Preeclampsia prediction via machine learning: a systematic literature review

Area of Science:

• Obstetrics and Gynecology
• Medical Informatics
• Computational Biology

Background:

• Preeclampsia is a serious pregnancy complication with unknown causes and risk factors.
• Early prediction of preeclampsia is crucial for timely intervention and improved maternal outcomes.
• Machine learning (ML) presents a powerful tool for developing predictive models for preeclampsia.

Purpose of the Study:

• To systematically review and analyze recent machine learning studies for preeclampsia prediction.
• To identify key features, algorithms, and geographic trends in ML-based preeclampsia research.
• To highlight limitations and future directions for enhancing ML models in preeclampsia detection.

Main Methods:

• Systematic literature review of studies published between January 1, 2013, and December 31, 2023.
• Searches conducted on Google Scholar and PubMed, identifying 183 studies, with 35 selected based on inclusion criteria.
• Analysis of common predictive features, ML algorithms, study locations, and dataset characteristics.

Main Results:

• Commonly used predictive features include maternal age, pregnancy history, body mass index, diabetes, hypertension, and blood pressure.
• Less frequently utilized features were medications, genetic data, and clinical imaging.
• Popular ML algorithms comprised Random Forest, Support Vector Machine, Logistic Regression, Decision Tree, and Naïve Bayes.
• Research is concentrated in China and the USA, with a recent surge in publications, but often relies on small, single-center datasets.

Conclusions:

• Machine learning shows significant potential for preeclampsia prediction, utilizing readily available clinical data.
• The current research landscape necessitates more diverse datasets and multi-center studies to improve model generalizability.
• Further research is essential to refine ML models for robust early detection and effective management of preeclampsia.