Related Experiment Video
Updated: Jul 5, 2026
04:18
Modeling Ascending Vaginal Infection, Preterm Birth, and Neonatal Morbidity in Mice
Published on: October 10, 2025
Development and validation of machine learning-based model for macrosomia and spontaneous preterm birth: A
Jiajun Chen1, Rong Liu2, Xuan Zhou3
1Department of Blood Transfusion, Shenzhen Baoan District Songgang People's Hospital, Shenzhen, China.
Summary
Machine learning models accurately predict macrosomia and spontaneous preterm birth (sPTB) using clinical data. These tools aid early identification and intervention for high-risk pregnancies.
Area of Science:
- Obstetrics and Gynecology
- Medical Informatics
- Machine Learning in Healthcare
Background:
- Predicting adverse pregnancy outcomes like macrosomia and spontaneous preterm birth (sPTB) is crucial for maternal and infant health.
- Existing prediction methods often lack comprehensive data integration and external validation.
Purpose of the Study:
- To develop and externally validate machine learning models for predicting macrosomia and sPTB.
- To identify key clinical predictors for these pregnancy complications.
Main Methods:
- Retrospective cohort study of 14,238 pregnant women (2018-2023).
- Utilized demographic, laboratory, and delivery outcome data.
- Developed predictive models using logistic regression and machine learning algorithms (Light GBM, AdaBoost, GBDT) with LASSO for feature selection.
Main Results:
- Key macrosomia predictors: pre-pregnancy BMI, 1-h postprandial glucose, maternal education, comorbidities.
- Key sPTB predictors: comorbidities, AST, uric acid.
- External validation showed AUC of 0.742 for macrosomia and 0.706 for sPTB (logistic regression models); models demonstrated good calibration.
Conclusions:
- Developed models effectively identify high-risk pregnancies for macrosomia and sPTB.
- Models facilitate early detection and targeted interventions.
- External validation underscores the need for prospective, multicenter studies to improve generalizability.
