Prognostic value of oocyte quality in assisted reproductive technology outcomes: a systematic review

Review Article

VOLUME 2, ISSUE 2, P120-139, APRIL 01, 2021


Nicole M. Fischer, M.P.H., Ha Vi Nguyen, M.D., Bhuchitra Singh, M.D., M.P.H., M.H.S., Valerie L. Baker, M.D., James H. Segars, M.D. 



To survey and assess modern methodologies used to test oocyte quality that have prognostic value in predicting assisted reproductive technology outcomes

Evidence Review

Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, we surveyed the English-language literature between January 1, 2010, and December 31, 2019, using PubMed, Scopus, and Embase databases. Two reviewers screened for articles focusing on oocyte quality markers that predict assisted reproductive technology outcomes, including embryo quality as well as fertilization, implantation, pregnancy, continued pregnancy, and live birth rates. Articles that did not mention oocytes or those that focused on nonhuman subjects, oocyte aging, oocyte maturation, embryo quality, interventions, or specific clinical diagnoses (endometriosis and polycystic ovarian syndrome) were deemed outside the scope of this analysis and excluded.


Twenty-six relevant articles were identified, including 19 prospective and 7 retrospective studies (n = 2,210 patients). We identified 3 general approaches for oocyte quality assessment: morphological evaluation (11 articles), genomics and proteomics (13 articles), and artificial intelligence (2 articles). Morphological assessment did not show a consistent pattern of predictive value of predicting in vitro fertilization outcomes (7 articles in favor of its predictive value, 4 against). A considerable proportion of genomic and proteomic articles identified promising biomarkers that may predict pregnancy and live birth (12 in favor, 1 against). Machine learning is a rapidly growing frontier that minimizes subjectivity while potentially improving predictive ability (2 in favor).


Although there remains a lack of consensus on optimal methods to predict reproductive success, machine learning and genomics demonstrate promise in improving the understanding of oocyte quality assessment and prognostication.