Development of a robust automated tool for the annotation of embryo morphokinetic parameters

Study question: Is it possible to automatically annotate human embryo development in time-lapse devices, with results comparable to manual annotation? Summary answer: We developed an automated tool for the annotation of embryo morphokinetic parameters having a high concordance with expert manual annotation in a large scale-study. What is known already: Morphokinetic parameters obtained with time-lapse devices are increasingly used for human embryo quality assessment. However, their annotation is time-consuming and can be operator-dependent, highlighting the need of developing automated approaches. Study design, size, duration: This monocentric pilot study was conducted using 701 blastocysts originating from 584 couples undergoing IVF with embryo culture in a time-lapse device and on 4 mouse embryos. Participants/materials, setting, methods: An automated annotation tool was developed based on grey level coefficient of variation and detection of the thickness of the zona pellucida. The timings of cellular events obtained with the automated tool were compared with those obtained manually by 2 expert embryologists. The same procedure was applied on 4 mouse preimplantation embryos obtained with a different device in a different setting. Main results and the role of chance: Although some differences were found when embryos were considered individually, we found an overall excellent concordance between automated and manual annotation of human embryo morphokinetics from fertilization to expanded blastocyst stage (r 2 =0.94). Moreover, the automated annotation tool gave promising results across species (human, mice). Limitations, reasons for caution: These results should undergo multi-centric external evaluation in order to test the overall performance of the annotation tool. Wider implications of the findings: Our system performs significantly better than the ones reported in the literature and on a bigger cohort, paving the way for high-throughput analysis of multicentric morphokinetic databases, providing new insights into the clinical value of morphokinetics as predictor of embryo quality and implantation. Study funding/competing interest(s): This study was partly funded by Finox Forward Grant 2016.