Oocyte quality in the era of AI: integration of morphology, metabolic activity and time-lapse imaging

Significant advances have been made with regards to quantifying the viability of the preimplantation-stage human embryo, through the analysis of morphokinetics, quantification of biomarkers and trophectoderm biopsy to facilitate preimplantation genetic testing for aneuploidies (PGT). However, there is still limited capacity to determine the inherent potential of gametes. In particular, assessing the viability of an oocyte has been elusive, and the standard means of determining oocyte quality remains assessing its morphology. Unfortunately, except for grossly abnormal features, such as giant eggs, oocyte morphology is not strongly associated with clinical outcomes. With the rise of oocyte donation and social oocyte freezing, there is an immediate need to improve the ability to determine gamete health and developmental potential to better counsel patients about their reproductive future. Furthermore, the outcomes of successive stimulation cycles can vary within an individual, heightening the need to ensure that the quality of a given oocyte cohort can be quantified and stimulation regimens optimized per patient. Developments in artificial intelligence and in media analysis and microscopies to assess metabolic state could provide valuable insights into oocyte health, thereby increasing the ability to not only improve oocyte health and selection, but also provide a much-needed tool to assist in optimizing ovarian stimulation.