FEMI: Deep learning-assisted analysis of single-oocyte trapping and maturation for enhanced on-chip IVM

Traditional in vitro maturation (IVM) involves manual handling and static culture conditions and lacks automated assessment. These limitations increase variability, introduce stress that may affect oocyte quality, and fail to recapitulate the dynamic in vivo microenvironment. This study presents a system that streamlines, automates, and potentially enhances IVM procedures through the capture, dynamic on-chip cultivation, and evaluation of single oocytes on-chip, termed FemaleTech Enhanced Maturation Intelligence (FEMI). The platform employs a novel microfluidic chip featuring hexagonal micropillars that efficiently trap individual oocytes across a broad size range while inherently excluding low-quality oocytes, ensuring stable positioning throughout culture. Numerical simulations and experimental validation confirmed the trapping principle and high capture efficiency of the proposed design. FEMI supports continuous microfluidic culture under controlled flow conditions, allowing partial recapitulation of physiological transport and mechanical cues while maintaining oocyte immobilization. By integrating deep learning-based image segmentation, the system automatically quantifies cumulus cell expansion, providing single-cell-level insights into oocyte development. The results suggest that FEMI may improve IVM efficiency and provide a useful platform for assisted reproductive technologies.