| Background: Current preoperative staging for lymph nodal status remains inaccurate.The purpose of this study was to build an ANN model to predict pathologic nodal involvement in clinical stage I-II ESCC patients and then validated the performance of the model.In subsequent research,we used bioinformatics technology based on public database data to find targets related to the aggressiveness of esophageal squamous cell carcinoma.Methods: A total of 523 patients(training set: 350;test set: 173)with perioperative information were evaluated.An artificial neural network(ANN)model was established for predicting pathologic nodal positive patients in the training set,which was validated in the test set.A receiver operating characteristic curve was also created to illustrate the performance of the predictive model.Survival curves were plotted both in the training set and test set.We analyzed the mi RNA expression profile analysis results of primary tumor tissues in two large public databases(Onco Mi R database,TCGA)to identify mi RNAs associated with lymph node metastasis in patients with esophageal squamous cell carcinoma.We compared the two sets of samples(pathologically node-negative patients and pathological lymph node-positive patients)in each data set to identify differentially expressed mi RNAs(DEM).We use Jonckheere’s Trend Test,with-log10(pvalue)>2.0 and |Rank Correlation| >0.2 as the initial criteria for screening mi RNAs.Only mi RNA signals that are differentially expressed in the two data sets at the same time are considered to be important differentially expressed mi RNA signals.Use the mi RPath3.0 version in Diana tools to predict mi RNA-related target genes and analyze signal pathway enrichment to clarify the biological functions of potential mi RNA signals.Mi Rwalk2.0 was used to screen potential target genes.We further used TCGAportal to screen out genes related to the prognosis of patients with esophageal squamous cell carcinoma.Results: Of the enrolled 523 patients with ESCC,41.3% of the patients were confirmed pathologic nodal positive(216/523).The ANN staging system identified the tumour invasion depth,tumour length,dysphagia,tumour differentiation and LVI as predictors for pathologic lymph node metastases.The C-index for the ANN model verified in the test set was 0.852,which demonstrated that the ANN model had a good predictive performance.According to the prediction results of the artificial neural network,all patients were divided into high-risk lesions group and low-risk lesions group.The 3-year OS rate of patients with high-risk lesions was worse than that of patients with low-risk lesions both in clinical stage I(training set: 24.4% vs 74.9%,P< 0.01;test set: 24.0% vs 75.0%,P< 0.05)and clinical stage II(training set: 22.4% vs 50.4%,P< 0.01;test set: 23.4% vs 61.8%,P< 0.01).Eleven mi RNAs were identified as potential mi RNAs.Respectively: 7 positively related mi RNAs: hsa-mir-192,hsa-mir-556,hsa-mir-552,hsa-mir-935,hsa-mir-375,hsa-mir-1266,hsa-mir-592.4 negatively related mi RNAs: hsa-mir-193 b,hsa-mir-205,hsa-mir-1910,hsa-mir-23 a.According to the prediction results of mi Rwalk,a total of 96 genes were jointly identified as target genes in the Target Scan,mi RDB,and mi RTarbase databases.Subsequently,we drew a Kaplan-Meier map for the above target genes based on the TCGA survival data,and finally screened out 6 key genes related to the prognosis of patients with esophageal squamous cell carcinoma: ALKBH5,AHDC1,TMEM170 A,PHF8,TRMT2 A,ZC3H7B.Conclusions: The ANN model presented good performance for predicting pathologic lymph node metastasis and added indicators not included in current staging criteria and might help improve the staging strategies.Using bioinformatics technology to identify potential targets for predicting lymph node metastasis can improve the accuracy of lymph node detection in patients with esophageal squamous cell carcinoma,thereby helping clinicians formulate more reasonable and appropriate treatment plans and improve the level of treatment. |
PDF Full Text Request