Analysis and Detection of Optimal Euploid Embryo Transfer Position

Embryo transfer (ET) is a critical procedure in assisted reproductive technology (ART) that significantly impacts the success rate of pregnancy. Despite extensive research, there is still no consensus on the optimal embryo transfer position. The aim of this study is to identify specific transfer positions from ultrasound images and use machine learning to predict pregnancy success rates based on these positions. We analyzed patient information and ultrasound images from 152 cases who underwent euploid embryo transfer, focusing on factors such as antl-Müllerian hormone (AMH) levels, endometrial thickness, and distances from the embryo transfer position to various points within the uterus. Using machine learning algorithms including Decision Tree, Gradient Boosting, Logistic Regression, Random Forest, and Support Vector Machine (SVM), we predicted pregnancy outcomes. The most accurate model achieved a prediction accuracy of 0.656, highlighting the importance of AMH, endometrial thickness, and distances from the transfer position.