JBRA Assist. Reprod. 2025;29(Suppl 1):47-47
POSTER PRESENTATION
doi: 10.5935/1518-0557.20250101
1Bios - Centro de Medicina Reprodutiva do Ceará
Objective: Direct cleavage, the division of an embryo from one cell to three or more, without the intermediate two-cell step, is associated with a lower potential for embryonic development. When cultivated in conventional incubators, this event goes unnoticed due to the punctual and static evaluation, while time-lapse incubators allow continuous monitoring of embryonic development. Artificial intelligence (AI) can automatically identify this and other anomalies, contributing to a more assertive selection of embryos with the greatest chance of achieving blastulation. The objective of this analysis is to evaluate the effectiveness of an AI system in detecting direct cleavage and correlating the occurrence of this phenomenon with the embryo blastulation rate.
Methods: A sample of 300 embryos grown in time-lapse incubators from patients with an average age of 37 years. The analysis was carried out between January 2024 and December 2024. AI software trained to identify morphokinetic events was used, with a focus on detecting direct cleavage. The blastulation rates of embryos with and without direct cleavage were compared.
Results: AI detected direct cleavage in 45 of the 300 embryos analyzed (15%). Of the 45 embryos with direct cleavage, only 7 (15.6%) reached the blastulation stage. In contrast, of the 255 embryos without direct cleavage, 158 (61.9%) reached the blastulation stage. The difference between blastulation rates was statistically significant (p<0.05).
Conclusion: The presence of direct cleavage, identified automatically by AI, is strongly correlated with a lower blastulation rate. The implementation of this technology in the laboratory environment can improve embryo selection, reducing the subjectivity of conventional analysis and increasing the chances of success in assisted reproduction treatments.