Advancing ICSI Robotics: Follow-Up on the Transition from Remote to Fully Automated Intracytoplasmic Sperm Injection
Mendizabal-Ruiz et al., American Society for Reproductive Medicine (ASRM) 2025 Scientific Congress & Expo
Objective
To evaluate the performance and clinical efficacy of a fully automated, AI-driven robotic system for intracytoplasmic sperm injection (ICSI) that requires only human supervision and minimal intervention.
Materials and Methods
The automated ICSI system integrates an inverted microscope, motorized stage, micromanipulators, laser system, piezo-actuator, and microinjectors, all governed by proprietary AI-driven software. All ICSI steps (sperm selection, laser immobilization, oocyte positioning, zona pellucida breaching, piezo-mediated oolemma rupture, and sperm deposition) are executed upon a single command, with human supervision and ability to pause, correct, and resume when necessary.
This single-center pilot study enrolled six eligible IVF/ICSI patients (18-39 years) who provided informed consent (ClinicalTrials.gov NCT06074835). Exclusion criteria included uterine pathology, severe male infertility (e.g., azoospermia), genetic conditions, and suboptimal stimulation. Sibling oocytes were randomly assigned to experimental (n=29) or control (n=24) groups. The experimental group underwent ICSI with a fully automated AI-driven robotic system, while the control group received conventional manual ICSI. Embryological outcomes were compared by Fisher’s Exact Test. The automation percentage was defined as the proportion of cases where each step was executed fully autonomously with no operator intervention.
Results
The automated system achieved varying levels of automation across different procedural steps: sperm selection (100%), sperm immobilization (89%), sperm loading (38%), egg holding (93%), laser-assisted partial zona pellucida ablation (93%), and oolemma penetration with sperm deposition (66%). The mean injection time was 4.6 minutes per oocyte.
Post-ICSI survival rates were 90% (26/29) in the experimental group and 83% (20/24) in the control group (p=0.69). Normal fertilization occurred in 66% (19/29) of the experimental group and 71% (17/24) of the control group (p=0.77). The usable blastocyst rate was 53% (10/19) in the experimental group and 65% (11/17) in the control group (p=0.52). Good quality blastocyst formation rates, according to Gardner's score (BB or superior), were 32% (6/19) vs. 41% (7/17) (p=0.73). Early clinical outcomes include three ongoing pregnancies from six embryos transferred in the experimental arm.
Conclusions
The fully automated ICSI system demonstrated success rates comparable to conventional manual ICSI while significantly reducing operator dependency. Although human supervision remains necessary for quality control and occasional error correction, this system improves automation from robot assistance to task autonomy, representing a significant advancement in assisted reproductive technology.
Impact Statement
This study demonstrates the first successful clinical application of a fully automated, AI-driven ICSI system, potentially increasing accessibility to high-quality ICSI worldwide by reducing reliance on highly skilled embryologists while improving procedural consistency and efficiency.
Support
Financial SUPPORT: Institutional