Warning: file_get_contents(/home/clients/3220c9f5d3e264c4b48dd5661e2a9f12/sites/river-journal.com/wp-content/uploads/journal/2025/2.3/2025.2.3_month.txt): failed to open stream: No such file or directory in /home/clients/3220c9f5d3e264c4b48dd5661e2a9f12/sites/river-journal.com/wp-content/themes/divi-child/functions.php on line 337

VOLUME 2 - NUMBER 3 - 2025

Deep learning enhances Day 3 embryo evaluation: validation of new software in IVF laboratories

  • Marco Toschi, Lucia Maresca, Lorena Bori, Camilla Coianiz, Giovanni Pizzi , Matteo Verdiglione, Galliano Daniela, Marcos Meseguer
  • Original Article, 30-33
  • Full text PDF

  • Introduction: During day 3 of development, embryos are classified according to morphological parameters: symmetry of cell divisions, fragmentation degree, division rate, number of cells. However, this practice is subject to intra and inter-operator variability. The need for more objective and standardised tools has stimulated interest in artificial intelligence (AI). New Software (Vitrolife’s iDA v2.0) algorithm is an AI system designed to evaluate embryo image. Materials and methods: This single-center retrospective study analyzed 1,418 embryos from 245 IVF cycles (own and donor oocytes, with or without PGT) with the use of Embryoscope®️ system (Vitrolife, Denmark). Results: Higher iDAScore on day 3 are significantly linked to increased blastocyst formation (5.46±1.44 vs. 3.91±1.51; p<0.001). Blastocyst formation rates increased across higher iDAScores: ≤3.1 (15.7%), 3.2–4.6 (36.6%), 4.7–6.1 (60.6%), and ≥6.2 (74.1%) (p<0.01). Elevated iDAScores on day 3 also correlated with better inner cell mass (ICM) and trophectoderm (TE) quality. A moderate positive correlation was observed between Day 3 iDAScores and Day 5/6 (Pearson’s r = 0.52, p<0.001), with higher Day 3 scores predicting higher Day 5/6 scores. Conclusion: This suggest that Day 3 iDAScore may be as a promising predictor of balstocyst formation and later embryo quality, offering more objective and standardized laboratory assessments.

  • KEY WORDS: Embryo selection, artificial intelligence (AI), IDAScore v2.