05-07-2025
AI helps couple achieve pregnancy after 18 years of failed IVF cycles: Fertility expert explains how tech can make this happen
After 20 years, a couple in the US is expecting to be parents as the woman has achieved pregnancy after a series of unsuccessful attempts of in-vitro fertilisation (IVF), where the eggs and sperm are fertilised outside the body in a laboratory and the resulting embryo is transferred into the woman's uterus. That was because the man had very low sperm. And even the little that was there was hidden and undetectable by an embryologist.
But now thanks to an artificial intelligence system, previously undetectable sperm can not only be identified but harvested for use. The STAR (Sperm Tracking and Recovery) system, developed at Columbia University Fertility Centre in New York, is expected to help men with azoospermia, a condition where no sperm is easily found in the semen. This technology is a breakthrough because it can harvest hidden semen, has a non-invasive extraction method and can offer an option to men battling fertility issues. This method costs under $3,000 (Rs 2.6 lakh).
'This is an example where technology multiplies human efforts. Even embryologists currently do the same thing. But now AI can assist them in identifying the sperm even better,' says Dr Anjali Malpani, IVF specialist and founder of Malpani Infertility Clinic.
A typical semen sample usually has millions of sperm. But some men have really low counts, so much so that they can't be seen even after centrifugation (spinning the semen sample to concentrate the sperm) and hours of meticulous searching under a microscope. This is different from ejaculatory duct disorders, which prevent sperm from being ejaculated despite being produced in healthy numbers in the testes, usually due to a blockage. Then we use a needle to extract the healthy sperm.
Azoospermia is diagnosed when no sperm can be found in at least two separate semen samples analyzed after centrifugation. So we study multiple ejaculates, which means collecting and analyzing multiple semen samples to confirm the absence of sperm. We usually collect the samples in quick succession, within an hour or two. This is done because sperm production can fluctuate, and a single semen analysis might not accurately reflect the overall situation. Multiple samples help account for these variations. Then an embryologist studies them to see if they can be used for IVF or not.
Now the semen sample is placed on a chip under a microscope. AI connects to the microscope through a high-speed camera and with high-powered imaging technology scans it. The STAR system takes more than 8 million images in under an hour to find what it has been trained to identify as a sperm cell. It can isolate sperm faster in the sample, allowing embryologists to recover cells that they could not have just with their eyes. So AI is amplifying the embryologist's vision.
Of course, each IVF case has different complexities which require the experience of a trained embryologist. But in resource-scarce settings and in the absence of a good embryologist, AI can scan sperm and eggs based on the dataset of thousands of samples that have been fed into them. Still I would say that an embryologist can grade embryos — assessing the quality and developmental potential of embryos — better.
