Artificial Intelligence-Based Prediction of Preeclampsia Using First-Trimester Biomarkers.

Area of Science:

• Maternal-fetal medicine
• Artificial intelligence in healthcare
• Biomedical data science

Background:

• Preeclampsia (PE) is a major global cause of maternal and perinatal mortality.
• Current first-trimester PE detection methods lack sufficient sensitivity.
• Early identification is crucial for timely intervention and improved outcomes.

Purpose of the Study:

• To develop and evaluate an AI-based predictive framework for early preeclampsia detection.
• To integrate maternal demographics, biophysical, and biochemical markers for enhanced prediction.
• To compare AI model performance against traditional first-trimester screening algorithms.

Main Methods:

• Development of machine learning models (SVM, RF, DNN) using a first-trimester patient dataset.
• Integration of maternal demographic, biophysical, and biochemical data.
• Rigorous cross-validation, feature selection, and regularization to prevent overfitting.

Main Results:

• The deep neural network (DNN) model achieved 93.4% accuracy in predicting preeclampsia.
• Key predictive features included placental growth factor (PlGF), PAPP-A, and mean arterial pressure (MAP).
• AI framework performance surpassed traditional first-trimester screening algorithms.

Conclusions:

• AI-driven predictive analytics show significant potential for early preeclampsia risk assessment.
• The proposed framework can aid in personalized prenatal management strategies.
• Further validation and clinical implementation are necessary to assess generalizability and utility.