O-077 Time-lapse imaging analysis of segmental aneuploid embryos: a multicenter study identifies morpho-kinetic patterns associated with chromosomal mosaicism

Abstract

Abstract Study question Do segmental aneuploid (SA) embryos display unique morpho-kinetic patterns at early developmental stages? Summary answer The first three cell division cycles of embryos carrying segmental aneuploidies are significantly slower than euploid and whole-chromosome aneuploid (WCA) embryos. What is known already Segmental aneuploidies in blastocyst-stage human embryos prevalently originate (i.e., 70% of cases) from mitotic errors. Their clinical management is challenging because it is impossible to establish their meiotic/mitotic origin from PGT-A analysis of a single trophectoderm biopsy. Morpho-kinetic analysis based on Time-lapse microscopy (TLM) has been proposed as a valuable tool for systematic embryo evaluation, although it is unclear whether this approach alone can predict chromosomal abnormalities. This study aims to pinpoint morphokinetic patterns associated with the occurrence of segmental aneuploidies and/or mitotic errors, possibly improving DNA-based diagnostic interpretation and their clinical management. Study design, size, duration This is a retrospective multicenter study including a total of 7,693 embryos from 2,370 IVF cycles and 2,127 couples, cultured between 2016 and 2021 in 4 European IVF centers. Embryos with no more than 4 chromosomal alterations were considered in the analysis, resulting in 3,288 euploids, 3,155 WCA, 715 SA, and 535 complex aneuploids (CA). Overall, the dataset contained 3,742 distinct euploid-SA embryo sibling pairs. Participants/materials, setting, methods Standard morpho-kinetic features were annotated using various TLM systems. Blastocysts were subjected to comprehensive chromosomal screening via PGT-A. Morpho-kinetic timings across different embryo groups were compared using Kolmogorov-Smirnov (KS) test and binomial test, and associations with cleavage features were assessed via two-sided Fisher Exact (FS) test. Multi-center and center-specific logistic regression models were calibrated using a subset of data (70%), and their predictive performance was evaluated on independent test data using Area-Under-ROC curve (AUROC) metrics. Main results and the role of chance SA embryos cleaved at significantly slower rate than their euploid siblings in the first 3 cell cycles, from the appearance of the second blastomere t2 (average delay delta=0.49h, KS p = 0.006, fraction of pairs in which SA was delayed compared to euploid sibling F = 54.2%, binomial p = 0.002) up to t8 (delta=2.16h, KS p = 4e-4, F = 57.9%, binomial p = 1e-8). The developmental delay was partially compensated at full blastocyst stage tB (delta=1.71h, KS = 0.0008, F = 57.5%, binomial p = 5e-8). SA timings also differed from WCA ones: SA embryos were significantly slower than their WCA siblings during early cell cycles, with maximum delay reached at t8 (delta=1.67h, KS p = 0.009, F = 55.4%, binomial p = 0.003), but they fully caught-up at tB, with no significant difference observed. These morpho-kinetic patterns were consistently observed in each center separately, but with different effect sizes. The presence of segmental aneuploidies was significantly associated with multinucleation (OR = 1.99, FS test p = 0.05) and morula cell exclusion (OR = 2.69, p = 0.006). A logistic model based on morpho-kinetic data and cleavage features from a single center dataset and regressed against embryo aneuploidy class (euploid vs. SA embryos) displayed adequate predictive performance on independent data from the same center (AUROC=0.70), although predictivity diminished when tested on data from other centers (AUROC=0.52-0.55). Limitations, reasons for caution We cannot rule out inter-center variability in the annotation procedures of morpho-kinetic parameters. Standardized and automated annotation would increase the performance of predictive models and their transferability. Further work is needed to assess whether the morpho-kinetic signal is driven by mitotic errors and if an association with clinical outcomes exists. Wider implications of the findings These findings might help uncovering biological events contributing to preimplantation mosaicism. Moreover, the developed predictive framework might help improving decision-making in PGT-A cycles, providing interpretation for segmental aneuploidies and possibly helping in the evaluation of embryos showing intermediate copy number values for whole chromosomes that are suggestive of mosaicism. Trial registration number Not applicable