Extended in vitro culture of human embryos demonstrates the complex nature of diagnosing chromosomal mosaicism from a single trophectoderm biopsy

STUDY QUESTION What is the accuracy of preimplantation genetic testing for aneuploidies (PGT-A) when considering human peri-implantation outcomes in vitro? STUDY ANSWER The probability of accurately diagnosing an embryo as abnormal was 100%, while the proportion of euploid embryos classified as clinically suitable was 61.9%, yet if structural and mosaic abnormalities were not considered accuracy increased to 100%, with a 0% false positive and false negative rate. WHAT IS ALREADY KNOWN Embryo aneuploidy is associated with implantation failure and early pregnancy loss. However, a proportion of blastocysts are mosaic, containing chromosomally distinct cell populations. Diagnosing chromosomal mosaicism remains a significant challenge for PGT-A. Although mosaic embryos may lead to healthy live births, they are also associated with poorer clinical outcomes. Moreover, the direct effects of mosaicism on early pregnancy remain unknown. Recently, developed in vitro systems allow extended embryo culture for up to 14 days providing a unique opportunity for modelling chromosomal instability during human peri-implantation development. STUDY DESIGN, SIZE, DURATION A total of 80 embryos were cultured to either 8 (n = 7) or 12 days post-fertilisation (dpf; n = 73). Of these, 54 were PGT-A blastocysts, donated to research following an abnormal (n = 37) or mosaic (n = 17) diagnosis. The remaining 26 were supernumerary blastocysts, obtained from standard assisted reproductive technology (ART) cycles. These embryos underwent trophectoderm (TE) biopsy prior to extended culture. PARTICIPANTS/MATERIALS, SETTING, METHODS We applied established culture protocols to generate embryo outgrowths. Outgrowth viability was assessed based on careful morphological evaluation. Nine outgrowths were further separated into two or more portions corresponding to inner cell mass (ICM) and TE-derived lineages. A total of 45 embryos were selected for next generation sequencing (NGS) at 8 or 12 dpf. We correlated TE biopsy profiles to both culture outcomes and the chromosomal status of the embryos during later development. MAIN RESULTS AND THE ROLE OF CHANCE Of the 73 embryos cultured to 12 dpf, 51% remained viable, while 49% detached between 8 and 12 dpf. Viable, Day 12 outgrowths were predominately generated from euploid blastocysts and those diagnosed with trisomies, duplications or mosaic aberrations. Conversely, monosomies, deletions and more complex chromosomal constitutions significantly impaired in vitro development to 12 dpf (10% vs. 77%, P < 0.0001). When compared to the original biopsy, we determined 100% concordance for uniform numerical aneuploidies, both in whole outgrowths and in the ICM and TE-derived outgrowth portions. However, uniform structural variants were not always confirmed later in development. Moreover, a high proportion of embryos originally diagnosed as mosaic remained viable at 12 dpf (58%). Of these, 71% were euploid, with normal profiles observed in both ICM and TE-derived lineages. Based on our validation data, we determine a 0% false negative and 18.5% false positive error rate when diagnosing mosaicism. Overall, our findings demonstrate a diagnostic accuracy of 80% in the context of PGT-A. Nevertheless, if structural and mosaic abnormalities are not considered, accuracy increases to 100%, with a 0% false positive and false negative rate. LIMITATIONS REASONS FOR CAUTION The inherent limitations of extended in vitro culture, particularly when modelling critical developmental milestones, warrant careful interpretation. WIDER IMPLICATIONS OF THE FINDINGS Our findings echo current prenatal testing data and support the high clinical predictive value of PGT-A for diagnosing uniform numerical aneuploidies, as well as euploid chromosomal constitutions. However, distinguishing technical bias from biological variability will remain a challenge, inherently limiting the accuracy of a single TE biopsy for diagnosing mosaicism. STUDY FUNDING, COMPETING INTEREST(S) This research is funded by the Ghent University Special Research Fund (BOF01D08114) awarded to M.P., the Research Foundation-Flanders (FWO.KAN.0005.01) research grant awarded to B.H. and De Snoo-van't Hoogerhuijs Stichting awarded to S.M.C.d.S.L. We thank Ferring Pharmaceuticals (Aalst, Belgium) for their unrestricted educational grant. The authors declare no competing interests. TRIAL REGISTRATION NUMBER N/A.