Marcos Meseguer

BACKGROUND: Time-lapse observation presents an opportunity for optimizing embryo selection based on morphological grading as well as providing novel kinetic parameters, which may further improve accurate selection of viable embryos. The objective of this retrospective study was to identify the morphokinetic parameters specific to embryos that were capable of implanting. In order to compare a large number of embryos, with minimal variation in culture conditions, we have used an automatic embryo monitoring system. METHODS: Using a tri-gas IVF incubator with a built-in camera designed to automatically acquire images at defined time points, we have simultaneously monitored up to 72 individual embryos without removing the embryos from the controlled environment. Images were acquired every 15 min in five different focal planes for at least 64 h for each embryo. We have monitored the development of transferred embryos from 285 couples undergoing their first ICSI cycle. The total number of transferred embryos was 522, of which 247 either failed to implant or fully implanted, with full implantation meaning that all transferred embryos in a treatment implanted. RESULTS: A detailed retrospective analysis of cleavage times, blastomere size and multinucleation was made for the 247 transferred embryos with either failed or full implantation. We found that several parameters were significantly correlated with subsequent implantation (e.g. time of first and subsequent cleavages as well as the time between cleavages). The most predictive parameters were: (i) time of division to 5 cells, t5 (48.8-56.6 h after ICSI); (ii) time between division to 3 cells and subsequent division to 4 cells, s2 (≤ 0.76 h) and (iii) duration of cell cycle two, i.e. time between division to 2 cells and division to 3 cells, cc2 (≤ 11.9 h). We also observed aberrant behavior such as multinucleation at the 4 cell stage, uneven blastomere size at the 2 cell stage and abrupt cell division to three or more cells, which appeared to largely preclude implantation. CONCLUSIONS: The image acquisition and time-lapse analysis system makes it possible to determine exact timing of embryo cleavages in a clinical setting. We propose a multivariable model based on our findings to classify embryos according to their probability of implantation. The efficacy of this classification will be evaluated in a prospective randomized study that ultimately will determine if implantation rates can be improved by time-lapse analysis.

BACKGROUND: An efficient oocyte cryopreservation method is mandatory to establish a successful egg-banking programme. Although there are increasing reports showing good clinical outcomes after oocyte cryopreservation, there is still a lack of large controlled studies evaluating the effectiveness of oocyte cryo-banking. In this study, we aimed to compare the outcome of vitrified-banked oocytes with the gold standard procedure of employing fresh oocytes. METHODS: A randomized, prospective, triple-blind, single-centre, parallel-group controlled-clinical trial (NCT00785993), including 600 recipients (alpha = 0.05 and power of 80% for sample-size calculation) selected among 1032 eligible patients from November 2008 to September 2009, was designed to compare the outcome of vitrified-banked oocytes with the gold standard procedure of employing fresh oocytes. The study was designed to establish the superiority of the ongoing pregnancy rate (OPR) of fresh oocytes over that of vitrified oocytes, by performing a likelihood ratio test in a logistic regression analysis expressed as odds ratio (OR) with 95% confidence interval (CI). A limit of 0.66 for OR of vitrified versus fresh groups was defined to set up a possible conversion from superiority to non-inferiority. Randomization was performed 1:1 based on a computer randomization list in vitrification (n = 300) or fresh groups (n = 300). The primary end-point was the OPR per randomized patient i.e. intention-to-treat population (ITT). Secondary end-points were clinical pregnancy (CPR), implantation (IR) and fertilization rates, respectively. Additionally, embryo developmental characteristics were recorded. RESULTS: There were no differences in donor ovarian stimulation parameters, demographic baseline characteristics for donors and recipients, ovum donation indications or male factor distribution between groups (NS). The OPR per ITT was 43.7 and 41.7% in the vitrification and fresh groups, respectively. The OR of OPR was 0.921 in favour of the vitrification group. Nevertheless, the 95% CI was 0.667-1.274, thus the superiority of fresh group with respect to OPR was not proven (P = 0.744). Non-inferiority of the vitrified group compared with the fresh group was shown with a margin of 0.667, which was above the pre-established non-inferiority limit of 0.66. CPR per cycle (50.2 versus 49.8%; P = 0.933) or per embryo-transfer (55.4 versus 55.6% ; P = 0.974), and IR (39.9 versus 40.9%; P = 0.745) were similar for patients receiving either vitrified or fresh oocytes. The proportion of top-quality embryos obtained either by inseminated oocyte (30.8 versus 30.8% for Day-2; and 36.1 versus 37.7% for Day-3, respectively) or by cleaved embryos (43.6 versus 43.8% for Day-2 and 58.4 versus 60.7% for Day-3, respectively) was similar between groups (NS). CONCLUSIONS: This controlled-randomized, clinical trial confirmed the effectiveness of oocyte cryo-storage in an ovum donation programme, failing to demonstrate the superiority of using fresh oocytes with respect to the use of vitrified egg-banked ones in terms of OPR. Instead, the non-inferiority of vitrified oocytes was confirmed. These findings involve highly relevant issues that may open a new range of possibilities in ART.

Visual morphology assessment is routinely used for evaluating of embryo quality and selecting human blastocysts for transfer after in vitro fertilization (IVF). However, the assessment produces different results between embryologists and as a result, the success rate of IVF remains low. To overcome uncertainties in embryo quality, multiple embryos are often implanted resulting in undesired multiple pregnancies and complications. Unlike in other imaging fields, human embryology and IVF have not yet leveraged artificial intelligence (AI) for unbiased, automated embryo assessment. We postulated that an AI approach trained on thousands of embryos can reliably predict embryo quality without human intervention. We implemented an AI approach based on deep neural networks (DNNs) to select highest quality embryos using a large collection of human embryo time-lapse images (about 50,000 images) from a high-volume fertility center in the United States. We developed a framework (STORK) based on Google's Inception model. STORK predicts blastocyst quality with an AUC of >0.98 and generalizes well to images from other clinics outside the US and outperforms individual embryologists. Using clinical data for 2182 embryos, we created a decision tree to integrate embryo quality and patient age to identify scenarios associated with pregnancy likelihood. Our analysis shows that the chance of pregnancy based on individual embryos varies from 13.8% (age ≥41 and poor-quality) to 66.3% (age <37 and good-quality) depending on automated blastocyst quality assessment and patient age. In conclusion, our AI-driven approach provides a reproducible way to assess embryo quality and uncovers new, potentially personalized strategies to select embryos.