Welcome back to Neural Notes, a weekly column where I look at some of the most interesting AI news of the week. In this edition: Australian medtech startup NeoGenix Biosciences has just achieved its first pregnancy through the use of ‘SpermSearch’ AI – which is used to help men with infertility issues.
According to NeoGenix Biosciences, there’s been a 50% decrease in male sperm count worldwide over the past 40 years. The company says this is due to a variety of factors, including pollution, diet and lifestyle choices.
SpermSearch aims to help combat this through the use of AI. The end-to-end development of SpermSearch AI took NeoGenix almost two years, which included assistance from the UNSW Founders Health 10x program, which provided networks and funding to expedite the process.
According to NeoGenix co-founder and clinical embryologist at IVF Australia, Dale Goss, AI has huge potential to assist in the reproductive health space. With male infertility in particular, the process can be labor-intensive and time-consuming.
“Traditional sperm identification is completely manual, where clinicians will spend upwards of five hours searching for sperm on a microscope. That process is lengthy, expensive, and diminishes sperm quality as it dies outside of the human body,” Goss said to SmartCompany.
Comparatively, NeoGenix claims that SpermSearch AI reduces search time by 75% and finds more sperm than unassisted embryologists.
“With AI-assisted technology, one embryologist found 12 sperm in one-third of the time it took two embryologists without AI to find only three sperm,” Goss said.
The AI was trained using thousands of micrograph images, initially using donor sperm, blood cells, and cancer cell lines. After receiving ethical approval, the team obtained images of testicular samples from men undergoing biopsies. This extensive dataset ensured the model could handle the variety of sperm found in clinical settings.
“Our model was trained using representative micrograph images of what we would normally see in the clinic,” Goss said.
Unsurprisingly, the variability in sperm and differences in images from various microscopes posed challenges for the company.
“Another major hurdle is that the biopsies tissue samples can vary a lot between surgeries, especially in men with severe infertility,” Neogenix co-founder and CEO Steven Vasilescu said to SmartCompany.
“Because of this, we worked hard to create a strong dataset with tens of thousands of different sperm to make sure the model can handle the variety of sperm found in the clinic.”
“This affects how the AI identifies sperm in the images. To fix this, we took pictures using various microscopes and made many adjustments to the images, like flipping, inverting, changing contrast, blurring, and altering colours. These adjustments help the model work better and can be easily set up in any clinic.”
Despite the hurdles, SpermSearch AI has now seen its first successful pregnancy working with one of the most severe male cases of infertility in Australia.
“It was a very difficult case and without the use of SpermSearch AI and support from IVFAustralia, we don’t think the couple would’ve been able to conceive,” Vasilescu said.
“Thanks to our cutting-edge AI technology, we were able to identify four times more sperm than unassisted embryologists. This led to the creation of two embryos, one of which was transferred and resulted in a clinical pregnancy.”
At the time of writing it was still early days, with the pregnancy sitting at seven weeks and the due date set for mid-2025.
“We are as overjoyed as they are,” Vasilescu said.
AI technologies like SpermSearch are expected to play an increasingly important role in fertility treatments.
“We believe that AI will help analyse the data we have in clinical settings and standardise the industry. Our overall goal in the next 5-10 years would be to see SpermSearch AI be used for all procedures where sperm selection is needed,” Goss said.
According to NeoGenix both the medical community and patients have responded positively to SpermSearch and it has been receiving interest from clinics worldwide.
“Many embryologists are really excited about using this technology. It can greatly reduce the time it takes to do a very repetitive task and improve the chances of success for couples who often have very low odds of having their own biological children,” Goss said.
Never miss a story: sign up to SmartCompany’s free daily newsletter and find our best stories on LinkedIn.
Senior Technology Journalist
A multi-award-winning journalist, Tegan has been reporting on technology, telecommunications and gaming for the past 10 years. She regularly appears on mainstream media channels as an expert tech commentator. Her previous roles include editor of Gizmodo Australia and global reviews editor at Finder. She has also served as a reporter and editor across Business Insider, Lifehacker and Kotaku Australia. Tegan was the co-host of one of the top technology podcasts in Australia, Queens of the Drone Age. You can find her on Twitter @Tegan_Writes.
