The Application of Artificial Intelligence in the success rate of In Vitro Fertilization

Background and aims: In vitro fertilization (IVF) is a complex series of procedures used tohelp with fertility or prevent genetic problems and assist with the conception of a child. DuringIVF, mature eggs are collected from ovaries and fertilized by sperm in a lab. Some of the factorsthat affect the success rate of IVF are age, previous pregnancy, previous miscarriage, BMI, andlifestyle. However, the most effective factors that can help predict the best conditions for IVF areoptimal embryo selection and implantation-based factors. On the other hand, artificial intelligence(AI), or machines that mimic human intelligence, have been gaining traction due to their potentialto improve outcomes in medicine. Artificial Intelligence represents the combination of machinelearning, and a moderation and self-adapting prediction model. Artificial Intelligence can aid inselecting the best oocyte and sperm combination as well as predicting embryo quality and implantationtime. Furthermore, Artificial Intelligence has the potential to assist clinicians in developingan optimal patient-specific treatment regimen to improve IVF success.Method: In this review article, the keywords “In Vitro Fertilization,” “Implantation,” “Embryo,”“Convolutional Neural Networks,” “Artificial Intelligence,” and “Spiking Neural Networks” havebeen searched in international databases of articles such as Pubmed, Google Scholar, Science Direct,Elsevier, Scopus, and proper articles were extracted and reviewed.Results: Implantation-based factors that need to be considered in IVF are maternal age, embryotransfer day, endometrial thickness, total gonadotrophin dose, and estradiol concentration. Also,embryo-based factors like embryo viability, morphology, euploid/aneuploid status, developmentalstage, the metabolic and proteomic profile, and the number of embryos are some other factorsthat should be assessed in order to predict the optimal embryo for a perfect IVF with a highsuccess rate. Selecting the optimal embryo is a complex task. The transferred embryos must becarefully selected among others based on the above-mentioned factors. For this purpose, Machinelearning methods have been used to predict implantation and rank the most promising embryos.Machine learning solutions usually combine ranking embryos’ steps by optimizing for implantationprediction and using the same model for ranking the embryos within a cohort. So, A machinelearning–based decision support system would be useful in improving the success rate of IVFtreatment. Also, Deep Learning can replace human assessment of embryonic developmental potentialand thus contribute to implementing a single-embryo transfer methodology. Furthermore,Artificial Neural Networks can detect the best embryos from a euploid cohort which can lead toa higher IVF success rate.Conclusion: In IVF treatments, early identification of embryos with high implantation potentialis required to shorten the time of pregnancy while avoiding clinical complications to the newbornand the mother caused by multiple pregnancies. In conclusion, the wider use of AI in preciselyassessing patient characteristics, such as ovarian reserve, age, endocrine status, and clinical diagnostictests, will undoubtedly increase the efficiency of IVF. All mentioned findings suggestmachine learning algorithms based on age, BMI, and clinical data have an advantage over logisticregression for the prediction of IVF outcomes and therefore can assist fertility specialists.

کلیدواژه ها:

In Vitro Fertilization ، Implantation ، Embryo ، Convolutional Neural Networks ، Spiking Neural Networks ، Artificial Intelligence