Models, Applications, and Open Issues in the Intelligent Prediction of Preeclampsia: A Review

Abstract

Preeclampsia (PE) is a placental dysfunction charac­terized by proteinuria and hypertension that affects approximately 10% of pregnancies worldwide. Because PE is multifactorial and pathogenic through multiple pathways and mechanisms, there is currently no effective treatment. Early detection and intensification remain the main clinical strategies. This review collects 99 related papers, discusses the challenges faced by the traditional methods of diagnosing and predicting PE, introduces the ML techniques related to predicting PE, and summarizes the standards, indices, and mechanisms that can address these challenges. This review aims to bridge the knowledge of Machine Learning (ML) and PE by discussing and evaluating the prediction of PE from the perspective of both medical practitioners and ML researchers. It also compares the traditional PE prediction approaches with the ML-based PE prediction approaches, highlights the state of the art in PE prediction, and suggests open issues for future research.