Preeclampsia and its prediction: traditional versus contemporary predictive methods

Area of Science:

• Obstetrics and Gynecology
• Medical Informatics
• Biomedical Engineering

Background:

• Preeclampsia (PE) is a significant threat to maternal and perinatal health.
• Early prediction, prevention, and management are vital to mitigate adverse pregnancy outcomes.
• Understanding PE's epidemiology, etiology, pathophysiology, and risk factors is essential.

Purpose of the Study:

• To review the epidemiology, etiology, pathophysiology, and risk factors of PE.
• To discuss the emerging role of Artificial Intelligence (AI) deep learning (DL) in PE prediction.
• To advance the understanding and clinical application of early PE prediction methods.

Main Methods:

• A narrative review was conducted.
• Examined traditional PE prediction models.
• Evaluated AI deep learning technology for PE prediction.

Main Results:

• PE involves risk factors like poor uterine artery remodeling, immune response, endothelial dysfunction, and imbalanced angiogenesis.
• Traditional PE prediction models have limited sensitivity and specificity (30-50% detection rates).
• AI models achieve approximately 70% detection rates, showing superior predictive capabilities for late-onset PE.

Conclusions:

• AI deep learning technology shows promise in revolutionizing PE prediction and management.
• AI-based approaches address shortcomings of traditional models for effective risk assessment.
• Further research should focus on expanding databases and validating AI performance in diverse populations.