AI and fertility: An artificial intelligence system has helped a US couple conceive after nearly two decades of infertility by locating two viable sperm cells in a sample previously deemed azoospermic, according to a correspondence in The Lancet.
The couple, a 39-year-old man and a 37-year-old woman, had undergone multiple IVF cycles and two surgical sperm-retrieval attempts without success. “A semen sample can appear totally normal, but under the microscope, you see a sea of cellular debris, with no sperm visible,” said senior author Zev Williams, director of the Columbia University Fertility Center.
Also Read | Are IVF babies as healthy as naturally conceived ones?
Researchers tested a new method called Sperm Tracking and Recovery (STAR), which pairs high-powered imaging with AI to perform high-speed, real-time searches for rare sperm. In this case, a 3.5-mL semen sample was processed and scanned; while manual examination found none, the STAR system analysed 2.5 million images in about two hours and identified seven sperm (two motile, five non-motile). Microfluidic channels isolated the target fraction, and a robotic micromanipulator extracted the cells.
The two motile sperm were injected into two mature oocytes, leading to embryo development. On day three, the embryos were transferred; 13 days later, the patient had her first positive pregnancy test, which progressed to a confirmed clinical pregnancy. At eight weeks, the ultrasound showed normal fetal development with a heartbeat of 172 bpm.
Azoospermia affects men whose ejaculate contains little or no sperm; surgical retrieval is often invasive and not always successful. The authors say the single-case result demonstrates feasibility and could reduce reliance on surgery for some patients, though larger clinical trials are underway to determine how broadly STAR can help. “Many couples with male-factor infertility are told they have little chance of having a biological child,” Williams noted, adding that AI may offer a new path forward for at least a subset of them.
