Columbia’s STAR AI Detects Rare Sperm, Enables Conception After 19 Years

Post by : Bianca Suleiman

A new application of artificial intelligence has produced a notable advance in reproductive medicine. Columbia University researchers developed the Sperm Tracking and Recovery (STAR) platform to locate extremely rare sperm cells in men who had been considered infertile.

In its first reported success, STAR helped a U.S. couple achieve pregnancy after 19 years of unsuccessful attempts. The husband, aged 39, and his 37-year-old wife had undergone multiple IVF cycles and two surgical sperm retrievals with no viable sperm found.

STAR combines AI analysis with a non-invasive microfluidic device, scanning at a rate of over a million microscopic images per hour. When a semen sample initially appeared to contain no sperm, the system continued scanning and analysed more than 2.5 million images over two hours before identifying two viable sperm cells.

Men affected by conditions such as azoospermia—where semen shows no sperm—or cryptozoospermia—where sperm are present only in vanishingly small numbers—often face years of invasive procedures and few options beyond donor sperm or adoption.

The STAR system routes semen through a specially designed microchip while AI algorithms inspect the flow in real time, isolating candidate sperm into tiny sealed compartments. Those isolated cells can then be used in intracytoplasmic sperm injection (ICSI), where a single sperm is injected directly into an egg.

Compared with conventional manual searches and surgical extraction methods, STAR’s closed, automated workflow is faster and reduces handling. Components are single-use to minimise contamination risk and preserve sterility throughout the process.

If larger clinical trials confirm these early results, the technology could alter how clinics manage severe male infertility, potentially expanding opportunities for biological fatherhood previously thought unattainable.