COPENHAGEN, Denmark — Tracking down the issues contributing to male infertility may get a whole lot easier soon. Researchers have presented a new artificial intelligence (AI) tool that’s capable of finding viable sperm within severely infertile men in seconds. Applying the AI algorithm could give hope to men who want a biological child but have no success in conceiving naturally.
The current process normally takes six hours to find and isolate sperm from human tissue. It involves undergoing a procedure where a portion of their testes is removed. Embryologists then take out sperm manually from this sample to fertilize their partner’s eggs through intracytoplasmic sperm injection (ICSI) treatment.
The long hours it takes to extract sperm can play a role in successfully identifying viable samples as embryologists might experience mental and physical fatigue. Embryologists need to shred tissue samples and tease them apart with fine needles. Any sperm available is then placed into a special liquid in a petri dish. Using a microscope, the clinician searches through droplets of the liquid to examine the sperm. However, contamination from other particles and cells can make it possible for an embryologist to miss a sperm. Additionally, the longer the process takes, the greater the chance the sperm is no longer usable.
In the current study, researchers used AI to see if it could speed up the process. The tool is called SpermSearch and was installed on a computer in an IVF clinic over a five-month period. They first trained the algorithm to identify sperm by showing thousands of still microscope photographs. Each picture contained sperm and high amounts of cells and debris, but only the sperm was highlighted.
The AI would eventually learn through images what a sperm looks like in different settings. The team used healthy sperm and samples of testicular tissue from seven patients between 36 and 55 years-old. All were diagnosed with non-obstructive azoospermia where there is no sperm in their semen. Around one percent of men have this severest form of infertility and the condition is seen in around five percent of couples looking for fertility treatments.
The AI tool and an embryologist — whose accuracy in examining sperm was considered 100 percent — took the test at the same time. Findings showed the algorithm was more accurate than an experienced clinician. Although there were some instances where sperm was only spotted by the embryologist and vice versa.
The AI tool did find 60 more sperm and was five times more accurate at identifying sperm in the special liquid. The data from this study shows that AI can spare healthcare providers this laborious process. This gives more time for the embryologist to evaluate whether the sperm is a good sample for ICSI treatment.
“This tool has the ability to give patients who have very little chance of fathering their own biological children an increased chance,” says lead author Dale Goss, a researcher at the University of Technology Sydney, in a media release. “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.”
While an amazing feat, the authors caution this was a proof-of-concept test and further clinical trials are needed to prove the technique is useful in a real-life setting. Additionally, the algorithm would need to be tested on other forms of infertility or on people who are undergoing other surgical approaches.
“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,” says Carlos Calhaz-Jorge, a professor at the Northern Lisbon Hospital Center and the Hospital de Santa Maria in Lisbon, Portugual who was not involved in the research.
The study was presented at the 39th annual meeting of the European Society of Human Reproduction and Embryology.