JBRA Assist. Reprod. 2025;29(Suppl 1):39-39
POSTER PRESENTATION
doi: 10.5935/1518-0557.20250093
1FERTILIA - Reproductive Medicine
Objective: To evaluate whether the score of an artificial intelligence (AI) algorithm for evaluating oocyte quality (OAI) correlates with the blastocyst development. To evaluate whether the implementation of the OAI allows the equitable distribution of oocytes in the same donors with and without use it.
Methods: Oocytes were analyzed by an artificial intelligence algorithm using 2D photographic analysis (score from 0 to 10). A higher score indicates a greater likelihood of an oocyte reaching the blastocyst stage (Bl). All the oocytes were freshly granted and cultivated individually after intracytoplasmic sperm injection (ICSI). For the first objective, a retrospective cohort study consisting of 286 oocytes derived from 61 egg donation cycles were assessed. Specificity, sensibility and area under the curve (AUC) examination was performed. Results were compared with data published by the OAI provider. For the second objective, 8 donors were selected and grouped in: previous cycle without OAI (cycle 1) and the next cycle with OAI (cycle 2). The percentage of blastocysts obtained were compared between cycles 1 and 2.
Results: The blastocyst distribution regarding the OAI scores were similar between the provider (16%, 37%, 44%, and 53%) and our own oocytes (22%, 36%, 49%, 58%). The AUC observed was 0.66 with a sensitivity 66% and specificity 57%. The blastocysts rate distribution between cycle 1 (0Bl 3%, 1Bl 32%, 2Bl 38%, 4Bl 27%) and cycle 2 (0Bl 0%, 1Bl 11%, 2Bl 72%, 4Bl 17%) was significantly different.
Conclusion: A correlation was observed between OAI scores and the blastocyst development rate. The application of OAI allowed to optimize the equitable distribution of blastocysts among recipients.