Aplicación de la inteligencia artificial en el laboratorio de reproducción asistida. Trabajo de revisión

Abstract

Introduction

The aim of this review is to learn about the current situation of artificial intelligence for clinical use in assisted reproduction based on a study of the limitations that have been found.

Search Methods

An exhaustive review is carried out through databases such as PubMed, Elsevier and the library of scientific societies, searching for original articles, using a combination of keywords such as Artificial intelligence, FIV, ART.

Results

There are algorithms capable of analysing different seminal parameters, among which the concentration, motility and morphology of human sperm stand out, and it is on this last parameter where the fundamental role of artificial intelligence is focused. Moreover, algorithms are being developed with static images or time-lapse sequences both at different specific points of embryo development and at specific periods.

Conclusions

Much of the literature is retrospective, so most of the algorithms appear to be in the early stages. In the commercial world, there is also a need for corroborative studies. The limitations often encountered are, in addition to the nature of the study, getting an explainable algorithm, introducing other parameters that also affect the outcome and the data set. It is therefore necessary to conduct randomised controlled trials in different clinics in which explainable algorithms are presented that through the analysis of different parameters achieve reliable results for clinical practice.