The definition of mosaic embryos and their management after PGT-A

Published 12 November 2021

Another day, another controversy in PGT-A, as leading groups appear in dispute over the definition of ‘mosaic’ based on a chromosomal copy number threshold, the idea that mosaic embryos might ‘self-correct’, and how the degree of mosaicism is related to clinical outcome.

Twenty years after preimplantation genetic testing for aneuploidy edged into routine use, the technique as a means to improve live birth rates remains controversial. The latest committee opinion from the ASRM concludes that its value as a ‘universal screening test for all IVF patients has yet to be determined’; in the UK the HFEA awards PGT-A a red light in its assessment of add-ons, meaning there’s no evidence from RCTs to demonstrate a benefit.

While this controversy runs and runs, another is now raising its head. This, as reflected in two reports in the November issue of Fertility and Sterility, concerns mosaicism as detected by PGT-A in preimplantation blastocysts and its consequences for clinical outcome. In one of the two F&S ‘views and reviews’, PGT specialist Nathan Treff lists three subjects which he considers of contention in this controversy: ambiguity in defining ‘the true frequency’ of chromosomal mosaicism; whether aneuploidies can ‘self-correct’; and how clinics should manage those embryos diagnosed as ‘mosaic’.(1)

According to the second of the two papers, today’s state-of-the-art platforms for PGT-A now offer the high resolution needed for the detection of mosaicism, which is based on copy number of chromosomes.(2) The authors explain that a PGT-A result with a chromosomal copy number of 2 would indicate normal disomy, and a copy number of 3 would indicate trisomy, and so on. Thus, add the authors, a chromosomal copy number of 2.4 would suggest that 40% of the cells are trisomic and 60% are disomic, indicating a ‘mosaic trisomy at the 40% level’.

Before the introduction of multicellular blastocyst biopsy, which only became a reasonable alternative after the introduction of such comprehensive technologies as next generation sequencing (NGS), a single blastomere analysed by PGT-A could not make such distinctions – the cell was usually deemed either normal or abnormal. But now, mosaicism exists in the details of today’s PGT-A - and presents a clinical challenge for those managing such embryos.

The application of a copy number threshold for the detection of mosaicism is one of Treff’s ‘misconceptions’ in PGT-A, described by him as an ‘arbitrary’ definition and an inaccurate predictor of true mosaicism. He and a colleague write: ‘The misconception that mosaicism can be accurately predicted in the preimplantation embryo and misinterpretation of simple copy number profiles have led to significant embryo wastage, loss of financial resources, confusion among physicians and embryologists, and, most importantly, harm to patients.’ They go even further, recommending that the practice ‘should be avoided’ and that labs should limit their predictions to simply euploid or aneuploid. Moreover, those studies which have analysed outcomes following the transfer of ‘mosaic’ blastocysts ‘should reconsider the validity of the original mosaicism diagnoses’.

However, such a back-to-the-future move as this is described as ‘obsolete’ by Viotti et al, basing much of their opinion on a study report from several high-profile groups on the clinical outcomes of transferring 1000 mosaic blastocysts and how those outcomes might help formulate a ranking system for mosaicism in PGT-A. Not only did the study include such a large number of blastocysts (1000 mosaic and 5561 euploid controls) but also used a standardised NGS platform and a previously proposed mosaic threshold level (20-80%) to define results. As expected, the mosaic transfers were associated with a lower LBR than the euploid, but there was also among them various ‘traits’ of mosaicism which affected outcomes in a statistically significant way. Thus, wrote the authors, combining this mosaic level with embryo morphology ‘revealed the order of subcategories regarding likelihood of positive outcome’ and ‘enabling the formulation of an evidence-based prioritization scheme for mosaic embryos in the clinic’. Not surprisingly, the title of the second F&S review which draws heavily on this study begins: ‘Let the data do the talking.’

Included in the centres of the 1000 mosaic embryo study (and of the F&S review) was that of Francesco Fiorentino in Rome who in a 2015 letter to the New England Journal of Medicine had described six healthy deliveries in a small series of 18 women for whom embryo screening had found no chromosomally normal embryos.(4) Until then, mosaic embryos were not usually transferred in IVF because they (like all other aneuploid embryos) were considered abnormal. Even in their NEJM letter, the Rome investigators noted that ‘it is reasonable to assume that mosaicism reduces the likelihood of success of IVF’ – which indeed the 1000 mosaic embryo study confirmed. The latter also seemed to confirm later findings from Fiorentino’s group that outcome (LBR and miscarriage) following the transfer of mosaic blastocysts would largely depend on the proportion of chromosomally abnormal cells. It was from such studies that the idea of embryonic self-correction arose and that the mosaic embryo might now be considered a ‘distinct category’ in terms of potential to implant.

The argument from their F&S review is therefore that mosaicism in general and its specific features as detected by PGT-A are indeed associated with variable clinical outcomes – and that ‘mosaicism should be considered for more informed and improved embryo selection in the clinic’. This, however, remains a long way from the view of Treff and Marin, who argue that, until better methods in the accuracy of predicting mosaicism are available, embryos formerly deemed ‘mosaic’ should be now reclassified as ‘putative mosaic’ or ‘consistent with possible mosaicism’ – or ‘misclassified’ when actually uniformly euploid or uniformly aneuploid.

As ever with PGT-A, this looks like a debate poised to run and run, with a heavy emphasis from conflicting interests. To help us along, an overview editorial from F&S declines to take sides, but does accept that ‘mosaic embryo transfer can lead to live birth of healthy infants and that we must rethink the implications of calling an embryo mosaic’.(5) Moreover, the editorial adds, as the use of NGS for PGT-A increases, ‘centers should discuss the potential for the mosaic embryo as a diagnosis during the informed consent process’ and ensure they have ‘written policies on the selection of embryos for transfer’.

This debate seems sure to be revisited in ESHRE’s forthcoming recommendations for good practice on chromosomal mosaicism in ART, which ESHRE’s SIG Reproductive Genetics is presently finalising. The recommendations will be drawn from a survey on current practice and from published data. The draft of the recommendations will be available for stakeholder review via the ESHRE website in early 2022.

1. Treff NR, Marin D. The ‘mosaic’ embryo: misconceptions and misinterpretations in preimplantation genetic testing for aneuploidy, Fertil Steril 2021; 116: 1205-1211.
2. Viotti M, McCoy RC, Griffin DK, et al. Let the data do the talking: the need to consider mosaicism during
embryo selection. Fertil Steril 2021; 116; 1212-1219.
3. Viotti M, Victor AR, Barnes FL, et al. Using outcome data from one thousand mosaic embryo transfers to formulate an embryo ranking system for clinical use. Fertil Steril 2021; 115: 1212-1224.
4. Greco E, Minasi MG, Fiorentino F. Healthy babies after intrauterine transfer of mosaic aneuploid blastocysts, N Engl J Med 2015; 373: 2089-2090.
5. Feinberg E. Planting the seed of doubt: the diagnosis and management of mosaic embryos. Fertil Steril 2021; 116: 1203-1204.

Get notified of new articles with our ESHRE newsletter.

Sign up and never miss an update