Preoperative Prediction Of TNM Stages Based On Conventional Clinical Indicators In Gastric Cancer

Objectives The aims of this study were to analyze the preoperative prediction value of clinical indicators alone and in combination for the TNM staging of gastric cancer and provide the basis for the clinical evaluation,diagnosis and treatment of gastric cancer.Methods 1.Collection of clinical data: The preoperative data of newly diagnosed gastric cancer patients hospitalized in our hospital for operation from January 2015 to December 2020 were retrospectively collected,including gastroscopy and biopsy pathology data,CT examination data,laboratory blood test data,and surgical and postoperative pathological data.2.Determination of TNM stage of gastric cancer patients: According to the surgical and postoperative pathological data,T stage,N stage,M stage and TNM stage of each patient were determined.3.Analysis of preoperative clinical characteristics: Clinical indicators were statistically analyzed in gastric cancer patients.The measurement data were represented by Mean ± SD and compared by t test or variance analysis between groups.Counting data were expressed by frequency and percentage and compared by Chi-square test between groups.4.Univariate analysis of the predictive value of preoperative indicators for TNM staging of gastric cancer: According to T1 ~ 4 stage,N0 ~ 3 stage,M0 ~ 1 stage,TNM I ~ IV stage stratification statistics of each index,t test,variance analysis or Chi-square test to compare the statistical significance of the difference between each period.The predictive value evaluation methods of each index for each stage of gastric cancer are as follows: the second classification index is evaluated by the area under the receiver operating characteristic(ROC)curve(AUROC);the multi-classification indexes were evaluated by the classification accuracy of discriminant analysis.5.Multivariate analysis of the predictive value of preoperative indicators for TNM staging of gastric cancer: T staging,N staging,M staging and TNM staging were used as dependent variables,and clinical indicators were used as independent variables.Multivariate binary logistic regression analysis or multivariate multi-classification logistic regression analysis were performed.The value of multi-index combination for TNM staging of gastric cancer was evaluated by AUROC of the prediction probability of the regression model.Results: 1.Clinical data: A total of 441 patients with gastric cancer were included in the analysis,including 293 males and 148 females,with an average age of 60.9 ± 10.7 years.T1 ~ 4 were 84 cases,41 cases,93 cases and 223 cases,N0 ~ N3 were 161 cases,62 cases,68 cases and 150 cases,M0 ~ M1 were 413 cases and 28 cases.TNM stages I,II,III and IV were 102 cases,78 cases,233 cases and 28 cases,respectively.2.Preoperative prediction of T staging of gastric cancer: Some individual clinical indicators have certain predictive value,among which the prediction accuracy of lymph node enlargement under abdominal CT,tumor size under gastroscopy,Borrmann classification,Laurence classification and total cholesterol is more than 40 %.The total accuracy of multi-index combined prediction of T staging of gastric cancer was about 70 %.The indicators entering the regression equation were alpha-fetoprotein,hematocrit,endoscopic gastric cancer staging,endoscopic tumor size and anti-thrombin III activity.3.Preoperative prediction of N stage of gastric cancer: Some individual clinical indicators also have certain predictive value,among which the accuracy of preoperative prediction of lymph node enlargement and gastric wall state,obstruction,tumor size under gastroscopy,Borrmann classification,Lauren classification and cholinesterase under abdominal CT is more than 40 %.The accuracy of multi-index combination in predicting N stage of gastric cancer was about 60 %.The indexes entering the regression equation were thrombin time,platelet distribution width,coefficient of variation of red blood cell distribution width,standard deviation of red blood cell distribution width,tumor location under endoscopy,family history of tumor,CT showed lymph node enlargement,gastric cancer stage under endoscopy,total bile acid and hemoglobin.4.Preoperative prediction of lymph node metastasis in gastric cancer: endoscopic tumor size has moderate predictive value(AUROC = 0.774).The AUROC of multi-index joint prediction of lymph node metastasis was 0.856,and the accuracy was about 80%.The independent predictors of lymph node metastasis risk included lymph node enlargement under preoperative CT,family history of gastric cancer,tumor size under endoscopy,advanced gastric cancer,and total bilirubin,homocysteine,mean red blood cell volume and D-dimer level.5.Preoperative prediction of distant metastasis of gastric cancer: endoscopic tumor size has a certain predictive value(AUROC = 0.698).Multi-index combined prediction of distant metastasis AUROC was 0.921,independent predictors were : age,CT found distant metastasis,hypertension,upper gastrointestinal bleeding,tumor size,ABO blood group,indirect bilirubin,uric acid,high density lipoprotein,basophil ratio,neutrophils,mean platelet volume,CA125.6.Preoperative prediction of early and advanced gastric cancer: patients with gastric cancer were divided into early stage(I + II)and advanced stage(III + IV)according to TNM staging.It is of great value to distinguish the two in individual indicators by endoscopic tumor size,with AUROC reaching 0.781.The total accuracy of multi-index combination for preoperative prediction of early and advanced TNM stage of gastric cancer was 80.0 %.The indexes entering the multi-classification Logistic regression equation were as follows: CT showed gastric wall changes,endoscopic tumor size,endoscopic tumor staging,total bile acid,homocysteine,red blood cell count,eosinophil count and mean red blood cell volume.7.Preoperative Prediction of TNM staging of gastric cancer: preoperative abdominal CT lymph node status,endoscopic tumor size accuracy of more than 40 %.The accuracy of multi-index combined prediction of TNM staging was 82.8 %.The indexes entering the multi-classification Logistic regression equation are CA153,CA125,prothrombin activity,reticulocyte count,eosinophil count,neutrophil count,eosinophil proportion,erythrocyte count,white blood cell count,average red blood cell volume,platelet hematocrit,total bile acid,serum calcium,high-density lipoprotein cholesterol,uric acid,indirect bilirubin,globulin,alkaline phosphatase,ABO blood group,biopsy histological type,CT showing distant metastasis,CT showing lymph node enlargement,pyloric obstruction,endoscopic tumor staging,and endoscopic tumor size.Conclusions 1.Some clinical indexes have certain value in preoperative prediction of T stage,N stage,M stage and TNM stage of gastric cancer,but the predictive value of single index is generally limited.2.The combination of indicators with independent predictive value screened from single clinical indicators has moderate value for preoperative prediction of T stage and N stage of gastric cancer,and has good value for preoperative prediction of M stage and TNM stage.3.Clinical indicators combined have good value in preoperative prediction of lymph node metastasis and distant metastasis in patients with gastric cancer.