Abstract: With the gradual emergence of population issues and the increasing incidence of infertility, the role of assisted reproductive technology (ART) is becoming increasingly important. Accurately, objectively, and comprehensively assessing the blastocyst formation rate and predicting delivery outcomes are urgent problems to be solved. This study retrospectively analyzed the clinical data of patients who underwent ART at Renmin Hospital, Hubei University of Medicine, from June 2017 to January 2020. A regression model for the blastocyst formation rate was established through regression analysis. The results showed that late-stage indicators in ART had a significant impact on the blastocyst formation rate, with top-quality embryos having the greatest effect. A discriminant analysis model for delivery outcome was established. It correctly classified 80.4% of the original grouped cases and 79.6% of the cross-validated grouped cases. The area under the receiver operating characteristic curve (AUROC) was 0.887, indicating that the discriminant model has a relatively high predictive diagnostic value. Analysis revealed that blood hCG (human chorionic gonadotropin) results play a crucial role in the discriminant analysis model. In addition, the number of top-quality embryos and the blastocyst formation rate also have a significant impact on the accurate prediction of delivery outcomes. Through analyzing the regression model, we propose exploring early blood markers that can predict the blastocyst formation rate. Simultaneously, a new discriminant model that can directly predict the final delivery outcomes was established.

Lay summary: Against the backdrop of gradually emerging population issues and a rising incidence of infertility, assisted reproductive technology (ART) is playing an increasingly vital role. Accurately, objectively, and comprehensively evaluating embryo quality and predicting transplantation success rates represent pressing challenges that need to be addressed. This study retrospectively analyzed clinical data from patients who underwent ART. We developed a model that uses statistical methods to assess the formation of blastocysts (early-stage embryo). The results indicate that certain late-stage indicators in ART significantly influence the blastocyst formation rate. Therefore, we propose a new direction: exploring early blood markers that can predict blastocyst formation rate. Simultaneously, a new method was established to predict delivery outcomes, demonstrating strong predictive performance and high accuracy.