Artificial intelligence algorithms for optimizing assisted reproductive technology programs: A systematic review

Abstract

Artificial intelligence (AI) has been experiencing rapid growth in recent years, and numerous applications are improving the single-step efficiency of the whole assisted reproductive technology (ART) procedure. In this review, we collected all the algorithms supplying ART and selected those supporting the clinical assistance to the procedure up to the successful attempt. Those with a clear role in improving ART performances were further selected. We found a questionnaire-based algorithm identifying patients at risk for endometriosis with early management and better fertility outcome. An algorithm can detect the values of simple gamete production (male) and reservoir (female) according to gradual scale allocation, and display themas normal or abnormal, spontaneousor stimulated gamete production. This can provide significant benefits for infertile couples undergoing diagnostic and therapeutic journeys. The calculators for the starting dose of gonadotropins and the trigger timing during controlled ovarian stimulation make clinical management more efficient. With the application of AI in ART, the ability to determine the optimal number of metaphase II oocytes required for blastocyst formation and number of oocytes needed for embryo production has been significantly improved. The calculation of the implantation rate as proposed in different calculators, using the ultrasound of endometrial vascularization or the age and euploidy of the embryo transferred, may provide further advancement in managing the ART procedure with more participation from the couples to increase the efficacy of the procedures. Finally, the calculator of presumptive success with an ART program based on couples or medical center profiling and efficiency is of tremendous comfort to couples. In conclusion, algorithms and machine learning development in human reproduction are growing daily with evident benefits. Infertility treatments by in vitro fertilization (IVF) are assisted by several algorithms that improve the efficiency of each procedure step, making IVF program’s management more effortless.