Blastocyst development using sequential media versus one-step media in embryoscope and planer incubators

Abstract

OBJECTIVE: To correlate the different categories provided by an automatic diagnostic test (Eeva) with the blastocyst formation rate and to assess possible confounding factors associated to such correlation. DESIGN: Observational, retrospective, multi-center cohort study. MATERIALS AND METHODS: A total of 3002 embryos from patients undergoing egg donation cycles were imaged with the Eeva System. The Eeva system provides dark-field image and automatic cell-division tracking without the intervention of the embryologist. On Day 3 of development, the system utilizes information on specific division times (defined as P2 and P3) to automatically classify embryos into HIGH, MEDIUM, and LOW according to their probability of becoming a blastocyst. In addition, embryologists classified the same embryos according to standard morphological criteria defined by the Spanish Association of Embryologists (ASEBIR).RESULTS: There was a straight correlation between the Eeva categories and the blastocyst formation rates. More specifically: blastocyst formation rate was 75.20% for embryos categorized as HIGH, 58.50% for embryos categorized as MEDIUM and 51.50% for embryos categorized as LOW. In addition, optimal blastocysts formation rate was 35.7% for HIGH, 27.2% for MEDIUM and 17.3% for LOW. A further logistic regression analysis was developed taking into account possible confounding factors thought as having a potential effect on clinical pregnancy including: number of oocytes retrieved and embryo morphology according to ASEBIR. Results of the model indicate that an embryo labeled as HIGH develops to blastocyst 2.014 times more than an embryo labeled as LOW (Table 1). Analysis of area under the curve revealed values of 0,728 CI95% (0.707-0.749) for blastocyst prediction according to Eeva and of 0.717 CI95% (0.703-0.732) according to ASEBIR classification. When Eeva and ASEBIR were combined the AUC value was of 0.788 (IC95% 0.768-0.808).CONCLUSIONS: This study demonstrates in a large data set that there is a direct link between the Eeva categories and the blastocyst formation rate and quality. Embryos HIGH have a higher probability of becoming a blastocyst than embryos LOW and this is independent of the presented confounding factors. Both classification methods are independent from each other and have similar predictive value. However, when both are combined, the predictive value increases.

Development of a dynamic machine learning algorithm to predict clinical pregnancy and live birth rate with embryo morphokinetics

“…From September 2015 onwards, embryos were cultured in Sage 1-Step media. Both culture conditions were compared, and no changes in blastocyst development were observed locally ( 45 ), similar to findings at other centers ( 46 ).…”