SNS 기사보내기
(Credit: Getty Images)
(Credit: Getty Images)
In a significant advancement for reproductive medicine, researchers at Seoul National University Bundang Hospital (SNUBH) have developed an artificial intelligence system that substantially improves the success rate of in vitro fertilization (IVF) treatments. The technology, which enhances embryo selection accuracy, offers new hope to countless couples struggling with infertility.
Led by Professor Lee Jung-ryeol from the Department of Obstetrics and Gynecology, the research team collaborated with Kai Health to create this innovative digital therapeutic device. Their achievement has been recognized as one of the top 10 representative initiatives for 2025 under the prestigious government-led "Pan-Ministerial Full-Cycle Medical Device Research and Development Project."
The AI system addresses one of the most critical challenges in IVF treatment -- identifying which embryos have the highest potential to result in a successful pregnancy. Conventional embryo selection methods, which rely on clinical embryologists using microscopes, achieve a pregnancy prediction accuracy of approximately 37 percent. The new AI-powered technology dramatically increases this to about 65 percent.
"As infertility and delayed pregnancies become increasingly prevalent societal concerns, we have successfully advanced AI-powered embryo selection technology to a level suitable for commercialization," said Professor Lee. "We hope this technology will significantly reduce the physical, psychological, and financial burden of IVF for many couples."
How the technology works
Unlike previous solutions that were limited to specific cultivation systems, this new AI approach offers broader applicability. The system works with embryo images taken from standard microscopes, making it highly versatile across different clinical settings.
Professor Lee Jung-ryeol of SNUBH
Professor Lee Jung-ryeol of SNUBH
When asked about the decision-making process behind the AI's evaluations, Professor Lee explained that the exact analytical methods used by the AI aren't entirely transparent.
"It is clear that it analyzes forms similar to what research technicians observe and judge,” Lee said. “During this form of analysis, it appears to draw conclusions by analyzing additional elements that humans cannot analyze."
The technology significantly improves accuracy and delivers rapid results, dramatically reducing the time needed for embryo assessment compared to traditional methods.
However, Professor Lee notes that some time is still required for capturing embryo images, suggesting that "automation of this aspect is necessary to apply the embryo selection solution more quickly and conveniently."
Beyond selection: comprehensive benefits
The benefits of AI implementation extend beyond mere accuracy improvements. Professor Lee highlights that, like AI applications in other fields, this technology offers consistent results and significant time savings. Most importantly, it creates a more level playing field in fertility treatment.
"When beginners use AI assistance, they can achieve results equal to or higher than those of highly experienced professionals," Professor Lee said. "Even for experienced professionals, AI can help reduce possible mistakes or be referenced in decisions to provide more accurate determinations."
Despite the impressive results, Professor Lee cautions that the technology has inherent limitations. "Pregnancy in in-vitro fertilization is not determined solely by embryo quality, so 100 percent pregnancy prediction is impossible with embryo selection alone," he said. "Therefore, it is necessary to add other clinical variables to further increase the pregnancy prediction rate."
The research team acknowledges that different hospitals have varying environments and embryo cultivation conditions.
To address this variability, Professor Lee recommends calibrating the system for each hospital before its initial use.
He further speculated that as AI selection systems become more widely used in everyday practice, "the embryo cultivation conditions may be standardized to match the embryo selection solution."
Helping high-risk groups
For high-risk groups such as women with decreased ovarian function or those who have experienced repeated implantation failures, AI technology offers specific benefits. While Professor Lee clarifies that "artificial intelligence cannot improve ovarian function or embryo quality," it can identify embryos with higher pregnancy potential among those generated.
"By prioritizing the transplantation of these embryos and excluding the transplantation of embryos with significantly lower pregnancy potential, ultimately, the time to pregnancy can be shortened," he said.
This is particularly crucial for patients with decreased ovarian function who produce fewer embryos, making it essential to understand each embryo's pregnancy potential. Similarly, for patients experiencing repeated implantation failures, selecting quality embryos plays a vital role in preventing further unsuccessful attempts.
Having successfully completed clinical trials and obtained regulatory approval from the Ministry of Food and Drug Safety, the technology has also secured certifications in Europe and Singapore. These achievements pave the way for imminent clinical adoption and real-world application.
tweet
Related articles
SNUBH rises to 68th in Newsweek’s 2025 Global Hospital Rankings, strengthening global reputation
SNUBH Professor Jeong named associate editor of Translational Stroke Research
SNUBH professor participates in study to redefine obesity as a chronic disease
Smartphone app improves anticoagulant adherence in atrial fibrillation patients
Lee Han-soo corea022@docdocdoc.co.kr
See Other Articles
Copyright © KBR Unauthorized reproduction, redistribution prohibited
News That You Haven't Seen