Key points are not available for this paper at this time.
Abstract Study question In the context of ICSI, how does the Sperm Tracer Deep Neural Network analysis compare to the embryologist’s sperm selection technique? Summary answer Sperm Tracer AI system significantly improves fertilization rates, offering a standardized, objective approach to sperm selection, but nuanced correlations highlight complexities in sperm parameters. What is known already Traditional sperm selection during ICSI relies on subjective, manual methods, introducing variability and affecting pregnancy success. In response, AI applications, aim to enhance objectivity, potentially improving pregnancy rates. Study design, size, duration Between September and December 2023, the study examined 26 fresh ICSI cycles in a total of 268 sibling oocytes. Men aged 18 years and above with 34.4 mean women’s age. The Sperm Tracer group underwent single embryo culture, while the Embryologists' group underwent group embryo culture. The study analyzes statistical differences in fertilization rates, day 3 and day 5 embryo development. The study duration aligns with the analyzed cycles. Participants/materials, setting, methods The study compared real-time AI-assisted sperm evaluation with the Sperm Tracer system to the traditional ICSI method by embryologists. Correlation analysis was performed for the following kinematic parameters: Amplitude of Lateral Head displacement (ALH), Dance (DNC), Linearity (LIN), Straightness (STR), Wobble (WOB), Curvilinear Velocity (VCL), Linear Velocity (VSL) and Average Path Velocity (VAP). Statistical analysis involved Odds Ratios (OR) with 95% confidence intervals, Mann-Whitney U tests and correlation analyses by Pearson’s coefficient. Main results and the role of chance The Sperm Tracer system exhibited a noteworthy enhancement in fertilization rates (OR: 0.46, 95% Confidence Intervals (CI) 0.23, 0.89, p = 0.0224) but no significant differences in cleavage stages, blastocyst formation, or the production of good-quality embryos compared to conventional methods. For correlation analysis used two groups of sperm population. Group A comprises spermatozoa subjected to ICSI with the Sperm Tracer system, resulting in unsuccessful oocyte fertilization and Group B consists of spermatozoa that achieved successful fertilization. In Group A, correlation analysis identified significant negative associations between DNC and linear parameters (LIN, STR, WOB) (p 0.001), highlighting the clinical relevance in sperm motility. Group B displayed a distinct negative correlation between DNC and LIN (p = 0.03941), indicating potential effects on sperm linear velocity. LIN in Group A exhibited strong negative correlations with WOB (p = 0.0001), moderate negative correlations with STR (p = 0.0028), and VCL (p = 0.0773). Moreover, ALH analysis in Group A unveiled robust negative correlations with LIN, STR, and WOB, suggesting clinical significance in sperm motility. These intricate findings underscore the Sperm Tracer system's primary advantage in optimizing fertilization rates, offering valuable insights for refining sperm selection strategies in assisted reproduction. Limitations, reasons for caution The study faces limitations due to a small sample, impacting generalizability. Divergence in culture embryo strategies introduces bias, necessitating validation in larger studies. The system includes morphology analysis, but no data analysis has been conducted. Exploring the correlation between morphology, considering embryo culture variations, and motility could provide valuable insights. Wider implications of the findings This study highlights AI applications like Sperm Tracer, showcasing their potential to improve assisted reproduction outcomes. Recognizing AI's impact on IVF success is vital for adoption. Integrating AI into clinic protocols may boost IVF efficiency, requiring training for embryologists to ensure effective utilization. Trial registration number Not applicable
Mantas et al. (Mon,) studied this question.