Pregnancy Rates Are Significantly Higher when Selecting Embryos for Transfer According to the Classification Provided by an Automatic Diagnostic Test.

To correlate pregnancy rates with morphological and time-lapse classifications (standard morphological criteria on day 3 defined by the Spanish Association of Embryologists (ASEBIR) and time-lapse classification provided by an automatic diagnostic test (Eeva). Observational, retrospective, multi-center cohort study. A total of 3002 embryos from cycles performed in our egg donation program were incubated in a conventional incubator with the Eeva system. This time-lapse system has an automatic software that classifies embryos according to the timings of P2 (t3-t2) and P3 (t4-t3) into HIGH, MEDIUM or LOW according to their probability of reaching blastocyst stage. These embryos were also classified by embryologist using morphological criteria defined by the Spanish Association of Embryologists (ASEBIR) into A, B, C or D. A stratified and multivariable analysis was performed to analyze the effect of both methods of selection on ongoing pregnancy. Patients where no HIGH embryos were transferred (n=75) had an ongoing pregnancy rate (OPR) of 46.70% and those patients where at least one HIGH embryo was transferred (n=109) significantly increased OPR to 67% (p=0.009). Those patients where no A ASEBIR embryos were transferred (n=108) presented an OPR of 60.20%, while patients with at least one A ASEBIR embryo (n=81) was 59.30% (p=0898). Combining both categorizations: Patients where no A or HIGH embryos were transferred had an OPR of 60.0% while those without A ASEBIR but at least one HIGH had 71.20% (p=0.025). Patients where at least one embryo was A but no one was HIGH had OPR of 33.30%, and those with at least one A and at least one HIGH was 60% (p=0.084). In the logistic regression analysis we observed whether at least one of the embryos is labelled as HIGH, OPR is 2.567 (CI95%1.305-5.632) times higher than a cycle where no HIGH embryos are transferred (p=0.006). In the multivariable model no effect is observed when at least one embryo categorized as A ASEBIR is transferred (OR=0.563 0.268-1.182) or according to day of transfer (OR=0.795 0.244-2.597). To our knowledge, this is the largest data set of patients which embryos were evaluated by the Eeva system. Our results demonstrated a significant increase in the chances of achieving an OPR when Eeva is selecting a HIGH embryo for transfer showing higher accuracy than standard morphology for embryo selection.