Overview
- The STAR (Sperm Tracking and Recovery) system, detailed in The Lancet by Zev Williams’s team at Columbia University, integrates high-speed imaging, microfluidics and deep-learning detection.
- In the case report, a sample read as sperm-free by manual microscopy was processed as STAR analyzed about 2.5 million images in roughly two hours and detected seven sperm, two of them motile.
- The two motile sperm were injected into two mature oocytes via ICSI, leading to embryos that produced a positive test and a confirmed clinical pregnancy.
- Reports describe STAR’s throughput in the millions of frames per hour and a robotic step that extracts candidate sperm in milliseconds for immediate use or freezing.
- Broader clinical studies are underway to assess performance and safety across more patients, with clinicians emphasizing AI as a complement to skilled embryologists.