Gastric adenocarcinoma prognosis marker and clinical prognosis prediction model

Abstract

The invention discloses a prognostic marker for gastric adenocarcinoma. The prognostic marker is composed of eight genes, namely F5, SLC5A1, PHYHD1, FNDC1, NFE2L3, SCUBE2, CBS and CTHRC1. A risk assessment model constructed on the basis of the eight genes is beneficial for better predicting the prognosis condition of a gastric adenocarcinoma patient, effectively evaluating the prognosis risk of different patients and performing early intervention, and brings great help for guiding clinical treatment of the disease, so that the five-year survival rate of the gastric adenocarcinoma patient is increased. The risk assessment model is high in prediction accuracy, has universality, is remarkably superior to other gastric adenocarcinoma prognosis methods in the prior art, and is also remarkably superior to other models constructed in the research process.

Description

Technical Field

[0001] The invention belongs to the field of biomedical technology, and specifically relates to gastric adenocarcinoma prognostic markers and a clinical prognosis prediction model. Background Art

[0002] Gastric cancer is the fifth most common malignant tumor in the world and the third most lethal of all cancers, with its incidence varying from region to region. Surgery remains the cornerstone of gastric cancer treatment. In recent years, with the continuous introduction of immune and molecular targeted drugs, the modes of perioperative and advanced treatment have undergone significant changes and progress, which has also increased the chances of survival. Although the prognosis of gastric cancer patients has been improved by radical surgery combined with postoperative adjuvant therapy, about 50% of patients still experience tumor recurrence and metastasis within 2 years after surgery and eventually die. In addition, due to the high heterogeneity of tumors, the treatment response and prognosis at different stages are inconsistent. Therefore, studying the molecular mechanisms of tumor invasion, metastasis, occurrence and prognosis from a genomic perspective may provide highly sensitive treatment methods, which also provides more hope and possibilities for the determination of new prognostic and diagnostic indicators and therapeutic targets. We aim to find new and effective prognostic risk markers for gastric adenocarcinoma and establish a new risk assessment model to provide a basis for further diagnosis and treatment of gastric adenocarcinoma patients.

[0003] F5 gene: The gene encoding coagulation factor V, whose encoded product is an important auxiliary factor in the blood coagulation cascade. This factor circulates in the plasma and is converted to an active form by the activation peptide released by thrombin during the coagulation process, generating a heavy chain and a light chain, which are fixed together by calcium ions. The activated protein is a cofactor that works with activated coagulation factor X to activate prothrombin to thrombin. Defects in this gene can lead to autosomal recessive bleeding diseases or autosomal dominant thrombotic diseases, namely activated protein C resistance.

[0004] SLC5A1 gene: the gene encoding solute carrier family 5 member 1. This gene encodes a member of the sodium-dependent glucose transporter (SGLT) family. The encoded integral membrane protein is the major mediator of dietary glucose and galactose uptake in the intestinal lumen. Mutations in this gene are associated with glucose-galactose malabsorption. Diseases associated with SLC5A1 include glucose / galactose malabsorption and renal glycosuria. Its related pathways include digestion and absorption and transmembrane transporter disorders.

[0005] PHYHD1 gene: The gene encoding Phytanoyl-CoA dioxygenasedomain-containing 1, the gene expression product of which is a 2-oxidogluconic acid (2OG)-dependent dioxygenase, has a potential role in Alzheimer's disease, certain cancers and immune cell function.

[0006] The function of the FNDC1 (Fibronectin Type III Domain Containing 1) gene involves multiple biological processes, such as cellular response to hypoxia, cardiomyocyte apoptosis, and protein phosphorylation. Studies have found that overexpression or knockdown of FNDC1 can promote or inhibit tumor occurrence and metastasis, respectively. For example, in colorectal cancer and non-small cell lung cancer, upregulation of FNDC1 accelerates the survival of tumor cells to 5-FU or radiotherapy, while its inhibition makes tumor cells more sensitive to chemotherapy and radiotherapy.

[0007] NFE2L3 gene: Belongs to the cap'n' collar region leucine zipper family, mainly located in the endoplasmic reticulum and nuclear membrane. As an important regulator of cellular stress response, it is expressed in various tissues of the human body. It participates in the regulation of various biological and cellular processes, such as cell cycle, cell differentiation or inflammation, and is significantly associated with epigenetics, cell cycle regulation, calcium signaling pathway and steroid hormone biosynthesis.

[0008] SCUBE2 gene: The encoded protein is a secretory or membrane-bound protein originally discovered in endothelial cells (ECs). It is involved in gene expression regulation and post-translational control, participates in the regulation of the HH signaling pathway, promotes endochondral ossification, and promotes angiogenesis. SCUBE2 is considered a tumor suppressor and is generally downregulated in several tumors, including gastric cancer. In patients with dyslipidemia and type 2 diabetes, high expression of SCUBE2 contributes to the integrity of the blood-brain barrier, and several studies have linked it to autoimmune diseases.

[0009] CTHRC1 gene: encodes a 28 kDa secreted glycosylated protein. Many studies have confirmed that CTHRC1 is expressed in a variety of human solid tumors, and it has been reported that the expression of CTHRC1 can increase the expression of activated HIF-1α / CXCR4 signaling pathway, which may lead to gastric cancer cell migration and invasion.

[0010] CBS gene: Located at 21q22, it participates in key steps in the sulfur amino acid metabolic pathway in vivo. It has multiple transcription start sites and can participate in multiple metabolic pathways, especially regulating the bioenergetic activity of cells by producing H2S. It is upregulated in a variety of cancers. Its expression is closely related to tumor angiogenesis and the increased demand of tumor cells for sulfur amino acids. It is an important factor in tumor growth and metastasis. Summary of the invention

[0011] The present invention aims at solving the above technical problems in the prior art and provides a prognostic marker for gastric adenocarcinoma.

[0012] The present invention first provides a prognostic marker for gastric adenocarcinoma, which consists of eight genes: F5, SLC5A1, PHYHD1, FNDC1, NFE2L3, SCUBE2, CBS and CTHRC1.

[0013] The present invention further provides the use of the gastric adenocarcinoma prognosis marker as a detection marker in the preparation of a clinical prognosis prediction model for gastric adenocarcinoma. Preferably, the construction of the clinical prognosis prediction model for gastric adenocarcinoma comprises the following steps: S1: Obtaining the expression data of the eight genes in the gastric adenocarcinoma prognostic markers in the samples from gastric adenocarcinoma patients used for model construction, as well as the prognostic data of gastric adenocarcinoma patients, the prognostic data being the survival status and survival time; S2: The prognosis data is used as the target variable and the random forest algorithm is used for training to obtain a trained model, which is the clinical prognosis prediction model for gastric adenocarcinoma. During the training process, each tree is constructed using different feature subsets and sample subsets; the basis for tree splitting is the information gain of each gene on the prognostic target. After training each tree, the model will integrate the prediction results of these trees and obtain the final prediction result through the average value; during the model building process, each tree is limited to split the data into 10 subsets, and the number of trees is set to 1000, so as to optimize the hyperparameters and finally obtain the model with the best performance.

[0014] The present invention further provides the use of a reagent for detecting the gene expression level of the gastric adenocarcinoma prognosis marker in the preparation of a gastric adenocarcinoma prognosis prediction product. Preferably, the gastric adenocarcinoma prognosis prediction comprises the following steps: (1) Detect the expression levels of eight genes, including F5, SLC5A1, PHYHD1, FNDC1, NFE2L3, SCUBE2, CBS, and CTHRC1, in gastric adenocarcinoma tumor tissue samples from the tested individuals; (2) The expression data of the eight genes obtained in step (1) are used as input into the trained gastric adenocarcinoma clinical prognosis prediction model; a risk score is calculated, where the higher the risk score, the worse the prognosis; The construction of the trained gastric adenocarcinoma clinical prognosis prediction model includes the following steps: S1: Obtaining the expression data of the eight genes in the gastric adenocarcinoma prognostic markers in the samples from gastric adenocarcinoma patients used for model construction, as well as the prognostic data of gastric adenocarcinoma patients, the prognostic data being the survival status and survival time; S2: The prognosis data is used as the target variable and the random forest algorithm is used for training to obtain a trained model, which is the clinical prognosis prediction model for gastric adenocarcinoma. During the training process, each tree is constructed using different feature subsets and sample subsets; the basis for tree splitting is the information gain of each gene on the prognostic target. After training each tree, the model will integrate the prediction results of these trees and obtain the final prediction result through the average value; during the model building process, each tree is limited to split the data into 10 subsets, and the number of trees is set to 1000, so as to optimize the hyperparameters and finally obtain the model with the best performance.

[0015] The present invention also provides a gastric adenocarcinoma prognosis prediction product, comprising a reagent for detecting the gene expression level of the gastric adenocarcinoma prognosis marker.

[0016] Preferably, the reagent is a reagent for detecting the expression level of each molecular marker. More preferably, the reagent is a primer for detecting the expression level of each molecular marker.

[0017] The present invention also provides a gastric adenocarcinoma prognosis evaluation model, comprising a detection data input module, an analysis module and a result output module, wherein: Detection module: used to detect samples from individuals to be tested and obtain the expression levels of eight genes, namely, F5, SLC5A1, PHYHD1, FNDC1, NFE2L3, SCUBE2, CBS and CTHRC1, in the samples; Analysis module: including a trained gastric adenocarcinoma clinical prognosis prediction model, the analysis module is used to analyze the expression of each gene detected by the detection module, and the obtained expression data is used as input to the trained gastric adenocarcinoma clinical prognosis prediction model. The construction of the trained gastric adenocarcinoma clinical prognosis prediction model includes the following steps: S1: Obtain the expression data of 8 genes in gastric adenocarcinoma prognostic markers in samples from gastric adenocarcinoma patients for model construction, as well as the prognostic data of gastric adenocarcinoma patients, the prognostic data are survival status and survival time; S2: The prognosis data is used as the target variable and the random forest algorithm is used for training to obtain a trained model, which is the clinical prognosis prediction model for gastric adenocarcinoma. During the training process, each tree is constructed using different feature subsets and sample subsets. The basis for tree splitting is the information gain of each gene on the prognostic target. After training each tree, the model will integrate the prediction results of these trees and obtain the final prediction result through the average value. During the model building process, each tree is limited to split the data into 10 subsets, and the number of trees is set to 1000, so as to optimize the hyperparameters and finally obtain the model with the best performance. Result output module: Calculate the patient risk score. The higher the patient risk score, the worse the prognosis.

[0018] The present invention uses the overall survival data of gastric adenocarcinoma patients in the TCGA and GETx2 databases, screens genes F5, SLC5A1, PHYHD1, FNDC1, NFE2L3, SCUBE2, CBS, and CTHRC1 as gastric adenocarcinoma prognostic markers, and constructs a risk assessment model based on F5, SLC5A1, PHYHD1, FNDC1, NFE2L3, SCUBE2, CBS, and CTHRC1, and uses external data sets to verify the accuracy of the model, and finds that the risk score is significantly correlated with the presence of relatively abundant stromal cells in its tumor microenvironment. The present invention helps to better predict the prognosis of gastric adenocarcinoma patients, effectively evaluate the prognostic risks of different patients, intervene early, and bring great help to guide the clinical treatment of the disease, thereby improving the 5-year survival rate of gastric adenocarcinoma patients.

[0019] The random survival forest model selected by the present invention has the following advantages: first, high prediction accuracy: it has lower prediction error rate and prediction error than the traditional Cox regression model, especially when dealing with high-dimensional data and complex interactions; second, strong variable identification ability: the model can effectively identify key variables that affect prognosis, and can screen variables by methods such as minimum depth, which helps to simplify the model and improve interpretability; third, strong adaptability: there is no need to assume that the influence of variables on the risk function is linear, and it is applicable to various types of survival data; fourth, good generalization ability: by adjusting parameters multiple times to find the optimal setting, the model can maintain high prediction performance on different data sets; in addition, the model also uses the VIMP method and the minimum depth method in combination. The former has the advantages of rapid response to demand changes and using the UML modeling method to guide design. The latter can improve the quality of local minima, without the need for over-parameterization assumptions, and can gradually approach the global optimal solution. Therefore, the model of the random survival forest method combined with the VIMP method and the minimum depth method has many advantages as mentioned above, which reflects the incomparable superiority of the model to a certain extent.

[0020] The risk assessment model of the present invention has high prediction accuracy and universal applicability, and is significantly superior to other gastric adenocarcinoma prognosis methods in the prior art, and is also significantly superior to other models constructed during the research process. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a univariate COX regression analysis diagram.

[0022] Figure 2 This is the VIMP analysis diagram for screening tumor marker genes for gastric adenocarcinoma.

[0023] Figure 3 VIMP values ​​of 8 genes screened by random survival forest.

[0024] Figure 4 Plots of prediction error rates for models with different numbers of survival trees.

[0025] Figure 5 The surv_cutpoint function is used to divide the data in the training set into high-risk group and low-risk group.

[0026] Figure 6 Box plots of scores for patients with different survival status in the training set.

[0027] Figure 7 To compare the differences in survival between the high-risk group and the low-risk group in the training set.

[0028] Figure 8 This is a scatter plot of risk in the training set.

[0029] Fig. 9 is the difference in survival between high-risk and low-risk patients in the training set.

[0030] Fig.10 ROC curve analysis showed the reliability of the prediction model in the training set in predicting prognosis.

[0031] Fig.11 is the time-dependent ROC curve of the training set.

[0032] Fig.12 To verify the difference in survival rate between high-risk and low-risk patients.

[0033] Fig.13 Time-dependent ROC curve for the validation set.

[0034] Fig.14 The figure shows the difference in risk scores between the high and low groups.

[0035] Fig.15 The figure shows the difference in tumor purity between the high group and the low group.

[0036] Fig.16 The difference results of tumor matrix between high group and low group are shown in the figure, where p=2e-08 means p=2 -8 . DETAILED DESCRIPTION

[0037] Example 1 1. Methods 1. Data collection.

[0038] The gene expression profile data of gastric adenocarcinoma patients were analyzed using the GEPIA2 online website, and differentially expressed genes in gastric adenocarcinoma were obtained as a training set. A total of 81 gastric adenocarcinoma patients treated in Zhejiang Cancer Hospital were included, and transcriptome sequencing of the tumors was performed as a validation set.

[0039] 2. Screening of prognostic markers.

[0040] Using the data in the training set, we first excluded non-coding RNAs, set the screening condition to |log2FC|>2, and screened out 714 differentially expressed genes with adjusted p value<0.05. We further screened prognosis-related genes using univariate COX regression in the training set, with the threshold set to p<0.1. The results are as follows: Figure 1 As shown, 184 prognosis-related genes were obtained for subsequent analysis. The figure screened for p<0.01 to show that hazard ratios greater than 1 indicate that the gene is associated with a poor prognosis, and less than 1 indicates that the gene is associated with a good prognosis.

[0041] We combine the VIMP method and the minimum depth method to determine the threshold. Figure 2 The VIMP analysis diagram for screening tumor marker genes for gastric adenocarcinoma. The VIMP value of each gene is calculated by the gg_vimp function. The VIMP value indicates the contribution of the gene to the prediction accuracy of the model. Genes with a VIMP value greater than 0 indicate that they have a positive impact on the model, while variables with a VIMP value less than 0 may reduce the prediction accuracy of the model. The gg_minimal_depth function is used to evaluate the importance of variables using the minimum depth method. The minimum depth indicates the first split position of a variable in the tree. The shallower the depth, the more important the variable. Eight genes with a high correlation with the prognosis of gastric adenocarcinoma were screened and used as risk markers.

[0042] 3. Model establishment.

[0043] Prepare the screened characteristic gene expression data, take the sample prognosis data (survival status and survival time) as the target variable, and use the random forest algorithm for training. During the training process, each tree is constructed using different feature subsets (randomly selected characteristic genes) and sample subsets (bootstrap sampling method). The basis for tree splitting is the information gain of each gene on the prognostic target. After training each tree, the model will integrate the prediction results of these trees and obtain the final prediction result through the average value (regression task). During the model building process, each tree is limited to split the data into 10 subsets, and the number of trees is set to 1000, so as to optimize the hyperparameters and finally obtain the model with the best performance.

[0044] 4. Risk division method.

[0045] The eight risk marker genes were substituted into the random survival forest prognostic model, the number of trees and the error rate curve were plotted, and the risk score of each sample was calculated. Figure 3 The VIMP values ​​of the eight genes screened by random survival forest show the importance of these eight genes. Figure 4 This is a model prediction error rate graph for different numbers of survival trees. As the number of survival trees increases, the prediction error rate decreases significantly. When the number of survival trees increases to a certain number, the error rate curve tends to be stable. The number of decision trees is set to 1000.

[0046] The high and low risk groups are determined according to the surv_cutpoint function. Figure 5 The surv_cutpoint function is used to divide the data in the training set into high-risk group and low-risk group. Figure 6 (where **** indicates p<0.0001) is a box plot of the scores of patients with different survival status in the training set. It can be seen that there is a difference in the overall scores between the patients who survived (0) and the patients who died (1) in the training set. Figure 7 To compare the differences in survival between the high-risk group and the low-risk group in the training set. Figure 8 The risk scatter plot for the training set shows the score distribution of the high-risk group and the low-risk group. Fig. 9 is the difference in survival between high-risk and low-risk patients in the training set.

[0047] 5. Evaluate the relationship between the risk model and overall survival in the training set and verify the accuracy of the model.

[0048] The Kaplan-Meier curve was drawn to compare the survival rate between the high-risk and low-risk groups. The ROC curve of the patients was drawn using the R package "survivalROC" and the area under the curve was calculated to verify the accuracy of the model. Fig.10 ROC curve analysis showed the reliability of the prediction model in the training set in predicting prognosis. Fig.11The figure is the time-dependent ROC curve, which shows the accuracy of the model in predicting survival rate after 2, 3, and 4 years of training set.

[0049] 6. Verify the model accuracy in the validation set.

[0050] The sample components in the validation set were divided into high-risk group and low-risk group according to the high- and low-risk boundaries confirmed in the training set. The Kaplan-Meier curve was drawn to compare the survival rate differences between the high- and low-risk groups. The ROC curve of the patients was drawn using the R package "survivalROC" and the area under the curve was calculated to further verify the accuracy of the model in the validation set. Fig.12 To verify the difference in survival rate between high-risk and low-risk patients. Fig.13 This is the Kaplan-Meier survival curve, showing the difference in survival rate between the high and low risk score groups in the validation set.

[0051] 7. Correlation analysis between risk score and other factors.

[0052] First, we determined that the risk score was significantly associated with prognosis ( Fig.14 ), and then compared the risk score with tumor purity ( Fig.15 ), the relationship between tumor stroma ( Fig.16 ).

[0053] 2. Results 1. The randomForest machine learning method was applied to the training set, and the VIMP method was combined with the minimal-depth method to determine the threshold, and the 8 prognosis-related genes in Table 1 were screened.

[0054] Table 1. Depth and VIMP values ​​of genes Gene Depth value VIMP value F5 8.380 0.002 SLC5A1 8.644 0.007 PHYHD1 8.848 0.001 FNDC1 8.877 0.003 NFE2L3 8.899 0.003 SCUBE2 8.929 0.001 CBS 8.965 0.002 CTHRC1 8.988 0.003 2. Risk scoring of training set.

[0055] The surv_cutpoint function finds the best position to split the data by traversing all possible cutoff points of the continuous variable. Its calculation process is based on the log-rank test statistic, and the formula is: , where Z: risk score; Q1: number of observed events (e.g., deaths) in the first group; E1: expected number of events in group 1 (based on estimated survival rates in all groups); V: The variance of the test.

[0056] The surv_cutpoint function was used to calculate the threshold of 20.63 as the risk score. The expression levels of the eight genes F5, SLC5A1, PHYHD1, FNDC1, NFE2L3, SCUBE2, CBS and CTHRC1 were measured by the kit and brought into the model to calculate the risk score of each patient. The risk score of gastric adenocarcinoma patients in the training set was divided into high-risk group and low-risk group with 20.63 as the threshold. Kaplan-Meier curve analysis found that the high-risk group had a lower survival rate than the low-risk group ( Fig. 9 , p<0.001). The overall ROC curve showed an AUC value of 0.86 ( Fig.10 ), the areas under the ROC curve for predicting 2-year, 3-year, and 4-year survival rates were 0.789, 0.783, and 0.828, respectively ( Fig.11 ), further indicating that the model has good predictive ability. At the same time, we also verified the risk score and survival information in the training set and found that the survival rate of the high-risk group was lower.

[0057] 3. Validate the risk model in the validation set.

[0058] To verify the universality of the risk model, the validation set was divided into a high-risk group and a low-risk group using the threshold determined in the training set. Kaplan-Meier curve analysis found that the survival rate of the high-risk group was lower than that of the low-risk group ( Fig.12 , p = 0.032), which is consistent with the findings in the training set. The areas under the ROC curve for predicting 2-year, 3-year, and 4-year survival rates were 0.757, 0.788, and 0.776, respectively ( Fig.13 ), indicating that the model also has good predictive ability in the validation set.

[0059] 4. The relationship between risk score and tumor purity and tumor stroma.

[0060] The tumor purity and tumor matrix score of each sample were calculated by the ESTIMATE algorithm in R language, and it was found that the tumor purity and tumor matrix score of the high-risk group in the training set were lower ( Fig.15 , Fig.16 ).

Claims

1. A prognostic marker for gastric adenocarcinoma, characterized in that: It consists of eight genes: F5, SLC5A1, PHYHD1, FNDC1, NFE2L3, SCUBE2, CBS and CTHRC1.

2. Use of the gastric adenocarcinoma prognosis marker according to claim 1 as a detection marker in the preparation of a clinical prognosis prediction model for gastric adenocarcinoma.

3. The use according to claim 2, characterized in that: The construction of the clinical prognosis prediction model for gastric adenocarcinoma includes the following steps: S1: Obtaining the expression data of the eight genes in the gastric adenocarcinoma prognostic markers in the samples from gastric adenocarcinoma patients used for model construction, as well as the prognostic data of gastric adenocarcinoma patients, the prognostic data being the survival status and survival time; S2: The prognosis data is used as the target variable and the random forest algorithm is used for training to obtain a trained model, which is the clinical prognosis prediction model for gastric adenocarcinoma. During the training process, each tree is constructed using different feature subsets and sample subsets; the basis for tree splitting is the information gain of each gene on the prognostic target. After training each tree, the model will integrate the prediction results of these trees and obtain the final prediction result through the average value; during the model building process, each tree is limited to split the data into 10 subsets, and the number of trees is set to 1000, so as to optimize the hyperparameters and finally obtain the model with the best performance.

4. Use of a reagent for detecting the gene expression level of the gastric adenocarcinoma prognosis marker according to claim 1 in the preparation of a gastric adenocarcinoma prognosis prediction product.

5. The use according to claim 4, characterized in that: Prognosis prediction for gastric adenocarcinoma involves the following steps: (1) Detect the expression levels of eight genes, including F5, SLC5A1, PHYHD1, FNDC1, NFE2L3, SCUBE2, CBS, and CTHRC1, in gastric adenocarcinoma tumor tissue samples from the tested individuals; (2) The expression data of the eight genes obtained in step (1) are used as input into the trained gastric adenocarcinoma clinical prognosis prediction model; a risk score is calculated, where the higher the risk score, the worse the prognosis; The construction of the trained gastric adenocarcinoma clinical prognosis prediction model includes the following steps: S1: Obtaining the expression data of the eight genes in the gastric adenocarcinoma prognostic markers in the samples from gastric adenocarcinoma patients used for model construction, as well as the prognostic data of gastric adenocarcinoma patients, the prognostic data being the survival status and survival time; S2: The prognosis data is used as the target variable and the random forest algorithm is used for training to obtain a trained model, which is the clinical prognosis prediction model for gastric adenocarcinoma. During the training process, each tree is constructed using different feature subsets and sample subsets; the basis for tree splitting is the information gain of each gene on the prognostic target. After training each tree, the model will integrate the prediction results of these trees and obtain the final prediction result through the average value; during the model building process, each tree is limited to split the data into 10 subsets, and the number of trees is set to 1000, so as to optimize the hyperparameters and finally obtain the model with the best performance.

6. Gastric adenocarcinoma prognosis prediction product, characterized in that: The invention comprises a reagent for detecting the gene expression level of the gastric adenocarcinoma prognosis marker according to claim 1.

7. The gastric adenocarcinoma prognosis prediction product according to claim 6, characterized in that: The reagent is a reagent used to detect the expression level of each molecular marker.

8. The gastric adenocarcinoma prognosis prediction product according to claim 7, characterized in that: The reagents are primers used to detect the expression levels of various molecular markers.

9. A clinical prognosis prediction model for gastric adenocarcinoma, characterized in that: It includes a detection data input module, an analysis module and a result output module, wherein: Detection module: used to detect samples from individuals to be tested and obtain the expression levels of eight genes, namely, F5, SLC5A1, PHYHD1, FNDC1, NFE2L3, SCUBE2, CBS and CTHRC1, in the samples; Analysis module: including a trained gastric adenocarcinoma clinical prognosis prediction model, the analysis module is used to analyze the expression of each gene detected by the detection module, and the obtained expression data is used as input to the trained gastric adenocarcinoma clinical prognosis prediction model. The construction of the trained gastric adenocarcinoma clinical prognosis prediction model includes the following steps: S1: Obtain the expression data of 8 genes in gastric adenocarcinoma prognostic markers in samples from gastric adenocarcinoma patients for model construction, as well as the prognostic data of gastric adenocarcinoma patients, the prognostic data are survival status and survival time; S2: The prognosis data is used as the target variable and the random forest algorithm is used for training to obtain a trained model, which is the clinical prognosis prediction model for gastric adenocarcinoma. During the training process, each tree is constructed using different feature subsets and sample subsets. The basis for tree splitting is the information gain of each gene on the prognostic target. After training each tree, the model will integrate the prediction results of these trees and obtain the final prediction result through the average value. During the model building process, each tree is limited to split the data into 10 subsets, and the number of trees is set to 1000, so as to optimize the hyperparameters and finally obtain the model with the best performance. Result output module: Calculate the patient risk score. The higher the patient risk score, the worse the prognosis.

Images