“This tool has the ability to give patients who have very little chance of fathering their own biological children an increased chance.”
Results presented at the 39th annual meeting of the European Society of Human Reproduction and Embryology (ESHRE) unveil a groundbreaking artificial intelligence (AI) tool capable of rapidly identifying sperm in severely infertile men within seconds, a process that traditionally takes scientists hours.
The algorithm developed by the study authors brings hope to men desiring biological children but experiencing a lack of sperm in their semen.
Presently, such patients undergo a biopsy procedure where a portion of their testes is extracted to aid in their journey to fatherhood. Embryologists manually extract sperm from the biopsy sample to fertilize the partner’s eggs using Intracytoplasmic Sperm Injection (ICSI) treatment.
This laborious process of locating and isolating sperm in human tissue can consume up to six hours, which may hinder the embryologist’s ability to identify sperm due to mental and physical fatigue.
However, data from this recent study conducted by Australian experts reveal that trained AI can alleviate clinicians from this challenging task. The findings demonstrate that the AI tool, called SpermSearch, swiftly detects the presence of sperm, enabling the embryologist to determine its viability for ICSI. Moreover, the results indicate that the algorithm outperforms an experienced clinician in terms of accuracy.
“This tool has the ability to give patients who have very little chance of fathering their own biological children an increased chance,” adds Mr. Dale Goss, the lead author from the University of Technology Sydney.
“The algorithm improves antiquated approaches that have not been updated in decades. It will ensure the rapid identification of sperm in samples, which will not only increase the chance of a couple conceiving their own biological children, but also reduce stress on sperm and increase efficiency in the laboratory.”
Non-obstructive azoospermia (NOA), the most severe form of male infertility where no sperm is present in the semen, affects approximately one percent of all men and is encountered by around five percent of couples seeking fertility treatment.
To locate sperm for Intracytoplasmic Sperm Injection (ICSI), embryologists currently employ manual techniques involving the partial dissection of tissue samples and meticulous examination using forceps or fine needles. Any detected sperm is then released into a specially prepared liquid in a petri dish.
With the help of a microscope, clinicians painstakingly search through minute droplets of the liquid, a challenging task due to potential contamination from other cells and particles. If the embryologist overlooks the presence of sperm, it diminishes the patient’s chances of parenthood, and the longer this process takes, the higher the likelihood that the sperm may not be viable.
The objective of the study was to investigate whether AI could expedite this process. The research was conducted over five months in two phases at an IVF clinic in Sydney, utilizing AI software installed on a computer. Initially, the algorithm was trained by exposing it to thousands of microscope images, featuring highlighted sperm amidst a backdrop of various cells and debris.
This training enabled the AI tool to gradually learn, through image analysis, to independently recognize the characteristics of sperm, employing its own evaluation system to refine and optimize its performance.
Lead author Mr. Goss and his team utilized both healthy sperm samples and testicular tissue samples from seven patients, aged 36 to 55, diagnosed with NOA, who had previously undergone surgical sperm retrieval at the clinic. The patients contributed tissue remnants prepared for sperm retrieval but remained unused.
The subsequent test involved a simultaneous comparison between the algorithm and an embryologist with a recognized precision level of 100%. The researchers assessed the time taken by both the AI tool and the embryologist to identify sperm, as well as the accuracy of their respective identifications.
Results revealed that the AI tool detected a greater overall number of sperm, although some were identified solely by the embryologist, and others solely by the AI tool.
The embryologist identified 560 sperm, the AI tool detected 611, and collectively they identified a total of 688 sperm. The algorithm required less than a thousandth of the time taken by the embryologist to identify sperm in each examined droplet area.
Moreover, the AI tool demonstrated higher accuracy and precision in identifying sperm, discovering an additional 60 sperm compared to the embryologist and achieving a five percent higher accuracy per viewable droplet area.
During their conference presentation, the authors emphasized that the study represents a proof-of-concept test and emphasized the necessity of conducting a clinical trial. The purpose of the trial would be to validate the utility and effectiveness of the technique in accurately detecting sperm.
They further noted the importance of conducting similar research among men experiencing different forms of severe infertility and undergoing alternative surgical procedures, such as sperm retrieval from different regions of the testes.
Professor Carlos Calhaz-Jorge, the chair of ESHRE, who was not involved in the research, commented on the study: “For men diagnosed with non-obstructive azoospermia, ICSI with sperm retrieved from the testicles is the only realistic chance of having biological children. This is a preliminary study on the use of AI for finding healthy sperm in men experiencing this type of infertility.
“Finding healthy sperm under the microscope in fragments of testicular biopsies can be an arduous process. The prospect of using AI to make the process quicker and more accurate is very interesting. We need to see more research to build on these results.”
Image Credit: Shutterstock