Establishment Of Risk Prediction Model For Vascular Invasion Of Early Gastric Cancer

Objective:To explore the risk factors of vascular invasion in patients with early gastric cancer,to establish a risk prediction model to evaluate the risk of vascular invasion in patients with early gastric cancer,and to analyze whether patients with early gastric cancer treated by endoscopy need further radical surgery or strengthen postoperative intervention to improve the prognosis of patients.Methods:The clinical data of 415 patients with early gastric cancer who underwent radical gastrectomy in the Department of Gastrointestinal surgery of the second affiliated Hospital of Nanchang University from January 2017 to January 2022 were collected to analyze the correlation between vascular invasion and preoperative inflammatory index,preoperative gastroscopic index,tumor marker and postoperative clinicopatho-logy.ROC curve was used to determine the best cutoff value(cut-off value)of various continuous variables.The significant variables of univariate analysis were included in multivariate Logistic regression analysis,and the stepwise backward regression method was selected to analyze the independent risk factors of vascular invasion in early gastric cancer.A line chart model was constructed and verified internally by Bootstrap method.The consistency index was calculated to evaluate the accuracy and clinical practicability of the model.Results:The probability of positive vascular invasion in patients with T1 gastric cancer was 10.36%(43/415).Univariate analysis showed that sex,age,hypertension,diabetes,BMI,tumor location,AFP,CEA,CA125,LMR,FPR and PNI were not associated with vascular invasion.The depth of tumor invasion(T),lymph node metastasis(N),nerve invasion,tissue classification,maximum tumor diameter,tumor type,gastroscopic tumor shape,CA199,NLR,SII,PLR,FAR and SIRI were correlated with vascular invasion.Logistic multivariate regression analysis showed that PLR>138.27,nerve invasion,tissue classification,T and N were independent risk factors for vascular invasion of early gastric cancer.Using the above independent risk factors to construct a line chart model,the area under the ROC curve of the model was 0.900,and the sensitivity and specificity were 92.5% and 75.6%,respectively.In this study,the consistency index of the line chart prediction model is 0.900,and the calibration curve shows that the results of predicting vascular invasion of early gastric cancer are highly consistent with the actual observation,which shows that the model has strong clinical practicability.Conclusion:The results of this study showed that PLR>138.27,nerve invasion,tissue classification,T and N were independent risk factors of vascular invasion in early gastric cancer.Based on this,a model for predicting the risk of vascular invasion in patients with early gastric cancer is established,which can provide clinical reference for individualized treatment of early gastric cancer.