Constructing LASSO regression model for early prediction of fetal growth restriction based on assisted reproductive pregnancy umbilical cord blood flow parameters and placental function indicators
CSTR:
Author:
Affiliation:

Clc Number:

R714.5;R714.8

Fund Project:

  • Article
  • |
  • Figures
  • |
  • Metrics
  • |
  • Reference
  • |
  • Related
  • |
  • Cited by
  • |
  • Materials
    Abstract:

    Objective:To investigate the early predictive value of combining umbilical cord blood flow dynamics parameters with placental function indicators in the second trimester of pregnancy for fetal growth restriction(FGR)in assisted reproduc-tive technology(ART)pregnancies,and to develop and validate a predictive model.Methods:A total of 317 singleton pregnant women who successfully achieved ART-assisted pregnancies and maintained ongoing pregnancies were enrolled and followed up until delivery.Based on the occurrence of FGR,the participants were categorized into the FGR group(n=41)and the non-FGR group(n=276).At 20~24 weeks of gestation,baseline characteristics,umbilical artery hemodynamic parameters[peak systol-ic velocity(S),end-diastolic velocity(D),S/D ratio,resistance index(RI),pulsatility index(PI),peak systolic velocity(Vs),end-diastolic velocity(Vd)]and placental function indicators[24-hour urinary estriol,serum human placental lactogen(hPL),fetal movement counts within 2 hours]were systematically collected.Independent predictors were screened using LASSO-Lo-gistic regression to construct a combined prediction model.The model's discriminative ability was evaluated using the receiver operating characteristic(ROC)curve.Internal validation assessing stability was performed using the Bootstrap method.Clinical decision curve analysis(DCA)was employed to evaluate clinical utility.Results:Through LASSO-logistic regression analysis,the optimal model was achieved with λ=0.074.The findings indicated that elevated S/D ratios,increased Resistance Index(RI),higher Pulsatility Index(PI),reduced serum hPL levels,low 24-hour urinary estriol levels,and fewer than 6 fetal move-ments within two hours were significant risk factors influencing FGR following successful assisted reproductive pregnancies(P<0.05).The nomogram model was internally validated using the Bootstrap method,with the area under the ROC curve for the LASSO-Logistic prediction model being 0.871.Furthermore,the Hosmer-Lemeshow test demonstrated that the mean abso-lute error between the actual values and the calibration curve was 0.03(χ2=0.162,P=0.872).Conclusion:The LASSO re-gression model,constructed based on mid-trimester umbilical artery hemodynamic parameters(S/D ratio,RI,PI),serum hPL,24-hour urinary estriol,and fetal movement counts<6 within 2 hours,demonstrates good early predictive value for FGR in ART pregnancies.This model exhibits stability and strong clinical utility,providing significant evidence-based support for indi-vidualized fetal surveillance and early intervention in ART pregnancies.

    Reference
    Related
    Cited by
Get Citation

汪文雁;马钦玲;.基于辅助生殖妊娠脐血流动力学参数及胎盘功能指标构建LASSO回归模型对胎儿生长受限的早期预测[J]. Journal of North Sichuan Medical College,2025,40(11):1417-1421+1445.

Copy
Related Videos

Share
Article Metrics
  • Abstract:
  • PDF:
  • HTML:
  • Cited by:
History
  • Received:
  • Revised:
  • Adopted:
  • Online: December 22,2025
  • Published:
Article QR Code