L26/P-102 An autonomous Artificial Intelligence (AI) – Assisted Robotic System for Sperm Selection Resulting in Clinical Pregnancy
Human Reproduction · 01.07.2026 · 41 (Supplement_1)
DOI: 10.1093/humrep/deag083.438 ↗Kurz gefasst
Fachartikel in der Zeitschrift Human Reproduction, online erschienen am 08.07.2026. Verfasst von L. Ruoso und 3 weiteren Autor:innen.
Abstract (Originalsprache)
Study question Can an AI-integrated microrobotic system improve the accuracy and consistency of ideal sperm selection while reducing operator-dependent variability in intracytoplasmic sperm injection (ICSI)? Summary answer An autonomous AI-driven microrobotic system can identify, select, and isolate spermatozoa based on strict morphology criteria, micrometric size, motility, and epigenetic profile in clinical practice. What is known already Sperm selection is a critical step in ICSI and requires extensive embryologist training. Because the process relies on subjective visual assessment, outcomes can vary between operators and even from day to day. This variability, combined with time pressure and modest inter-observer agreement, can negatively affect results. Selecting a poor-quality spermatozoon may lead to abnormal early embryo development, fewer good-quality blastocysts, failed pregnancies, or, in rare cases, congenital anomalies. A fully automated workflow that replicates embryologist decision-making could reduce operator-dependent risks and improve the consistency and safety of sperm selection. Study design, size, duration To evaluate a device clinically within a program to automate ideal sperm selection for ICSI. Thousands of high-magnification sperm images were collected to train the AI, alongside the development of a microrobotic console replicating the standard procedure with sub-micron precision. After full-cycle validation runs and sperm videos offline evaluation by three embryologists, the clinical phase enrolled ICSI patients, using device-selected sperm to fertilize oocytes. Participants/materials, setting, methods The system utilizes supervised deep-learning models to detect and classify motile spermatozoa in real time. It was trained on over 20,000 high-magnification frames collected across multiple laboratories, each independently annotated by three senior embryologists. These algorithms also coordinate precision motors that drive proprietary robotic components with sub-micron accuracy. The platform integrates a micromanipulator, an X-Y stage, an AI-assisted high-definition camera with automated focusing, and a bidirectional pump. Main results and the role of chance Algorithms assessing morphology, size, and progressive motility enabled automated normal/abnormal sperm classification at different sperm orientations within a frame rate of 30 frames/second, with high reliability and precision values. To date, three couples with recurrent implantation failure (female ages 32, 37, 41) have undergone treatment using the device following ovarian stimulation with an antagonist. After operator confirmation of the required number of spermatozoa, the system autonomously performed sperm tracking, selection, aspiration, and relocation to a clean droplet without manual manipulation. Selected spermatozoa were subsequently verified under high magnification for strict morphology criteria and progressive motility before immobilization and ICSI. All injected oocytes were fertilized. Two patients underwent fresh embryo transfer, and one patient had two high-quality blastocysts vitrified. One clinical pregnancy of the 37-year-old female was reached, while the 41-year-old female did not conceive. These preliminary findings demonstrate the technical feasibility, safety, and clinical applicability of AI-driven robotic sperm selection, supporting consistent fertilization and embryo development in ICSI routine. Limitations, reasons for caution This report presents a clinical application of the device and documents the pregnancy achieved with AI-assisted sperm selection. A multinational randomized controlled trial is now underway to statistically evaluate its impact on fertilization, blastocyst development, and pregnancy rate outcomes. Wider implications of the findings An AI-driven autonomous robotic device may enable standardized, operator-independent sperm selection. By integrating morphology, motility, and epigenetic profile, this approach has the potential to improve reproducibility and clinical outcomes in ICSI routine. Trial registration number No
Zitieren
APA
Ruoso, L., Ferber, B. M., Shulman, A., & Cassuto, N. (2026). L26/P-102 An autonomous Artificial Intelligence (AI) – Assisted Robotic System for Sperm Selection Resulting in Clinical Pregnancy. Human Reproduction, 41(Supplement_1). https://doi.org/10.1093/humrep/deag083.438
Vancouver
Ruoso L, Ferber BM, Shulman A, Cassuto N. L26/P-102 An autonomous Artificial Intelligence (AI) – Assisted Robotic System for Sperm Selection Resulting in Clinical Pregnancy. Human Reproduction. 2026;41(Supplement_1). doi:10.1093/humrep/deag083.438
BibTeX
@article{Ruoso2026,
author = {Ruoso, L. and Ferber, B. Meiri and Shulman, A. and Cassuto, N.G.},
title = {L26/P-102 An autonomous Artificial Intelligence (AI) – Assisted Robotic System for Sperm Selection Resulting in Clinical Pregnancy},
journal = {Human Reproduction},
year = {2026},
volume = {41},
number = {Supplement_1},
doi = {10.1093/humrep/deag083.438}
}
EUVIMED