JBRA Assisted Reproduction 2025;29(Suppl.2 SBRA 2025):45
Poster Presentation
29^th Annual Congress of the SBRA. São Paulo/SP - Brazil, 2025
doi: 10.5935/1518-0557.20263584

P-33. Beyond Gardner and KIDScore: Fragmentation of the Inner Cell Mass and Trophectoderm as the Strongest Predictor of Embryo Aneuploidy

Hamilton de Martin¹, Eduardo Gomes Sá¹, Ellayne Cavalcanti Queiroz¹, Andrea Mesquita Lima¹, Gleicyane Sousa Santos Alam¹, Renata Reis Pimentel Castro¹, Maria Clara Parente Torquato¹, Fabio Eugênio Magalhães Rodrigues¹, Eduardo Paula Miranda¹, Sebastião Evangelista Torquato¹

¹Bios Centro de Medicina Reprodutiva – Fortaleza - CE - Brasil

Objective: To determine whether the presence and extent of fragmentation in the inner cell mass (ICM) and trophectoderm (TE), assessed using the Blastocyst Fragmentation Indicator (BFI) by time-lapse imaging, provides superior prediction of embryo aneuploidy when compared to conventional morphological grading (Gardner system) and automated morphokinetic scoring (KIDScore).
Methods: This retrospective cohort study was conducted at a private IVF center employing continuous time-lapse culture and routine PGT-A. A total of 1,182 embryos cultured individually in time-lapse incubators and biopsied for PGT-A between September 2022 and December 2024 were analyzed. Embryos were classified by the Gardner system into groups A (comprising AA, AB, BA), B (BB, BC, AC), and C (CA, CB, CC). KIDScore categories (KIDScore D5 v3.1) were defined as A (KIDScore ≥7), B (KIDScore 4–7), and C (KIDScore <4). BFI categories were defined as NN (no fragmentation), FN (fragmentation of ICM only), and FF (fragmentation of both ICM and TE). Aneuploidy rates were compared across groups, and predictive performance was assessed using logistic regression, odds ratios, and receiver operating characteristic (ROC) analysis. Cross-classification analyses were performed to evaluate the independence and overlap of BFI, Gardner, and KIDScore assessments. This analysis represents a focused segment of a larger, ongoing investigation using the same time-lapse embryo cohort, with additional morphodynamic parameters being addressed in parallel studies.
Results: Fragmentation assessed by BFI was a significantly stronger predictor of embryo aneuploidy than either Gardner or KIDScore. The AUC for BFI was 0.791 (95% CI 0.76–0.82), significantly higher than both the Gardner system (AUC 0.688) and KIDScore (AUC 0.697). BFI sensitivity and specificity for aneuploidy prediction were 0.854 and 0.637, respectively. Euploidy rates progressively declined with increasing fragmentation: 62.5% in NN, 20.1% in FN, and only 6.4% in FF embryos. Fragmentation was associated with markedly lower odds of euploidy (OR for FN vs. NN: 0.157 [95% CI 0.11–0.23]; FF vs. NN: 0.047 [0.03–0.09]), stronger than the corresponding effect sizes for KIDScore (C vs. A: OR 0.278 [0.20–0.39]) and Gardner (C vs. A: OR 0.335 [0.24–0.47]). Cross-classification revealed that while most Gardner/KIDScore A embryos were NN, a considerable proportion of B and C embryos were FN or FF. This indicates that BFI identifies chromosomal risk not captured by traditional scores. Multivariable analyses confirmed BFI as the strongest independent predictor of aneuploidy, even after adjusting for maternal age and standard morphological parameters.
Conclusion: Fragmentation of the inner cell mass and trophectoderm, objectively assessed using the BFI, outperforms conventional Gardner and KIDScore systems in predicting embryo aneuploidy. These findings challenge the reliance on established morphological and morphokinetic criteria alone for embryo deselection. Incorporating fragmentation analysis into embryo assessment protocols provides greater accuracy for chromosomal risk stratification, supporting a shift toward multidimensional embryo selection strategies in assisted reproduction. Further studies are warranted to elucidate the biological mechanisms underlying fragmentation, as well as to explore the potential of integrating BFI into AI-driven embryo selection algorithms.