Artificial Intelligence Enhanced Prediction of Live Birth Outcomes for Informed Clinical Decision-Making: A Comprehensive Survey of Data-Driven Models, Predictive Factors, and Reproductive Healthcare Systems

Abstract

Live birth is considered as an important clinical endpoint in the reproductive medicine and assisted reproductive technology (ART). Conventional statistical approaches are struggling in capturing the nonlinear relationships and its associated hidden patterns in the large-scale fertility datasets. This survey comprehensively reviews the recent studies published between 2020 and 2026 that is focusing on AI-enhanced live birth prediction systems. The reviewed literature includes conventional machine learning algorithms such as Logistic Regression, Random Forest, support vector Machine (SVM), Extreme Gradient Boosting (XGBoost), LightGBM, CatBoost, and ensemble learning models, where the well as deep learning architectures including Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), Transformer-based networks, and explainable AI frameworks. The survey is categorizing predictive variables into demographic, clinical, hormonal, embryological, genetic, laboratory, and lifestyle factors while comparing model performance across the diverse fertility datasets. Analysis of the reviewed literature is indicating that AI-based approaches is consistently performing than the conventional statistical models. Logistic Regression models reported predictive accuracies ranging from the 62% to 78%, while Random Forest and Gradient Boosting techniques achieved accuracies between the 75% and 89%. Advanced ensemble models showing Area under the Curve (AUC) values between the 0.82 and 0.93. Deep learning architectures achieved prediction accuracies ranging from the 85% to 94%, with the Transformer-based and hybrid ensemble frameworks reporting AUC values exceeding 0.95 in several large-scale studies. Across the surveyed works, maternal age, embryo quality, anti-Müllerian hormone (AMH), antral follicle count (AFC), endometrial thickness, fertilization rate, blastocyst development score, body mass index (BMI), and lifestyle indicators emerged as most of the influential predictors.