Preimplantation genetic testing for aneuploidy: A review of published blastocyst reanalysis concordance data

Abstract Preimplantation genetic testing for aneuploidy (PGT‐A) reduces miscarriage risk, increases the success of IVF, shortens time to pregnancy, and reduces multiple gestation rates without compromising outcomes. The progression of PGT‐A has included common application of next‐generation sequencing (NGS) from single nucleotide polymorphism microarray, quantitative real‐time PCR, and array comparative hybridization platforms of analysis. Additional putative advances in PGT‐A capability include classifying embryos as mosaic and predicting the presence of segmental imbalance. A critical component in the process of technical validation of these advancements involves evaluation of concordance between reanalysis results and initial testing results. While many independent studies have investigated the concordance of results obtained from the remaining embryo with the original PGT‐A diagnosis, compilation and systematic analysis of published data has not been performed. Here, we review results from 26 primary research articles describing concordance in 1271 human blastocysts from 2260 pairwise comparisons. Results illustrate significantly higher discordance from PGT‐A methods which utilize NGS and include prediction of mosaicism or segmental imbalance. These results suggest caution when considering new iterations PGT‐A.


| INTRODUCTION
Preimplantation genetic testing for aneuploidy (PGT-A) is a proven intervention in the treatment of infertility, with decreased clinical miscarriage risk, increased delivery rates from the first embryo transfer attempt, and reduced risk of multiple gestation without compromising success rates. [1][2][3][4][5] 5 However, the diagnostic predictive value of PGT-A, when considering embryonic mosaicism, [6][7][8] or segmental imbalance 9,10 remains controversial and has involved considerable resources to investigate.
Several commercially available PGT-A methods involve predicting the presence of mosaicism from a single trophectoderm biopsy. These methods primarily rely upon classifying intermediate chromosome copy numbers as falling in a 'mosaic range'. 7 Mosaicism may primarily originate during embryonic mitotic cell division and chromosome nondisjunction, leading to daughter cells with different chromosomal constitution (trisomy, disomy, and/or monosomy). 6 Contemporary methods of quantitation of chromosomes from a multicell sample (ie, trophectoderm) may identify mosaicism by observing intermediate copy number (ie, 2.6 instead of 3 in the case of mosaic trisomy) indicating a mixture of euploid and aneuploid cells. Different methods of classification have been proposed, including 20% to 80% (2.2-2.8 mosaic range) and 30% to 70% (2.3-2.7 mosaic range). 11,12 Several studies suggest that embryos classified as mosaic have reduced reproductive potential when compared to embryos classified as euploid. [13][14][15][16][17] These clinical outcome data have largely been used as evidence for the validity of mosaicism diagnoses. Others

| RESULTS
Thirty-seven studies published in peer-reviewed journals were retrieved after the search was performed. From these, one study performed a retest in leftover amplified DNA, six performed a retest in embryo spent culture media only, two involved testing products of conception, and two studies failed to provide necessary data for analysis. As a result, 26 published studies including 24-chromosome PGT-A and analysis of two or more blastocyst-derived samples were identified (Table 1). 9,16,27-50 A total of 1271 embryos were included in this review. In addition, 147 embryos were from patients who had an abnormal karyotype were used for analysis of positive controls.  Table 2 and Figures 1 and 2), which were all significantly different (P < .05). Data compiled from each study for concordance analysis is available in Table S2.
A series of three subset analyses were performed ( Table 3). The first subset analysis involved evaluation of concordance data from published studies using NGS in the original biopsy compared to studies not using NGS in the original biopsy (

| Whole embryo concordance analysis
Although a chromosomal mosaicism diagnosis is often provided after analysis of a single trophectoderm biopsy, more rigorous and stringent criteria for this categorization involve the evaluation of more than one biopsy from the same embryo. 6    One perspective on these results is that this is completely expected. mosaic, it may be possible that they were actually uniformly aneuploid.
Transferring these embryos and observing reduced reproductive potential cannot rule out this interpretation without direct reanalysis of the remaining embryo. In the present study, approximately one-third of mosaic embryo was found to possess a full aneuploidy upon reanalysis. This is Conversely, in cases where an embryo is classified as 'low range' mosaic, it may be possible that they were actually uniformly euploid.
Transferring these embryos and observing healthy deliveries cannot rule out this interpretation without direct reanalysis of the remaining embryo. In the present study, approximately one-third of mosaic embryos were found to possess uniform euploidy upon reanalysis. This is consistent with clinical outcomes, where approximately onethird lead to healthy deliveries. 17 Blastocyst mosaicism is undoubtedly a real phenomenon. While the use of intermediate copy numbers from a single biopsy alone may be insufficient to accurately predict the remaining embryo, there may be several useful innovations to consider. Given that mitotic nondisjunction is a commonly observed mechanism leading to mosaicism, 51,52 methods that can distinguish meiotic and mitotic origins of aneuploidy (from a single biopsy) may improve specificity of mosaicism predictions. In addition, given that mosaicism originates during post-zygotic cell division events and is commonly associated with micronuclei, time-lapse morphokinetics may also be instrumental in improving specificity. 53 Significantly different discordance rates were also observed when evaluating the use of segmental imbalance predictions. Very little data on the clinical predictive value of segmental imbalance has been developed.
Given the challenges already faced with regard to interpretation of clinical outcomes of putative mosaic embryos, and based on this review, more