Predicting preeclampsia in early pregnancy using clinical and laboratory data via machine learning model

Area of Science:

• Obstetrics and Gynecology
• Medical Diagnostics
• Computational Biology

Background:

• Preeclampsia (PE) is a serious pregnancy complication.
• Early identification of PE is crucial for maternal and fetal health.
• Existing prediction methods may benefit from enhanced laboratory data integration.

Purpose of the Study:

• To identify laboratory test indicators associated with preeclampsia development.
• To develop prediction models for early-onset preeclampsia (EOPE), late-onset preeclampsia (LOPE), and preterm preeclampsia (Preterm PE).
• To evaluate the performance of machine learning models using clinical and laboratory data.

Main Methods:

• Retrospective recruitment of 144 EOPE, 363 LOPE, 231 Preterm PE, and 1458 healthy participants.
• Utilized clinical and laboratory data from routine prenatal visits.
• Developed prediction models using ensemble machine learning algorithms.

Main Results:

• Identified specific laboratory markers for EOPE, LOPE, and Preterm PE subtypes.
• Ensemble models incorporating clinical and laboratory data outperformed clinical-only models.
• EOPE and LOPE models demonstrated good sensitivity and specificity in predicting PE subtypes.

Conclusions:

• Key laboratory indicators for predicting PE were identified.
• Developed prediction models show significant potential for assessing PE risk and severity.
• Clinical and laboratory data integration enhances PE prediction accuracy.