Construction Of A Risk Score Prediction Model For Gastric Cancer Based On Genes Related To The TGF-β Pathway And Analysis Of Clinical Relevance

Objective:According to the latest statistics reported by the International Agency for Research on Cancer(IARC),stomach cancer remains one of the top five cancers in the world and poses a great threat to human health.There is a large body of research on the mechanisms of the transforming growth factor beta(TGF-β)pathway,with findings suggesting that the TGF-β pathway plays a key role in a variety of malignancies.However,further research is needed regarding the potential role of the TGF-β pathway in the prognosis and characterization of the tumour microenvironment in gastric cancer.We therefore used the genetic data of gastric cancer patients from The Tumour Genome Atlas public database(TCGA)to construct a predictive model of gastric cancer risk score based on genes related to the transforming growth factor β pathway,and used this model to correlate the tumour microenvironment of gastric cancer.Methods:Firstly,gene expression data from the Tumour Genome Atlas public database(TCGA)for gastric cancer and normal paracancerous tissues as well as clinical data were downloaded,while TGF-βpathway-related genes that have been investigated and identified so far were summarised from several databases and relevant literature.Differentially expressed genes(DEGs)between cancerous and normal paracancerous tissues were extracted using R language software,and then the DEGs were combined with TGF-β pathway-related genes to obtain differentially expressed TGF-β pathway-related genes,followed by analysis of the correlation between differentially expressed TGF-β pathway-related genes and GC prognosis using the Gene Expression Profiling Interactive Analysis(GEPIA)online platform.The risk score prediction model was then constructed based on the prognosis-significantly associated genes obtained by screening the prognosis-significantly associated genes using univariate and multifactorial COX regression analysis,and the risk coefficients were then calculated using the constructed model,and the median was used as the cut-off value to classify the gastric cancer samples into two groups: low risk and high risk.The correlation between prognostic index and clinical prognosis was then verified by chi-square test,single-factor COX regression analysis and multi-factor COX regression analysis of relevant clinical indicators,to carry out validation of the risk score prediction model as an independent prognostic factor for patients with gastric cancer,and to assess the validity of the risk score prediction model.We then analysed the role of differentially expressed genes in the tumour microenvironment between high and low risk groups and the functional enrichment results based on the high and low risk groups distinguished by this model,and performed functional analyses related to ESTIMATE scores,immune cell infiltration,etc.Results:1.GC gene expression data was downloaded from the TCGA database and 46 differentially expressed TGF-β pathway-related genes were obtained after screening.Validation by the GEPIA online analysis platform revealed that a total of 17 genes were associated with the prognosis of GC patients,and by univariate and multifactorial COX analysis,only COL1A2,GDF7,INHBB,among the 17 genes were found to be SERPINE1 were significantly associated(p<0.05).2.A risk score prediction model was developed based on COL1A2,GDF7,INHBB and SERPINE1,with the prognosis index(PI)formula: Prognosis index(PI,Prognosis index)=(0.415 x expression of COL1A2)+(0.432 x expression of GDF7)+(0.422 x expression of INHBB)+(0.628 ×expression of SERPINE1).After calculating the median value of the prognostic index PI,a total of 159 patients were included in the high-risk group and 158 patients in the low-risk group.k-M analysis showed that patients in the high-risk group had a poorer prognosis.The results of the chi-square test showed that the prognostic index PI was statistically different from the survival status and T of GC patients(p<0.05).One-way COX regression analysis showed that age,Stage and prognostic index were all associated with OS.Multi-factor COX regression results showed that age,prognostic index and M stage were all independent influences on the prognosis of GC patients.3.The results of immune cell infiltration analysis of differentially expressed genes in the high and low risk groups showed that dendritic cells,macrophages and neutrophils were more infiltrated in the high risk group than in the low risk group.The ESTIMATE score analysis showed a significantly higher abundance of stromal cell infiltration in the high-risk group than in the low-risk group.4.By searching the literature and guidelines related to the treatment of gastric cancer,the common chemotherapeutic drugs used in the treatment of gastric cancer were screened and the results of chemotherapy drug sensitivity analysis were conducted,which showed that the low-risk group had higher drug sensitivity than the high-risk group.Conclusion:We constructed a predictive model based on four TGF-β pathway-related gene risk scores,COL1A2,GDF7,INHBB and SERPINE1,from the public database TCGA for gastric cancer gene expression data,which can be used as an independent influencing factor to evaluate the prognosis of GC patients.In addition,an epidemic cell infiltration analysis was performed.A higher abundance of stromal cell infiltration was found in the high-risk group than in the low-risk group.As well as chemotherapeutic drug sensitivity analysis,the low-risk group was more sensitive to chemotherapeutic drugs than the high-risk group,providing some direction for the development of individualized protocols and targeted treatment for gastric cancer patients in the future.