Artificial Intelligence in IVF: A Systematic Review of Predictive Models for Live Birth Outcomes across Clinical, Imaging, and Genomic Domains

Background: Artificial intelligence has emerged as a transformative tool in reproductive medicine, offering new avenues for predicting live birth outcomes in in vitro fertilization. While numerous models have been developed, their clinical relevance, methodological rigor, and generalizability remain variable. Objective: This systematic review aims to critically evaluate AI-based models designed to predict live birth in IVF, with a deliberate focus on clinically meaningful endpoints rather than intermediate surrogates. Methods: A structured search and appraisal were conducted using the Checklist for Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies, Prediction model Risk Of Bias ASsessment Tool, and Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis frameworks across six electronic databases PubMed (since ۱۹۹۶), Scopus (since ۲۰۰۴), Web of Science (since ۱۹۹۷), Embase (since ۱۹۴۷), MEDLINE (since ۱۹۶۶), and the Cochrane Library (since ۱۹۹۶) covering all available publications up to September ۲۰۲۵. Inclusion criteria targeted original studies that developed or validated artificial intelligence models with live birth as the primary outcome. Included studies were categorized by input modality (clinical-only, imaging-based, genetic, and multimodal) and algorithmic architecture (e.g., logistic regression, convolutional neural networks, generative adversarial networks, and transformer-based models). Subgroup-specific models targeting phenotypes such as polycystic ovary