Automation of sperm preparation, oocyte retrieval and denudation, ICSI, and vitrification in IVF: insights from a proof-of-concept study

STUDY QUESTION Can key embryological procedures, including sperm preparation, oocyte retrieval and denudation, ICSI, and blastocyst vitrification, be automated? SUMMARY ANSWER Individual automated systems enabling sperm preparation, oocyte retrieval and denudation, ICSI, and embryo vitrification demonstrated feasibility in a small proof-of-concept study. WHAT IS KNOWN ALREADY Automated systems for specific IVF procedures have shown promise, with recent reports of successful births following automated gamete isolation and ICSI. STUDY DESIGN, SIZE, DURATION Prospective proof-of-concept study evaluating five automated systems nicknamed ‘Pearls’. Each system automated one of the following procedures: sperm preparation via swim-up (Pearl 1-SU); sperm preparation by a dish-based approach called horizontal swim-out (Pearl 1-HSO), oocyte retrieval, and denudation (Pearl 2); sperm selection and laser immobilization and piezo-ICSI (Pearl 3); and blastocyst vitrification (Pearl 5). The study enrolled patients undergoing treatment at Hope IVF, Guadalajara, Mexico, between October 2023 and October 2024, with data collection on pregnancy outcomes being completed in October 2025. PARTICIPANTS/MATERIALS, SETTING, METHODS Thirty-two patients or oocyte donors each underwent a single egg retrieval following minimal or mild ovarian stimulation. One patient was excluded from the study due to poor ovarian response. Oocytes were prospectively allocated to treatment with a single automated system: Pearl 1-SU (n = 20), Pearl 1-HSO (n = 37), Pearl 2 (n = 34), Pearl 3 (n = 57), with sibling oocytes treated manually per standard protocol. In the case of the automated vitrification system (Pearl 5), sibling blastocysts produced following manual ICSI were assigned to either manual or automated (n = 10) vitrification. MAIN RESULTS AND THE ROLE OF CHANCE Pearl 1-SU and Pearl 1-HSO automated sperm preparation systems achieved fertilization rates of 90% (18/20) and 73% (27/37), and sibling oocytes injected with manually prepared sperm achieved 60% (6/10) and 75% (18/24) fertilization. Usable blastocyst rates were 55.5% (10/18) and 70.4% (19/27) for the automated systems, and 50% (3/6) and 66.7% (12/18) for manual preparation. Pearl 2 automated oocyte retrieval and denudation resulted in a fertilization rate of 79.4% (27/34) and a usable blastocyst rate of 48.1% (13/27); manual procedures achieved 66.7% (26/39) and 76.9% (20/26) fertilization and blastocyst formation rates, respectively. Pearl 3 automated ICSI resulted in a fertilization rate of 59.6% (34/57), while this rate in sibling oocytes injected manually was 80% (36/45). Usable blastocyst formation rates were 58.8% (20/34) for automated and 61.1% (22/36) for manual ICSI. Thirty-five single embryo transfers from the automation group resulted in 16 clinical pregnancies and 16 live births in 26 patients (Pearl 1-SU = 3/6; Pearl 1-HSO = 4/7; Pearl 2 = 4/7; Pearl 3 = 4/13; Pearl 5 = 1/2). Two extremely preterm deliveries in the Pearl 1 group, associated with severe pre-eclampsia and vaginal/urinary infection, respectively, led to neonatal and infant deaths. No transfers involved embryos from the manual group for Pearl 1. One embryo transfer from the manual group of Pearl 2 resulted in a negative pregnancy. In the Pearl 3 manual group, four single blastocyst transfers in four patients resulted in four live births; in the Pearl 5 manual control group, three transfers in three patients resulted in two negative pregnancy results and one early pregnancy loss. LIMITATIONS, REASONS FOR CAUTION The small sample size and pilot design of this study do not allow statistical comparisons between sibling oocytes from automated and manual treatment groups. The automated systems did not achieve complete autonomy, and several tasks required direct operator intervention or execution through digital control. Pearl 5 encountered technical limitations that require further development before additional clinical evaluation. Although the two extremely preterm births that led to neonatal and infant deaths occurred in the context of clear obstetric complications, they compel collection and analysis of extensive technical and biological data from the automated systems in future studies and close monitoring of all pregnancies and births resulting from these novel procedures. WIDER IMPLICATIONS OF THE FINDINGS These results show the feasibility of integrating individual automated systems into an IVF laboratory as experimental standalone tools, allowing a stepwise evaluation. The modular approach, once shown to be efficacious, may provide a practical pathway toward broader laboratory automation and enable laboratories to address specific bottlenecks and standardize operations. FUNDING The study was sponsored by Conceivable Life Sciences, New York, NY, USA. DISCLOSURES A.C.-B. is an employee, shareholder, and company officer at Conceivable Life Sciences and a shareholder in IVF 2.0 and Hope IVF Mexico. A.V.A. has no conflicts of interest to declare. G.M.-R. is a stock option holder in Conceivable Life Sciences and has received consulting fees from Conceivable Life Sciences and IVF 2.0. A.F.-S.