Artificial Intelligence in Sperm Selection for Assisted Reproductive Medicine

recently fertility and infertility Researchers summarize the available evidence for the application of artificial intelligence (AI) and machine learning in sperm selection.

All relevant papers were obtained from PubMed Central, Web of Science and the MEDLINE-Academic database. A total of 261 articles were found in the initial search. However, 34 papers were selected based on the inclusion criteria.

study: Artificial intelligence (AI) for sperm selection – a systematic review. Image credit: Inna Dodor / Shutterstock.com

Importance of sperm selection

Over 100 million people worldwide face infertility problems, of which up to 50% are due to male factors.

Semen analysis is associated with the investigation of sperm morphology, motility, and DNA integrity and is essential for the diagnosis and subsequent treatment of male factor infertility. Embryologists face the daunting task of selecting a single sperm from millions of sperm in a sample based on various parameters. This is a laborious process with a high risk of selection error.

Semen parameters are strong prognostic indicators of fertilization and pregnancy outcome. If analysis indicates that semen parameters are suboptimal, apply assisted reproductive technology (ART) to allow sperm to overcome barriers in the female reproductive tract, thereby increasing the chances of conception. can do.

The success rate of ART remains relatively low globally due to lack of proper sperm selection. Despite technological advances, final sperm selection is largely performed manually by embryologists according to World Health Organization (WHO) standards. Sperm selection is critical, as intracytoplasmic sperm injection (ICSI) requires a single sperm.

WHO provides guidance for selecting appropriate sperm based on motility as well as morphology such as sperm head length, presence of vacuoles, and circularity. However, embryologists do not have enough time to comprehensively evaluate the whole sperm, which may affect the success of ART. Here, it may be possible to apply AI to improve the efficiency of sperm selection.

Sperm selection with AI and machine learning

Previous studies have shown that AI can consistently and effectively identify embryos with optimal developmental and implantation potential. AI can also reduce embryologist time and effort associated with visual assessment and manual embryo grading.

Machine learning algorithms can handle large datasets, as well as the large amounts of data evaluated during embryo evaluation. Therefore, this technique can be applied to automate the sperm selection process by combining genetic and visual data. Implementing AI and machine learning algorithms into his ART lab could greatly improve an embryologist’s ability to select sperm.

The preferred sperm morphology is a smooth oval head, lack of large/multiple vacuoles, acrosome covering 40-70% of the head, thinness of the medium piece, and a head size of up to 3 minutes. characterized by residual cytoplasm of up to 1 in . AI algorithms can standardize and accelerate sperm analysis based on available models. Furthermore, sperm morphology combined with deep learning algorithms can be evaluated with approximately 98% accuracy.

The performance of AI and machine learning algorithms depends on the quality of the training dataset images. To obtain higher accuracy with these systems, the system should be trained using larger scale and higher quality sperm image data.

In some cases, spermatids are highly susceptible to sperm head damage, resulting in chromosomal aberrations, DNA fragmentation, and telomere shortening. Male fertility is inversely correlated with DNA fragmentation index (DFI), which is crucial for sperm selection. Techniques such as single-cell gel electrophoresis (SCGE), terminal deoxynucleotidyl transferase UTP nick-end labeling (TUNEL), and sperm chromatin structure assays are used to detect DNA fragmentation.

Scientists developed a machine learning algorithm by training the system with sperm images linked to relevant DFI values. This standardized system can accurately assess single sperm quality based on a trained dataset, thus eliminating concerns about human subjectivity.

Embryologists determine sperm motility using holographic imaging, computer-assisted sperm analysis (CASA), and microfluidic platforms. CASA is a high-throughput method that assesses large numbers of sperm at the sample level, but does not analyze simple sperm motility.

This can be resolved by developing a mathematical model based on the three-dimensional (3D) spiral motion of the tail beat. High-resolution holographic imaging techniques allow scientists to assess the tail beating patterns of free-swimming sperm.

Multiple data on sperm motility associated with CASA, microfluidic chips, and holographic imaging were used to train an AI system in combination with other male fertility parameters to select the best sperm for ART. will be Therefore, the application of AI and machine learning techniques has significantly improved the conception rate and pregnancy success rate after ART.