Construction of a Nomogram Prediction Model for Late-onset Fetal Growth Restriction Based on Lasso-logistic Regression
Objective:To analyze the predictive value of a nomogram model based on ultrasound Doppler blood flow parameters and placental thickness for late-onset fetal growth restriction(FGR).Methods:This prospective study included 330 singleton pregnant women who underwent antenatal examination and delivered in our hospital from October 2022 to May 2024,with a gestational age range of 32-40 weeks.Among them,143 cases were in the late-onset FGR group(observation group),and 187 cases were in the non-FGR group(control group).Basic information of the pregnant women in both groups was collected,along with blood flow spectral parameters of the fetal ductus venosus,umbilical artery,and middle cerebral artery,and placental thickness was measured.Differences in various parameters between the two groups were compared using univariate analysis.Further,Lasso-Logistic regression analysis was performed to identify independent predictors of FGR occurrence,based on which a nomogram prediction model was constructed.To evaluate the predictive value of the model.Receiver operating characteristic(ROC)curves,calibration curves,and decision curves were plotted.Results:Univariate analysis revealed that the umbilical artery systolic-to-diastolic velocity ratio(UA-S/D),umbilical artery resistance index(UA-RI),and umbilical artery pulsatility index(UA-PI)values were significantly higher in the observation group compared to the control group,while the middle cerebral artery systolic-to-diastolic velocity ratio(MCA-S/D),middle cerebral artery resistance index(MCA-RI),middle cerebral artery pulsatility index(MCA-P1),ductus venosus acceleration(DV-a),ductus venosus peak velocity(DV-S),and cerebroplacental ratio(CPR)values were significantly lower,accompanied by thinner placental thickness.These differences were statistically significant(P＜0.05).Lasso-Logistic regression analysis further identified UA-S/D,MCA-S/D,MCA-RI,CPR,and placental thickness as the most valuable independent predictors of late-onset FGR.Based on these factors,a nomogram prediction model for late-onset FGR was constructed,with an area under the curve(AUC)of 0.992,indicating high predictive accuracy.The calibration curve demonstrated good agreement between the predicted probabilities and actual probabilities.Additionally,clinical decision curve analysis showed that using this model could yield positive net clinical benefits.Conclusion:This study demonstrates that the multi-factor prediction nomogram model based on Lasso-Logistic regression can effectively identify the risk of late-onset FGR,providing a visual reference for early clinical intervention and contributing to the prevention of adverse pregnancy outcomes.
fetal growth restrictionDoppler ultrasoundnomogramprediction model
万方数据
维普
