Application of artificial intelligence algorithms in in vitro fertilization -embryo transfer to improve the efficiency and outcomes by selecting the optimal embryo: a systematic review abstract
Introduction:
In vitro fertilization -embryo transfer (IVF -ET) typically evaluates embryo quality through visual morphological techniques to identify viable embryos for transfer. However, the success rate of IVF -ET is 30% - 70% due to human errors in the selection process. The main challenge is absence of efficient models and clinical instruments capable of predicting the results of IVF -ET and selecting the optimal embryo for transfer. The utilization of artificial intelligence (AI) algorithms and machine learning (ML) in the field of Reproductive and Infertility medicine and assisted reproductive technology (ART) has achieved significant advancements over the recent years. AI has been used broadly in clinical assessments of follicular development, determining the ideal timing for transplantation, and forecasting pregnancy outcomes. This review is focused on discussing AI capabilities to predict the outcomes of IVF -ET and ultimately improve the success rate through the application of AI algorithms and ML Techniques. Search Strategy: Based on PRIMSA checklist a systematic search was conducted by two people in PubMed, Embase and Scopus for English articles published from 2018 to 2024. Google scholar was utilized for reviewing literatures. Research terms included: in vitro fertilization -embryo transfer, in vitro fertilization, assisted reproductive technology, artificial intelligence, machine learning and their associated terminology. Results: 372 articles were found which 116 of them were selected for this review. The results imply that most of the applied AI algorithms and ML techniques have significant accuracy and calculation speeds which can assist embryologists in predicting IVF -ET outcomes and facilitating a more objective selection process for oocytes and embryos prior to transfer. Conclusion and Discussion: There are emerging techniques aimed at advancing embryo selection, including: Three-dimensional (3D) ultrasound, Fluorescence lifetime imaging microscopy (FLIM) and optical lenses. Combinations of these technologies with AI could result in better outcomes. However, utilization of AI as a significant asset in IVF -ET necessitates a strong collaborative effort among ART specialists, AI experts, and clinical staff. Additionally, more data from randomized controlled trials and studies are needed to assess the validity of the algorithms before finalizing AI systems application.