Adaptive Agentic AI-Driven Women's Digital Twin for Personalized High-Risk Pregnancy Prediction and Clinical Decision Support

<jats:p>High-risk pregnancy remains a major global public health challenge due to its association with increased maternal and neonatal morbidity and mortality. Existing prediction methods primarily rely on static clinical data and conventional machine learning techniques, limiting their ability to support continuous monitoring, dynamic risk assessment, and personalized clinical decision-making. This study proposes an Adaptive Agentic AI-Driven Women’s Digital Twin framework for intelligent prediction and clinical decision support in high-risk pregnancy. The proposed framework integrates multimodal maternal healthcare data, including maternal clinical parameters and Cardiotocography (CTG) signals, to create a continuously updated digital representation of the patient. It employs an adaptive multi-agent architecture comprising monitoring, prediction, reasoning, recommendation, explanation, and feedback agents that collaboratively analyze patient data, estimate pregnancy risk, generate personalized recommendations, and continuously refine decision-making. Multiple machine learning and deep learning models are evaluated using Accuracy, Precision, Recall, F1-score, ROC-AUC, Sensitivity, Specificity, and Matthews Correlation Coefficient (MCC). Model transparency and clinical interpretability are enhanced through SHAP-based explainable artificial intelligence (XAI), enabling clinicians to understand key factors influencing predictions. Experimental results demonstrate that integrating multimodal data fusion with Women's Digital Twin technology and Agentic AI significantly improves prediction accuracy, continuous patient monitoring, individualized risk assessment, and clinical decision support.</jats:p>