AI-Driven Prediction of Preterm Birth via Uterine Bioelectrical Signal Analysis (EHG) towards Personalized Tocolysis Management

Preterm birth remains a leading cause of neonatal morbidity and mortality worldwide, affecting approximately ۱۵ million infants annually. Current diagnostic methods, including tocodynamometry and cervical length assessment, exhibit limited predictive accuracy and fail to provide personalized risk stratification. Electrohysterography (EHG), which non-invasively records uterine myoelectrical activity, offers a superior alternative with higher sensitivity and robustness. This congress paper synthesizes recent advances in artificial intelligence-driven EHG signal analysis for preterm birth prediction. Machine learning and deep learning architectures, including multi-head attention networks, multivariate empirical mode decomposition, and hybrid deep learning frameworks, have achieved classification accuracies exceeding ۹۶%, with sensitivities up to ۹۸.۵%. Beyond prediction, EHG-based AI models enable dynamic monitoring of tocolytic therapy response, facilitating personalized tocolysis management through real-time assessment of uterine contractility patterns. Despite promising results, challenges persist, including dataset imbalance, lack of external validation, and absence of standardized protocols. This review establishes an integrated framework linking EHG signal processing, AI-based prediction, and personalized clinical decision-making, and outlines future directions toward precision obstetrics.

کلیدواژه ها:

Preterm Birth ، Electrohysterography ، Artificial Intelligence ، Machine Learning ، Tocolysis.