72092, were obtained. Among them, the TCGA cohort was used as the training set for endoplasmic reticulum stress score and prediction model construction, and the GEO cohort was used as the validation set for external validation of the prediction model. Table 1 shows the clinical characteristics of patients in the four cohorts.

[0048] Table 1 Clinicopathological characteristics of lung adenocarcinoma patients in the TCGA cohort and GEO cohort

[0049]

[0050] (2) A total of 799 endoplasmic reticulum stress-related genes with a correlation coefficient > 7 were downloaded and screened from the GeenCards website ( https: / / www.genecards.org / );

[0051] (3) Differential expression analysis of endoplasmic reticulum stress-related genes was performed using the limma package in R language

[17] , and 157 endoplasmic reticulum stress genes with significant differential expression between tumor tissues and adjacent tissues were obtained (Figure 1(A));

[0052] (4) Univariate cox regression screening of the above endoplasmic reticulum stress-related genes was performed using the survival package in R language

[18] , and a total of 153 genes significantly related to the overall survival of patients were obtained (Figure 1(B));

[0053] (5) By taking the intersection of Venn diagrams, 45 endoplasmic reticulum stress genes that were both significantly related to the overall survival and significantly differentially expressed in tumor tissues were obtained (Figure 1(B));

[0054] (6) LASSO regression was performed using the glmnet package in R language

[19] , and the above genes were further screened to obtain 18 key genes of endoplasmic reticulum stress ( Figure 1(C) - Figure 1(D) ), and each gene was assigned a weight coefficient ( Figure 1(E) to Figure 1(F) ), and an endoplasmic reticulum stress score (ERSS, endoplasmic reticulum stress score) was established. The formula is as follows:

[0055] ERSS = λ1 * SERPINH1 expression level + λ2 * DSG2 expression level + λ3 * GPR37 expression level + λ4 * PCSK9 expression level + λ5 * TRPA1 expression level + λ6 * F2 expression level + λ7 * CDKN3 expression level - λ8 * DMD expression level - λ9 * NR3C2 expression level - λ10 * CFTR expression level - λ11 * CYP1A2 expression level - λ12 * MAPT expression level - λ13 * SYT2 expression level - λ14 * CYP2D6 expression level - λ15 * SCN4A expression level - λ16 * NUPR1 expression level - λ17 * PIK3CG expression level - λ18 * DERL3 expression level; The value ranges of λ1 to λ18 are from 0 to 1, and the optimal values of each coefficient are: λ1 is 0.13882291, λ2 is 0.09216343, λ3 is 0.07294183, λ4 is 0.06978185, λ5 is 0.06473978, λ6 is 0.03490076, λ7 is 0.02606376, λ8 is 0.01049731, λ9 is 0.01071941, λ10 is 0.01443196, λ11 is 0.01872414, λ12 is 0.01915136, λ13 is 0.03551274, λ14 is 0.03613574, λ15 is 0.0400695, λ16 is 0.0432351, λ17 is 0.07008271, λ18 is 0.0722234;

[0056] In the above formula: SERPINH1, serine protease inhibitor H1; DSG2, desmoglein 2; GPR37, G protein-coupled receptor 37; PCSK9, proprotein convertase subtilisin / kexin type 9 protein; TRPA1, transient receptor potential cation channel subfamily A member 1; F2, coagulation factor II recombinant protein; CDKN3, cyclin-dependent kinase inhibitor 3; NR3C2, mineralocorticoid receptor gene; DMD, dystrophin; CFTR is a cystic fibrosis transmembrane conductance regulator; CYP1A2, cytochrome oxidase; MAPT, microtubule-associated protein Tau; SYT2, synaptotagmin 2; CYP2D6 is a member of the cytochrome P450 family; SCN4A, alpha subunit of voltage-gated sodium channel; NUPR1, nuclear protein 1 transcriptional regulator.

[0057] (7) After dividing the patients into ERSS high-expression group (252 cases in total) and ERSS low-expression group (252 cases in total) according to the median value of ERSS, Kaplan-Meier survival analysis was performed using the survival package in R language

[18] To verify the survival prognostic significance of ERSS, it was found that the OS (overall survival) of patients with high ERSS was significantly shorter (Figure 2(A));

[0058] (8) Perform multivariate Cox proportional hazards regression using the survival package in R language

[18] , and it is clear that ERSS and TNM staging are independent risk factors for the survival of patients with lung adenocarcinoma, while age, gender, and smoking history have no significant correlation with survival (Figure 2 (B));

[0059] (9) Exclude age, gender, and smoking history, and establish a combined survival prediction model for ERSS and TNM staging through Cox regression. The calculation formula of the model is as follows:

[0060] Prognostic index = β1 * TNM staging + β2 * ERSS;

[0061] Among them, when the TNM staging is I, it is assigned a value of 1, when it is II, it is assigned a value of 2, when it is III, it is assigned a value of 3, and when it is IV, it is assigned a value of 4. The value ranges of β1 and β2 are from 0 to 5, and the optimal values of each coefficient are: β1 takes 1.58167, and β2 takes 3.41433.

[0062] Figure 2 (C) shows the nomogram visualizing the model through the rms package in R language

[20] , aiming to facilitate clinical application: The first row in the figure is the score scale, the second row is ERSS, with a range of 0 to 3.5, corresponding to scores of 0 - 100 points, the third row is TNM staging, with a range of 1 - 4, corresponding to scores of 0 - 32 points, the fourth row is the total score, which is the sum of the scores of each factor, the fifth row is the model's prediction of the 1-year survival rate of patients, the sixth row is the model's prediction of the 3-year survival rate of patients, and the seventh row is the model's prediction of the 5-year survival rate of patients. Specific examples of the application of the model will be shown in the subsequent content.

[0063] In the nomogram, the range of TNM staging is 1 to 4, and the range of ERSS is 0 to 3.5. The Nomogram coefficient conversion formula for TNM staging and ERSS is:

[0064] Nomogram TNM分期 = log(β1) × (4 - 1) = 1.37544;

[0065] Nomogram ERSS = log(β2) × (3.5 - 0) = 4.29793;

[0066] The Nomogram coefficient of ERSS is the largest, indicating that it has the greatest impact on the prediction result. Therefore, the score range of ERSS is set to 0 - 100, and the score range of TNM stage is set to 100×(1.37544 / 4.29793)≈32. The score of ERSS in the nomogram = 100*(ERSS / 3.5), and the score of TNM stage in the nomogram = 32*(TNM stage / 4). Table 2 shows the sorting and scoring results after coefficient conversion.

[0067] Table 2 Sorting and scoring results after coefficient conversion

[0068]

[0069] The predictive efficacy of ERSS, TNM, and the combined model for the 1-year, 3-year, and 5-year survival of patients was evaluated by ROC curves respectively ( Figure 2(D) to Figure 2(F) ).

[0070] In 3 GEO cohorts, the survival prediction value of ERSS was verified by Kaplan-Meier survival analysis and multivariate Cox regression respectively, and the prediction value of the combined model was verified by ROC curves.

[0071] In the GSEA30219 cohort, the overall survival of patients with high ERSS was significantly shorter (P = 0.005, Figure 3(A)), and ERSS was the only independent risk factor in this cohort (Figure 3(B)). The ROC curve indicated that the combined model had a good predictive effect on the overall survival (1-year AUC = 0.775, 3-year AUC = 0.675, 5-year AUC = 0.658; Figure 3(C)).

[0072] In the GSEA31210 cohort, the overall survival of patients with high ERSS was shorter (P = 0.059, Figure 3(D)), and ERSS and TNM stage were independent risk factors for the overall survival of patients (Figure 3(E)). The ROC curve indicated that the combined model had a very good predictive effect on the overall survival (1-year AUC = 0.919, 3-year AUC = 0.799, 5-year AUC = 0.717; Figure 3(F)).

[0073] In the GSEA72094 cohort, the overall survival of patients with high ERSS was significantly shorter (P = 0.0035; Figure 3(G)), and ERSS and TNM stage were independent risk factors for the overall survival of patients (Figure 3(H)). The ROC curve indicated that the combined model had a good predictive effect on the overall survival (1-year AUC = 0.695, 3-year AUC = 0.710, 5-year AUC = 0.739; Figure 3I )

[0074] The following examples illustrate the specific methods of applying the risk model (prediction model), such as Figure 4As shown: If a patient's ERSS is 1.5 and the TNM stage is stage IV, draw vertical lines from the second row (ERSS) and the third row (TNM stage) to the first row (score scale) respectively to obtain the scores of ERSS and TNM stage in the nomogram, which are approximately 43 and 32 points respectively. The sum of the two gives the total score, approximately: 43 + 32 = 75 points. Draw vertical lines from the fourth row (total score) to the fifth, sixth, and seventh rows respectively to obtain the 1-year, 3-year, and 5-year survival rates of the patient, which are approximately 78%, 56%, and 36% respectively.

[0075] References

[0076] [1] GAO S, LI N, WANG S, et al. Lung Cancer in People's Republic of China[J]. J Thorac Oncol, 2020, 15(10): 1567 - 76.

[0077] [2] GANTI A K, KLEIN A B, COTARLA I, et al. Update of Incidence, Prevalence, Survival, and Initial Treatment in Patients With Non - Small CellLung Cancer in the US[J]. JAMA Oncol, 2021.

[0078] [3] HWANG J K, PAGE B J, FLYNN D, et al. Validation of the Eighth EditionTNM Lung Cancer Staging System [J]. J Thorac Oncol, 2020, 15(4):649 - 54.

[0079] [4] YUN J K, LEE G D, KIM H R, et al. Validation of the 8th edition of theTNM staging system in 3,950 patients with surgically resected non - small celllung cancer[J]. J Thorac Dis, 2019, 11(7):2955 - 64.

[0080] [5]YOUNG R M,ACKERMAN D,QUINN Z L,et al.Dysregulated mTORC1 renders cells critically dependent on desaturated lipids for survival under tumor-like stress[J].Genes Dev,2013,27(10):1115-31.

[0081] [6]WEI Y,WANG D,TOPCZEWSKI F,et al.Saturated fatty acids induce endoplasmic reticulum stress and apoptosis independently of ceramide in liver cells[J].Am J Physiol Endocrinol Metab,2006,291(2): E275-81.

[0082] [7]SHIMIZU Y,HENDERSHOT L M.Oxidative folding:cellular strategies for dealing with the resultant equimolar production of reactive oxygen species[J].Antioxid Redox Signal,2009,11(9):2317-31.

[0083] [8]DONG L,KREWSON E A,YANG L V.Acidosis Activates Endoplasmic Reticulum Stress Pathways through GPR4 in Human Vascular Endothelial Cells[J].International journal of molecular sciences,2017,18(2).

[0084] [9]WALTER P,RON D.The unfolded protein response:from stress pathway to homeostatic regulation[J]. Science,2011,334(6059):1081-6.

[0085]

[10] CHEN X,CUBILLOS-RUIZ J R.Endoplasmic reticulum stress signals inthe tumour and its microenvironment[J].Nat Rev Cancer,2021,21(2):71-88.

[0086]

[11] CHAI X,DING X,LYU X,et al.POU4F3 Acts as a Tumor Suppressor inLung Adenocarcinoma via the Endoplasmic Reticulum Stress Signaling Pathway[J].J Cancer,2022,13(2):554-64.

[0087]

[12] SONG X,XU X,LU J,et al.Lamprey Immune Protein Mediates Apoptosisof Lung Cancer Cells Via the Endoplasmic Reticulum Stress Signaling Pathway[J].Front Oncol,2021,11:663600.

[0088]

[13] GARCíA-NAVAS R,GAJATE C,MOLLINEDO F.Neutrophils drive endoplasmicreticulum stress-mediated apoptosis in cancer cells through arginase-1release[J].Sci Rep,2021,11(1):12574.

[0089]

[14] WANG X,CHEN B,XU D,et al.Delicaflavone Reverses CisplatinResistance via Endoplasmic Reticulum Stress Signaling Pathway in Non-SmallCell Lung Cancer Cells[J].Onco Targets Ther,2020,13: 10315-22.

[0090]

[15] YAMASHITA M, OGASAWARA M, KAWASAKI Y, et al. Deficiency of protein-L-isoaspartate (D-aspartate) O-methyl-transferase expression under endoplasmic reticulum stress promotes epithelial mesenchymal transition in lung adenocarcinoma[J]. Oncotarget, 2018, 9(17): 13287 - 300.

[0091]

[16] HEAGERTY P J, LUMLEY T, PEPE M S. Time-dependent ROC curves for censored survival data and a diagnostic marker[J]. Biometrics, 2000, 56(2): 337 - 44.

[0092]

[17] RITCHIE M E, PHIPSON B, WU D, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies[J]. Nucleic Acids Res, 2015, 43(7): e47.

[0093]

[18] LOPRINZI C L, LAURIE J A, WIEAND H S, et al. Prospective evaluation of prognostic variables from patient-completed questionnaires. North Central Cancer Treatment Group[J]. J Clin Oncol, 1994, 12(3): 601 - 7.

[0094]

[19] FRIEDMAN J,HASTIE T,TIBSHIRANI R.Regularization Paths for Generalized Linear Models via Coordinate Descent[J].J Stat Softw,2010,33(1):1-22.

[0095]

[20] PARK S Y.Nomogram:An analogue tool to deliver digital knowledge[J].J Thorac Cardiovasc Surg,2018, 155(4):1793。

Claims

1. A method for prognostic evaluation of the survival period of patients with lung adenocarcinoma, characterized in that, The prognostic index is calculated using the following formula: Prognostic index = β1 * TNM stage + β2 * ERSS; where β1 and β2 are weight coefficients, 0 < β1 < 5, 0 < β2 < 5; when the TNM stage is I, it is assigned a value of 1, when the stage is II, it is assigned a value of 2, when the stage is III, it is assigned a value of 3, and when the stage is IV, it is assigned a value of 4; ERSS is the endoplasmic reticulum stress score; ERSS = λ1 * SERPINH1 expression level + λ2 * DSG2 expression level + λ3 * GPR37 expression level + λ4 * PCSK9 expression level + λ5 * TRPA1 expression level + λ6 * F2 expression level + λ7 * CDKN3 expression level - λ8 * DMD expression level - λ9 * NR3C2 expression level - λ10 * CFTR expression level - λ11 * CYP1A2 expression level - λ12 * MAPT expression level - λ13 * SYT2 expression level - λ14 * CYP2D6 expression level - λ15 * SCN4A expression level - λ16 * NUPR1 expression level - λ17 * PIK3CG expression level - λ18 * DERL3 expression level; The value ranges of λ1 to λ18 are all 0 to 1; The death risk of patients with lung adenocarcinoma is evaluated using the said prognostic index.

2. The method for evaluating the prognosis of the survival period of lung adenocarcinoma patients according to claim 1, wherein β1 takes the value of 1.58167, and β2 takes the value of 3.41433.

3. The method for evaluating the prognosis of the survival period of lung adenocarcinoma patients according to claim 1, characterized in that, λ1 takes the value of 0.13882291, λ2 takes the value of 0.09216343, λ3 takes the value of 0.07294183, λ4 takes the value of 0.06978185, λ5 takes the value of 0.06473978, λ6 takes the value of 0.03490076, λ7 takes the value of 0.02606376, λ8 takes the value of 0.01049731, λ9 takes the value of 0.01071941, λ10 takes the value of 0.01443196, λ11 takes the value of 0.01872414, λ12 takes the value of 0.01915136, λ13 takes the value of 0.03551274, λ14 takes the value of 0.03613574, λ15 takes the value of 0.0400695, λ16 takes the value of 0.0432351, λ17 takes the value of 0.07008271, and λ18 takes the value of 0.0722234.

4. The method for evaluating the prognosis of the survival period of a lung adenocarcinoma patient according to any one of claims 1 to 3, characterized in that, The calculation process of the endoplasmic reticulum stress score ERSS includes: 1) Screening N1 endoplasmic reticulum stress-related genes with a correlation coefficient greater than 7 points; 2) Performing differential expression analysis on the N1 endoplasmic reticulum stress-related genes through the limma package in R language to obtain N2 endoplasmic reticulum stress-related genes with significant differential expression between tumor tissues and para-carcinoma tissues; 3) Performing univariate cox regression screening on the N2 endoplasmic reticulum stress-related genes through the survival package in R language to obtain N3 genes significantly related to the overall survival of patients; 4) Taking the intersection of the N2 endoplasmic reticulum stress-related genes and the N3 genes significantly related to the overall survival of patients through a Venn diagram to obtain N4 endoplasmic reticulum stress genes that are both significantly related to the overall survival and have significant differential expression in tumor tissues; 5) Perform LASSO regression using the glmnet package in R language to further screen the N4 genes, obtain 18 key genes of endoplasmic reticulum stress, and assign a weight coefficient to each gene to obtain the endoplasmic reticulum stress score ERSS.

5. A method for evaluating the prognosis of the survival period of patients with lung adenocarcinoma based on endoplasmic reticulum stress score, characterized in that, Calculate the endoplasmic reticulum stress score ERSS using the following formula: ERSS = λ1 * expression level of SERPINH1 + λ2 * expression level of DSG2 + λ3 * expression level of GPR37 + λ4 * expression level of PCSK9 + λ5 * expression level of TRPA1 + λ6 * expression level of F2 + λ7 * expression level of CDKN3 - λ8 * expression level of DMD - λ9 * expression level of NR3C2 - λ10 * expression level of CFTR - λ11 * expression level of CYP1A2 - λ12 * expression level of MAPT - λ13 * expression level of SYT2 - λ14 * expression level of CYP2D6 - λ15 * expression level of SCN4A - λ16 * expression level of NUPR1 - λ17 * expression level of PIK3CG - λ18 * expression level of DERL3; Among them, the value ranges of λ1 to λ18 are all 0 to 1; Use the endoplasmic reticulum stress score ERSS to evaluate the death risk of lung adenocarcinoma patients.

6. The method for evaluating the survival prognosis of lung adenocarcinoma patients based on endoplasmic reticulum stress score according to claim 5, wherein λ1 takes 0.13882291, λ2 takes 0.09216343, λ3 takes 0.07294183, λ4 takes 0.06978185, λ5 takes 0.06473978, λ6 takes 0.03490076, λ7 takes 0.02606376, λ8 takes 0.01049731, λ9 takes 0.01071941, λ10 takes 0.01443196, λ11 takes 0.01872414, λ12 takes 0.01915136, λ13 takes 0.03551274, λ14 takes 0.03613574, λ15 takes 0.0400695, λ16 takes 0.0432351, λ17 takes 0.07008271, λ18 takes 0.0722234.

7. The method for evaluating the prognosis of the survival period of lung adenocarcinoma patients based on the endoplasmic reticulum stress score according to claim 5, characterized in that, The calculation process of the endoplasmic reticulum stress score ERSS includes: 1) Screen N1 endoplasmic reticulum stress-related genes with a correlation coefficient greater than 7 points; 2) Perform differential expression analysis on the N1 endoplasmic reticulum stress-related genes through the limma package in R language to obtain N2 endoplasmic reticulum stress-related genes with significant differential expression in tumor tissues and adjacent tissues; 3) Perform univariate cox regression screening on the N2 endoplasmic reticulum stress-related genes through the survival package in R language to obtain N3 genes significantly related to the overall survival of patients; 4) Obtain the intersection of the N2 endoplasmic reticulum stress-related genes and the N3 genes significantly related to the overall survival of patients through a Venn diagram to obtain N4 endoplasmic reticulum stress genes that are both significantly related to the overall survival and significantly differentially expressed in tumor tissues; 5) Perform LASSO regression using the glmnet package in R language to further screen the N4 genes, obtain 18 key genes of endoplasmic reticulum stress, and assign a weight coefficient to each gene to obtain the endoplasmic reticulum stress score ERSS.

8. A survival prognosis system for patients with lung adenocarcinoma, characterized in that, It includes a nomogram, and the first row of the nomogram is a score scale, and the score range of the score scale is 0 to 100; The second row is the score scale of the endoplasmic reticulum stress score (ERSS), the ERSS score range is 0 to 3.5, and the length of the ERSS score scale is the same as the length of the score scale. The ERSS score of 0 corresponds to the score of 0 on the score scale, and the ERSS score of 3.5 corresponds to the score of 100 on the score scale; The third row is the TNM staging scale, the TNM staging value range is 1 to 4, the TNM staging value of 1 corresponds to the score of 0 on the score scale, and the TNM staging value of 4 corresponds to the score of 32 on the score scale; The fourth row is the total score scale, the score range of the total score scale is 0 to 140, and the length of the total score scale is the same as the length of the score scale; The left endpoints of the score scale, the ERSS score scale, the TNM staging scale, and the total score scale are aligned; The right endpoint of the score scale, the right endpoint of the ERSS score scale, and the left endpoint of the total score scale are aligned; The fifth row is the probability of survival greater than 1 year, the probability value range is 0.95 to 0.05, the probability value of 0.95 corresponds to the score of 38 on the total score scale, and the probability value of 0.05 corresponds to the score of 132 on the total score scale; The sixth row is the probability of survival greater than 3 years, the probability value range is 0.95 to 0.05, the probability value of 0.95 corresponds to the score of 18 on the total score scale, and the probability value of 0.05 corresponds to the score of 112 on the total score scale; The seventh row is the probability of survival greater than 5 years, the probability value range is 0.9 to 0.05, the probability value of 0.95 corresponds to the score of 22 on the total score scale, and the probability value of 0.05 corresponds to the score of 99 on the total score scale; Add the value of the score scale corresponding to the ERSS score and the value of the score scale corresponding to the TNM staging to obtain the total score value. The probability values in the fifth to seventh rows corresponding to the total score value are the survival probabilities of patients with lung adenocarcinoma for more than 1 year, more than 3 years, and more than 5 years, respectively; The endoplasmic reticulum stress score (ERSS) is calculated by the method described in any one of claims 5 to 7.

9. A computer device, comprising a memory, a processor, and a computer program stored on the memory; characterized in that, The processor executes the computer program to implement the steps of the method described in any one of claims 1 to 4, or to implement the steps of the method described in any one of claims 5 to 7.

10. A computer program product, comprising a computer program / instructions; characterized in that, When the computer program / instructions are executed by the processor, the steps of the method described in any one of claims 1 to 4 are implemented, or the steps of the method described in any one of claims 5 to 7 are implemented.

11. A computer-readable storage medium having a computer program / instructions stored thereon; characterized in that, When the computer program / instructions are executed by the processor, the steps of the method described in any one of claims 1 to 4 are implemented, or the steps of the method described in any one of claims 5 to 7 are implemented.

Images