Columbia University has developed an artificial intelligence system that helps doctors locate sperm in men previously diagnosed with none, offering new hope for biological parenthood.
The method, called Sperm Track and Recovery (STAR), uses AI to scan semen or tissue samples for extremely rare sperm cells missed by standard lab tests. It targets azoospermia, affecting about 1% of all men and 10% of infertile men.
STAR combines imaging, AI, and robotics. Samples flow through microfluidic chips while a camera captures 300 images per second. A machine learning algorithm analyzes these images in real time to spot sperm among debris. A robot isolates the sperm within milliseconds, avoiding damage from spinning methods.
Dr. Zev Williams, director of the Columbia University Fertility Center, says the system found sperm in nearly 30% of patients tested who were previously told they had no chance. It identified 40 times more sperm than manual searches, achieving 100% sensitivity.
The first pregnancy using STAR was confirmed in 2025. The couple, Samuel and Penelope, had been trying to conceive for over two years. Samuel had Klinefelter syndrome, a genetic condition causing low sperm production. Penelope said, "It's starting to feel really real now."
Dr. Williams described the moment the team found sperm: "Everyone was just jumping up and down with joy. Now there's a baby girl and hopefully many more."