Correct assessment and interpretation of results determines the accuracy of any diagnostic test, and PGT-A is no exception

Sir, Barad et al. (2022) report live births from blastocysts with ‘abnormal’ results, which were previously denied transfer. While the authors suggest that ‘mosaic’ embryos should be transferred, in using the term ‘abnormal’, they inappropriately blur the lines between diagnoses of uniform whole chromosome aneuploidy, segmental aberrations and ‘mosaicism’. Indeed, six of the eight live births in their cohort resulted from the transfer of blastocysts classified as ‘mosaic’, another was from a segmental aneuploid blastocyst, and only one from a blastocyst diagnosed with uniform whole chromosome aneuploidies by preimplantation genetic testing for aneuploidy (PGT-A) (Barad et al., 2022). The live birth potential of ‘mosaic’ blastocysts has been reflected in a previous report of over 1000 ‘mosaic’ embryo transfers, even showing that ‘mosaic’ blastocysts with better morphology can fare better than euploids with poor morphology (Viotti et al., 2021). Transfer of ‘mosaic’ blastocysts in the absence of an euploid alternative is already the recommended approach since 2016 (Preimplantation Genetic Diagnosis International Society, 2016; Practice Committee and Genetic Counseling Professional Group (GCPG) of the American Society for Reproductive Medicine, 2020). True mosaicism is only one possible explanation for the observation of ‘intermediate copy numbers’ by PGT-A (Paulson and Treff, 2020). The positive predictive value of a positive test result depends on the underlying prevalence of the condition, and since the prevalence of true biological mosaicism is expected to be <1% in human preimplantation blastocysts (Treff and Marin, 2021), it is possible that only a fraction of blastocysts diagnosed with intermediate copy numbers by PGT-A are true mosaics. The publication of only one single case of postnatally confirmed mosaicism diagnosed by PGT-A attests to this fact (Kahraman et al., 2020). Moreover, differences in diagnosing mosaicism across PGT-A providers raise considerable concerns regarding the accuracy of ‘mosaicism’ predictions. High variability in putative mosaicism rates across providers, ranging from 2.9% to as high as 23.9% (Popovic et al., 2022) points to technical artifacts, and highlights varying practices in interpreting and reporting ‘mosaic’ results. Accordingly, reporting mosaicism may primarily come at the expense of euploid diagnoses.