Comparison of the Performance of Decision Tree,Random Forest Model and Regression Nomogram Model in Predicting Ad-verse Perinatal Outcomes in Pregnant Women with Fetal Growth Restriction
YUAN Yuan
LIU Yi
XU Yetao
SUN Lizhou
PAN Yi
Abstract：Objective To establish a predictive model that includes multiple risk factors for fetal growth restriction(FGR),explore the related risk factors for FGR and provide reference for prenatal counseling and postpartum care.Meth-ods A retrospective analysis was conducted on the clinical information of 359 FGR pregnant women who gave birth in the obstetrics department of the author's hospital from January to December 2023,according to the perinatal outcome,the pregnant women were divided into poor prognosis group(n=151)and good prognosis group(n=208).Three methods,including decision tree,random forest model and regression nomogram model,were used to construct the prediction model of adverse perinatal outcomes in FGR pregnant women.The performance of each method was analyzed by evaluating the area under the curve(AUC)of the receiver operating characteristic(ROC)curve,accuracy,precision and other parame-ters.Results Compared with the good prognosis group,the maternal age,gravidity,body weight,body mass index(BMI),cesarean section rate,neonatal intensive care unit(NICU)transfer rate,intrapartum hemorrhage,postpartum hemorrhage within 24 h and abnormal fetal blood flow signal rate of the pregnant women in poor prognosis group all in-creased(all P＜0.05),gestational age at admission,gestational age at diagnosis,gestational age at delivery,placental weight,neonatal birth weight,Apgar scores(1 min and 5 min)all decreased(all P＜0.05).Among the three models,random forest model had the highest performance in predicting FGR,with an AUC of 0.914(95％CI:0.854-0.975).Conclusion Random forest model can accurately predict the occurrence of FGR.Strengthening the monitoring and treatment of early warning factors can help to predict the adverse outcome of FGR early,and take intervention measures in advance to improve the prognosis of mother and child.
Keywords：Fetal growth restrictionRisk predictionAdverse perinatal outcomeRandom forest modelDecision treeRegression nomogram model
Publication Date：2025-11-28
Online Publishing Date：2026-01-14(First online date of this platform, not the publication date of the document)
Pages：8( 947-953,1008 )