Columbia researchers report first successful pregnancy using AI-guided sperm recovery method

Researchers at the Columbia University Fertility Center reported the first successful pregnancy using an AI-guided method they developed to recover sperm in men with azoospermia, in which ejaculate contains little or no sperm. 

The case is described in a research letter published in The Lancet. 

Male factors account for approximately 40% of couples with infertility. Of those, about 10-15% of men with infertility have azoospermia. 

A semen sample can appear totally normal, but when you look under the microscope you discover just a sea of cellular debris, with no sperm visible. Many couples with male-factor infertility are told they have little chance of having a biological child." 

 Zev Williams, senior author of the paper and Director of the Columbia University Fertility Center

Men with azoospermia may undergo a procedure to have sperm surgically extracted from the testes, but the procedure is often unsuccessful and can cause vascular problems, inflammation, or a temporary decrease in testosterone levels.   

A few specialized labs employ technicians to manually inspect semen samples - a lengthy and expensive process - after they have been processed with a centrifuge or other agents that can damage sperm. 

"The field has really been challenged to find a better way to identify and retrieve viable sperm cells in men with exceedingly low sperm counts," Williams says. 

A STAR is born 

Williams assembled a team of researchers and clinicians to develop a new method that combines a variety of technologies to identify and retrieve rare sperm cells from men with azoospermia. 

"Our team included experts in advanced imaging techniques, microfluidics, and reproductive endocrinology to tackle each individual step required to find and isolate rare sperm," says Hemant Suryawanshi, an assistant professor of reproductive sciences at Columbia University Vagelos College of Physicians and Surgeons and project leader. 

The STAR (Sperm Tracking and Recovery) method, unveiled earlier this year, employs high-powered imaging technology to scan through a semen sample from men with azoospermia, taking over 8 million images in under an hour. AI is used to identify sperm cells in the sample, and a microfluidic chip with tiny, hair-like channels isolates the portion of the semen sample containing the sperm cell. Within milliseconds, a robot gently removes the sperm cell so that it can be used to create an embryo or frozen and stored for future use. 

First successful pregnancy using STAR 

STAR was tested in a patient that had been trying to start a family for nearly 20 years, including multiple IVF cycles at other centers, several manual sperm searches, and two surgical procedures to extract sperm.

The patient provided a 3.5 mL semen sample. In about two hours, STAR scanned 2.5 million images, identifying 2 viable sperm cells, which were then used to create two embryos and start a pregnancy.

The findings, though based on one case, show the feasibility of this technology to overcome long-standing barriers in helping men with azoospermia. 

"You only need one healthy sperm to create an embryo," Williams says. 

Larger clinical studies are underway to evaluate the efficacy of STAR in broader patient populations.