Analysis And Discussion On Prognosis Related To Upper Tract Urothelial Carcinoma

Objective: This study aims to explore the factors affecting the prognosis of patients with upper tract urothelial carcinoma(UTUC)from clinical and genetic perspectives.By constructing a clinical prediction model,we predict the prognosis of UTUC patients at 1,3,5,and 10 years,while filling the gap in data on UTUC patients in tropical regions of China.In addition,by screening ferroptosis genes,we establish a UTUC ferroptosis gene regulatory network to provide direction and theoretical basis for subsequent gene-related functional studies.Methods: 1.Clinical data of UTUC patients from the Department of Urology of Hainan Affiliated Hospital of Hainan Medical University for nearly 10 years were collected,and the clinical data and survival data of 149 patients were retrospectively analyzed.Machine learning methods were used to build random survival forests,and clinical prediction models were established by combining the results of univariate and multivariate COX analysis.In this study,the consistency parameter C-index,the integrated discrimination improvement(IDI),and the net reclassification improvement(NRI)were used to evaluate the model.Finally,the model results are presented in Nomogram.2.The gene expression profiles(GEPs)of UTUC patients were extracted from the GEO public database,and the ferroptosis genes extracted from the ferroptosis genes database were cross-screened to screen out the UTUC ferroptosis genes.Through GO and KEGG pathway analysis,the functional enrichment results of UTUC ferroptosis genes were explored.The PPI regulatory network of the UTUC ferroptosis genes was mapped using the String online database.Finally,combined with prognosis-related genes,a risk prediction model was constructed.Results: 1.We collect 187 patients from the Department of Urology of Hainan Affiliated Hospital of Hainan Medical University and 149 cases were finally included in the study after removing the information of lost visits.Among them,the male to female ratio was about 3:2,the age of onset was 38-88 years(median age 67 years),and the survival time was about 21-4010 days(median days 706 days).Through machine learning to build a random survival forest,it was found that tumor stage had the highest importance score in the model,while the scores of variables such as T stage,whether surgery or not,surgical method and tumor size gradually decreased.Univariate COX analysis showed that the age of onset,whether surgery or not,surgical method,tumor size,pathological diagnosis,tumor stage,T,N,tumor grade,NLR,PLR,LMR,SII,SIRI were all risk factors affecting the prognosis of UTUC(P<0.05).Multivariate COX analysis showed that age of onset,surgery or not,and tumor stage were independent risk factors affecting the prognosis of patients with UTUC.Then,we use these factors as predictors to build a clinical predictive models.Finally,the calibration curve found that the actual value was similar to the predicted value.The C-index of the new model is 0.845,the result of the IDI is 0.284,and the result of the NRI is 0.800.Compared with the traditional TNM staging model(C =0.705),it can be considered that the new model has better prediction effect and more accurate prediction results.2.A total of 33 UTUC gene expression profiles were screened from the GEO public database.A total of 764 ferroptosis genes were downloaded from the three major genetic databases of Gene Cards,NCBI and Ferr Db.Set the adj.p value<0.05 and | log2 FC |>1.0 as the critical values for screening differential genes,and 952 differential genes were screened.Finally,50 ferroptosis differential genes were obtained.GO enrichment analysis showed that these genes were significantly involved in the regulation of amino acid transmembrane input,DNA-binding transcription factor activity,and the structural components of cytoskeleton and extracellular matrix,which were related to the activities of various transmembrane transport proteins.KEGG pathway analysis showed that these genes were mainly enriched in ferroptosis,atherosclerosis,cancer pathway,cancer central carbon metabolism,and hippocampus signaling and other related pathways.Then,univariate COX regression analysis was used to screen the prognostic genes of UTUC,and combined with the ferroptosis differential genes,a total of five UTUC ferroptosis prognostic differential genes(RGS4,KRT16,SCARA5,LRFN5,SLC7A5)were screened.The risk prediction model was constructed based on these 5 ferroptosis genes.Conclusions: 1.Through statistical analysis of the clinical information of 149 patients with UTUC,this study uses the methods of machine learning and univariate COX regression analysis to build a clinical prediction model that predicted the 1-,3-,5-and 10-year survival rates of UTUC patients.Through the evaluation of the model,it is found that the prediction accuracy of the new model is higher,which can provide a certain reference value for clinicians to guide the prognosis of patients.2.In this study,5 ferroptosis genes were screened,and the risk prediction model was constructed after removing the overfitting genes using Lasso.According to the risk curve and survival curve,it can be found that there is a significant difference in the prognosis between the two groups of patients with high and low risk.