Related Experiment Video
Updated: Jun 12, 2026
Semiconductor Sequencing for Preimplantation Genetic Testing for Aneuploidy
Published on: August 25, 2019
Noninvasive PPROM risk stratification with explainable AI using routine antenatal CRP and albumin
Shidong Tan1,2, Sihan Wu3, Shengnan Wu1
1Clinical and Translational Research Center, Department of Integrated Traditional Chinese Medicine (TCM) and Western Medicine, Shanghai Key Laboratory of Maternal Fetal Medicine, Shanghai Institute of Maternal Fetal Medicine and Gynecologic Oncology, Shanghai First Maternity and Infant Hospital, School of Medicine, Tongji University, Shanghai, China.
An AI model accurately predicts preterm premature rupture of membranes (PPROM) risk using routine antenatal data. This noninvasive tool identifies inflammation and oxidative stress markers, enabling early risk stratification for better obstetric care.
Area of Science:
- Obstetrics and Gynecology
- Artificial Intelligence in Medicine
- Maternal-Fetal Medicine
Background:
- Premature rupture of membranes (PROM) and preterm PROM (PPROM) are major causes of preterm birth and neonatal complications.
- Current methods for identifying high-risk pregnancies for PPROM are limited.
- Intrauterine infection is a primary driver, but early detection remains challenging.
Purpose of the Study:
- To develop and validate a noninvasive artificial intelligence (AI) predictive framework for late-gestation PPROM risk.
- To integrate multimodal antenatal data for enhanced predictive accuracy.
- To provide a tool for precision obstetric care and early intervention.
Main Methods:
- Development of a multimodal AI model using a large dataset (114,601 pregnancies).
- Integration of longitudinal maternal demographics, laboratory results, and ultrasound data.
- Optimization of an XGBoost algorithm and utilization of SHAP analysis for interpretability.
Main Results:
- The AI model achieved high performance with an AUC of 0.952.
- Key predictive features included elevated C-reactive protein (CRP) and decreased albumin levels.
- Analysis revealed dynamic risk patterns across gestation, highlighting inflammation and oxidative stress.
Conclusions:
- The AI framework offers a precise and noninvasive method for PPROM risk stratification in late pregnancy.
- Elevated CRP and low albumin are significant indicators of PPROM risk.
- This approach supports the translation of AI into precision obstetric care for improved outcomes.