A novel artificial intelligence (AI) application has demonstrated the capacity to pinpoint sperm within profoundly infertile men in moments, a significant improvement over the several hours traditionally required by researchers, as disclosed at the 39th annual assembly of the European Society of Human Reproduction and Embryology (ESHRE).
Image Credit: Explode / Shutterstock
The researchers express optimism in the algorithm they've engineered, promising a glimmer of hope to males desiring to father their own children, yet lack any trace of sperm in their semen.
At present, these patients must undergo an operation where a segment of their testicles is surgically removed to facilitate them in becoming fathers. Embryologists manually retrieve sperm from this biopsy sample to fertilize the female partner's eggs via Intracytoplasmic Sperm Injection (ICSI) treatment. This painstaking process of locating and isolating sperm in human tissue can take as long as six hours, which can exhaust the embryologist's mental and physical capacities, potentially impairing sperm identification.
Information from this groundbreaking study by Australian specialists illustrates that AI, once trained, can relieve clinicians of this laborious task.
Evidence suggests that the tool, named SpermSearch, instantly identifies sperm, subsequently leaving it to the embryologist to verify the presence of sperm and its suitability for ICSI. Outcomes show that the algorithm exhibits greater accuracy than a seasoned clinician.
The study's primary author, Mr. Dale Goss of the University of Technology Sydney, commented: "This tool grants the potential for those with minuscule prospects of fathering their biological children to have an enhanced chance.
"The algorithm refines outdated methodologies that have remained stagnant for decades. It promises swift identification of sperm in samples, augmenting not only the likelihood of a couple conceiving their biological offspring, but also lessening the strain on sperm and augmenting efficiency in the laboratory."
Approximately one percent of all men exhibit no sperm in their semen, a manifestation of the most severe type of infertility known as non-obstructive azoospermia (NOA). This affliction is prevalent in around five percent of couples seeking fertility treatments.
Embryologists, to pinpoint sperm for ICSI, must partially dissect tissue samples and delicately separate them using forceps or thin needles. Any present sperm is then transferred into a specially prepared liquid, which is situated in a petri dish.
Using a microscope, the clinician combs through droplets of this solution, inspecting tiny fractions at a time. However, this task becomes increasingly challenging due to interference from other cells and particles. Missing the sperm decreases the patient's chance to become a parent, and the longer this process extends, the more likely the sperm will become nonviable.
The goal of the research was to investigate if AI could expedite this process. The study was implemented at an IVF clinic in Sydney in two stages spanning five months using AI software installed on a computer. The team initially trained the algorithm by presenting it with thousands of static microscope photographs, only highlighting the sperm amidst a high concentration of other cells and debris.
This guidance allowed the AI application to eventually learn to distinguish sperm via image analysis using its performance evaluation and adjustment system.
Mr. Goss and his team employed healthy sperm and samples of testicular tissue from seven patients aged between 36 and 55 years. All had received an NOA diagnosis and had previously undergone surgical sperm retrieval at the clinic.
The patients donated leftover tissue post-treatment, prepared for sperm retrieval but ultimately not required.
The algorithm and an embryologist, whose precision was assumed to be 100%, concurrently underwent a test. The researchers compared the time each took to identify sperm and their accuracy levels.
The results revealed that the AI located more sperm in total, although some were only detected by the embryologist and some solely by the AI.
The embryologist discovered 560 sperm, the AI located 611, and collectively they found 688. The algorithm spotted sperm in each droplet area it analyzed in less than a 1000th of the time taken by an embryologist.
It also proved more accurate and precise in identifying sperm – the AI tool located 60 more sperm and was five percent more accurate than the embryologist per observable droplet area.
In their conference presentation, the authors emphasize that the study is predicated on a proof-of-concept test and necessitates a clinical trial. This is required to validate the technique's utility and successful execution of sperm detection.
They suggest that further research should be conducted among males with other severe infertility forms and those undergoing different surgical procedures, like sperm collections from varying testes regions.
ESHRE's chair, Professor Carlos Calhaz-Jorge from the Northern Lisbon Hospital Centre and the Hospital de Santa Maria in Lisbon (Portugal), unassociated with this research, remarked: "For males diagnosed with non-obstructive azoospermia, ICSI with sperm retrieved from the testicles presents the only realistic possibility of fathering biological children. This preliminary study on AI's utilization in locating healthy sperm in men grappling with this form of infertility is intriguing.
"Identifying healthy sperm under the microscope in testicular biopsy fragments can be a taxing process. The potential of employing AI to hasten and enhance the process's accuracy is quite intriguing. We anticipate more research to build on these findings."
AI Reveals EU Tourists Prioritize Environmental Over Social Sustainability in Travel Choices