山西医科大学学报2025，Vol.56Issue(12)：1319-1327,9.DOI:10.13753/j.issn.1007-6611.2025.12.001
中医药干预胃癌术后恢复:逻辑回归模型构建
Traditional Chinese medicine intervention in postoperative recovery of gastric cancer:construction of a Logistic regression model
摘要
Abstract
Objective To develop a prediction model for postoperative recovery of gastric cancer based on a Logistic regression algo-rithm,explore the effect of traditional Chinese medicine(TCM)intervention on postoperative recurrence,and identify the risk factors for gastric cancer prognosis.Methods Clinical data were retrospectively collected from 408 gastric cancer patients in the First Hospital of Shanxi Medical University from January 1,2019 to March 12,2025.The data were randomly divided into a training set(285 cases)and a test set(123 cases).Statistical analysis was performed using SPSS 26,and the data modeling was conducted using RStudio 2024.12.1.Variables were screened using univariate and multivariate Logistic regression and a forest plot was generated.Significant variables(P<0.05)were incorporated into the final model,and a nomogram was constructed.Likelihood ratio test,Wald test,and Hosmer-Lemeshow goodness-of-fit test were used to evaluate the model.Bootstrap resampling was performed 1 000 times for validation,and the C-index was calculated along with a calibration plot.Finally,the ROC curve and AUC values were plotted using an external validation set,and the clinical utility was assessed using decision curve analysis(DCA).The effect of TCM intervention was further analyzed using Kaplan-Meier(K-M)curves.Results Multivariate Logistic regression analysis incorporated seven variables:TCM intervention,anemia,abnormal liver function,dyslipidemia,body mass index,serum albumin,and abnormal stool.TCM intervention was the most significant risk factor(P<0.05).The Hosmer-Lemeshow test indicated an acceptable model fit(P>0.05).The AUC values for the external validation set were 0.737 and 0.678 for the training and testing sets,respectively.Decision curve analysis showed high clinical net benefit within a threshold probability range of 0.05-0.5.K-M analysis demonstrated that TCM intervention significantly improved recurrence-free survival in gastric cancer patients.Conclusion The predictive model for postoperative recurrence of gastric cancer demonstrates good performance,and traditional Chinese medicine intervention is identified as an independent protective factor associated with a reduced risk of postoperative recurrence.关键词
胃癌/中医药干预/预测模型/危险因素/列线图/Bootstrap法/决策曲线分析Key words
gastric cancer/traditional Chinese medicine intervention/predictive model/risk factors/nomogram/Bootstrap method/decision curve analysis分类
医药卫生引用本文复制引用
WU Ke,GUO Hongbo..中医药干预胃癌术后恢复:逻辑回归模型构建[J].山西医科大学学报,2025,56(12):1319-1327,9.基金项目
山西省社会科学界联合会科研专项课题(DJKZXKT2023043) （DJKZXKT2023043）
山西省中医药管理局课题(2022ZYYC073) （2022ZYYC073）
