🧬 PART FIVE REPRODUCTIVE CYTOGENETICS

584  REPRODUCTIVE CYTOGENETICS 3. reproductive failure following the natural loss, by miscarriage, of a confirmed pregnancy; and 4. reproductive failure due to fetal death in later pregnancy. BIOLOGY Embryology A knowledge of embryology underpins an understanding of certain processes of chromosomal pathology, and the counselor should be familiar with some basic terms. While not every detail of anatomy of the embryo is requisite knowledge, it is a minimum requirement to appreciate the different roles of the inner cell mass/embryoblast, and of trophoblast/chorionic villi, in their separate paths of development giving rise to the embryo proper and to the placenta, respectively. The counselor should understand the evolution of gametes → zygote → cleavage-stage embryo → morula → blastocyst → embryo → fetus, according to the following broad stages and as illustrated in Figures 20–2 and 20–3.4 • The zygote is the cell that arises from fertilization of an egg by a sperm, and this cell can be called a conceptus (“that which is conceived”). This cell then undergoes a series of cell divisions, and this cell mass, by day 2–3, is the cleavage-stage embryo; the expression conceptus can continue to be applied. The bunch of cells at day 4 is a morula. Figure 20–1.  The Iceberg of Chromosomal Pregnancy Loss. 4 Strictly speaking, in utero life is divided into three periods: pre-embryonic (the first 2 weeks after conception), up to formation of the primitive streak; embryonic (to the end of the 8th week) when the body forms and organs are largely constructed; and fetal (from 8 weeks to term), characterized by growth and changes in proportion rather than the appearance of new features. In clinical and everyday parlance, pregnancy timing is dated from the last menstrual period, and so, with conception usually happening mid-cycle, the first 2 weeks of a 40-week pregnancy are “non-pregnant,” and a normal pregnancy is really only of 8½ months. REPRODUCTIVE FAILURE  585 • At days 5 to 12 post-conception, the conceptus is a blastocyst. The “cyst” in that name refers to the fluid-filled cavity within a sphere of cells (Figure 20–2 left, Day 6), this outer layer of cells comprising the trophoblast,5 which will eventually develop into chorionic villi, forming the placenta. A group of cells within the cyst, attached to the inner surface of the sphere, is the inner cell mass/embryoblast, and it is from these cells that the actual embryo will arise. During about day 7, implantation occurs: that is, the attachment of the conceptus to the lining of the uterus. • After implantation, the embryonic tissue types (ectoderm, endoderm, mesoderm) differentiate from the epiblast (Figure 20–2 right, Germ layers), and then the basic body and organ formation is laid down, through to the 8-week mark. • From 8 weeks to birth is the fetal stage. Gametic Cytogenetics The logical starting point in an assessment of fertility is the gamete, and from a chromosomal point of view it is the process of meiosis, the final act of gametogenesis, to which we address our main focus. Many more sperm are made than eggs, by orders of magnitude, and one might have expected a higher standard of meiotic fidelity in the scarcer gamete (Hunt and Hassold 2002); but in fact, it is the other way round, and so it is the egg that commands most of our attention in terms of the practical relevance of gametic chromosomal pathology. OÖCYTES AND POLAR BODIES In vitro fertilization (IVF) is widely applied in the management of infertility, and one consequential benefit of this has been the access afforded to study of the oöcyte and its minor partner, the polar body. Many eggs sampled at IVF prove to be surplus to the requirements of the couple, and they are often willingly donated for research. Several studies addressing the chromosomal status of oöcytes of 46,XX women have shown that a remarkably substantial fraction of these gametes are aneuploid, and the rate is markedly related to the woman’s age (Samura et al. 2023). Comparing euploidy rates to age gives an inverted U-shaped curve, with aneuploidies seen more in the very young and in older women, and euploidy peaking at 75% in the late 20s (Figure 20–4). In women in their mid-30s, about a half of eggs are disomic or nullisomic, whereas by around age 40 the fraction rises to a remarkable three-quarters, and with an astonishing high of up to 90% in the mid-40s. These rates are influenced by the woman’s ovarian reserve, as measured by the anti-Müllerian hormone levels: the higher the level, the better the odds for euploidy (La Marca et al. 2022). The typical frequencies of aneuploidies per individual chromosome in the eggs of women of older childbearing age6 are depicted in Figure 20–5. It is the smaller chromosomes that are the more vulnerable to error, as the tail to the right of the graph attests. In the data of Tschare et al. (2023), analyzing oöcytes from women of age range 25 to 42, the chromosomes most often seen in aneuploid state were, by a substantial margin, numbers 15, 16, 21, and 22. The molecular basis of the maternal age link—and knowing 5 This is also the tissue—trophectoderm—that is taken at the time of preimplantation genetic testing (PGT) biopsy, typically on day 5–6. 6 Older women are more likely to seek IVF, and so this is where most data have come from. 586  REPRODUCTIVE CYTOGENETICS Figure 20–2.  The timeline of embryogenesis, and the development of the different tissues. Note: “Embryoblast” (day 6) is synonymous with Inner Cell Mass, and these cells will give rise to the actual embryo. The unlabeled black and white cells shown at days 8 and 9 (upper right in each panel) represent decidua placentalis. Source: From R Essers et al., Prevalence of chromosomal alterations in first-trimester spontaneous pregnancy loss, Nat Med 29:3233–3242, 2023. Courtesy A Salumets and MZ Esteki, and with the permission of Nature Medicine. REPRODUCTIVE FAILURE  587 Figure 20–2.  Continued.

588  REPRODUCTIVE CYTOGENETICS that the first meiotic division remains in a “holding pattern” for decades, from birth to ovulation—may lie in a vulnerability of, inter alia, the aging kinetochore and the spindle mechanism (Zielinska et al. 2019; Mikwar et al. 2020). This is strikingly illustrated in the work of Battaglia et al. (1996): their images of oöcytes show how the structural integrity of the meiotic apparatus declines as a woman gets older (Figure 20–6). Meiosis is under genetic control, and ~20 known Mendelian conditions are associated with mutation in relevant genes that cause oöcyte maturation arrest (Sang et al., 2021; Zhang et al. 2022; Zhang et al. 2024b; Lin et al. 2025). SPERM The gamete whose chromosomes are more readily accessible to analysis is the spermatozöon, and very large numbers can be analyzed. In most men, 1%–2% of sperm are chromosomally imbalanced (Bell et al. 2020) (Figure 20–7). In general, there is no Figure 20–3.  The windows of observation available to the researcher, along the timeline of chromosomal interest from gametes to child. Only the 3-8 week embryonic period remains substantially unstudied. Note: Evidently, time on the x axis is depicted asymmetrically. FDIU, fetal death in utero. Figure 20–4.  Aneuploidy Rates in Oöcytes by Age. Note: These data were derived from ovarian tissue taken from girls and women at the time of “fertility-insurance” storage, ahead of their undergoing potentially damaging chemotherapy; and from eggs donated from patients attending an IVF clinic. Note that the best chance for euploidy is seen in the late twenties. See also Figure 13-10 for more detail. Source: Adapted from JR Gruhn and ER Hoffman, Errors of the egg: The establishment and progression of human aneuploidy research in the maternal germline, Annu Rev Genet 56:369–390, 2022. Courtesy ER Hoffmann, and with the permission of Annual Reviews. REPRODUCTIVE FAILURE  589 discernible correlation of either full aneuploidy or deletion with paternal age (Bonus et al. 2022; Samarasekera et al. 2023). As with the egg, a number of Mendelian loci are known, mutation at which leads to meiotic arrest and non-obstructive spermatogenic failure (Kasak and Laan 2021; Xie et al. 2022). It might have been thought that fathers of aneuploid children, in whom the aneuploidy had been of paternal origin, could have represented a group with a particular susceptibility to producing aneuploid sperm, but in fact this appears usually not to be the case (Hixon et al. 1998; Blanco et al. 1998). Nevertheless, at least in the case of some fathers of Down syndrome or Turner syndrome children, a propensity to produce aneuploid or disomic sperm has been noted, and it may be that that a minority of normally fertile men are predisposed to meiotic errors at spermatogenesis, whether generalized or restricted to one chromosome (Martin 2008). A case is proposed for chromosome 9 heteromorphism as a susceptibility factor (Mottola et al. 2023), but not supported in Hernandez-Nieto et al. (2021). Cytogenetics of the Preimplantation and Embryonic Periods An aneuploid gamete (nullisomic or disomic) will lead to an aneuploid conceptus (monosomic or trisomic). A diploid gamete combining with a normal gamete will give rise to a triploid conceptus. With many oöcytes and some sperm aneuploid, and simplistically supposing equal fertilizing/fertilizable capacity, the expectation is for a very Figure 20–5.  Aneuploidies in oöcytes of women of ages 36-40. Disomies per chromosome are shown above the 0% baseline, nullisomies below. Note: The mirror-like appearance, above and below the line, reflects the agency of meiotic error, with essentially equal likelihoods of the missegregating chromosome going to the egg (for disomy), or to the polar body (for nullisomy); the similar observation of equal numbers of disomies and nullisomies in seen also in Oberle et al. (2024). The smaller chromosomes are the more susceptible to aneuploidy. Light green = chromatid gain, darker green = 2-chromatid chromosome gain; Light red = chromatid loss, darker red = 2-chromatid chromosome loss. These findings are based on inferences from the analysis of the first and second polar bodies in 676 oöcytes, from women undergoing PGT-A (PGT for aneuploidy). Source: From P Verdyck et al., Aneuploidy in oocytes from women of advanced maternal age: analysis of the causal meiotic errors and impact on embryo development, Hum Reprod 38:2526–2535, 2023. Courtesy P Verdyck, and with the permission of Oxford University Press, and the European Society of Human Reproduction and Embryology. 590  REPRODUCTIVE CYTOGENETICS considerable fraction of conceptions to be chromosomally unbalanced. On top of this, dispermy (two sperm fertilizing the one ovum) can cause triploidy. An abnormal post-zygotic cell division can give rise to mosaicism, and this is a common happening. These several possibilities all add up to a substantial potential for chromosome abnormality in the very early conceptus, in the first days (even hours) of existence. NON-IMPLANTATION AND “OCCULT ABORTION” Although the natural in vivo circumstance might differ from the extrapolated observations in vitro, it is nevertheless a fair assumption that unbeknown, a substantial fraction of all human conceptions have a lethal chromosomal burden and will not implant. Transient implantation may be associated with little or no perturbation of the menstrual cycle, although the woman may fleetingly feel pregnant as a hormonal response is briefly elicited. This is “occult abortion”. It becomes a semantic question whether the existence Figure 20–6.  Eggs of Younger and Older Women at Meiosis II. Notes: Shown here is the disposition of the chromosomes at meiosis II in the oöcytes from younger and older women, illustrating what may be the physical basis of the maternal age effect. The microtubules of the spindle stain green, and the chromosomes stain orange. The chromosomes are well organized at the metaphase plate at the equator of the cells in the younger women (the 22-year-old’s oöcyte, on the upper left, is viewed on a tilt). In contrast, the 40-year-old’s oöcyte shows the chromosomes in disarray. The 42-year-old woman’s oöcyte has one chromosome, at the top, dislocated from the metaphase plate, and the disposition of the other chromosomes at the equator is not as regular as in the younger women. Source: From Battaglia et al., Influence of maternal age on meiotic spindle assembly in oöcytes from naturally cycling women, Hum Reprod 11: 2217–2222, 1996. Courtesy DE Battaglia, and with the permission of Oxford University Press, and the European Society of Human Reproduction and Embryology.

REPRODUCTIVE FAILURE  591 Figure 20–7.  The Relative Frequencies of Aneuploidy per Chromosome in Sperm. Notes: Disomies per chromosome are shown above the 0‰ baseline, nullisomies below. In most sperm, there is a single abnormality; some may have more than one. Overall, around 0.3% of sperm in this study were nullisomic, and 0.2% disomic; this excess of nullisomy over disomy is also seen in other such studies (García-Mengual et al. 2019). Additionally, 0.4% of sperm were diploid. 99% were normal haploid. No readily discernible pattern presents itself (cf. eggs, Figure 20-5 above). Note the orders of magnitude difference in the y axis measurements of eggs (Figure 20-5) cf. sperm: the egg data is shown in whole numbers %, vs. the sperm data in fractions ‰ (per thousand). These data are derived from ten normozoöspermic men, from whom near a quarter-million sperm were analyzed. Source: Drawn from data in Table 3 in S Zhu et al., Comprehensive chromosome FISH assessment of sperm aneuploidy in normozoospermic males, J Assist Reprod Genet 39:1887–1900, 2022. of a non-implanting conceptus could be described as a pregnancy and whether its loss could be considered an abortion. THE CLEAVAGE EMBRYO (“PRE-EMBRYO”) (DAYS 1–3) In the first three days of existence, the conceptus is a simple sphere, a number of cells contained within an outer shell (the zona pellucida) (Figure 20–8 a to c). The population background rate of abnormality of the day 1–3 embryo can be assessed in the study of those whose fertility is not in question, namely in couples, one of whom is a heterozygote for an X-linked Mendelian condition, presenting at IVF for PGT for specific mutation testing. Pellicer et al. (1999), assessing some or all of chromosomes 13, 16, 18, 21, 22, X, and Y, studied 16 Mendelian heterozygous women, 10 of whom were of age range 30–36 years and six of 37–41 years. The younger women had a total of 12 chromosomally abnormal embryos out of 62 tested (19%), but a considerably higher figure (46%) was observed in the older mothers. Similar findings, from larger numbers, are due to Rubio et al. (2003; Table 20–1). These FISH-based studies did not include data on the remaining 17 chromosomes and are thus necessarily underestimates, and substantially so, of the total abnormal burden. The “atypical” group of patients otherwise presenting to the IVF clinic are of course a population of clinical interest, and thus the observations gained from study of them, however unrepresentative they might possibly be of the general population, are very germane to the agenda of the counselor. Demko et al. (2016) studied more than 22,000 day-3 embryos tested using SNP microarray (thus looking at all 24 chromosomes) and found that the proportion of aneuploid embryos remained fairly constant at 65% between maternal ages 24 and 35 years, but increased markedly at older ages. With similar numbers studied, Matorras et al. (2024) observed 86% of day-3 embryos to be aneuploid from mothers of age above 35 years. Mosaicism of the very early embryo (Figure 20–9) has been one of the more startling discoveries to emerge from the laboratory, and it is fair to imagine a similar picture may apply in vivo. Many IVF embryos are aneuploid or diploid/aneuploid mosaics, 592  REPRODUCTIVE CYTOGENETICS reflecting the fact that the first few post-zygotic divisions are particularly susceptible to errors in chromosomal distribution7 (Voullaire et al. 2000b; Currie et al. 2022; McCoy et al. 2023). At younger maternal ages, mosaicism may surpass maternal meiosis as a type of day-3 abnormality (Figure 20–10). The main mechanism may be chromosome loss due to anaphase lag,8 but chromosome gain is also frequent; mitotic nondisjunction (inferred from observing different cells with the corresponding monosomy and trisomy) occurs in a small minority (Daphnis et al. 2008). When the mosaicism is extensive, exhibiting several different karyotypes, these embryos may be referred to as being chromosomally “chaotic.” Blennow et al. (2001), studying embryos from translocation carriers that had been diagnosed as aneuploid on PGT, demonstrated in some that every cell could be different: the absolute maximum mosaicism. However, lesser degrees of mosaicism might not necessarily predestine the embryo to failure, and the aneuploid cells may be “sequestered off ” from those that will become the inner cell mass, these being the cells that will give rise to the embryo proper (Liñán et al. 2018; Griffin et al. 2023). THE MORULA AND BLASTOCYST (DAYS 4–5) IVF methodology has largely moved on to the slightly later timeframe of the blastocyst (day 5–6), and the more extensive data now available from this period may be of more current relevance. The short period (days 4 to 5) during which the cleavage embryo Figure 20–8.  The cleavage-stage embryo and morula. Days 1 to 3, a to c, the 2-cell, 4-cell, and 8-cell cleavage stages. Day 4 to 5, d and e, morula. Notes: z.p., zone pellucida; p.gl., polar bodies. Source: H Gray, Anatomy of the Human Body, 1918. 7 The very earliest cause of failure of the embryo may apply at the very earliest mitoses, from as early as mitosis number 1 (Bell et al. 2020). These initial mitoses require integrity of, among others, the TLE6 and PADI6 genes; mutation in these genes leads to embryonic arrest at this stage (Maddirevula et al. 2017). 8 Anaphase lag at mitosis no. 1 is dramatically illustrated in time-lapse photography in Currie et al. (2022). REPRODUCTIVE FAILURE  593 advances through the morula stage (Figure 20–8, d to e) and into the blastocyst may be an important hurdle at which the development of many chromosomally abnormal pre-embryos can arrest. Not all may fall at the hurdle: some mosaic aneuploid day-3 embryos may have converted, by day 5, into euploidy (discussed below). Figure 20–9.  Scenarios whereby Mosaicism may Arise. Note: In these examples, (above) a mitotic error at the morula stage produces aneuploid cells (orange), which then go on to involve the inner cell mass (ICM), but the trophectoderm (TE) (in blue) is normal. (Below) a newly aneuploid lineage (purple) arises at the 4-cell stage, following error during the second round of mitosis, and this lineage generates the TE, while the ICM (green), the presumptive embryo, is euploid. Source: A Lee and AA Kiessling, Early human embryos are naturally aneuploid—can that be corrected? J Assist Reprod Genet 34:15–21, 2017. Courtesy AA Kiessling, and with the permission of Springer Nature. Table 20–1.  Frequencies (%) of Aneuploidies for Certain Chromosomes in the Day-3 Embryos of a Cohort of 28 Presumed Normally Fertile Women, According to Maternal Age CHROMOSOME <37 YEARS ≥37 YEARS 13 7 21 16 7 16 18 5 15 21 9 30 22 5 14 X,Y 7 15 Notes: Figures are percentages, based upon an analysis of 215 embryos. In addition to these aneuploidies, small numbers (3%) of haploid and polyploid embryos were observed, too few to allow a useful age comparison. The cohort comprised women having preimplantation genetic testing for an X-linked Mendelian condition. Source: From C Rubio et al., Chromosomal abnormalities and embryo development in recurrent miscarriage couples, Hum Reprod 18:182–188, 2003.

594  REPRODUCTIVE CYTOGENETICS A very substantial fraction of blastocysts are aneuploid. Franasiak et al. (2014), for example, analyzed more than 15,000 blastocysts of parents presenting at IVF/PGT for biopsy for a variety of reasons, using 24-chromosome methodology, and could thus derive rather precise estimates of aneuploidies according to maternal age. A small increase above baseline is seen in the youngest mothers; the least rate was in the 25- to 34-year-old age group, progressively increasing steeply from age 35 years and then into the mid-40s (as the graph in Figure 20–11 attests) when well over a half, and finally up to 80%–90% at age 43–45 are aneuploid. The shape of this graph reflects the similar tendency to aneuploidy in the oöcyte (Figure 20–4). The reader will note also that the aneuploidy graphs in Figures 20–5 (oöcytes) and 20–12 (blastocysts) quite closely overlap, at least with respect to the trisomies.9 This observation of somewhat equivalent fractions per chromosome, of trisomy and monosomy in oöcytes and blastocysts, is consistent with the following interpretations: that meiotic nondisjunction in oöcytes produces essentially equal numbers of disomic and nullisomic gametes, which have equal chances of fertilizing/being fertilized; and that most trisomies and monosomies can survive through to the morula (day 4) stage. As the morula morphs into a blastocyst (days 4–5) a lethality begins to operate, and the aneuploidy rate overall falls (Matorras et al. 2024). This is most notable in the monosomies, at least for the larger chromosomes, and monosomies are mostly seen in slightly lesser fractions than are the trisomies (Figure 20–12, upper). A day or so later, when implantation would normally happen, almost all autosomal monosomies will have succumbed (De Munck et al. 2021). Figure 20–10.  Meiotic cf. Mitotic Mosaicism in Day-3 Embryos. Notes: If the dotted line were smoothed, mitotic abnormality could be seen as being practically constant, at a little under half of all blastomeres, irrespective of maternal age. The abrupt take-off during the mid 30s in the maternal meiotic data is very evident (the dip at 45-46 may reflect smaller numbers at this age). The paternal meiotic data would probably, in actuality, reduce to a near-straight line barely lifting above the baseline, at 1%–2%. These data are derived from an analysis of single blastomeres. Source: Drawn after RC McCoy, Mosaicism in preimplantation human embryos: When chromosomal abnormalities are the norm. Trends Genet 33:448–463, 2017. Courtesy RC McCoy, and with the permission of Elsevier. 9 A notable exception is trisomy 3, one of the more common aneuploidies at oögenesis, but infrequent in the blastocyst and rare at miscarriage (cf. Figure 20–22). And monosomy 3 is scarcely observed. Presumably, this chromosome is gene-dense with respect to very early-developmental critical loci. REPRODUCTIVE FAILURE  595 Mosaicism of the blastocyst of various categories is very commonly seen, and this is necessarily of mitotic origin (Figures 20–9, 20–13, and 20–14). Making a diagnosis of mosaicism in clinical practice will naturally depend upon how closely the sampled tissue (trophectoderm) is representative of the whole embryo, and what fraction of cells are abnormal. True mosaicism of the embryo, studied at biopsy of trophectoderm, will likely be missed if the level of mosaicism in the biopsied tissue is less than about 20% (Figure 20–15). If it is a relatively simple mosaicism (e.g., a single trisomy; the diploid-aneuploid mosaic in Figure 20–13), and with the coexistence of a normal euploid lineage, there may often be an optimistic outlook as the blastocyst stage proceeds. Similarly as with the day-3 embryo discussed above, Chavli et al. (2024) describe how abnormal cells may be “sequestered off ” to the trophectoderm and become a part of the placenta, or fall by the wayside. Such aneuploid mosaic embryos may thus undergo progressive normalization that can then, in due course, go on to produce a euploid fetus and a normal child (Greco et al. 2015; Liñán et al. 2018). And yet, exceptions are on record: for example, Greco et al. (2023) report two cases of mosaicism of the blastocyst—one for trisomy 21 and the other for a segmental loss of 1p36—going on to the same mosaicism being identified at follow-up prenatal diagnosis. THE “EMBRYO PROPER” (WEEKS 3–8) The embryo proper, in the sense that the body plan is beginning to be laid out, takes form from about the middle of the third week post-conception and is barely 1½ mm long; by the end of the eighth week post-conception (10 weeks by maternal dates4), it will be 3 cm. (To the embryologist, this is Carnegie stages 8 through 23.) This is a timeframe, now well beyond the gaze of the IVF laboratory, that is not easily studied (Figure 20–3). First, this appears to be a period during which the threshold for natural abortion is relatively Figure 20–11.  Aneuploidy Rates in Blastocysts, According to Maternal Age. Notes: The regression curve (red) is derived from the actual data points. Note how this curve closely reflects that of oöcyte data (Figure 20–4 above). Source: From JM Franasiak et al., The nature of aneuploidy with increasing age of the female partner: A review of 15,169 consecutive trophectoderm biopsies evaluated with comprehensive chromosomal screening, Fertil Steril 101: 656–663.e1, 2014. Courtesy JM Franasiak, and with the permission of Elsevier. 596  REPRODUCTIVE CYTOGENETICS high, and many abnormal embryos seem able to maintain existence (however imperfect that existence might be). Second, the practicalities of collecting intact embryos from very early spontaneous miscarriage present obvious drawbacks: the embryos are scarcely or not at all discernible among the products of conception otherwise (chorionic Figure 20–12.  Aneuploidies in Day 5 Blastocysts. Notes: (Upper) The range of aneuploidies observed in day 5–6 blastocysts, and their frequencies as a percentage of all aneuploidies. Note that trisomy and monosomy for each chromosome rather closely mirror their relative frequencies. Note also the broad similarity with the spread of data in oöcytes (Figure 20–5). Here, as in other studies, chromosomes 16 and 22 are seen to be the most susceptible to aneuploidy. (Lower) Comparing the distributions of aneuploidies in Miscarriage vis a vis Blastocysts. The spread of trisomies is fairly similar, whereas the autosomal monosomies have all but disappeared, attesting to the differential early pregnancy survival of the two aneuploidy classes. Source: From J Rodriguez-Purata et al., Embryo selection versus natural selection: How do outcomes of comprehensive chromosome screening of blastocysts compare with the analysis of products of conception from early pregnancy loss (dilation and curettage) among an assisted reproductive technology population? Fertil Steril 104:1460–1466, 2015. Courtesy J Rodriguez-Purata, and with the permission of Elsevier.

REPRODUCTIVE FAILURE  597 villi the main component). Nevertheless, a very few studies are on record concerning observations at therapeutic abortion, at “missed abortion,” and at early first-trimester miscarriage. Observation at induced abortion offers the earliest window to this period. An example is the day-25 trisomy 8 embryo illustrated in Figure 20–16, showing a gross distortion of Figure 20–14.  Mosaicism in Day-5 Blastocysts. Notes: The normal figure of only 11% might seem disconcerting, but several of the diploid-aneuploid mosaics could subsequently “self-correct”, and become euploid. Abn., uniformly abnormal. The material comprised 55 cleavage-stage embryos or morulae, cultured in vitro through to the blastocyst stage, from donors of median maternal age 34 (range 22-42), and analyzed by single-cell sequencing. Source: Drawn after EA Chavli et al., Single-cell DNA sequencing reveals a high incidence of chromosomal abnormalities in human blastocysts, J Clin Invest 2024:e174483, 2024. Courtesy EB Baart, and with the permission of the American Society for Clinical Investigation. Figure 20–13.  Categories of Mosaicism of the Embryo. Notes: The aneuploid mosaic has monosomic (-1), and doubly monosomic (-2) cells. The diploid-aneuploid mosaic has normal (clear) and trisomic (+1) cells. The ploidy mosaic has normal and triploid (+23) cells. The chaotic mosaic has several different monosomies, and a few normal cells. Source: From RC McCoy, Mosaicism in preimplantation human embryos: When chromosomal abnormalities are the norm. Trends Genet 33:448–463, 2017. Courtesy RC McCoy, and with the permission of Elsevier. 598  REPRODUCTIVE CYTOGENETICS normal anatomy, as can be appreciated alongside the photograph of a normal embryo of similar gestation. It is a fair assumption that some other trisomies, particularly of chromosomes containing crucial early-developmental genes, may be of similarly devasting effect and lead to very early natural miscarriage. Missed abortion refers to embryonic death without expulsion of the products of conception. Philipp et al. (2003, 2004) were able to study by endoscopy the anatomy of embryos in women prior to uterine evacuation for missed abortion. Some embryos showed no recognizable external structures (e.g. trisomy 3; Figure 20–17), while in some an outline embryonic form could be Figure 20–15.  Mosaicism of the Embryo. Notes: The detection of mosaicism will depend upon the fraction of trisomic cells, and their distribution within the biopsied trophectoderm. In the lower image, only four cells are trisomic, out of 529; such a low level could well be missed at routine PGT. Source: From EA Chavli et al., Single-cell DNA sequencing reveals a high incidence of chromosomal abnormalities in human blastocysts, J Clin Invest 2024:e174483, 2024. Courtesy EB Baart, and with the permission of the American Society for Clinical Investigation. Figure 20–16.  Trisomy 8 in a Very Early Embryo. Notes: Midline sagittal section of a day-25 embryo with non-mosaic trisomy 8 (left), due to meiotic I nondisjunction. A normal embryo is shown (right) for comparison. The trisomic embryo is devastatingly malformed. Normal brain structure, the prosencephalon (pros) and rhombencephalon (rh), and regular somite (s) development are clear to see in the normal embryo, compared with the gross deformity in the trisomic embryo. da, dorsal aorta; h, heart; nt, neural tube; pa1, first pharyngeal arch; pe, pharyngeal endoderm; tb, tail bud; ys, yolk sac. Source: From C Golzio et al., Cytogenetic and histological features of a human embryo with homogeneous chromosome 8 trisomy, Prenat Diagn 26:1201–1205, 2006. Courtesy HC Etchevers, and with the permission of Wiley-Blackwell. REPRODUCTIVE FAILURE  599 recognized (e.g. trisomies 4 and 7). We have seen a triploid embryo from a missed abortion (Figure 20–18). In a Thai study of early first trimester abortion, based upon chorionic villus analysis and which included a fraction of losses happening during weeks 5 to 7+6— in other words, within the embryonic period—these workers were able to chromosomally analyze 17 cases, of which 9 were aneuploid, with trisomies of 15, 18, 19, 20, 21, and 22 (Parinayok et al. 2022). The practice of noninvasive cell-free DNA prenatal screening, now becoming widespread, has given clearer insight into the chromosomal picture of early pregnancy, at the time at which the blood sample would have been drawn (usually at the 10-week by-dates mark, or shortly thereafter, as the embryonic period segues into the fetal stage). Pertile et al. (2017) analyzed findings due to just under 90,000 samples from ongoing pregnancies, mostly from late first trimester, with a minority from second, thus representing in particular the transition from the embryo into the early fetal developmental stage. Some would have presented due to advanced maternal age, but ascertainment in the majority was, in all likelihood, essentially unbiased. Of these, 306 (0.34%) had a “rare autosomal trisomy”; that is, a trisomy of other than 13, 18, 21, or sex chromosomal. And while the spread of abnormality was somewhat comparable to that seen at miscarriage, with trisomies 15, 16, and 22 prominent, the largest single contributor was actually trisomy 7. These pregnancies, if left alone, would surely have gone on to later miscarriage or intrauterine fetal death. Cytogenetics of Very Early Pregnancy Failure RECURRENT IMPLANTATION FAILURE AT IVF Couples 46,XX and 46,XY. A failure of implantation following natural (in vivo) fertilization cannot be recognized as such; but in vitro, the whole process can be followed. In Figure 20–17.  Trisomy 3 in an Embryo at Missed Abortion. Notes: View by hysteroscopy prior to evacuation of products of conception. No normal anatomic structures are discernible, only an irregular tissue mass. Source: From T Philipp et al., Abnormal embryonic development diagnosed embryoscopically in early intrauterine deaths after in vitro fertilization: a preliminary report of 23 cases, Fertil Steril 82:1337–1342, 2004. Courtesy T Philipp, and with the permission of the American Society for Reproductive Medicine and Elsevier.

600  REPRODUCTIVE CYTOGENETICS the case of recurrent implantation failure (RIF10) following IVF, more than one cause may apply, and a distinction is to be made between maternal (“uterine receptivity”) and embryo-related characteristics. Maternal causes notwithstanding, a substantial fraction of RIF is due to chromosomal abnormality of the embryo (Ma et al. 2022). Matorras et al. (2024) observed that women having had more previous aneuploidies at embryo biopsy, and proceeding to subsequent IVF, had fewer embryos that were euploid compared with women whose embryos at previous PGT had mostly been euploid (Figure 20–19). Xi et al. (2022) applied the methodology of noninvasive chromosome screening (NICS) of the embryo in IVF patients presenting with previous RIF: those having had NICS achieved a clinical pregnancy rate of 47%, compared with 29% in those in whom choice of embryo transfer was judged purely on blastocyst morphology. Observations such as these lead to the assumption that, in at least a considerable number of cases, recurring errors at cell division may be the basis of the RIF. Figure 20–18.  A Triploid (69,XXY) Embryo. Notes: The face has no landmarks other than eyes and a single opening. The anterior trunk is open, with the heart and liver visible. The disrupted tissue at the neck was the site of biopsy for the cytogenetic analysis. 10 The European Society of Human Reproduction and Embryology defines RIF thus: RIF describes the scenario in which the transfer of embryos considered to be viable has failed to result in a positive pregnancy test sufficiently often in a specific patient to warrant consideration of further investigations and/or interventions (ESHRE Working Group on Recurrent Implantation Failure, 2023). REPRODUCTIVE FAILURE  601 A question: at which cell division? A vulnerability might in fact lie as much in very early mitosis rather than at meiosis. Voullaire et al. (2007) compared the frequency of aneuploidy at PGT in embryos from women with RIF and without such history and found a higher rate of complex chromosome abnormality, which they defined as aneuploidy of three or more chromosomes, in the RIF group. McCoy et al. (2015) studied large numbers of day-3 embryos and day-5 blastocysts, and identified previous IVF failure to be associated slightly, but nevertheless significantly, with mosaicism due to mitotic error. Kort et al. (2018), analyzing blastocyst data, calculated an odds ratio of 1.36 cf. fertile controls, that a couple having suffered RIF might have yet another failure due to aneuploidy, independent of maternal age. In their review, Kim et al. (2024) determined that PGT-A (PGT for aneuploidy) improved the implantation rate (42% cf. 22% in those not having had PGT-A) and livebirth rate (47% cf. 29%) in women having suffered previous RIF, notwithstanding that the PGT-A group were a little older (average 38) than the non–PGT-A group (35). Thus it may be that in at least some of these RIF couples, mitosis rather than meiosis presents the more hazardous stumbling block (McCoy et al. 2015). An underlying susceptibly may obtain, in embryos of theirs, to undergo post-zygotic chromosomal error during the first two or three rounds of mitosis—an effect that might relate to either maternal or paternally inherited factors and be unrelated to parental age (Figure 20–10). Recurrent meiotic error may be marginally less contributory. Translocation Carrier Couples. As for the small category of those with heterozygosity for an autosomal translocation, in one series of 65 women who had had ≥15 failed Figure 20–19.  The Likelihood of Euploidy According to History of Previous Embryos. Note: A history of previous aneuploid embryos portends a higher subsequent risk for aneuploidy. Source: From R Matorras et al., Lessons learned from 64,071 embryos subjected to PGT for aneuploidies: results, recurrence pattern and indications analysis, Reprod Biomed Online 49:103979, 2024. Courtesy R Matorras, and with the permission of Elsevier. 602  REPRODUCTIVE CYTOGENETICS transfers, 8% were translocation carriers, two being sisters with the same translocation (Raziel et al. 2002). On somewhat larger numbers, Stern et al. (1999) noted the rate of balanced rcp and rob carriers to be 3% in 219 couples (both partners tested) who had failed more than 10 embryo transfers, but 9% in 130 couples who had three or more consecutive first-trimester abortions. Cytogenetics of Spontaneous Abortion and Later Pregnancy Failure CLINICAL MISCARRIAGE Most (80%) miscarriages happen in the first 12 weeks of pregnancy, and mostly in weeks 8 through 12. By this stage, tissue from “products of conception” is often obtainable and analyzable. Of all recognized pregnancies (recognized in the traditional way, that is), ~15% end in natural miscarriage, also referred to as spontaneous abortion (Quenby et al. 2021).11 If the products of conception (POC) are successfully karyotyped, a majority of abortuses are shown to have a chromosome abnormality (Finley et al. 2022; Essers et al. 2023; Kutteh et al. 2024; Xu et al. 2024a). Finley and colleagues studied 24,900 POC samples by SNP microarray methodology, and the abnormalities documented are set out in Figure 20–20; similarly, Kutteh and colleagues analyzed 65,333 POC samples, classifying by maternal age (Figure 20–21). A clinically significant chromosome abnormality was seen in these studies in ~60% of samples, and in practically all, this would have been the cause of the miscarriage. Thus, most miscarriages reflect “biology in action”: the natural elimination of an unsurvivable abnormality. Trisomies account for about two-thirds of all cytogenetic abnormalities identified at spontaneous abortion, and the full range of trisomies and their relative frequencies are depicted in Figures 20–20 and 22. The most commonly seen abnormal karyotypes are trisomy 16, monosomy X, and triploidy. Indeed, as many as 1% of all human conceptions may have trisomy 16 (Benn 1998); monosomy X and triploidy account each for approximately 10% of all abnormalities. Double trisomy (trisomy for two chromosomes) is infrequent, and triple trisomy is rarely seen. Pathogenic CNVs and unbalanced structural rearrangements constitute much of the remainder (Dai et al. 2024). The origin of the abnormality is, in most, an error at maternal meiosis I, and this includes most of the major trisomies: trisomies 13, 14, 16, 21, and 22, with trisomy 18 a possible exception. Robinson et al. (1999) analyzed the originating status of certain of the less studied karyotypes: trisomies for chromosomes 2, 4–10, 12, 15, 17, and 20. Around three-fourths showed three alleles for the trisomic chromosome, 11 The distinction between an embryo and a fetus (and see footnote 2) in this setting is not necessarily straightforward. Embryonic development may have arrested, and spontaneous abortion will be inevitable, but the pregnancy may continue for one or a few weeks (“missed abortion”) and using apparent gestational age would give a misleading impression. In this case, it is more useful to consider the developmental stage of the embryo in judging the effects of a particular aneuploidy (Philipp et al. 2003). For example, the triploid embryo (not fetus) shown in Figure 20–18 was retained in the uterus until 18 weeks, but development had arrested at the 7–8 week mark.

REPRODUCTIVE FAILURE  603 thus confirming a meiotic origin. Most of the remainder are presumed to have been due to somatic errors; some might have been mosaic but not detected as such. Trisomy 8 is notable in that all cases were due to a meiotic error, which stands in contrast to somatic errors being almost entirely the basis of the mosaic trisomy 8 syndrome (Warkany syndrome, p. 687) that is diagnosed post-natally. Uniparental disomy appears not to be an important causative factor in miscarriage (Xu et al. 2024a). There is a striking maternal-age effect in the risk, per pregnancy, for a miscarriage, increasing tenfold from late teens to 40s—an effect presumably due in large part to a chromosomal basis (Nybo Andersen et al. 2000; Cohain et al. 2017; Figure 20–20.  Chromosomal Findings in Products of Conception from Spontaneous Abortions. Notes: These abnormal cases comprised 54% of the total of 24,900 analyzed POCs. “Single” refers to trisomy for a single chromosome, with those chromosomes more frequently represented individually indicated by number. Note the substantial involvement of trisomies 16 and 22, reflecting their high frequencies at oögenesis and blastogenesis (see above), and their presumed relatively low critical-gene content, that has allowed them to survive through to at least this stage of pregnancy. Polyploidy and monosomy X are also prominent. Chr = chromosome; ROHs = regions of homozygosity; WC = whole chromosome; WG-UPD = whole-genome uniparental disomy. Source: From J Finley et al., The genomic basis of sporadic and recurrent pregnancy loss: a comprehensive in-depth analysis of 24,900 miscarriages, Reprod Biomed Online 45:125–134, 2022. Courtesy T Sahoo, and with the permission of Elsevier. 604  REPRODUCTIVE CYTOGENETICS Figure 20–21.  Chromosomal Findings in Products of Conception from Spontaneous Abortions, According to Maternal Age. Note: The maternal age effects are opposite in aneuploidies (essentially increasing with maternal age) cf. del/ dups and triploidy (falling with increasing age). Source: From WH Kutteh et al., Role of genetic analysis of products of conception and PGT in managing early pregnancy loss, Reprod Biomed Online 49:103738, 2024. Courtesy WH Kutteh, and with the permission of Elsevier. Figure 20–22.  The Relative Frequencies of Trisomies Observed in Products of Conception. Notes: The frequencies descend from the most often seen (trisomy 16) to the least (trisomy 1). These data are from the 1,872 abnormal results of a series of 5,457 consecutive samples of products of conception, on G-banding analysis. The curious pre-eminence of trisomy 16 is noted, and yet this aneuploidy, as a non-mosaic abnormality, never survives through to term (Peng et al. 2021). Unsurprisingly, the largest chromosome, no. 1, is the least represented, attesting to its extreme lethality in the trisomic state; but notably, one of the smallest, no. 19, is the second-most lethal, presumably reflecting its very high gene load (Table 1-1). The findings here are very similar to those of Rodriguez-Purata et al. shown in Figure 20–12, and near identical to those of Gu et al. (2021). Source: From BT Wang et al., Abnormalities in spontaneous abortions detected by G-banding and chromosomal microarray analysis (CMA) at a national reference laboratory, Mol Cytogenet 7:33, 2014. Courtesy BT Wang, and with the permission of Springer Nature. REPRODUCTIVE FAILURE  605 Figure 20–23). The influence of a slight paternal-age effect is open to argument (du Fossé et al. 2020). Twinning. If an abnormal twin dies, the normal twin may ensure continuation of the pregnancy, with only a parchment-like vestige (fetus papyraceus) remaining, preserved in the uterus along with the normal twin. A “vanishing twin” is plausibly proposed in the circumstance of two cell lines being identified at chorion villus sampling, 46,XX and 47,XY,+autosome, and a normal singleton baby subsequently born (Falik-Borenstein et al. 1994). A very rare observation is the trisomy that would otherwise lead inevitably to early miscarriage, but in which a monozygous euploid co-twin allows some ongoing in utero survival. These cases may result from a very early post-zygotic event that generates a trisomic cell line (or which generates a normal cell line from a trisomic conceptus), and the trisomic co-twin, among other grossly devastating defects, fails to form a heart (acardius). The normal euploid co-twin provides the blood circulation to the abnormal fetus (twin reverse arterial perfusion). This scenario has been reported with trisomies 2 and 11 (Blaicher et al. 2000; Mihci et al. 2009), and we have seen one case due to trisomy 3 (Figure 20–24). RECURRENT PREGNANCY LOSS Couples 46,XX and 46,XY. More than two miscarriages is the criterion for “recurrent pregnancy loss”12 (RPL). In principle, the causal error could have been with the sperm, with the egg, with the embryo, or with the mother (Klimczak et al. 2021). For those who are karyotypically normal, do some couples miscarry due to a predisposition to produce aneuploid conceptions? Or is the basis in a maternal rather than an embryonic cause: do certain factors influence uterine carrying capacity? Or, could recurrent aneuploid miscarriage, in the setting of a high background rate of aneuploidy in humans, Figure 20–23.  The Effect of Maternal Age upon Pregnancy Loss. Notes: These data are derived from over a million Danish women intending carrying to term in 1978-1992, according to their age. The maternal age effect is very evident. Source: Drawn after AM Nybo Andersen et al., Maternal age and fetal loss: population based register linkage study, BMJ 320:1708–1712, 2000. 12 Definition from the American Society for Reproductive Medicine (2008): “Recurrent pregnancy loss is a disease distinct from infertility, defined by two or more failed pregnancies. When the cause is unknown, each pregnancy loss merits careful review to determine whether specific evaluation may be appropriate. After three or more losses, a thorough evaluation is warranted.”

606  REPRODUCTIVE CYTOGENETICS simply reflect random biology, with increasing maternal age the only clear predisposing factor? A common event is common, and not uncommonly it may happen more than once. In fact, the major contributors may be these: uterine receptivity; and random biology, maternal-age related. Van den Berg et al. (2012) determined, on a review of the literature at that time, that the aneuploidy rate in products of conception following previous RPL (39%) was slightly less compared to the background rate of 45% in sporadic, one-time miscarriage. The more broadly focused (including del/dup) categories of chromosomal abnormality due to Li et al. (2024c) are listed in Table 20–2. In this dataset, from a younger (average age 30) cohort, chromosomal imbalances of all types occurred in a substantial fraction (59%) of pregnancies of those having had previous RPL. Narrowing the viewpoint to autosomal trisomies, and with the mechanism of recurrent single-chromosome meiotic error in mind, these, at 33%, contributed rather less to the overall POC chromosomal abnormality (albeit the largest single category). In other words, there appears not to be a particular maternal propensity to produce aneuploid eggs, as a cause of RPL. A comparative viewpoint is afforded in Ozawa et al. (2019), who studied aneuploidy rates in women of mostly an older maternal age (into their late 30s and 40s) and comparing those who had had just one, with those having had two or more miscarriages. Table 20–3 sets out the individual trisomies as identified in the notable fraction of 61% of POCs from those who had previously suffered multiple miscarriages, but this high proportion is still less than in those having had only one previous miscarriage, with an aneuploidy rate of 72%. The points emerging from the foregoing is that aneuploidy of the POC is less often observed in couples who have had a large number of miscarriages, sometimes into double figures, than among those who have had fewer; and that maternal age is a major factor behind much RPL. Returning to Li et al. (2024c), these workers followed prospectively women with an RPL history, having had one tested (“index”) POC: and those women whose index POC was euploid had a slightly lower eventual live birth outcome Figure 20–24.  Trisomic Co-Twin Survival. Notes: The euploid twin enabled survival of the trisomy 3 monozygous co-twin. The abnormal co-twin, identified at mid-trimester ultrasonography, was anencephalic and acardiac, but survived to that stage due to a shared circulation from the “pump twin”. A likely scenario is that a mitotic error in an initially normal conceptus at a very early stage, perhaps the second or third mitosis, led to the generation of a trisomic 3 cell line, with the embryo then splitting along a plane of separation between the euploid and aneuploid tissue. Case of G McGillivray et al. (2004). REPRODUCTIVE FAILURE  607 (67%) compared with the aneuploid POC group (72%), and a marginally higher miscarriage rate (27% cf. 23%). As well as the ploidy status of the index miscarriage, the rates of subsequent live birth were notably related to the number of previous miscarriages, as follows: those with two, three, or four or more previous miscarriages, the index miscarriage being euploid, had subsequent livebirth rates of 76%, 56%, and 44% respectively, these percentages quite markedly falling away with increasing numbers of miscarriages. The comparable figure for those in whom the index miscarriage was aneuploid were 74%, 71%, and 63%, these percentages showing a lesser decline (Figure 20–25). Table 20–2.  Range of Abnormalities in Products of Conceptions, from Women Having Suffered Previous Recurrent Loss CATEGORY (The abnormality in an “index POC”; see Notes) % % Single aneuploidy 41 Autosomal trisomy 33 Autosomal monosomy 0.5 Monosomy X 7 Sex chromosome trisomy 0.2 Multiple aneuploidy 2.5 Polyploidy 8 Triploidy 6 Hypotriploidy 0.7 Hypertriploidy 0.8 Tetraploidy 0.3 Partial aneuploidy (large CNVs) 5 Terminal deletion/duplication 1 Interstitial deletion/duplication 0.3 Terminal deletion + terminal duplication (suggestive of unbalanced translocation) 1.5 Multiple deletions and/or duplications 2 Pathogenic microdeletion/duplication 1.5 Terminal microdeletion/duplication 0.5 Interstitial microdeletion/duplication 1 Multiple microdeletions and/or duplications 0.1 UPD (whole-genome) 1 Total % Abnormal 59 Total % Normal 40 Notes: These data come from a cohort of 830 couples having had previous recurrent miscarriage, the mean number of previous miscarriages being 2.4, at mean gestational age 9.8 weeks. The average maternal age was 30. These couples went on to have a further miscarriage, the “index POC,” which was tested by CMA (chromosome microarray analysis) as in the data above, the figures shown as percentages (note that a majority of the POCs, 59%, were chromosomally abnormal). At follow-up for at least two years, 627 went on to have a live birth. Those having had a chromosomally abnormal index POC had a slightly higher livebirth rate than the euploid POC group (72% cf. 67%,) and a slightly lesser miscarriage rate (23% cf. 27%). Source: From Y Li et al., Reproductive outcomes in couples with recurrent pregnancy loss after embryonic chromosomal microarray analysis, J Assist Reprod Genet 41:161–170, 2024. 608  REPRODUCTIVE CYTOGENETICS Table 20–3.  Autosomal Trisomies in Products of Conception, Following Previous Recurrent Pregnancy Loss (RPL) or a Previous Single Miscarriage Trisomy Previous RPL (N, out of 175) Previous single miscarriage (N, out of 189) Trisomy Previous RPL (N, out of 175) Previous single miscarriage (N, out of 189) 1 0 0 15 14 20 2 2 2 16 10 22 3 0 1 17 2 1 4 4 0 18 4 7 5 1 4 19 0 0 6 0 3 20 3 4 7 2 1 21 11 29 8 1 4 22 25 16 9 2 3 Triploid 7 9 10 2 0 Mono X 3 10 11 1 0 Double 10 4 12 1 1 Complex 9 3 13 4 6 Other 11 9 14 7 5 Notes: A total of 175 and 189 POCs had been analyzed, from women having had previous multiple losses (RPL), or a single miscarriage, respectively. Overall, 146 (87%) of POCs in the RPL cohort and 164 (78%) of POCs in the single cohort had a chromosomal abnormality. This 9% difference is substantially accounted for by a lesser fraction of viable trisomies in the RPL group, 11%, cf. 22% in the single miscarriage group; there was little difference (44% cf. 46%) in the rates of non-viable trisomy. The excess of monosomy X in the single group, and an excess of trisomy 22 in the RPL group, is also to be noted, although the denominators are small. The maternal-age profile in this cohort was substantially skewed towards older age: a minority were <35 (87), 160 were 35–39, and 117 were 40 and over. Miscarriage happened prior to 12 weeks gestation. Source: From data in N Ozawa et al., Maternal age, history of miscarriage, and embryonic/fetal size are associated with cytogenetic results of spontaneous early miscarriages, J Assist Reprod Genet 36:749–757, 2019. Figure 20–25.  A Comparison of Livebirth Rates Following Previous Miscarriage, According to Ploidy Status of the POCs. Source: From data in Y Li et al., Reproductive outcomes in couples with recurrent pregnancy loss after embryonic chromosomal microarray analysis, J Assist Reprod Genet 41:161–170, 2024.

REPRODUCTIVE FAILURE  609 In other words, euploidy of a POC implies a worse outlook for a future pregnancy than does aneuploidy; and in that circumstance, an “inhospitable endometrium” may be the probable basis. For these women, a detailed endocrine and immunological13 work-up may be indicated (Kuroda 2024). If no cause is then shown, Eliwa et al. (2024) refer to these couples as having “truly unexplained RPL.” Otherwise, a combination of certain genetic factors may predispose to a risk for euploid miscarriage (Aminbeidokhti et al. 2024). These researchers speak of an “intolerome”: genes and variants that are subjects of biological intolerance and natural selection. Nevertheless, some couples might have a particular proneness to produce aneuploid gametes, and indeed, from the recurrence data with respect to the three viable autosomal trisomies discussed elsewhere (Chapter 13), this might seem, prima facie, logical. But the question: Is this indeed so? (Mumusoglu et al. 2025). Kort et al. (2018) found that women with a history of RPL were 1.3 times more likely to produce aneuploid embryos compared to fertile controls who had undergone PGT for sex selection. This increase applied to both meiotic and mitotic errors. Similarly, Liu et al. (2020) found that in women <35 years old with a history of RPL, 49% of embryos were aneuploid, compared to 37% aneuploidy rate in a control group having PGT for monogenic disorders. But these findings are not consistently seen: several other studies have found that a woman’s prior reproductive history, including previous miscarriage, shows no association with rate of euploidy in her embryos (Kornfield et al. 2022; Turgut et al. 2023; Cimadomo et al. 2021). What we may distill from these contradictory conclusions is that if there is indeed a susceptibility, it must be small and likely affecting only a minority. Couples with Abnormal Karyotypes. For the small group of people who are heterozygous for a chromosomal rearrangement, RPL may of course occur with a much higher frequency, and this briefly stated fact is the basis of much of what is written in this book. The more usually seen rearrangement is the reciprocal translocation, and this is the observation in around 2%–3% of all couples suffering loss of two or more pregnancies before 20 weeks’ gestation. Smaller fractions are due to Robertsonian translocations and inversions. Unsurprisingly, the rates are similar in whichever part of the world studies have been done (Elkarhat et al. 2019). These chromosomal rearrangements are usually of sufficient size to be readily detectable at standard karyotyping, and typically of sufficient size that the imbalanced combinations will very frequently lead to inevitable miscarriage. But the application of molecular methodology can sometimes discover a subtle parental rearrangement that previously would have passed muster—a possibility especially to be pursued in the event of recurrence of the same structural abnormality (Li et al. 2024b). Chapters 5 to 9 set out particular circumstances. The large study due to Park et al. (2022), based upon 19,000 couples presenting with recurrent spontaneous abortion, is representative. These researchers document a long list of chromosome abnormalities seen in 844 of these couples, and Figure 20–26 shows the broad distribution of abnormalities. Among these 844, balanced reciprocal translocation was the most common abnormality, in almost one-half. Otherwise, Robertsonian 13 “It is a wonder of pregnancy that the fetus is not rejected by the mother’s immune system,” as Kent (2009) comments. One genetic factor proposed to be central to this process is HLA-G, which “plays a mysterious role in the mechanism of maternal-fetal immune-tolerance” (Zhuang et al. 2021). 610  REPRODUCTIVE CYTOGENETICS translocations were seen in 14%, and inversions also in 14%. Sex chromosome aneuploidies comprised most of the rest, including Turner syndrome variants, Klinefelter syndrome and variants, 47,XXX and mosaic forms, 47,XYY and mosaic forms, and X deletions and duplications. Fetal Death In Utero, Stillbirth, and Perinatal Death Fetal Death in Utero. This expression, abbreviated FDIU (and also “in utero fetal demise” or “second-trimester fetal demise”), refers to fetal death occurring after the usual time of clinical miscarriage (c. 13 weeks) through to the time of potential viability (c. 26 weeks). A classical chromosomal abnormality is typically seen in 5%–10%. Gawron et al. (2013) studied 118 such cases, many with concomitant malformations, finding these aneuploidies: trisomies 13, 18, and 21, triploidy, monosomy X, and “rare chromosomal aberrations.” Others have found similarly, in small-for-gestational-age fetuses (Pasquini et al. 2023). Stillbirth and Perinatal Death. A stillbirth refers to fetal death occurring after the time of viability (c. 26 weeks) or, in lay parlance, “the birth of an infant that has died in the womb.” Some use 28 weeks as the cut-off, others ≥20 weeks. A perinatal death occurs from 28 weeks through to the first seven days of life, and thus includes stillbirth. From Irish data, malformations are seen in about one-third of such cases, and half of these (thus ~10% of all) are chromosomally related (Power et al. 2020). Similarly to FDIU, it is trisomy 21, trisomy 18, trisomy 13, and monosomy X that are most often seen, along Figure 20–26.  Chromosome Abnormalities carried by Couples who have had Recurrent Pregnancy Loss. Notes: These data come from a survey of 844 couples, they comprising 4.4% of a cohort of 19,000 couples presenting with recurrent pregnancy loss. Translocations (61%) are clearly the predominant category, and the relative proportions of the three different categories of translocation are shown in the “pie-within-pie”. Inv, inversions (14%); mos, mosaicisms (11%); mar, marker chromosomes (5%); sex, non-mosaic sex chromosome aneuploidy (5%); others (4%); rob, Robertsonian (14%); ccr; balanced complex chromosome rearrangement (0.6%); rcp; reciprocal translocations (47%). Source: Redrawn from SJ Park et al., Chromosomal abnormalities of 19,000 couples with recurrent spontaneous abortions: a multicenter study, Fertil Steril 117:1015–1025, 2022, with the permission of Elsevier.

REPRODUCTIVE FAILURE  611 with an occasional rare pathogenic del or dup (Reddy et al. 2012).14 A role for CNVs and other genetic variation in the etiology of stillbirth is the subject of ongoing research (Ernst et al. 2015; Stanley et al. 2020). Chromosomal microarray yields a result more often than does classical karyotyping, in which the nature of tissue sampled at postmortem can lead to culture failure (Giordano and Wapner 2024). Among liveborn babies, only 1 in 250 has an unbalanced chromosome abnormality on standard, classical karyotyping. Thus, there has been a very effective selection, at several stages along the prenatal developmental pathway, against those conceptions that were abnormal. Cytogenetics of Infertility Infertility is defined thus in the American Society for Reproductive Medicine (2012): “A disease, defined by the failure to achieve pregnancy after 12 months or more of regular unprotected intercourse.” Certainly it is common, affecting at least 15% of couples, and increases with increasing age (Figure 20–27). Of the many causes of infertility, a chromosomal basis is seen in a small minority, 1%–2% (Figures 20–28 and 20–29). The frequency of karyotypic abnormality in couples with infertility depends upon the criteria of ascertainment and the threshold for determining abnormality, and quite wide ranges of figures have been produced (Table 20–4). Sex chromosomal abnormalities include XXY and XXY/XY in the male, and Turner syndrome and its variants in the female. AZF deletions on the Y chromosome are common; the “XX male” and XY female are rare (Chapter 24). Autosomal rearrangements as a cause of infertility are noted in several chapters. The reciprocal translocation (especially when an acrocentric is involved) and the inversion may be associated, though Figure 20–27.  The Rates of Infertility According to Age. Notes: The fractions represent the proportions of persons by sex and age to have ever tried for 12 months or more to conceive, or to have sought fertility advice. These data were derived from a whole population study, the Dunedin Multidisciplinary Health and Development Study. Source: From T Van Roode et al., Cumulative incidence of infertility in a New Zealand birth cohort to age 38 by sex and the relationship with family formation, Fertil Steril 103:1053–1058.e2, 2015. Courtesy WR Gillett, and with the permission of Elsevier. 14 Data from the post prenatal-diagnosis era will presumably have been influenced by the practice of elective abortion. 612  REPRODUCTIVE CYTOGENETICS infrequently, with moderate to severe oligospermia (Chapters 5 and 9). Robertsonian translocations are occasionally associated with infertility in the male, or less often the female (Chapter 7). Translocation between a sex chromosome and an autosome is a rarely identified cause of infertility (Chapter 6). Complex rearrangements (Chapter 10) and rings (Chapter 11) usually present an insurmountable obstacle to cell division in the spermatocyte, resulting in azoöspermia; oögenesis is typically more robust. An association of infertility with chromosomal polymorphism is controversial. In a Hungarian study from Andó et al. (2022), the common inv(9)(p11q12)/inv(9) (p11q13) proved to be less frequent among infertile males, and no different among Figure 20–28.  The Fractions of Male and Female Infertility having a Chromosomal Basis. Note: These data were derived from couples attending fertility clinics, and who went on to take advantage of assisted reproductive technology (IUI, IVF, ICSI). The fractions in men and women in this study (2,646 men, 2,620 women) are remarkably similar, 1.5% cf. 1.3%. Source: From the data in A Riccaboni et al., Genetic screening in 2,710 infertile candidate couples for assisted reproductive techniques: results of application of Italian guidelines for the appropriate use of genetic tests, Fertil Steril 89:800–808, 2008. Figure 20–29.  The Categories of Chromosomal Abnormality seen in Male and Female Infertility. Note: This study included 40 men and 34 women (representing the “wedges” in Figure 20–28 above). inv = inversion, rcp = reciprocal translocation, rob = Robertsonian translocation. It is notable that Robertsonian translocations were seen twice as often in the male, and inversions twice as often in the female. The sex chromosomal abnormalities were, in the male, mostly Klinefelter syndrome, and in the female, X mosaic states; in some women, low-level X mosaicism may or may not have been relevant. Source: From the data in A Riccaboni et al., Genetic screening in 2,710 infertile candidate couples for assisted reproductive techniques: results of application of Italian guidelines for the appropriate use of genetic tests, Fertil Steril 89:800–808, 2008.

REPRODUCTIVE FAILURE  613 females, than in a control population. Yet, in Yuan et al. (2023), an abnormal sperm morphology was more often seen among inv(9) heterozygotes. INFERTILITY: CHROMOSOMAL FACTORS IN THE FEMALE Fertility in the 46,XX female begins to fall in the mid-30s as the ovarian reserve dwindles. The average maximum number (but a wide range) of 300,000 ovarian follicles is reached in mid-fetal life, reducing thereafter to less than 1,000 remaining at the age of menopause (Figure 20–30). A diminishing ovarian reserve, as indicated by a low level of anti-Müllerian hormone, may point to an increasing risk for aneuploidy (Gat et al. 2017). An important age-related factor may be a decline in the functional competence of the meiotic spindle, compromising chromosomal distribution and leading to the generation of aneuploid gametes (Figure 20–6). Premature ovarian insufficiency (POI), also referred to as premature ovarian failure (POF), can be primary or secondary, associated with primary or secondary amenorrhea. Several genes have been implicated (França and Mendonca 2022; Fan et al. 2023a); our focus here is limited to the chromosomal. X Chromosome. Classical and variant forms of Turner syndrome account for a number of cases of female infertility (Table 15–1). The variant forms include mosaic states which involve a 45,X cell line, and which may be in the company of a normal 46,XX cell line, a 47,XXX cell line, an isochromosome of the X long arm, or a ring of the X chromosome. In a Malaysian study, Ten et al. (1990) karyotyped 117 women with primary amenorrhea who had previously been investigated for other causes, and one-third had a sex chromosome abnormality. They were classified as follows: X aneuploidies (8%), Table 20–4.  An Overview of the Frequencies (%) of Chromosomal Abnormalities Associated with Couple Infertility Male Female All Chromosome aberration 2.5 1.6 2.0 Autosomal 1.1 1.1 1.1 Translocation 1.0 0.9 1.0 Reciprocal 0.6 0.7 0.7 Robertsonian 0.4 0.3 0.3 Inversion 0.1 0.1 0.1 Other 0.03 0.04 0.04 Sex chromosome 1.1 0.1 0.6 Numerical 1.0 0.1 0.5 Structural 0.1 0.05 0.1 Mosaicism 0.1 0.3 0.2 Marker chromosome 0.03 0.6 0.5 Notes: Percentages are rounded. Karyotyping was performed in 17,054 patients affected by infertility, which was defined as not being able to conceive a pregnancy after at least 12 months of unprotected intercourse. This study did not include analysis for Yq AZF deletion. Source: From J Yuan et al., Detection of chromosome aberrations in 17 054 individuals with fertility problems and their subsequent assisted reproductive technology treatments in Central China, Hum Reprod 38(Suppl ii):34–46, 2023. 614  REPRODUCTIVE CYTOGENETICS X structural abnormalities (7%), presence of a Y (14%), and presence of a gonosomal marker chromosome (2%). Six women were mosaic, all having a 45,X cell line. Similar findings from India, although of a lesser frequency (16%) overall, are due to Chandel et al. (2023) (Table 20–5). As for secondary infertility, Devi and Benn (1999) studied 30 women with unexplained secondary amenorrhea under the age of 40 years. Four (13%) had chromosomal abnormalities: an Xq isochromosome, Turner syndrome mosaicism (45,X/46,XX), an X-Y translocation, and an X-autosome translocation. Low-level X aneuploidy may or may not matter. The conundrum of how to interpret low-level X aneuploidy—for example, just one or two 47,XXX or 45,X cells15—is addressed in a newsletter from the European Cytogeneticsts Association (Madan and Lundberg 2015). The salient points listed in this article are worth noting: • The frequency of low-level X aneuploidy is correlated with age and gender but not with reproductive history. • There is no significant difference in the number of aneuploid cells between women with recurrent abortions (1.64%) and age-matched controls (1.78%). • Approximately 16% of women of reproductive age show X aneuploidy in 2%–10% of blood cells. • The finding of low-level X aneuploidy in peripheral blood of a mother is not a predictor of fetal aneuploidy. • Low-level X aneuploidy is found in blood but not in skin or bone marrow. Figure 20–30.  The Natural Fall-off in Oöcyte Reserve with Increasing Age. Notes: Numbers of egg cells (non-growing follicles) are shown at different ages. 95% prediction intervals are shown in brackets, and as the two outer limits of the graph. Ages are in months, prior to birth (0); ages after birth in years. Mid-fetal refers to 18-22 weeks post conception. The y axis is a logarithmic scale. Source: Redrawn from WHB Wallace and TW Kelsey, Human ovarian reserve from conception to the menopause, PLoS One 5:e8772, 2010. Courtesy WHB Wallace, and with the permission of the Public Library of Science. 15 See also Mosaic loss of the X, p. 471. REPRODUCTIVE FAILURE  615   Autosomal Translocation Carriers. Of the many possible causes of POI, an uncommon finding is an autosomal structural anomaly, mostly rcp or rob, accounting for only 2% of POI cases (Vichinsartvichai 2016; Chen et al. 2023e) and with only a single representative of such a case, 46,XX,der(10)t(10;15)(p15;q22.1), in Table 20–5. A most remarkable coincidence leading to infertility in a young woman is described in Kuechler et al. (2010). Her father was heterozygous for a mutation in the FSHR (follicle-stimulating hormone receptor) gene, which is located at 2p16.3; and her mother carried an apparently balanced translocation, t(2;8)(p16.3;p23.1), but which in fact had a microdeletion at the 2p16.3 breakpoint demonstrable on microarray. This microdeletion removed two exons of the FSHR gene. The daughter inherited this translocation, plus the paternal mutation, and in consequence the former “unmasked” the latter; no normal FSHR was produced, and folliculogenesis was arrested. INFERTILITY: CHROMOSOMAL FACTORS IN THE MALE Much couple infertility is associated with inadequate (quantitative and qualitative16) sperm production in the male partner, and a fraction of this, in particular where Table 20–5.  Chromosomal Observations (%) in a Population of 470 Women presenting with Primary Amenorrhea Numerical 46,XX 84% 45,X 4.3 46,XY (sex reversal) 3.4 Mosaicisms, numerical 45,X/46,XX/47,XXX and 45,X/46,XX 1.3 45,X/46,XY and 45,X/47,XXY and 46,XX/46,XY 0.6 Structural Del Xp and del Xq 0.9 Inv X 0.2 Autosomal rcp 0.2 Mosaicisms, structural 45,X/del X 0.9 45,X/iso X 0.9 45,X/46,X+mar and 47,XX+mar 0.9 45,X/46,derX from an X-X translocation 0.2 Note: Those abnormalities listed as 0.2% were seen in only single cases. Source: From D Chandel et al., Clinical profile and cytogenetic correlations in females with primary amenorrhea, Clin Exp Reprod Med 50:192–199, 2023. 16 Quantitative: A normal number of sperm must be produced, namely, 15–200 million per ml of semen. Oligospermia (“few sperm”) may be mild (10–15 million), moderate (5–10 million), or severe (<5 million). A complete absence of sperm is azoöspermia. Qualitative: Sperm should have a normal appearance and normal motility. Teratozoöspermia refers to an abnormal appearance of sperm: asthenozoöspermia is a lack of “swimming” capacity.

616  REPRODUCTIVE CYTOGENETICS obstructive causes are excluded, is associated with an abnormal karyotype. In men with demonstrable azoöspermia or oligospermia, numerical and structural gonosomal abnormalities (mostly XXY, and Y rearrangements) and structural autosomal abnormalities (mostly reciprocal and Robertsonian translocations) are identified in around 10% (Fu et al. 2023). On testicular histology, several of these chromosomal states are associated with a Sertoli-only syndrome (also known as del Castillo syndrome), in which germ cells are absent (Ghanami Gashti et al. 2021). Infertility in 46,XY men, the more so in those with severe teratozoöspermia, is associated with an increased risk to produce aneuploid sperm (Ioannou et al. 2019). ICSI Associations. A French study (Gekas et al. 2001) brought together all the ICSI (intracytoplasmic sperm injection) programs in France over a three-year period and included some 3,208 individuals—2,196 men and 1,012 women—who had come forward as candidate couples for ICSI. Each individual had at least 20 metaphases examined. Sex chromosome mosaicism at a level of <10% was categorized as “minor.” In the men, 6% showed a chromosomal abnormality, and in the women (excluding insignificant minor sex chromosome mosaicism) 2%, even though a basis of the infertility had been supposed to lie in the male partner (Table 20–6). Combining with data from 10 other similar series, consensus figures are derived of 5%–6% and 4%–5% for male and female karyotypic abnormality, respectively. As noted above, the male figure is higher (10%) if sperm abnormality is shown.   X Chromosome Abnormalities. The XXY state, Klinefelter syndrome, is the most frequently observed abnormal classical karyotype in infertile males. Some of these may be (and more especially in men with extreme oligospermia rather than azoöspermia) low-level XY/XXY mosaics (Madian et al. 2020). A rare variant form is iso(Xq) Klinefelter syndrome, 47,X,iso(Xq)Y (Simsek et al. 2019). Otherwise, mosaicism with a 45,X cell line—that is, a 45,X/46,XY karyotype—is often associated with infertility (Akinsal et al. 2018). Very rare forms include XX//XY chimerism (Sugawara et al. 2005; Higgins et al. 2014),17 the X-autosome translocation (Chapter 6), and the “XX male” (p. 769).   Y Isochromosomes. The long-arm and short-arm Y isochromosomes, the respective karyotypes 46,X,idic(Y)(q) and 46,X,idic(Y)(p)18 (Figure 15–7), may be seen in non-mosaic and in 45,X/46,X,i(Y) mosaic forms. While abnormal genital phenotypes may be associated with this karyotype, here we are discussing the otherwise normal male presenting with infertility. This is seen in either type, long-arm and short-arm. In the iYq, Yq-located AZF spermatogenesis loci are lost; and in the iYp, there is a double amount of Yq and the AZF loci (Figure 15–6) (Codina-Pascual et al. 2004). Complicated forms are due to mosaic states (Hemmat et al. 2013; Jiang et al. 2020b; Aftab et al. 2020). The typical clinical correlate of the iYq is a Sertoli-only syndrome (Van Cauwenberghe et al. 2020).   Yq AZF Microdeletions. While the initial discovery had been made by cytogeneticists (Tiepolo and Zuffardi 1976), these Y deletions are mostly not detectable cytogenetically and are routinely analyzed using molecular methodology (and are not listed in 17 A man with 46,XY[79]//46,XX[22] chimerism in Higgins et al. (2014) had presented due to couple infertility, albeit that his reproductive indices were normal; the only abnormal sign in himself was hypomelanosis of Ito. 18 Some inconsistency in nomenclature, according to Yp or Yq breakpoint, is noted in footnote 7 on p. 487. REPRODUCTIVE FAILURE  617 the Tables above). These deletions account for around 10% of male infertility (Fu et al. 2023). The deletions remove AZF19 spermatogenesis factors in Yq11.223 and Yq11.23, AZF a, b, and c; the commonest is AZFc (Figure 6–19). If some fertility is retained, or successfully provided in the IVF clinic, the defect is transmissible from father to son (Cram et al. 2000).   Translocation Carriers. In the setting of a balanced rearrangement, gametogenesis in the male heterozygote is vulnerable to the stumbling block imposed by a chromosomal abnormality, and infertility occasionally results (Table 20–4).20 The frequencies of asthenozoöspermia, oligospermia, and teratozoöspermia in rcp and rob subjects in Yuan et al. (2023) were higher than among the 46,XY infertile men in their series. Notwithstanding the above, it remains true that fertility is usually unimpaired, or scarcely impaired, in male translocation heterozygotes. In their study of just over 10,000 sperm donors, all of proven fertility, the frequency of men carrying reciprocal and Robertsonian translocations, and also pericentric inversions, did not differ with statistical significance from Table 20–6.  Chromosome Findings (Frequencies in %) in a Series of Candidate Couples for Intracytoplasmic Sperm Injection Karyotype Female Male Sex chromosomal * 47,XXY 2% 47,XXY mosaic forms 0.4 45,X/46,XY 0.4 47,XYY 0.4 Structural Y 0.4 X or Y autosome translocation 0.2 Autosomal Reciprocal translocation 0.7 0.8 Mosaic translocation 0.1 Complex translocation 0.1 Robertsonian translocation 0.7 0.8 Pericentric inversion 0.7 0.1 Unbalanced abnormality 0.2 Normal chromosomes 98 94 Note: There were 2,196 men and 1,012 women in this study. The pre-eminence of Klinefelter syndrome is to be noted. Some female partners not studied, since it had been assumed the infertility was in the male. *Excluding a variety of low-level X chromosome mosaicisms in 28 women, of doubtful significance. See low-level X aneuploidy above. Source: From J Gekas et al., Chromosomal factors of infertility in candidate couples for ICSI: an equal risk of constitutional aberrations in women and men, Hum Reprod 16:82–90, 2001. 19 AZF (azoöspermia factor) is also known as DAZ (deleted in azoöspermia). 20 An important element in this male vulnerability may be the integrity at meiosis of the X-Y bivalent, synapsing and recombining at the pseudoautosomal regions at the tips of Xp and Yp—the “sex vesicle.” Unpaired or aberrantly associating autosomal segments, particularly of the acrocentric chromosomes, might disturb this integrity, leading to disruption of spermatogenesis (Oliver-Bonet et al. 2005b; Pinho et al. 2005; Vialard et al. 2006a). Another element may be impaired synapsis of homologous segments in the normal and the rearranged chromosomes, which of itself prevents further progress in gametogenesis, and spermatogenesis may be more sensitive to this obstacle than is oögenesis (Oliver-Bonet et al. 2005a).

618  REPRODUCTIVE CYTOGENETICS the general population (Ravel et al. 2006a). The semen indices of these men were within normal ranges. This epidemiology indicates that while it is true that a few heterozygotes may have impaired fertility, their numbers are too small to sway the figures of a large carrier population into statistical significance. CONSIDERATIONS RELATING TO IN VITRO FERTILIZATION It was reasonable to have imagined that IVF-conceived babies might have been more likely to suffer a chromosomal abnormality, given the facts of a presumed parental infertility and the artificial circumstances of their conception. But the observation is that there is little, if any, such risk. One of the largest and most stringent studies addressing this question comes from Australia, in which 6,946 IVF babies born in the period 1991– 2004 were compared with 20,838 controls (Halliday et al. 2010). The rate of chromosome abnormality in the IVF babies was 0.99%, compared with 0.97% in the non-IVF babies: an insignificant difference. Karyotyping of the oligospermic man is prudently to be done before proceeding to ICSI (Bonduelle et al. 2002). Bofinger et al. (1999) provided ICSI to a couple, the husband having severe oligospermia and the wife being of older childbearing age. At amniocentesis, on the grounds of the mother’s age, a 45,X/46,X,r(Y) chromosome constitution was discovered, and belatedly, the same karyotype was found in the father. The experience of Veld et al. (1997) is equally telling concerning two men who, both having suffered reproductive misfortune following ICSI, turned out to have a Robertsonian translocation of chromosome 13. It has been hypothesized that the delicate interplay whereby the epigenetic reprogramming of chromosomes is applied, according to parent of origin, might be vulnerable in the artificial setting of IVF (De Rycke et al. 2002). This appears indeed to be the case with respect to Beckwith-Wiedemann syndrome due to epigenetic error, perhaps more so in the case of ICSI having been employed, and the risk is severalfold that of the general population (Johnson et al. 2018). No such effect is discerned in other imprinting syndromes, including Angelman, Prader-Willi, and Russell-Silver syndromes (Henningsen et al. 2020) (and see Chapter 19). Genetics and Pathogenesis of Hydatidiform Mole Hydatidiform mole is an abnormal pregnancy that in the most usual form can be considered, in a sense, as a male chromosomal disorder. Typically, there is either a completely paternal karyotypic origin (two haploid paternal sets, 2n = 46) or an additional male haploid set (two paternal and one maternal haploid sets, 3n = 69) (Florea et al. 2023). The presence of two paternal chromosomal complements is referred to as “diandry.” The chorionic villi undergo a degenerative change, forming fluid-filled sacs (hence hydatidiform; mole means mass). The characteristic appearance has long been recognized (Figure 20–31); the “snowstorm” appearance on ultrasound imaging is very distinctive. The phenotype is marked (“complete mole,” 46-count) when the genetic origin is completely paternal, and attenuated (“partial mole,” 69-count) in the presence of a maternal haploid contribution. Most cases are sporadic, but a fraction—more so in the complete mole—are recurrent. REPRODUCTIVE FAILURE  619 COMPLETE MOLE The complete mole typically comprises only placental tissue. The usual karyotype is 46,XX, looking at first sight like a normal female karyotype. This is due, in most, to a doubling (endoreduplication) of the chromosomal complement of a single 23,X sperm, while a few are dispermic. In either of these cases, there is no maternal chromosomal contribution. With the mole’s nuclear genome being of entirely paternal origin, there is a total uniparental paternal disomy (“uniparental diploidy”). Moles due to doubling of a sperm chromosomal complement are entirely homozygous; in other words, they have a complete uniparental isodisomy, and this extraordinary functional imbalance leads to the phenotype. A minority of cases are biparental, and these are mostly due to maternal NLRP7 or KHDC3L homozygosity (Nguyen et al. 2018; Rath et al. 2023). Complete mole occurs more often at the beginning and end of reproductive life in the female. It is more common in the early teenager and in women in their forties (Bagshawe and Lawler 1982). Recurrence is typically related to maternal homozygosity for one of the genes aforementioned. PARTIAL MOLE An additional paternal haploid chromosome set is the basis of most cases of partial mole. This is diandric triploidy, 69,XXX or 69,XXY (rarely 69,XYY), which may typically be the result of a normal ovum fertilized either with two sperm (dispermy) or with a diploid sperm. Partial moles usually present as threatened, incomplete, or missed abortion (Figure 20–18) during the late first or early second trimester, the mean at 12 weeks. There is hydatidiform change of some villi, and the placenta is abnormally large; evidence of a fetal component may or may not be discernible. In a twin pregnancy from two Figure 20–31.  The Appearance of Hydatidiform Mole, Quite Probably the First Recorded Depiction (Baillie 1799).

620  REPRODUCTIVE CYTOGENETICS zygotes, one diploid and the other triploid, a normal fetus can coexist with a molar placenta (Barinova et al. 2022). A similar picture can result from a post-fertilization mitotic error in an initially normal single zygote (Makrydimas et al. 2002). GENETIC COUNSELING Infertility, Failure to Conceive Infertility is common, and the role of the genetic counselor in its management is becoming central (Verpoest et al. 2023). We focus here on chromosomal causes of infertility, but the counselor will need to be aware of other well-known genetic causes (fragile X pre-mutation with POI in the female; absence of the vas deferens due to unrecognized cystic fibrosis in the male), and emerging “new” genes (including TUBB8 and WEE2 in the female, and TEX11 and KLHL10 in the male). Among the catalog of investigative tests that are available, a karyotype is well up on the list, for male and female equally, and depending upon the initial first-tier infertility investigations. Karyotyping should be routine in women presenting with primary ovarian dysfunction or recurrent pregnancy loss. In the case of men with non-obstructive azoöspermia and oligospermia, both karyotyping and AZF (Yq microdeletion) testing is appropriate (Kalantari et al. 2023). Advances in reproductive technology now enable many otherwise infertile couples to have children. In the case of men with poor sperm production, intracytoplasmic sperm injection (ICSI) at IVF is a means to get a single sperm into an egg. Success with IVF is not necessarily easy to achieve, neither is it a certain outcome, and counselors dealing with infertile couples need a particular awareness of the psychological and practical difficulties they may face (Boivin et al. 2001). Intrinsic fertility cannot be restored in men with persistent azoöspermia associated with seminiferous tubule failure, and neither in women who have had ovarian failure. The counselor will need to understand how disappointing and indeed devastating this may be to some couples (sometimes one of them more than the other) and to be prepared to deal with this. The option of gamete donation, which may enable one of the couple to be a genetic parent, will often “bring home a baby.” Infertility with a Parental Chromosome Abnormality If a chromosomal defect is discovered in one or other of the couple, this at least provides an explanation for the infertility and (according to the exact nature of the defect) may prevent the disappointment of undergoing pointless further investigation. In some, artificial reproductive technology may enable a normal/balanced gamete to be identified and retrieved and used at IVF. Where this is impossible, IVF using donor gametes offers an entrée to parenthood.   Women. Ovum donation may be an appropriate option for women whose infertility is due to a sex chromosomal abnormality. If the internal anatomy is intact, success may well follow, as is rather notably illustrated by the patient reported in Chen et al. (2003b), who had a Turner syndrome variant due to an isodicentric X and who produced triplets following ovum donation from a sister. In some cases, the woman’s own mother, with REPRODUCTIVE FAILURE  621 whom of course she shares half her genes, has been the donor (potentially via oöcytes or ovarian tissue taken proactively in case of her daughter’s future need); sisters, and as noted above, may also be willing (Ye et al. 2020). Translocations and other rearrangements need to be assessed on their merits, as extensively discussed elsewhere.   Men. In those in whom the chromosome defect leads to oligospermia rather than complete failure of spermatogenesis, IVF with ICSI is a possible means to achieve pregnancy, and PGT will often be appropriate. A small (but growing) number of cases of fatherhood in men with Klinefelter syndrome have resulted from ICSI (p. 475). Rare sex chromosome abnormalities are judged individually. As with women, translocations and other rearrangements need to be assessed on their merits. In men with complete spermatogenic arrest, gamete donation may be considered, and a brother or father might be available and willing. In the case of a Yq microdeletion/AZF abnormality, couples choosing the option of IVF with ICSI should know that a male child would be predicted to have, very probably, the same type of infertility (Foresta et al. 2001). Some might consider having PGT to ensure having a daughter; although Kim et al. (1998) comment that “interestingly, after genetic counseling, the decision to proceed with ICSI for the overwhelming majority of couples remains unchanged.” Nap et al. (1999) assessed 28 such couples, and they interviewed the 10 counselors who had seen them, in six clinics in the Netherlands and in Belgium. A considerable majority of couples (79%) chose to continue with plans for ICSI, with only a few choosing donor insemination (7%) or opting out altogether (14%). Infertility with Parental Chromosomes Normal Where the 46,XY male has oligospermia, and if IVF with ICSI is to be attempted, there are grounds (discussed above) for presuming a very slightly increased risk for de novo structural aberration or gonosomal aneuploidy. It may be prudent to offer prenatal diagnosis for an ICSI-produced pregnancy in the setting of a severe paternal oligospermia (Bashiri et al. 2023). However, given the immense investment couples will have made to achieve the pregnancy, there may be reservation about proceeding to an invasive prenatal diagnostic procedure, even being aware of a possibly increased genetic risk. Noninvasive prenatal testing of cell-free DNA in maternal blood is an attractive alternative in this setting. Recurrent Miscarriage Couples who have had one or perhaps two miscarriages generally do not come to a genetic clinic, and neither, as a rule, do they have cytogenetic analysis of the products of conception or an analysis of their own karyotypes. Their physician or obstetrician may have advised them that this loss will very likely be part of the 15% or so of all pregnancies that miscarry, and the chance of a successful pregnancy in the future should be good. But having had three or more miscarriages requires investigation (and actually, some would have two as a threshold). To use the jargon, such couples have had multiple abortions, recurrent miscarriage, or recurrent pregnancy loss (RPL) (or to put it in Latin, abortus habitualis). At this point, a chromosome analysis of the couple is indicated. Normal Chromosomes in the Couple. The considerable majority of couples having suffered RPL will type 46,XX and 46,XY. In many centers, cytogenetic analysis of abortus

622  REPRODUCTIVE CYTOGENETICS material is not routinely done, and so chromosomal normality or abnormality cannot usually be demonstrated (Stephenson et al. 2002 speak of “this unfortunate omission” compromising the management of couples presenting with recurrent miscarriage.) Analysis of products of conception can be useful in offering clearer advice to these women (Papas and Kutteh 2021), and the Royal College of Obstetricians and Gynaecologists (RCOG) advise that “Cytogenetic analysis should be offered on pregnancy tissue of the third and subsequent miscarriage(s) and in any second-trimester miscarriage” (Regan et al. 2023). Kutteh et al. (2024) advise that if the work-up proposed by the American Society of Reproductive Medicine is applied (including a search for autoimmune, endocrine, and uterine anatomical factors), alongside a chromosome study, an explanation for RPL may be forthcoming in a substantial majority of cases (Figure 20–32). Trisomy of the conceptus is a common underlying cause, more so in women of older maternal age, but recurring euploid miscarriage may also be observed and if so, imply a less promising outlook (Figure 20–25). Euploidy of the POC likely reflects an underlying maternal factor that would apply to all pregnancies, whereas aneuploidy at least offers the hope that better fortune might attend the next ovulation. Past history of having had a child is relevant: Cohain et al. (2017) advise that “parous women until their late 30s, who have experienced multiple miscarriages, can be counseled that if they keep trying, they will likely carry a pregnancy to term . . . but she should act expediently”. PGT in a subsequent pregnancy would be discretionary, and while we acknowledge the policy of the RCOG that “There are currently insufficient data to support the routine use of PGT for aneuploidies (PGT-A) for couples with [otherwise] unexplained recurrent miscarriage, while the treatment may carry a significant cost and potential risk,” we note also the belief of Kutteh et al. (2024) that “there is a clear role for PGT-A in RPL, especially in cases with recurrent POC aneuploidy.” Liang et al. (2023) rehearse various factors that may influence the aneuploidy recurrence risk: maternal age, number of previous pregnancy losses, the estradiol level on the ovulation trigger day, and the blastocyst formation rate. Figure 20–32.  The Combined Value of Chromosome Microarray and Full Clinical Evaluation in Explaining Pregnancy Loss. Notes: Pie graphs show the fractions of pregnancy loss due to chromosomal abnormality (left), and overall fractions explicable if the investigative criteria of the American Society of Reproductive Medicine are applied, along with chromosome analysis (right). Source: From WH Kutteh et al., Role of genetic analysis of products of conception and PGT in managing early pregnancy loss, Reprod Biomed Online 49:103738, 2024. Courtesy WH Kutteh, and with the permission of Elsevier. REPRODUCTIVE FAILURE  623 Chromosome Abnormality in the Couple. If a chromosomal rearrangement in one of the couple is identified, this will very probably be the underlying cause, but the possibility of a fortuitous discovery is not to be discounted. The precise nature of the rearrangement (consult the appropriate chapter), the reproductive history of any others in the family who have it, and the presence or absence of gynecological pathology allow one to make a judgment of its role in the etiology of the abortions. An offer of PGT at IVF, at a future planned pregnancy, may be very appropriate in order to reduce the likelihood of another miscarriage (Li et al. 2024c). Recurrent Implantation Failure A “failed embryo transfer” following IVF may be considered as a form of pregnancy loss not unlike that of the natural miscarriage of a wanted pregnancy. A distinction needs to be made between a parental versus an embryonic cause. A detailed format is offered by the European Society of Human Reproduction and Embryology (2023), which includes advice on immune, endocrine, hematologic, and other factors in the couple. Concerning the chromosomal question, while there is some controversy, this expert group nevertheless comments that “ . . . where embryo testing is conducted by either array comparative genomic hybridization or NGS approaches on blastocyst biopsies . . . PGT for aneuploidies could be considered a good strategy for women with RIF, as a reduced number of embryo transfers [may be] required to achieve pregnancy and live birth”—a conclusion broadly supported in Mei et al. (2024). In other words, PGT is discretionary and potentially helpful, but with some reservation (Liu et al. 2024d). Fetal Death In Utero, Stillbirth Microarray analysis is useful in identifying a chromosomal cause (Martinez-Portilla et al. 2019). If such is discovered, appropriate genetic counseling can be provided according to the nature of the chromosomal finding. Hydatidiform Mole Having had one complete mole, the recurrence risk is 1%. But having had a second mole, the risk then rises to 1 in 4 (Eagles et al. 2015). Those suffering a second or third mole will often represent the rare category of a genetic predisposition due to maternal homozygosity for one of two known genes, NLRP7 and KHDC3L (Rath et al. 2023). This latter group is recognized by every pregnancy being molar (Kalogiannidis et al. 2018). These repeating cases typically show biparental inheritance, in contradistinction to the androgenetic basis of the majority of moles. Ovum donation offers the best chance for parenthood. The risk for recurrence of a partial mole is much less: “having a partial mole does not appear to significantly increase the risk of further hydatidiform mole in later pregnancies” (Eagles et al. 2015). If there is a recurrence, it will more likely be of the same type. In essence, the risk is largely that of a recurrent triploidy. Very rare exceptions with multiple recurrence exist, and presumably reflecting a different, yet to be clarified genetic basis (Filges et al. 2015; Guha Sarkar et al. 2022; Salima et al. 2023).

21 PRENATAL TESTING PROCEDURES THE MEANS TO DIAGNOSE the fetal karyotype provides medical cytogenetics with one of its major areas of application. The discovery of an abnormality allows the option of termination of the pregnancy or, later in gestation, a more suitable obstetric management. The main indications for prenatal cytogenetic diagnosis are the following: (1) fetal anomaly detected on ultrasonography; (2) increased risk on maternal screening test; (3) parental heterozygosity for a chromosome rearrangement; (4) the birth of a previous child with a chromosome defect; (5) in the absence of other screening, the pregnant woman being of older childbearing age; and (6) opportunistic testing when prenatal sampling has occurred for single gene testing. PRENATAL LABORATORY DIAGNOSTIC PROCEDURES Prenatal diagnosis (PND) of chromosome disorders began in the early 1970s by the culturing of amniotic fluid cells obtained from amniocentesis at about 16 weeks of pregnancy, and analyzed according to classical methodology. A number of other approaches to PND have since been developed, ranging from preimplantation genetic testing (following in vitro fertilization) through chorion villus sampling (CVS), to the testing of cell-free DNA in maternal plasma. Naturally, parents-to-be are anxious to have results as early as possible. A desire for an early result needs to be balanced against a number of considerations, which can include complexity of the procedure, both clinically and in the laboratory; procedural complication; reliability of results; cost; and the prior risk for a fetal abnormality. Four particular analytical procedures have enabled a faster return of results for common aneuploidies, compared with classical cytogenetics, namely FISH, QF-PCR, MLPA, and microarray analysis (and see also Chapter 2). Fluorescence In Situ Hybridization Fluorescence in situ hybridization (FISH) using multiple colored probes and targeting the chromosomes most prone to survivable aneuploidy (13, 18, 21, X, and Y) bypasses the need for culture, whether the cells are from amniotic fluid or CVS, and the result can be given within the space of one working day (Morris et al. 1999). By way of example, Feldman et al. (2000) applied amniotic fluid cell FISH to high-risk pregnancies (that is, with ultrasonographic abnormalities). They detected 14 cases of trisomy 21, 10 of trisomy 18, three of trisomy 13, four of monosomy X, and one triploid, in 4,193 samples over the period 1996–1998, for a total abnormality rate of 11%, and at a 24-hour turnaround. 626  REPRODUCTIVE CYTOGENETICS As have others, Weremowicz et al. (2001) note the usefulness of the FISH approach in being able to provide a rapid answer, particularly when there are grounds for suspecting an abnormality or if the pregnancy is more advanced; but they also emphasize the need for careful counseling so that patients are aware of the limitations. With respect to trisomy 21, Witters et al. (2002) had an encouraging record: In a study comprising 5,049 amniotic fluid samples, in which interphase FISH was applied in parallel with conventional karyotyping, all 70 cases of trisomy 21 were detected, and no false-positive result arose. Just one false positive is on record, probably due to technical aspects of probe hybridization (George et al. 2003). On the question of mosaicism, Van Opstal et al. (2001) note that FISH on uncultured cells may provide a more accurate picture than on cultured cells, the latter possibly being subject to selective pressure in vitro, with the abnormal cells more prone to fail in culture. On the other hand, the class of amniocyte that grows preferentially in culture (namely amniotic mesoderm) might, according to the reinterpretation of Robinson et al. (2002), more closely reflect the true embryonic state. Focused FISH can be applied in specific circumstances. The ultrasound discovery of a cardiac outflow tract abnormality would, for example, point to the need for 22q11 analysis. A rapid diagnosis is particularly to be desired in the setting of parental heterozygosity for a chromosome rearrangement in which there may be a high risk for abnormality, and FISH can provide this. Thus, Cotter and Musci (2001) used subtelomeric probes for 5pter, 5qter, and 14qter to enable rapid diagnosis for a pregnant woman with the karyotype 46,XX,t(5;14)(p14.2;p13), she having had a previous child with cri du chat syndrome. Similarly, Pettenati et al. (2002) applied this approach in the circumstance of parental heterozygosity for a number of reciprocal and Robertsonian translocations. Focused FISH can also be used to follow up NIPT results that show increased risk for a rare autosomal trisomy. FISH tests only that segment of the chromosome to which the probe binds. Inferentially, the complete chromosome is present; but this is not necessarily so. With chromosome 18, it is the centromere which the FISH probe recognizes. We have seen a case in which amniocentesis was done on the basis of a maternal age of 40 years, albeit that the risks based on serum screening were lowered for age (trisomy 21, 1 in 164; trisomy 18, 1 in 8,030). FISH showed three chromosome 18 signals. Fetal growth and morphology on ultrasonography were normal. The couple considered whether they might request termination, but wanted to await the result of karyotyping. This showed a supernumerary minute marker—barely a speck—which appeared to comprise only chromosome 18 centromere. The karyotype was 47,XX,+mar.ish der(18) (D18Z1+)dn[13]/46,XX[4]‌ (case of MD Pertile). The pregnancy continued. The child subsequently born was followed up at age three years, and while she manifested a familial shortness, her cognitive and personality development were entirely normal (S Fawcett, personal communication). Quantitative Fluorescence Polymerase Chain Reaction Quantitative fluorescence polymerase chain reaction (QF-PCR) relies on the use of molecular markers that display a high level of heterozygosity such that the presence of three alleles—that is, a trisomy—can reliably be detected. Similarly to FISH, when compared to karyotyping, QF-PCR has the advantages of lower cost, shorter turnaround of

PRENATAL TESTING PROCEDURES  627 results, and lower technical complexity. The major disadvantage is that QF-PCR will only identify common aneuploidies. Most deletions and duplications, along with balanced rearrangements, will not be detected. Cirigliano et al. (2009) assessed the relative performances of QF-PCR (for chromosomes 13, 18, 21, X, and Y) and standard karyotyping in a series of 43,000 amniocentesis samples tested over a 9-year period. The QF-PCR assay detected 95% of all clinically relevant chromosome abnormalities, and there were no false-positive results. In a smaller study, de la Paz-Gallardo et al. (2015) suggested that if karyotyping were restricted to samples where there was an ultrasound abnormality, and QF-PCR were used for lower-risk samples, laboratory costs could be halved, with minimal reduction in the detection of clinically significant abnormalities. Nevertheless, viewed against noninvasive prenatal testing (see below), the advantages of QF-PCR become less compelling. Multiplex Ligation-Dependent Probe Analysis Multiplex ligation-dependent probe analysis (MLPA) is a PCR-based assay that can combine probes to many chromosomal loci, and which again has the advantage of a short turnaround time (Schouten et al. 2002; Shaffer and Bui 2007). In a comparison with standard chromosome analysis of 4,585 amniocentesis specimens, MLPA had 100% sensitivity and specificity for identifying common aneuploidies, but it failed to pick up 26 other abnormalities that were detected by karyotyping, including 12 with potential clinical significance (Boormans et al. 2010). A targeted application enables the diagnosis of microdeletion syndromes that would otherwise escape detection (Konialis et al. 2011). Microarray Chromosome microarrays have been used in prenatal diagnosis since the mid-2000s (Le Caignec et al. 2005; Rickman et al. 2006), but their use only became widespread after the publication in 2012 of two large prospective studies that demonstrated their superior diagnostic yield compared to conventional karyotyping (Shaffer et al. 2012; Wapner et al. 2012). Microarray analysis can be applied to CVS and amniocentesis samples, with or without prior culture. Results can be delivered within a few days, although in practice it may take longer due to the efficiency benefits for laboratories afforded by “batching” samples. The additional diagnostic yield is greatest (~10%) in fetuses with multiple ultrasound abnormalities, but benefit is also apparent, at a level of approximately 1%, in lower-risk women such as those of advanced maternal age. A possible drawback is that microarrays can show CNVs of uncertain significance or low penetrance for phenotypes not able to be assessed in the prenatal setting; such findings are found in ~1% of tests (Hillman et al. 2013). A pragmatic response is for CNVs with mild phenotypes or low penetrance not routinely to be reported in the prenatal setting (Armour et al. 2018; Durkie et al. 2024).1 1 The increased application of genomic testing has led to greater acceptance of different reporting policies for prenatal compared to post-natal samples, particularly in relation to variants of uncertain significance.

628  REPRODUCTIVE CYTOGENETICS Low Pass-Genome Sequencing Low pass-genome sequencing (LP-GS) refers to genome sequencing performed at a lower average depth of coverage, specifically below 15× (Raca et al. 2023). LP-GS can be used to detect chromosome deletions and duplications with a resolution at least equivalent to microarray (Wang et al. 2020a), and with advantages in terms of workflow efficiency and scalability (Mighton et al. 2024). Importantly, the use of genomic sequencing for copy number detection provides the opportunity for a single streamlined platform for the detection of a broad range of genetic variants. PRENATAL DIAGNOSTIC CLINICAL PROCEDURES Chorionic Villus Sampling Chorionic villus sampling (CVS) is typically a first-trimester procedure, the usual time being at 10–11 weeks gestation. The normal approach is transabdominal (transvaginal CVS was formerly used). The operator inserts a needle through the lower abdominal wall, under ultrasound guidance, to penetrate to the placental tissue; with gentle negative pressure on the syringe, a small amount of chorionic villus tissue is aspirated. (The expression “placental biopsy” could also be applied, although in practice this term is used when the testing is done in later pregnancy; see below). Compared with amniocentesis, the earlier period of diagnosis permitted by CVS may be seen as more useful in the setting of a higher genetic risk. If a genetic abnormality is identified and abortion is chosen prior to 14 weeks, this can be a more private matter, and the termination procedure is an operative intervention (curettage or suction evacuation of the uterus) (Verp et al. 1988). There is potential in CVS for diagnostic difficulty due to the occasional detection of confined placental mosaicism (which may, for some chromosomes, carry a risk also for uniparental disomy). Non-mosaic results for the common aneuploidies are, however, highly reliable (Smith et al. 1999). In experienced hands, there is a high degree of safety: A meta-analysis concluded that the risk of procedure-related miscarriage in women who had undergone CVS was only 0.22%, about 1 in 500 procedures (Akolekar et al. 2015). Direct, Short-Term, and Long-Term Chorionic Villus Sampling.  Chorionic villi can be analyzed directly (same day), after short-term culture (next day or two), or after long-term (a week or so) culture. For microarray analysis, DNA extracted from whole villi (uncultured cells) are typically used as the source of DNA, but “backup” cultures can be established that are available if the microarray fails, or if a microarray finding requires additional confirmation. For karyotype analysis, most laboratories offer only long-term CVS in which the mesenchymal core of the villus is the source of the analyzed cells. Trophoblast is the source of the cell population studied at direct and short-term CVS culture; these cells are no longer extant after the first few days in culture. DNA from uncultured CVS is likely to be derived from both cytotrophoblast and mesenchyme, and if mosaicism is detected in these samples, confirmation on cultured cells or amniocentesis should be considered. In the early 1990s, there were disconcerting reports of an increased incidence of transverse limb deficiencies and tongue and jaw defects—“oromandibular-limb PRENATAL TESTING PROCEDURES  629 hypogenesis”—following early CVS (before 10 weeks, and especially up to 8 weeks). The association appeared likely to be causal, and one line of circumstantial evidence was that the rate of anomalies fell with increasing gestational stage from 9 to 11 weeks (Firth 1997). Various mechanisms were proposed: oligohydramnios, bradycardia, hypovolemia, thromboembolism, vasoconstriction, antibody-mediated reaction, and increased apoptosis following disruption of end arteries (Luijsterburg et al. 1997). Given these observations, it became normal practice for CVS not to be done earlier than 10 weeks. Amniocentesis Transabdominal amniocentesis from 15 weeks’ gestation onwards, with culture of cells for chromosome analysis, has been the standard cytogenetic prenatal diagnostic procedure since the 1970s. It has a high degree of safety to both mother and fetus: Maternal complications or fetal injury due to direct trauma are practically unknown. The risk for maternal Rhesus immunization (Rh-negative mother, Rh-positive fetus) can be circumvented by administering an antibody injection. In their meta-analysis, Akolekar et al. (2015) concluded that the risk of procedure-related miscarriage in women who had undergone amniocentesis was 0.11%, about 1 in 1,000 procedures. The cytogenetic results are very reliable. The biological sources of error are, first, that maternal rather than fetal cells, or a mixture of both, are sampled. In practical terms, this rarely causes a problem. Second, fetal mosaicism may go undetected, since only a limited number of cells can feasibly be examined. Very few examples of this error are recorded. Amniotic fluid culture has a high success rate. If amniotic cells fail to grow for no obvious reason, there may be a substantial risk for fetal aneuploidy. Reid et al. (1996) followed up 42 failures (1%) among 4,134 amniocenteses. Complete information could be obtained on all but one of these 42 cases. Karyotyping was ultimately done in most (78%) of these failed cases, and of these, 19% revealed an abnormality (compared with a 4% abnormality rate in the whole material). The clear lesson from these studies is that women having had a failed amniocentesis culture should be offered careful review and retesting. Prior cell culture is not necessary for microarray analysis but, as with CVS, “backup” cultures may be established for use only if the microarray fails or if an abnormal microarray result requires confirmation or additional analysis. As with CVS, DNA from uncultured cells may provide a more accurate picture than from cultured cells, due to chromosomally abnormal cells being selected-against in vitro or being more prone to fail in culture. The obvious disadvantage of amniocentesis is that the results are not available until about 16–18 weeks. If the reason for the amniocentesis had been an abnormality on second-trimester ultrasound, the procedure may not be done until 18–20 weeks, aggravating this difficulty. The outlook for the long-term health of the child does not differ between CVS and amniocentesis (Schaap et al. 2002). Noninvasive Prenatal Testing Using Cell-Free DNA from Maternal Blood Noninvasive prenatal testing (NIPT) for fetal aneuploidy detection, using cell-free DNA from maternal blood, has been used in the clinical setting since 2012. Although in reality

630  REPRODUCTIVE CYTOGENETICS NIPT is a screening test (some write NIPS) rather than a diagnostic test, in practical terms it is such a good test that we are considering it separately from other screening modalities. Cell-free DNA (cfDNA) testing takes advantage of the fact that everyone has free DNA (i.e., not within the nucleus of cells) circulating in the bloodstream, and in pregnant women a proportion of that DNA is derived from the fetus. Or more specifically, it is derived from the placenta, and in fact from apoptosing cells of the cytotrophoblast. This is the “fetal fraction” of the maternal sample which, in the late first trimester of pregnancy, typically comprises approximately 10% of the whole (Becking et al. 2024). cfDNA can be analyzed by short-read sequencing, using either whole-genome analysis or by methods that target specific chromosomes, and can also be quantified by microarray (Stokowski et al. 2015). Although it is possible to distinguish fetal DNA from maternal DNA through single nucleotide polymorphisms (SNPs) within the cfDNA sample, most methodologies simply measure the total amount of DNA (i.e., maternal and fetal contribution) derived from each chromosome, and compare each chromosome with other chromosomes within the sample. For example, in a euploid pregnancy with fetal fraction of 10%, for each chromosome the total amount of DNA is 90% maternal and 10% fetal. In the setting of a Down syndrome pregnancy, the fetal contribution from chromosome 21 increases by 50%, as there are now three copies of fetal chromosome 21 instead of the normal two. The relative excess in chromosome 21 cfDNA fragments will be (0.9 × 2) + (0.1 × 3) = 2.1, compared with the normal for a euploid fetus of (0.9 × 2) + (0.1 × 2) = 2.0. The relative increase in the number of chromosome 21 counts is therefore only 5% (2.1/2.0 = 1.05). The fact that such a small difference can be measured accurately is testament to the power of short-read sequencers as a molecular counting tool. NIPT for the Common Trisomies.  NIPT using cfDNA for trisomies 21, 13, and 18 has very high sensitivity and specificity (Table 21–1). A related and possibly more relevant measure of test accuracy is the positive predictive value (PPV), which equates to the proportion of women with an abnormal result who actually do have an aneuploid fetus. The PPV incorporates the pretest likelihood that the aneuploidy is present, and therefore increases with maternal age. Even when an NIPT test has very high sensitivity and specificity, in a young woman with low pretest risk of having an aneuploid fetus, the PPV for a rare aneuploidy such as trisomy 13 will be low. Estimates of PPV for common aneuploidies detected by NIPT at different maternal ages are presented in Table 21–2, and they serve as a useful reminder that high-risk NIPT results should always be followed up Table 21–1.  Sensitivity and Specificity of NIPT for Common Aneuploidies ANEUPLOIDY SENSITIVITY (%) SPECIFICITY (%) Trisomy 21 99.4 99.9 Trisomy 18 97.7 99.9 Trisomy 13 90.6 100 Source: The meta-analysis of A Mackie, Using cell-free fetal DNA as a diagnostic and screening test: current understanding and uncertainties, BJOG 124:47, 2017. PRENATAL TESTING PROCEDURES  631 with an invasive test. Any decision to terminate a pregnancy should not be based upon positive NIPT results alone. While early trials focused on women from high-risk groups, more recent studies have demonstrated the clinical utility of NIPT in low-risk and average-risk women. Although the use of NIPT in low-risk women is associated with a lower PPV than for higher-risk women, leading to more false-positive results, NIPT still outperforms conventional screening by a very considerable margin (Gregg et al. 2016). NIPT for Other Trisomies.  It is technically straightforward to expand NIPT to autosomes other than 13, 18, and 21, and in fact NIPT methods that use whole-genome sequencing already collect the necessary sequencing data. But the clinical utility of this information is not clear. Non-mosaic whole chromosome aneuploidy, other than the three common aneuploidies, typically results in fetal loss. In an ongoing pregnancy, the trisomic cells are often mosaic and confined to the placenta, and detection of these abnormalities may lead to parental anxiety and unnecessary diagnostic procedures. That is not to say that the detection of these “Rare Autosomal Trisomy” (RATs) might not be without clinical benefit. Finding a RAT at NIPT has been seen with more pregnancy complication (Lannoo et al. 2024) and, in rare instances, discovery of certain aneuploidies might have direct implication for the pregnancy—one example being the detection of trisomy 15 in cfDNA leading to the diagnosis of Prader-Willi syndrome due to uniparental disomy of chromosome 15 (Hong et al. 2023). Nevertheless, as of 2023, the American College of Medical Genetics and Genomics has recommended against screening for autosomal aneuploidies other than 13, 18, and 21 (Dungan et al. 2023). NIPT for Sex Chromosome Aneuploidies.  NIPT for sex chromosome aneuploidies has a detection rate of >90% and a PPV of approximately 50%, similarly to the other common aneuploidies (Shear et al. 2023) (Table 21–3). But caution is needed. Sex chromosome aneuploidies are more common than autosomal, in one series being detected in 1 in 272 amniocentesis samples (Forabosco et al. 2009). In consequence, inclusion of sex chromosome aneuploidies in an NIPT test will increase the overall number of invasive tests and the number of samples yielding a false-positive result. The phenotypes in the common sex chromosome abnormalities are typically much less severe than autosomal aneuploidies, and in some cases may not be clinically apparent (as discussed in the following Chapter 22). On the other hand, early identification of sex chromosome abnormalities might allow post-natal management to be optimized, with potential long-term benefit to the affected child (Martin et al. 2023). Careful pretest counseling should accompany NIPT for sex chromosome abnormalities, and some pregnant women may choose to restrict testing to autosomal trisomies. Table 21–2.  Positive Predictive Values of NIPT Results at Different Maternal Ages PPV (%) AT MATERNAL AGEs ANEUPLOIDY 25 YEARS 30 YEARS 35 YEARS 40 YEARS Trisomy 13 7 10 21 50 Trisomy 18 15 21 39 69 Trisomy 21 51 61 79 93 Source: The online calculator at htpp://secure.itswebs.com/nsgc/niptcalculator/index.html which uses sensitivity and specificity from Gil et al. (2015).

632  REPRODUCTIVE CYTOGENETICS NIPT for Microdeletions, Microduplications, and Copy Number Variants.  NIPT can be extended to detect microdeletions such as those of 1p36, 15q11q13, and 22q11.2. These microdeletions are associated with severe phenotypes, and on that basis they are appropriate for prenatal screening. NIPT can detect some microdel/dups with high sensitivity and low false-positive rate; but a concern is that, false-positive results will accumulate as greater proportions of the genome are analyzed, leading to unnecessary invasive tests (Gregg et al. 2016). Improved sequencing and analytical techniques may assuage this concern. Lefkowitz et al. (2016) retrospectively analyzed 1,166 NIPT samples, including 42 known to harbor CNVs: All but one of the 42 CNVs were detected and there was only one false-positive result, attesting to the feasibility of whole-genome CNV analysis using cfDNA. Zaninović et al. (2022) found rather wide ranges of reported sensitivity (from 20% to 100%), specificity (from 82% to 100%), and PPV (from 3% to 100%). Reflecting these uncertainties, the American College of Medical Genetics and Genomics recommended in 2023 offering NIPT for 22q11.2 deletion syndrome, but not for other CNVs (Dungan et al. 2023), whereas the American College of Obstetricians and Gynecologists (2020) does not support NIPT for any CNV. False-Positive and False-Negative NIPT Results.  Cell-free fetal DNA is, as noted above, placental in origin, and so NIPT by cfDNA shares the major pitfall of CVS, namely being susceptible to confined placental mosaicism (CPM; see Chapter 22). False-positive results due to CPM cannot be overcome by technical improvements (Brady et al. 2016). Accepting that NIPT results are, in principle, concordant with CVS findings, amniocentesis may be necessary to confirm that the aneuploidy is (if this is the case) confined to the placenta. In the event that NIPT shows trisomy for an imprintable chromosome, testing for uniparental disomy should be considered if amniocentesis is chosen. Early and undetected co-twin demise (“vanished twin”) may lead to a false-positive test result (Brady et al. 2016). Chromosome abnormalities are common in vanished twins, and the involuting placenta of an aneuploid vanished twin may continue to release aneuploid DNA into the maternal bloodstream for weeks after demise. Using an SNP-based technology to identify fetal haplotypes, Curnow et al. (2015) were able to identify haplotypes from a vanished twin in 0.18% of pregnancies tested, with fetal Table 21–3.  Sensitivity, Specificity and PPV of NIPT for Sex Chromosome Aneuploidies ANEUPLOIDY SENSITIVITY (%) SPECIFICITY (%) PPV (%) Monosomy X 98.8 99.4 14.5%* 47,XXY 100 100 97.7 47,XXX 100 99.9 61.6 47,XYY 100 100 100 Note: These data are derived from pregnancies at high pretest risk for aneuploidy, and may not be generalizable to average-risk pregnancies. PPV = positive predictive value. *The use of a paired-end cfDNA sequencing assay differentiates short (placental) from long (maternal) cfDNA fragments, and can improve the PPV for monosomy X to >60%, by removing false-positive calls associated with maternal monosomy X mosaicism (Scarff et al. 2023). Source: The systematic review and meta-analysis of MA Shear et al., A systematic review and meta-analysis of cell-free DNA testing for detection of fetal sex chromosome aneuploidy, Prenat Diagn 43:133–143, 2023. PRENATAL TESTING PROCEDURES  633 demise having occurred between two weeks and eight weeks prior to maternal blood sampling. CNVs or mosaic aneuploidy in the mother may affect test interpretation. Unidentified maternal sex chromosome mosaicism (45,X/46,XX or 47,XXX/46,XX) is an important cause of false-positive NIPT results for sex chromosome abnormalities (Wang et al. 2014b). We have also found Y chromosome material in maternal DNA, which has led to a high-risk result for sex chromosome aneuploidy in the fetus. Less commonly, mosaicism in the mother for other chromosomes or chromosome segments can mislead (Brady et al. 2016). If NIPT is extended to CNVs, the possibility of unidentified CNVs being present in the mother and detected by NIPT will need to be considered, both in pretest counseling and in interpretation of results. A particularly devastating cause of aneuploidy detectable in cfDNA is maternal malignancy. Ji et al. (2019) identified 41 cases of occult malignancy from a series of 1.93 million NIPT samples. Maternal malignancies were typically associated with a “tumor-like” aneuploidy profile, comprising copy number gains and losses across multiple chromosomes (a pattern that is unlikely to be mistaken for fetal aneuploidy). Heesterbeek et al. (2024) provide guidelines for the management of pregnant women who receive these challenging results. Follow-Up Procedures Following Abnormal NIPT Results.  Invasive testing is recommended to confirm a high-risk cfDNA finding, with CVS the preferred modality, in order to allow for a more timely return of results. In a minority of cases, the CVS result will be mosaic, and this will necessitate a second procedure (namely amniocentesis) in order to distinguish CPM from true fetal mosaicism. If mosaicism is likely, there is an argument for bypassing CVS and using amniocentesis as the first-line invasive test. Grati et al. (2015) address this question by estimating the frequency with which a CVS, done after a high-risk cfDNA result, would need a follow-up amniocentesis due to suspected placental mosaicism. The authors did not actually do cfDNA testing, but rather modeled data based on results from more than 50,000 CVS karyotypes obtained from cytotrophoblast (direct preparation) and mesenchyme (long-term culture), and followed by confirmatory amniocentesis. Central to this modeling was the assumption that the fetal fraction of cfDNA originates mainly from the cytotrophoblast layer of the chorionic villus; and so, as noted above, the cfDNA test result should be concordant with direct rather than long-term cultures. The findings, along with the likelihood of mosaic CVS results being confirmed by amniocentesis, are shown in Table 21–4. Based on these interpretations, CVS is the recommended procedure following a high-risk cfDNA result for trisomy 21 or trisomy 18, but with the caveat of a 2%–4% risk of an inconclusive result, which would require a follow-up amniocentesis. For trisomy 13 and monosomy X, the benefit of an early diagnosis by CVS is to be balanced against the likelihood of an inconclusive result with an amniocentesis then needed (and ultrasonography may in any event have been clarifying). NIPT for Twin Pregnancies.  Judah et al. (2021) reviewed NIPT data for nearly 8,000 twin pregnancies for a meta-analysis. The performance of cfDNA testing for trisomy 21 in twin pregnancy was slightly less than that reported in singleton pregnancy (Table 21–5), but was nonetheless clearly superior to that of the first-trimester combined test or second-trimester biochemical testing. For trisomies 18 and 13, the detection rates were slightly lower than for singleton pregnancies, but confidence intervals were wide due to the smaller number of cases. The authors conclude that parents can be counseled

634  REPRODUCTIVE CYTOGENETICS that cfDNA testing in twin pregnancy is mainly useful in screening for trisomy 21, but that there is uncertainty concerning trisomies 18 and 13. Experimental, Less Used, or Former Technologies EARLY AMNIOCENTESIS In the late 1980s, early (10–13 weeks) amniocentesis was proposed as an alternative to CVS. In a carefully controlled comparison, Nicolaides et al. (1994) found a 2%–3% additional fetal loss rate in early amniocentesis, and possibly a higher incidence of talipes among subsequently born children. Daniel et al. (1998), comparing 10–14 week procedures with 15 weeks and upward, observed that the early amniocentesis samples were not quite as satisfactory; multiple needle insertions were more often required; and the pregnancy loss rate was greater. On the whole, the differences were not great, other than the loss rate of 2.2% in the early group compared with 0.6% in the mid-trimester group. Similar figures were reported in Collins et al. (1998). In the Canadian Early and Midtrimester Amniocentesis Trial, the findings for 11wk 0d through to 13wk 6d were somewhat more disconcerting, with more complications and a higher culture failure rate (Delisle and Wilson 1999). The procedure is rarely undertaken now. Table 21–4.  Follow-Up Scenarios after a High-Risk cfDNA Result Likelihood of Mosaicism at CVS Following High-Risk cfDNA Results for Common Aneuploidies, and Likelihood of Mosaicism at CVS Being Confirmed by Amniocentesis ANEUPLOIDY LIKELIHOOD THAT A HIGH-RISK cfDNA RESULT WILL BE FOLLOWED BY DETECTION OF MOSAICISM AT CVS (%) LIKELIHOOD THAT A MOSAIC CVS RESULT WILL BE CONFIRMED BY AMNIOCENTESIS (%) Trisomy 21 2 44 Trisomy 18 4 14 Trisomy 13 22 4 Monosomy X 59 26 CVS, chorion villus sampling. Source: FR Grati et al., The type of feto-placental aneuploidy detected by cfDNA testing may influence the choice of confirmatory diagnostic procedure, Prenat Diagn 35:994–998, 2015. Table 21–5.  Performance of cfDNA testing for Common Trisomies in Twin Pregnancies AFFECTED PREGNANCIES (N) DETECTION RATE (%) FALSE POSITIVE RATE (%) Trisomy 21 137 99.0 0.02 Trisomy 18 50 92.8 0.01 Trisomy 13 11 94.7 0.10 Source: H Judah et al., Cell-free DNA testing of maternal blood in screening for trisomies in twin pregnancy: updated cohort study at 10–14 weeks and meta-analysis, Ultrasound Obstet Gynecol 58:178–189, 2021. PRENATAL TESTING PROCEDURES  635 FETAL BLOOD SAMPLING Fetal blood is aspirated by direct puncture of a blood vessel, usually in the umbilical cord (cordocentesis). The risk of fetal loss from the procedure is ~1% (Kosian et al. 2023). Before FISH analysis of uncultured cells (see above) came to be more widely used, cordocentesis was useful when speed of diagnosis was of the essence, in the setting of the detection of a fetal anomaly at ~18-week ultrasonography. The procedure once had a role in assisting resolution of mosaicism in amniotic fluid culture (Shalev et al. 1994), but this was largely replaced by the use of FISH. PLACENTAL BIOPSY In principle, this is the same as first-trimester CVS. The placenta is sampled by a transabdominal approach, and this is a straightforward procedure. The main application had been when a rapid result was needed, although newer methodologies have largely bypassed that imperative. An insufficient amount of amniotic fluid remains an indication. FETAL CELL ISOLATION FROM MATERNAL BLOOD In 1969, Walknowska et al. identified cells with a male karyotype during the cytogenetic analysis of lymphocyte cultures from pregnant women, and they recognized the potential to use these cells for prenatal diagnosis. The two important cell types that are released from the fetal tissue into the maternal circulation are the nucleated red blood cell and the trophoblast2 (Dhallan et al. 2007; Maron and Bianchi 2007). Fetal cells circulating in maternal blood are an obvious target for noninvasive prenatal diagnosis, yet with a concentration of only one fetal cell in each 2–3 ml of maternal blood (Kølvraa et al. 2016), isolation is difficult. For this reason, cell-based NIPT approaches have fallen somewhat out of favor, particularly with the successful implementation of NIPT approaches that use cfDNA. Nonetheless, researchers continue to explore the potential of fetal cell–based techniques. A number of different techniques have been used to isolate fetal cells from maternal blood, after which various genetic methodologies can be applied, including FISH, PCR, whole-genome amplification, and short-read sequencing. CELOCENTESIS The extra-embryonic celom, which exists only during the first trimester, is a source for (non-dividing) cells originating from extra-embryonic mesoderm. Given its anatomical continuity with the cytotrophoblast, Makrydimas et al. (2006) comment that it could be thought of as “a liquid extension of the placenta.” Celocentesis involves the transvaginal insertion of a needle into the celomic cavity and aspiration of 1mL fluid containing cells of fetal origin. The procedure has the attraction of an earlier timing (7–9 weeks) than CVS and NIPT, but so far its use has been limited by uncertainty about the safety of the technique, the low amount of DNA recovered, and by high levels of maternal cell contamination (Giambona et al. 2016). In practice, celocentesis has usually been applied for the early prenatal diagnosis of monogenic conditions, in particular hemoglobinopathies. Giambona et al. (2018) undertook celocentesis and isolated fetal cells from a large series of women whose pregnancies were at risk of thalassemia or sickle cell disease; QF-PCR 2 Fetal lymphocytes can also be isolated from the maternal circulation, but they can persist for many years, making them unsuitable for prenatal diagnosis.

636  REPRODUCTIVE CYTOGENETICS for chromosomes 13, 18, 21, X, and Y was done primarily for the purpose of detecting maternal cell contamination, but also diagnosed pregnancies that were affected by trisomy 21, trisomy 13, and triploidy. CYSTIC HYGROMA, PLEURAL EFFUSION, AND ASCITES Cystic hygroma and fetal hydrops have a strong association with fetal aneuploidy, especially monosomy X. A concomitant oligohydramnios may make amniocentesis difficult. Fluid from cystic hygroma, ascites, and pleural effusion contains lymphocytes, and these cells can be analyzed using cytogenetic or molecular techniques within the time frame of a few days. In one small series of 14 pregnancies affected by hydrops, samples were collected from pleural fluid (n = 4), cystic hygroma (n = 5), and ascites (n = 6), and two cases of monosomy X were detected (Gole et al. 1997). CERVICAL LAVAGE OR CYTOBRUSH Trophoblast cells can migrate from the confines of the uterine cavity and enter the endocervical canal, and they can be collected for molecular analysis by endocervical irrigation and aspiration (lavage), or by insertion of a “cytobrush” (Bischoff and Simpson 2006). The attraction, in principle, is of diagnosis as early as five weeks, and a (relatively) noninvasive procedure. Cells of maternal origin outnumber fetal cells at a ratio of 2000:1 (Imudia et al. 2009), but fetal cells can be isolated by incubating with specifically binding antibodies attached to magnetic particles. Fetal cells are then separated with a magnet and subjected to molecular analysis (Fritz et al. 2015). PROTEOMIC FINGERPRINTING Proteomic fingerprinting of amniotic fluid assesses the expression profile of proteins coded from specific chromosomes, or otherwise expressed in the context of a specific aneuploidy, and this could be considered as a functional assay for trisomy (Vasani and Kumar 2019). “PRIMUM NON NOCERE” AND PRACTICAL CONSIDERATIONS “First, do no harm” is a cornerstone of medical practice. Yet, almost inevitably, having a prenatal diagnostic procedure causes anxiety. Rothman (1988), in her book The Tentative Pregnancy, is particularly critical of what she viewed as a medicalized distortion of the normal process of being pregnant. Hodge (1989) describes her personal experience of Waiting for the Amniocentesis, and we reproduce her letter in full: I drafted the following letter to the editor one week before I expected to hear the results of my amniocentesis: “I am 40 years old and 19 weeks pregnant with what will presumably be my third child. I am on the basic science faculty of a medical school. When I teach medical students about amniocentesis, I occasionally mention the difficulty for the woman of having to wait until well into the second trimester to receive her results. “I am in that situation myself now, awaiting my results. And before experiencing it, I was unprepared for two phenomena. One was just how difficult the wait is. Pregnancy is always a time of waiting, but now time has slowed down to an extent PRENATAL TESTING PROCEDURES  637 I did not anticipate. The other, more disturbing phenomenon is how the waiting has affected my attitude toward the pregnancy. At many levels I deny that I really am pregnant “until after we get the results.” I ignore the flutterings and kicks I feel; I talk of “if ” rather than “when” the baby comes; I am reluctant to admit to others that I am pregnant. I dream frequently and grimly about second-trimester abortions. In some sense I am holding back on “bonding” with this child-to-be. This represents an unanticipated negative side effect of diagnostic amniocentesis. And all this, even though my risk of carrying a chromosomal abnormality is less than 2 percent. “I presume I am not alone in these reactions, yet I have not seen this problem mentioned in the literature, nor did my physician or genetic counselor discuss it with me. I am writing now to bring it to the attention of clinicians with pregnant patients undergoing diagnostic amniocentesis. I suggest to both clinicians and their patients that, when weighing the relative risk and benefits of prenatal diagnosis performed later (amniocentesis) as compared with earlier (chorionic villus biopsy), they not underestimate the negative effects of a 4½ month wait before the woman knows if she is ‘really’ pregnant.” The next day, before I had mailed this letter, I received the results, and unfortunately they were the dreaded ones: trisomy 21. I have since then had the grim second-trimester abortion. From my current perspective of grief and shock, I encourage clinicians to help their patients avoid the denial described in my letter. My husband and I spared ourselves no pain by holding back emotionally. It has become a cultural expectation that one will keep one’s pregnancy a secret until one has had the “all clear” from the amnio. One reasons, “If we get a bad result, we won’t have to tell anyone.” But I now believe that reasoning is wrong. After our bad result, my husband and I did tell everyone. Sympathy and support from our friends, family, and colleagues have helped us to survive the ordeal of aborting a wanted pregnancy. By keeping the loss a secret, we would have cut ourselves off from such support when the feared outcome did happen. Not every couple will react this way, some preferring to keep their personal affairs private, but many will. The counselor needs to acknowledge these criticisms and to rise to the challenge of providing a sympathetic and skillful service to clients/patients according to their varying responses to deciding to have, undergoing, and waiting for the results of prenatal diagnosis, and then supporting those who do get an abnormal result. Quite a few pregnant women are in any event against invasive prenatal testing. In a study of pregnant women (age 37 years or older) who had not undergone prenatal diagnosis in Victoria, Australia, 33% had actively declined, with the two main reasons being concern about the safety of the test and a conviction that they would not in any event have a termination (Halliday et al. 2001). Hill et al. (2016) studied preferences for prenatal tests for Down syndrome in nine countries, and found that pregnant women placed greatest emphasis on test safety and risk of miscarriage when choosing a prenatal test. In contrast, for health professionals, test accuracy was the most important factor in determining choice of prenatal test. A practical question is pain: The thought of insertion of a needle or of a catheter sufficiently deeply to sample a pregnancy would naturally be cause for some apprehension. Csaba et al. (2006) surveyed a number of women having prenatal diagnosis in New York, asking them to quantify their anxiety ahead of the procedure (transabdominal CVS, transcervical CVS, or amniocentesis), and their perception of pain immediately afterward.

638  REPRODUCTIVE CYTOGENETICS In each procedure, the pain was typically seen as “mild,” and three-fourths of the women thought it was the same or less painful than they had been expecting. Those who were more anxious—mostly the younger and nulliparous—felt the pain more keenly, and thus special reassurance should be given to these women. SCREENING FOR FETAL TRISOMY In broad terms, “screening” describes testing a whole population, or a whole segment of population, for a condition that in fact only a (typically small) fraction will have. This criterion applies to pregnancy screening for Down syndrome (DS) or other trisomy. All, or at least many pregnant women in a population may be tested, but only a very few will turn out actually to have an affected pregnancy. A requirement of a screened condition is that the condition be well understood, and that an intervention be available. More precisely, screening in this context should meet three criteria: It should identify women who are at increased risk, prior to their having a definitive diagnostic test; it should be offered systematically to pregnant women who are considered to be at only baseline population risk; and it should be viewed as beneficial to those who receive it, in terms of either choosing termination or being prepared for the birth of a child with DS (Weisz and Rodeck 2006). Until the implementation of NIPT, the screening tools used were the taking of a maternal blood sample and the performing of an ultrasonogram, and these methods remain in use. The methodology involves the analysis of data (maternal age, serum measurements, ultrasound findings) according to a sophisticated computed algorithm, in order to calculate a risk that the fetus is affected by DS. If the calculated risk is greater than that of a certain threshold risk figure (usually taken as 1 in 250), the pregnancy is regarded as being at “increased risk,” and definitive testing is then offered. Since other aneuploidies can also influence the measured indices, the test procedure in practice becomes broader than just a trisomy 21 screen. First-Trimester and Second-Trimester Biochemical Screening Certain biochemical markers in the mother’s serum may have altered concentrations, whether increased or decreased, if she is carrying a trisomic pregnancy; presumably, these differences reflect perturbation in the trisomic fetoplacental unit. An assessment is made of the degree to which each level differs from expectation, and these data are factored into an algorithm that takes into account the prior risk due to maternal age (Spencer 2007). Sophisticated computer packages are employed to calculate an overall risk figure. The two first-trimester analytes most commonly measured are the β component of human chorionic gonadotropin (β-hCG) and pregnancy-associated plasma protein-A (PAPP-A), the former typically high and the latter low in a DS pregnancy. PAPP-A levels are influenced by mode of conception, being lower in pregnancies that are conceived using in vitro fertilization and leading to a higher rate of false-positive results in first-trimester screening for DS in these pregnancies (Amor et al. 2009). In the second trimester, the analytes measured in many jurisdictions comprise α-fetoprotein (AFP), PRENATAL TESTING PROCEDURES  639 estriol, β-hCG, and inhibin-A (four analytes; hence, the “quadruple test”3). In trisomy 21, the AFP is low, hCG high, uE3 low, and inhibin-A high. With the increased uptake of noninvasive prenatal testing for DS, the question has arisen whether biochemical screening has a role in pregnancies that are also being screened by NIPT. Two arguments have been made in favor of retaining biochemical screening. The first is that although biochemical screening is aimed primarily at the detection of DS, in fact a range of other chromosome abnormalities may be seen, many of which would not be picked up on NIPT. The risk of a rare chromosome abnormality has been estimated to be 4% when PAPP-A levels are very low (<0.2 multiple of median [MoM]), and 7% when free β-hCG levels are very low (<0.2 MoM) (Petersen et al. 2014). Although it has also been argued, on the other hand, that the specific screening algorithms had not been set up for the purpose of finding rare chromosome abnormalities, and that using this screening to cast a wider net might have only a marginal benefit in terms of detection of other abnormalities but would yet imply a significant increase in the false-positive rate (Yaron et al. 2016). The second argument in favor of retaining biochemical screening is that abnormal serum biochemistry, and particularly low levels of PAPP-A and free β-hCG, can give warning of third-trimester pregnancy complications such as pre-eclampsia, intrauterine growth restriction, and preterm birth (Sovio et al. 2024; Peris et al. 2024). First-Trimester Ultrasonographic Screening Ultrasonographic scanning is applied during the window of 11–14 weeks inclusive. This particular parameter is assessed: the degree to which the skin at the neck is separated from the underlying tissue by fluid (Figure 21–1). Since this fluid does not reflect the sound wave on the scan, it is referred to as “nuchal translucency”; “nuchal thickening” is another expression. Bekker et al. (2006) propose that the underlying cause is anomalous development of the lymphatic system in the region of the neck, and it appears that this development is susceptible to a chromosomal imbalance. An increased nuchal translucency is associated with DS, and combined with maternal age the detection rate is 69%–75%, for a false-positive rate of 5%–8% (Rink and Norton 2016). An increase in nuchal translucency is not specific to DS, being also observed in trisomy 13, trisomy 18, monsomy X, and triploidy. Bardi et al. (2020) have assessed various risks in the setting of the specific finding of an increased nuchal translucency, and these are summarized in Table 21–6. Other ultrasonographic markers of DS are absence of the fetal nasal bone, tricuspid regurgitation, and abnormal blood flow in the ductus venosus (Rao and Platt 2016). A practical question is this: If, following the observation of an increased nuchal translucency a CVS or amniocentesis is done, and the chromosomes are normal, is there a residual risk for some other type of fetal abnormality? The main two risk categories are for single-gene disorders and fetal structural abnormalities. If the translucency resolves, and if no defects (with particular focus on the fetal heart) are seen at 14–16 weeks gestation, the prognosis is good, with a better than 95% chance of a baby with no major abnormalities. In fetuses with nuchal translucency ≥3.5 mm at 11–14 weeks gestation, normal microarray, and no other abnormalities on subsequent scans, the chance of a diagnostic variant on exome sequencing is only 2%, but is higher for nuchal translucencies ≥5 mm. 3 The “double test” uses AFP and estriol; the “triple test” is AFP, estriol, and β-hCG.

640  REPRODUCTIVE CYTOGENETICS In the presence of additional ultrasound findings, the likelihood of a diagnostic variant increases to >20% (Mellis et al. 2022). Noonan syndrome or other “RASopathy” require consideration, accounting for about half of single-gene disorders in fetuses with a large nuchal translucency and normal karyotype (Bardi et al. 2020). On the question of neurodevelopment, Hellmuth et al. (2017) followed up more than 220,000 euploid children who had been screened in the first trimester, and they found a sixfold increased risk of intellectual disability in children with a nuchal translucency above the 99th centile, although the absolute risk remained low (<1%). Buffin et al. (2021) followed up a cohort of 203 chromosomally normal infants with nuchal translucency >95th centile and 208 controls to age two years: They found that the developmental quotient (DQ) was slightly lower (108.6) in the increased nuchal translucency Figure 21–1.  Nuchal Translucency. Notes: Ultrasound image shows lateral view of the fetus. The two + markers delineate the degree of separation of the nuchal skin, in this case, 5.40 mm. Courtesy C Siles. Table 21–6.  Implications of an Increased Nuchal Translucency NUCHAL TRANSLUCENCY TRISOMY 13/18/21 (%) OTHER CHROMOSOMAL (%) SINGLE GENE (%)* STRUCTURAL ANOMALY (%)** ALL ABNORMAL FETUSES (%) 95–99th centile 13 2.2 0.6 5.9 21 >99th centile 34 12 3.3 12 62 3.5–4.9 mm 25 6 1.2 11 43 5.0–6.4 mm 44 17 5.5 11 77 6.5-7.9 mm 51 12 2.6 17 83 ≥8.0 mm 34 23 7.4 14 79 Notes: These data are based on 1901 pregnancies with NT ≥ 95th centile in the Netherlands, 2010–2016. * 50% of single-gene disorders were Noonan syndrome, and the other 50% were various rare syndromes. ** The most common structural anomalies were cardiac (42%), multiple congenital anomalies (16%), urogenital (11%) and skeletal (7%). Source: From F Bardi et al., Is there still a role for nuchal translucency measurement in the changing paradigm of first trimester screening? Prenat Diagn 40:197–205, 2020. PRENATAL TESTING PROCEDURES  641 group compared to controls (112.8); however, both results were well within the expected range for age, and only one child in the increased nuchal translucency group had a DQ in the significantly impaired range of <70. An increased nuchal translucency may also be detected after a low-risk NIPT result. Kelley et al. (2021) rehearse the key counseling points that apply here, noting that the chance of an adverse outcome increases with the nuchal translucency measurement size and the presence of other abnormalities on ultrasound. A summary of their approach is presented in Figure 21–2. First-Trimester Combined Ultrasonography and Biochemical Screening A better detection is achieved through a combination of first-trimester nuchal translucency assessment and the measurement of maternal serum-free β-hCG and PAPP-A. If the blood test is done first, these results can be held pending the ultrasound, and the combined figure can be available soon after the scan is done. Detection rates are Figure 21–2.  Interpreting an Increased Nuchal Translucency when NIPT has been Normal. Notes: A flow diagram of diagnostic pathway for increased nuchal translucency and low-risk non-invasive prenatal testing Source: Adapted from J Kelley et al., Increased nuchal translucency after low-risk noninvasive prenatal testing: What should we tell prospective parents? Prenat Diagn 41:1305–1315, 2021. 642  REPRODUCTIVE CYTOGENETICS typically >90%, for a false-positive rate of 5% or less (Santorum et al. 2017). The validity of this approach in more precisely targeting an increased-risk population is attested in the experience from Denmark, where a national program was put in place in 2004. The number of diagnostic procedures (amniocentesis or CVS) declined from 7,524 in 2000 to 3,510 in 2006; and yet, during the same period, the number of newborns with DS fell from approximately 50 to 30 per year (Ekelund et al. 2008). While the prime focus of screening is on trisomy 21, a side benefit is the detection of other, and typically more severe, chromosomal disorders. Trisomy 13 and trisomy 18 both show reduced levels of β-hCG and PAPP-A at first-trimester screening, more so in trisomy 18, along with increased nuchal translucency or frank cystic hygroma. Few other trisomic pregnancies proceed through to the time of screening. Unsurprisingly, those that do so display abnormalities at screening. For example, in trisomy 22 at first-trimester screening, the β-hCG is very elevated, PAPP-A somewhat reduced, and fetal growth restriction is typical (Sifakis et al. 2008). In non-mosaic trisomy 9, the biochemistry is similar to that of trisomy 18 (Priola et al. 2007), and the same may apply in the mosaic case. In a triploid pregnancy, the biochemical indices at first-trimester screening are also quite abnormal, and very differently so according to the category of triploidy, digynic or diandric (p. 343). In the digynic form, β-hCG and PAPP-A are both much reduced, whereas in the diandric type, β-hCG is greatly elevated and PAPP-A marginally reduced. Likewise, ultrasonography is distinctly different, with severe growth restriction in the digynic type, and nearer normal growth but with an enlarged and partially molar placenta in diandric triploidy (Kagan et al. 2008). Second-Trimester Ultrasonography A fetal anatomic survey in the second trimester may lead to the diagnosis of DS through the detection of a major structural abnormality, such as a heart defect or duodenal atresia. In addition, a number of “soft signs” on second-trimester ultrasonographic fetal assessment point to an increased likelihood for DS. An advantage is that this procedure is often done routinely as part of normal obstetric management, and thus a de facto DS screen is added on, essentially at no additional cost. However, the observations do not lend themselves to a ready analysis in terms of adjusting the level of risk; furthermore, the frequency of these “soft signs” in normal fetuses leads to a high false-positive rate. A more rigorous approach is to adjust the patient’s risk assessment using positive and negative likelihood ratios4 which are available for each “soft sign” (Table 21–7). Twin Pregnancies.  In the case of biochemical screening, two fetoplacental units lead to the production of twice as much of the particular biochemical substance, which is then conveyed into the maternal bloodstream. Muller et al. (2003) examined the second-trimester analyte levels, and Spencer et al. (2008) the first-trimester analyte levels, in cohorts of twin pregnancies; and the MoM values were essentially double those of singleton pregnancies. The valid MoMs for risk evaluation can thus be derived by dividing the observed result by approximately 2. Intriguingly, monochorionic (and presumably monozygous) twins at first trimester have a somewhat lower PAPP-A mean (1.6 MoM) 4 The Likelihood Ratio (LR) is the likelihood that a given test result would be expected in a patient with the target disorder (e.g., Down syndrome) compared to the likelihood that that same result would be expected in a patient without the target disorder.

PRENATAL TESTING PROCEDURES  643 than do the dichorionic (2.1 MoM) (Madsen et al. 2011); if chorionicity is distinguishable at ultrasonography, adjustment can be made by applying the appropriate PAPP-A divisor. A theoretical complicating factor, in the case of one (dizygous) twin being trisomic 21, is that the normal co-twin might “dilute out” the abnormal serum biochemistry and thus invalidate the test result. However, in a large French study addressing a second-trimester population, such an effect, if present, was marginal (and not significant statistically), and screening in this setting was considered to be effective (Garchet-Beaudron et al. 2008). Concerning the ultrasonography, nuchal translucency screening allows each twin to be assessed individually, and the detection rate for aneuploidy is similar to singleton pregnancies (Cleary-Goldman et al. 2005). For monochorionic twins, a single risk estimate can be calculated for the pregnancy using the average of the two nuchal translucency measurements, whereas for dichorionic twins, a specific risk is calculated for each twin. When first-trimester serum markers and nuchal translucency results are combined, a detection rate of 90% can be achieved, for a false-positive rate of 5.9% (Madsen et al. 2011). Garchet-Beaudron et al. (2008) point out other issues relating to twin pregnancies. Logically, the age-related risk for DS might be expected to double in a dizygous twin pregnancy. But such logic appears not to apply; actual observation does not record such an increase. The technical procedures in the event of an increased-risk result are more demanding: double amniocentesis, with each sac sampled separately; and, if one twin is trisomic and selective termination is sought, the normal twin is placed at risk. Screening in twin pregnancies requires special expertise. In the case of a “vanishing twin” at the first trimester, as manifest by a second, empty sac, the PAPP-A level is artificially elevated and should be omitted from the risk calculation (Chaveeva et al. 2020). As noted above, cfDNA testing is now the preferred screening methodology for twin pregnancies due to its superior performance (Judah et al. 2021), but first-trimester combined screening results can nonetheless be used to stratify twin pregnancies into low- and high-risk categories, with those at risk being offered follow-up cfDNA testing (Galeva et al. 2019). Table 21–7.  Positive and Negative Likelihood Ratios of Sonographic Markers for Trisomy 21 MARKER POSITIVE LR NEGATIVE LR LR ISOLATED MARKER* Intracardiac echogenic focus 5.8 0.80 0.95 Cerebral ventriculomegaly 27.5 0.94 3.81 Increased nuchal fold thickness 23.3 0.80 3.79 Echogenic bowel 11.4 0.90 1.65 Mild hydronephrosis 7.6 0.92 1.08 Short femur 4.8 0.74 0.78 Short humerus 3.7 0.80 0.61 Aberrant right subclavian artery 21.5 0.71 3.94 Absent/hypoplastic nasal bone 23.3 0.46 6.58 *Calculated by multiplying positive LR for given marker by negative LR for all other markers. LR = likelihood ratio. Source: Adapted from M Agathokleous et al., Meta-analysis of second-trimester markers for trisomy 21, Ultrasound Obstet Gynecol 41:247–261, 2013. 644  REPRODUCTIVE CYTOGENETICS Interpretation of Screening Results What do these various figures—detection rate, false-positive rate, positive predictive value—mean? A little epidemiology is in order. Imagine a group of 10,000 pregnant women, of all ages. Assuming a birth prevalence for DS of 1.2 per 1,000, we can take it that 12 women would otherwise give birth to a baby with DS. If the particular screening approach has a detection rate of, for example, 85%, 10/12 of these DS pregnancies would be recognized as being at increased risk, and they could be identified at prenatal diagnosis. The remaining 15% who are carrying a DS fetus (2/12) would fail to be recognized. If the false-positive rate is, for example, 4%, 400 women would have an increased-risk report from screening, but they would go on to receive a normal result from the following amniocentesis or CVS. Putting these figures in the conventional format, we have (Table 21–8): The detection rate (sensitivity) of the test is 10/12 (85%). Thus, 15% of women with a trisomic 21 fetus will be missed by the test. The positive predictive value of the test is only 10/410 (2.4%). Thus, 97.6% of women returning an “increased-risk” result will not have a baby with DS.5 The negative predictive value is 9,588/9,590 (99.98%); in other words, a “low-risk” result means a 99.9% chance for an unaffected baby. The false-positive rate is an important parameter. As noted earlier, this represents the fraction of women who will then go on to have an invasive definitive test that will return a normal chromosomal result. Clearly, the smaller this figure the better. The trade-off is this: the smaller the false-positive rate, the less the detection rate. To judge the effectiveness and acceptability of the screening, we can declare a false-positive rate that is desirable, and this would then determine what the detection rate will be; or, we can choose a preferred detection rate and accept the false-positive rate that this would incur. A typically desired false-positive rate is 5%; based upon this, the detection rate is noted as DR5. Desired detection rates of 85% or 90% would come at a cost of false-positive rates noted as FPR85 and FPR90 (Weisz and Rodeck 2006). THE UNDERSTANDING OF WOMEN WHO HAVE SCREENING The interpretation of a DS screening test result to the patient is fraught with potential for confusion. The major pitfall is that an “increased-risk” test result may sometimes be understood by the woman and her medical advisor to mean that the pregnancy is likely to be affected. As we showed earlier, many women who screen “positive” will go on to have a normal baby. Counselors doing this work need a clear awareness of these issues so that they can enable their patients to understand, intuitively or explicitly, the concept Table 21–8.  Sensitivity, Positive Predictive Value, and Negative Predictive Value of DS Screening TEST INTERPRETATION FETUS WITH DS FETUS NOT WITH DS TOTAL Test shows “increased risk” 10 400 410 Test shows “low risk” 2 9,588 9,590 Total 12 9,988 10,000 5 Thus emphasizing the point that the expression to use is merely “increased risk,” not “high risk.”

PRENATAL TESTING PROCEDURES  645 and relevance of a low positive predictive value. The counselor is referred to Macintosh’s (1994) essay “Perception of Risk” for a very readable and practical commentary upon these issues, and to Marteau and Dormandy (2001) for an overview of the complexity of the issues. The ideal is that those having a screening test for DS and other aneuploidies should have a basic awareness of the conditions and of the rationale of the screening procedure, and that their beliefs and perceptions and attitudes should be reasonably consonant with the aims and practice of the program. The ideal has not necessarily been met. Jaques et al. (2004), in a paper provocatively titled “Do Women Know That Prenatal Testing Detects Fetuses with Down Syndrome?” surveyed responses from pregnant women 37 years old or older in Victoria, Australia, in 1998–1999; and the answer to their question was, disconcertingly, that “Down syndrome” was not mentioned as a reason for undergoing pregnancy testing in almost 40% of respondents. Not every woman will respond “rationally” to an increased-risk interpretation, according to the view of rationality as seen by the providers of the screening program. Those who enter into a screening program without being properly aware of the implications may find themselves “in an untenable situation—anxious about a positive result, but unwilling to incur the risks of diagnostic testing” (Kuppermann et al. 2006). Depressive symptoms, and thus a reduced capacity to make clear decisions, may be exacerbated in those with a predisposition, and Hippman et al. (2009) see a role for the counselor in recognizing this. For those who are better informed, understanding is by no means a neutral matter, and Rapp (1999) refers to the role of women as “moral pioneers” in coming to terms with the ethical issues that readily available screening may, in these modern times, present. Susanne et al. (2006) assessed responses in women who had had what turned out to be a false-positive screening result, following them prospectively through the pregnancy and after the baby was born. Several declared that they had “withheld” their pregnancy and only returned to reacceptance after the normal chromosomal result from amniocentesis had been conveyed; nevertheless, most would have the same testing in a future pregnancy. Counselors need to be well attuned to these several complexities; and if a woman’s family physician can share in the decision-making process, this is typically well received (Légaré et al. 2011). NIPT differs from other prenatal tests by combining ease with high accuracy. Kristalijn et al. (2022) surveyed more than 4,500 women in the Netherlands and found that the vast majority of women who receive a low-risk result were positive about the experience and felt they had received sufficient information. However, those with increased-risk results were more likely to report negative feelings and anxiety, and to feel that they had not been sufficiently prepared for the result. This lack of preparedness is especially prominent for findings outside trisomies 21, 18, and 13. Lewit-Mendes (2023) report that people who received an increased chance of sex chromosome aneuploidy result at NIPT felt unprepared for the unanticipated complexity of the result, and were faced with making a time-sensitive decision for a condition they had not previously considered or even been aware of. Women who received increased-risk results for rare autosomal trisomies and structural chromosome abnormalities were similarly unprepared for these results (Bakkeren et al. 2024). Kater-Kuipers et al. (2020) argue that the ease and routineness of NIPT challenges the notion of informed decision-making in prenatal testing, and propose a three-step counseling model in which three decision moments are distinguished and differently addressed: (1) professionals explore women’s values concerning whether and why

646  REPRODUCTIVE CYTOGENETICS they wish to know whether their baby has a genetic disorder; (2) women receive layered medical-technical information and are asked to make a decision about screening; (3) during post-test counseling, women are supported in decision-making about the continuation or termination of their pregnancy. Concerning the facts about DS itself in the context of pregnancy screening, the National Society of Genetic Counsellors (USA) has published a practice resource on communicating a prenatal or post-natal diagnosis of DS (Sheets et al. 2011). The authors recommend elements of a diagnosis conversation including genetic causes of DS and recurrence risks, recommendations for sensitively delivering a diagnosis of DS, essential information for the initial discussion of a diagnosis of DS, discussions of options of pregnancy management after a prenatal diagnosis, and DS resources; and they add that information about life outcomes is particularly valued by prospective parents. Provision of a website or pamphlet is the simplest means of conveying information, and many clinics/jurisdictions produce their own material. It is a fine matter to judge what should be the level and tone of the information. Reviews of leaflets produced in the United Kingdom (Bryant et al. 2001) and Canada (Lawson et al. 2012) concluded that the viewpoints expressed were, in the main, weighted unduly negatively toward DS, and this is consistent with a broader criticism that genetic counseling primarily provides information consistent with the medical model of disabilities that focuses on problems to be fixed, as opposed to the social model, which prioritizes supports, services, and acceptance (Ijaz et al. 2025). It is true that information ought to be couched in such terms that it will be useful, in the fullest sense of that word, to the wide range of people for whom it is intended (and see p. 814). Equally, the comment can be made that attempting to neutralize negative aspects of DS may send a mixed message, since being given the option of abortion in order to avoid having a DS child rather plainly implies that having such a child may not be a desirable outcome. The view that is offered should be clear, accurate, and even-handed.6 Secular Trends in Prenatal Screening and Diagnosis of Aneuploidy The population prevalence of Down syndrome has been influenced by three factors over the past half century: changes in maternal-age distribution (Chapter 13), improved survival, and advances in prenatal screening technologies and policies (Figure 21–3). Birth prevalence has fluctuated according to the changing maternal-age profile and to the take-up of prenatal diagnosis (Figures 21–4 and 21–5). In Europe, Australia, and New Zealand, the livebirth prevalence for DS has fallen, while in the USA there has been a modest increase (Figure 21–5). Another possible point to factor in to the overall picture may be a changing attitude toward pregnancy termination for an aneuploidy (Jacobs et al. 2016). The increasing precision of screening has led to a reduction in the number of invasive procedures being done. Pynaker et al. (2024) analyzed data for the years 1976–2023 from Victoria, Australia, and noted that while the total number of invasive tests climbed 6 These matters are dealt with in considerable detail in a document from the National Health Service of the United Kingdom, “Psychological Aspects of Genetic Screening of Pregnant Women and Newborns: A Systematic Review” (Green et al. 2004).

PRENATAL TESTING PROCEDURES  647 steadily from 1976 to 2000, it then declined such that in 2023, fewer procedures were being done than at any time in the previous 35 years (Figure 21–6). At the same time, the number of prenatal diagnoses of DS was the highest recorded. This improved targeting reflected changes in the indications for invasive prenatal testing over the same period (Figure 21–7), and the number of invasive procedures performed per diagnosis of a major chromosome abnormality declined from 100 to two. Renshaw et al. (2013) have reported similar improvements in the United Kingdom, with the number of invasive procedures per syndrome diagnosis reducing from 46 to five between 1991 and 2010. FETAL ULTRASONOGRAPHIC ANOMALIES We have discussed above ultrasonography in the specific setting of targeted screening. Otherwise, a mid-trimester ultrasound examination is, of course, a routine part of Figure 21–3.  The Prevalence of Down Syndrome in the USA since the 1950s. Notes: With increasing survival over the timeframe 1950s-2010s, the numbers of older DS individuals have been progressively increasing (yellow and orange bands, above), while the prevalence in children from the 1960s has been fairly constant. During this period, the US population doubled, from 150 million, to 310 million; and thus the fraction of DS among the whole population of children has been falling. Source: From SE Antonarakis et al., Down syndrome, Nat Rev Dis Primers 6:9, 2020. Courtesy SE Antonarakis, and with the permission of Springer Nature. Figure 21–4.  Down Syndrome Pregnancy Outcomes in the USA 1974–2013. Notes: Livebirths and natural losses have increased fractionally over this timeframe, consonant with population growth. Prenatal diagnosis with termination expanded very substantially from the 1970s through to the turn of this century, but has remained fairly constant thereafter. Source: From SE Antonarakis et al., Down syndrome, Nat Rev Dis Primers 6:9, 2020. Courtesy SE Antonarakis, and with the permission of Springer Nature. 648  REPRODUCTIVE CYTOGENETICS standard obstetric management. The discovery of a fetal malformation in the course of a routine ultrasound is a common indication for a fetal chromosome study. In Victoria, Australia, for example, ~30% of prenatal chromosome tests between 2013 and 2016 were done on the grounds of ultrasound findings of a fetal malformation or of a marker of aneuploidy (Lostchuck et al. 2019). Certain major ultrasonographic defects are fairly specific: For example, holoprosencephaly predicts the likelihood of trisomy 13, fetal hydrops/cystic hygroma predicts monosomy X or trisomy 21, and an endocardial cushion defect or duodenal atresia predicts trisomy 21. Conotruncal defects are associated with the 22q11.2 deletion, and aortic narrowing suggests the 7q11.23 deletion of Williams syndrome (Krzeminska et al. 2009). Asymmetrical growth (head circumference vs. crown–rump length) may point to triploidy (Salomon et al. 2005). The acardiac fetus is often due to an otherwise Figure 21–5.  Livebirth Prevalence of Down syndrome in Europe, USA, Australia and New Zealand, over the 20th and Early 21st Centuries. Notes: Actual live birth prevalence is shown in blue, green shows modelled additional live birth prevalence as it would have been in the absence of selective pregnancy termination. A considerable fall-off in prevalence is to be seen in all four jurisdictions over the mid 20th century, a reflection of changing maternal age profiles during that era. Termination practice since the availability of prenatal diagnosis is similar in Australia, New Zealand, and Europe, but somewhat less applied in the USA. Source: From G de Graaf et al., Estimation of the number of people with Down syndrome in Australia and New Zealand, Genet Med 24:2568–2577, 2022. Courtesy BG Skotko, and with the permission of Elsevier and the American College of Medical Genetics and Genomics. PRENATAL TESTING PROCEDURES  649 Figure 21–6.  Uptake of Prenatal Diagnostic Procedures. Notes: These data are from Victoria, Australia, and cover the whole period during which prenatal chromosomal diagnosis has been available. The decline since 2000 reflects the increasing accuracy of screening tests, which have replaced advanced maternal age as the main indication for diagnostic procedures (Figure 21-7). Source: C Pynaker et al., Annual report on Prenatal Diagnosis in Victoria 2023, The Victorian Prenatal Diagnosis Database, Murdoch Children’s Research Institute 2024. doi.org/10.25374/MCRI.27895995 Figure 21–7.  Indications for Invasive Prenatal Diagnosis in Victoria, Australia, since 2000, as Percentages of All Tests. Notes: Notable observations are the falling away of advanced maternal age and first trimester screening as an indication, as the more precise approaches of ultrasonography and NIPT took hold. Source: From C Pynaker et al., Annual report on Prenatal Diagnosis in Victoria 2023, The Victorian Prenatal Diagnosis Database, Murdoch Children’s Research Institute 2024. doi.org/10.25374/MCRI.27895995

650  REPRODUCTIVE CYTOGENETICS unsurvivable autosomal trisomy, possibly tempered by mosaicism with a normal cell line but their existence being maintained by a (karyotypically normal) monozygous co-twin (Figure 20–24). Certain renal defects have a frequent association with fetal aneuploidy, as do cardiac malformations generally (Amor et al. 2003; Wimalasundera and Gardiner 2004; Carbone et al. 2011). Up to one-third of heart defects are associated with fetal aneuploidy, although in most there will be additional anomalies. Cysts of the choroid plexus (tissue within the cerebral ventricles) are a “soft marker” for trisomy 18, but not trisomy 21 (Walkinshaw 2000); they are otherwise harmless (DiPietro et al. 2011). On the specific question of rare autosomal abnormalities detected by conventional karyotyping (rare trisomies, deletions, duplications, supernumerary markers, various other structural rearrangements), a large European series based upon reports from malformation registers in several jurisdictions linked ultrasound findings to cytogenetic results (Baena et al. 2003). Nearly half of all rare autosomal abnormalities showed fetal anomalies on ultrasonography, with heart and brain defects and growth retardation more often seen with deletions, and cystic hygroma, hydrops, and nuchal translucency more typically associated with trisomies and duplications. These rare abnormalities comprised 7% of all chromosomally abnormal prenatal diagnoses. In which cases should a chromosome analysis be conducted, following the discovery of structural anomalies by ultrasound examination? Staebler et al. (2005) examined the karyotypes (on classical cytogenetics) in 428 fetuses with ultrasound-detected anomalies over a 10-year period. The karyotype was abnormal in 9% of cases with an isolated malformation, and in 19% of cases with multiple malformations. The following isolated defects were typically associated with a normal karyotype: hydronephrosis with high obstruction, unilateral multicystic dysplastic kidney, gastroschisis, intestinal dilatation, cystic adenomatoid malformation, pulmonary sequestration, vertebral anomaly, and tumor. Thus, one of these as a single malformation is not necessarily an indication, whereas, clearly enough, the presence of multiple malformations would warrant chromosome study. Daniel et al. (2003a) reviewed 1,800 cases in which an anomaly (an actual malformation or a minor marker of aneuploidy) had been detected at ultrasonography and assembled a table of risks of aneuploidy according to the findings (Table 21–9). The abnormal karyotypes included trisomies 13, 18, 21, triploidy, 45,X and mosaics, various autosomal and gonosomal duplications and deletions, rare trisomies, and de novo apparently balanced rearrangements. Lostchuck et al. (2019) examined trends of ultrasound-indicated invasive tests over a 30-year period and found that the diagnostic yield had remained stable at 18%, but with yield being higher at early gestation and lower after 16 weeks. Over time, there was a decline in the relative contribution of fetal ultrasound abnormalities to the detection of trisomy 21, reflecting the impact of NIPT; this was balanced by an increase in the detection of copy number variants following the introduction of prenatal microarray. The more precise tool of chromosome microarray provides a considerable additional yield above that which is detected by conventional (“classical”) karyotyping. Wapner et al. (2012), in a large prospective, blinded cohort study, showed that 6% of pregnancies with abnormal ultrasound findings and a normal (classical) karyotype had a clinically relevant CNV.7 The superiority of microarray was confirmed by subsequent meta-analysis, which demonstrated that in the setting of fetal anomalies, an additional 7% of 7 As we discuss in Chapter 18, the expressions microdeletion/microduplication may also be appropriate when a CNV is known to be pathogenic. PRENATAL TESTING PROCEDURES  651 abnormalities were revealed compared with conventional karyotyping (Hillman et al. 2013). Based on these findings, the American College of Obstetricians and Gynecologists (2013) recommended that microarray analysis be performed in patients with a fetus with one or more major structural abnormalities identified on ultrasound examination, and who are undergoing invasive prenatal diagnosis. In karyotypically normal fetuses, certain ultrasound abnormalities are associated with a higher frequency of CNVs. Mastromoro et al. (2022) reviewed the diagnostic yield of chromosome microarray across a broad range of single and multiple ultrasound findings, using as a comparison group women who were having prenatal diagnosis for advanced maternal age, and in whom the frequency of CNVs with known or suspected pathogenicity was 0.84%. The diagnostic yield was 4.5% for isolated structural anomalies and 12% for multiple structural anomalies, but rates for specific subcategories varied significantly, with cardiovascular and craniofacial anomalies implying the highest risk (Table 21–10). The Particular Association with Cardiovascular Defects The heart and the great vessels are the most frequently involved among all major malformations seen in fetal life, and almost 1% of all newborns will have a heart defect. If prenatal screening tests, serum, and ultrasound point to an increased chromosomal risk, the odds of seeing a cardiovascular defect are considerable, on the order of 20%. A Polish study from 2018 to 2021 looked at all pregnancies (n = 12,776) in the southeastern part Table 21–9.  The Likelihood of Discovering a Classical Chromosome Abnormality at Prenatal Diagnosis, According to the Pattern of Defects Identified at Fetal Ultrasonography, for All Maternal Ages DEFECTS LIKELIHOOD OF AN ANEUPLOIDY (%) CNS/cranial shape plus cardiac* 53 Key malformation,** singly or in combination 37 CNS ± other*** 21 Increased nuchal translucency, first trimester, ± other abnormality 25 Increased nuchal translucency, second trimester, ± other abnormality 13 Cardiac ± other abnormality 9 Pyelectasis/two vessel cord/echogenic bowel/ short femur 6 Other (singly or in combination) 3 Notes: Some percentages are considerably higher or lower for older and younger maternal ages, respectively. These data were obtained prospectively. *Excluding anencephaly/spina bifida. **Cystic hygroma/hydrops/exomphalos/severe IUGR/duodenal atresia/talipes. ***Excluding anencephaly/spina bifida/cardiac, including choroid plexus cysts. CNS = central nervous system; IUGR = intrauterine growth retardation. Source: From A Daniel et al., Prospective ranking of the sonographic markers for aneuploidy: Data of 2143 prenatal cytogenetic diagnoses referred for abnormalities on ultrasound. Aust N Z J Obstet Gynaecol 43:16– 26, 2003.

652  REPRODUCTIVE CYTOGENETICS of that country, in which screening tests led to the choice, in 1,005 of these women, of an invasive prenatal procedure (Wójtowicz et al. 2022). Of these thousand cases, in 202 a cardiovascular malformation was seen, either singly or along with other, extra-cardiac defects. A classic karyotypic abnormality was more often associated with a septal defect, whereas presumed pathogenic CNVs indicated the likelihood of a conotruncal abnormality. Unsurprisingly, chromosome abnormalities were more often seen where both cardiac and extra-cardiac defects were identified. Particular types of heart defect point to the likelihood of a chromosomal imbalance (Table 21–11) (see also p. 785). Table 21–10.  Chromosomal Diagnoses according to Ultrasound Findings INDICATION CHROMOSOME MICROARRAY YIELD (%) No indication (baseline risk) 0.79 Advanced maternal age 0.84 All isolated structural anomalies 4.5 Multiple structural anomalies 12 Individual isolated anomalies CNS 3.7 Musculoskeletal 4.7 Kidney/genitourinary 3.4 Gastrointestinal 4.7 Cardiovascular 5.2 Thorax/Respiratory 3.2 Craniofacial 5.2 Abdomen/body wall 3.3 Fetal growth restriction 3.3 Hydrops 4.9 Polyhydramnios 2.8 Cystic hygroma 3.8 All isolated soft markers 2.8 Multiple soft markers 3.6 Individual isolated anomalies Echogenic bowel 1.5 Absent/hypoplastic nasal bone 22 Intracardiac echogenic focus 1.0 Choroid plexus cyst 1.4 Mild ventriculomegaly 6.0 Short femur 6.6 Increased nuchal translucency 2.6 Notes: Listed here is the additional diagnostic yield of chromosomal abnormalities in fetuses undergoing microarray analysis, following particular ultrasound findings. Source: G Mastromoro et al., Molecular approaches in fetal malformations, dynamic anomalies and soft markers: Diagnostic rates and challenges—systematic review of the literature and meta-analysis, Diagnostics (Basel) 12:575, 2022. PRENATAL TESTING PROCEDURES  653 Twins. In the event of a twin pregnancy having been shown on ultrasonography, the question arises of a suitable prenatal diagnostic procedure, if this is considered appropriate. A point to make here is that although monozygous (MZ) twins would be expected to have the same karyotype, and almost always do, this does not invariably apply. The ability to interpret monozygosity is not perfect; and those MZ twins in which the split occurred soon after conception, prior to the differentiation of the extra-fetal tissues, may have the same ultrasound morphology of membranes (amnion and chorion) as would a dizygous pair. Thus, the advice is that dual amniocenteses rather than CVS may be the procedure of choice; and more especially so in the setting of discordance for an anatomical anomaly (Lewi et al. 2006). Fetal loss after CVS and amniocentesis in twin pregnancies happens similarly, at around 2%, and is not much above the background fetal loss rate in twin pregnancies (Di Mascio et al. 2020). Table 21–11.  The Association with Cardiovascular Defects CARDIOVASCULAR DEFECT FRACTION WITH A CHROMOSOME ABNORMALITY (%) Septal defect 70 Conotruncal defect 50 Left-sided obstructive 52 Aortic arch 30 Right-sided obstructive 63 Single umbilical artery 52 Notes: The figures here give likelihoods of identifying a chromosomal abnormality in the event that prenatal screening has indicated an increased risk for a chromosome abnormality, and that ultrasound has identified a cardiovascular anomaly. Septal defects: ventricular, atrial Conotruncal: double outlet right ventricle, tetralogy of Fallot, transposition of the great vessels Left-sided: hypoplastic left heart Aortic arch: aortic coarctation/stenosis/interruption Right-sided: hypoplastic right heart, Ebstein anomaly Source: Wójtowicz et al., Cardiovascular anomalies among 1005 fetuses referred to Invasive prenatal testing— A comprehensive cohort study of associated chromosomal aberrations. Int J Environ Res Public Health 19:10019, 2022.

22 CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS THE MAIN FOCUS of chromosomal prenatal diagnosis had been upon trisomy 21, usually in the context of older childbearing age or of an increased-risk screening test. Trisomy 21 does remain, for most women and couples, the prime concern—the condition that most people are aware of—but with the sophistication of 21st century technology, the great majority of chromosomal imbalances are, in principle, diagnosable. Noninvasive prenatal testing—no more than a blood test for the woman—widens access and increases uptake very considerably. Routine fetal ultrasonography can detect quite subtle malformation, and genetic testing will often follow such a discovery. The counselor can expect to deal with a broad spectrum of chromosomal abnormality, presenting in the prenatal clinic. DECISION-MAKING FOLLOWING PRENATAL DIAGNOSIS OF A CHROMOSOMAL ABNORMALITY To some extent, the possibility of “other abnormalities besides Down syndrome” should have been raised at pretest counseling. But when a chromosomal abnormality is actually discovered, it is of course necessary to discuss in detail with prospective parents the implications of this particular abnormality, and to help them decide on a suitable course of action. Outlines of the clinical consequences of these abnormalities follow, to serve as a basis for the decisions that these women and couples need to make. In transmitting the information, the counselor is obliged to be clear and accurate about the particular abnormality, and to take care that the couple’s autonomy in the decision-making process is not compromised. A decision for or against termination is the immediate one to be made. Hodgson et al. (2016) interviewed 102 women and men six weeks after diagnosis of fetal abnormality1 and identified the following considerations in the decision about whether to terminate the pregnancy: 1. Uncertainty about the diagnosis or results 2. Uncertainty about the prognosis or severity 3. Expected survival 4. Worry about baby experiencing pain 5. Personal views on abortion and disability 6. Self-perceived ability to parent a disabled child 7. Potential for long-term care needs 1 In this study the fetal abnormality was chromosomal in origin in 63% of affected pregnancies. 656  REPRODUCTIVE CYTOGENETICS 8. Societal treatment of people with disability 9. Impact on their other children 10. Impact on their career and finances 11. Potential maternal risks in pregnancy 12. Not the life they had imagined for their child 13. Fearing judgment of others 14. Societal views about abortion 15. Wondering what others would do in the same situation Shaffer et al. (2006) undertook a large retrospective review (1983–2003), analyzing parental decisions in 816 prenatal diagnoses of a major aneuploidy, at a San Francisco clinic. Termination was chosen in 86% of autosomal trisomy and in 60% with a sex chromosome aneuploidy. Of the latter, the rates of termination increased progressively from XXX (40%) to XYY (57%), 45,X (65%), and XXY (70%), in parallel with a perceived severity of phenotype. The rates did not differ significantly during the 21-year period, to the slight surprise of these authors. Jacobs et al. (2016) assessed responses to the prenatal diagnosis of an aneuploidy in a Scottish population, observing a fall-off in the numbers choosing termination during the period 2000–2011. They suggested that this may have reflected “societal changes in accepting greater diversity” but acknowledged that “this interpretation is of course purely speculative.” Finding a variant of uncertain clinical significance (VUS) is a more recent question, and while pregnancy termination is infrequently chosen, its discovery at prenatal diagnosis nonetheless imposes a considerable psychological burden on prospective parents (Lou et al. 2020; Libman et al. 2024). The simplicity of a blood test may have drawn many more into having a prenatal test, but this appears to have made little difference to the termination rates for DS, and indeed many parents have used NIPT for information, and have continued pregnancies in the context of a high risk of DS (Hill et al. 2016). Trisomy 21 Skotko et al. (2009) emphasize the need for the person conveying the news of a Down syndrome (DS) result to be well informed, whether that be a counselor, obstetrician, or other health professional (and of course this qualification is scarcely confined to a diagnosis of DS). Ideally, the news should be given in person; but where that is not feasible, a phone call or telehealth appointment should be at a prearranged time. Parents who decide to continue a trisomy 21 pregnancy, versus those who have chosen termination, would presumably come from different points of view. Skotko and colleagues note that contact with a DS support group might be useful for some couples in deciding the fate of a DS pregnancy, although they do observe that few studies have assessed the views of those who have terminated a trisomic pregnancy, from whom the other face of the decision could be given a hearing. A study of health professionals in Finland showed some inconsistency in comparing the points of view of midwives and public health nurses with the options available to their patients; and the acknowledgment was made that this difference could be viewed as a healthy sign, in recognizing that plurality of opinion is the way of the world (Jallinoja et al. 1999). Thus, most (79%) of these midwives and nurses agreed that all pregnant women should be offered a screening test, although only 44% personally accepted the

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  657 concept of genetic abortion. An acceptance of abortion correlated with education and with a professional experience with DS patients. In the United States, Britt et al. (2000) studied 142 women who had had a prenatal diagnosis of trisomy 21, seen in Detroit over the period 1989–1998. Those who had already had children, and where the diagnosis of trisomy 21 was made earlier in the pregnancy, were more likely to choose termination. In the Netherlands, Korenromp et al. (2007) found that among women who had chosen termination following prenatal diagnosis of DS, “child-related” motivations (Table 22–1) were the most prevalent, but concerns about burden to the family were also important. Lou et al. (2020) discuss how some women detach themselves from the pregnancy, for example by using the term “fetus” and acting as if they were not pregnant, whereas others maintain attachment, such as by referring to the fetus as “our child.” But these expressions of attachment and detachment sometimes coexist and may change over time. Sex Chromosome Abnormality NIPT has brought renewed focus on the prenatal diagnosis of sex chromosome aneuploidy (SCA). Prior to NIPT, very few SCAs were diagnosed prenatally, and these were mostly as incidental findings. Identifying SCA had not been an aim of prenatal screening; but sex chromosomes are now included in many NIPT tests (in part as a response to requests for fetal sex diagnosis). In Victoria, Australia, Loughry et al. (2023) reported a fivefold increase in the prenatal diagnosis of 47,XXY between 2005 and 2020, corresponding to the time of NIPT implementation. These matters now sharpen the question of what to tell prospective parents (White et al. 2023; Reimers et al. 2023) and of the associated ethical issues (Warton et al. 2023). Despite this, parents report being unprepared for the diagnosis of SCA when it actually happens (Lewit-Mendes et al. 2023). Even prior to NIPT, views about SCA were evolving, generally in the direction of a more conservative response to the news of a chromosomal abnormality. In Denmark, Nielsen et al. (1986) reported that ~80% of prenatal diagnoses of sex chromosome aneuploidy at that time were followed by the choice of abortion. In an English/Finnish study from the same period, termination (in ~60% overall) was more likely to be chosen in Table 22–1.  Reasons for the Choice of DS Termination The Five Most Acknowledged “Child-Related” Motivations of 71 Women Choosing Termination of a Trisomic 21 Pregnancy, from a Study Based in Eight Dutch Hospitals MOTIVATION STATEMENT PERCENTAGE OF WOMEN IN AGREEMENT WITH STATEMENT I believed the child would never be able to function independently. 92 I considered the abnormality too severe. 90 I considered the burden for the child itself too heavy. 83 I worried about the care of the child after my/our death. 82 I considered the uncertainty about the consequences of the abnormality too high. 78 Source: From MJ Korenromp et al., Maternal decision to terminate pregnancy in case of Down syndrome, Am J Obstet Gynecol 196:149.e1–11, 2007. 658  REPRODUCTIVE CYTOGENETICS the case of the XXY and 45,X karyotypes by younger parents with fewer previous children, and in all cases in which an ultrasonographic defect was identified (Holmes-Siedle et al. 1987). From a large survey of centers in five European countries covering the years 1986 to 1997, the rate of choice of termination with respect to XXY was 44% (Marteau et al. 2002). In a German study over a similar period, termination was chosen by a much smaller fraction, only 13%, among parents who had been given a prenatal diagnosis of 47,XXX, 47,XXY, or 47,XYY (in contrast, just 2% of parents at the same clinic decided to continue a pregnancy with trisomy 21) (Meschede et al. 1998). This may in part have reflected the practice of this clinic to emphasize the point that “the mean global IQ of around 90 falls well within the normal range and is compatible with a productive and socially well-adjusted life.” More recently, half of couples in a Chinese study elected to continue the pregnancy, but the choice to continue was very much dependent on the specific sex chromosome aneuploidy: 89% for 47,XYY, 85% for 47,XXX, 35% for 47,XXY, and just 7% for 45,X (Luo et al. 2024). In the specific case of Turner syndrome, from 19 registries in 10 countries across Europe from 1996 to 1998, 79% of parents chose termination if morphological abnormality (and in particular cystic hygroma) had been seen on ultrasonography, versus 42% in which the diagnosis had not been led into by an ultrasonographic defect (Baena et al. 2004). The poor post-natal survival associated with prenatally diagnosed cystic hygroma (Levy et al. 2021) is likely to have a major influence on decision-making. The way in which information is given has an important impact, and counselors need to be well aware of the weight that parents, in some emotional turmoil at the news they have just received, may put upon the news given them. Consider the example of 47,XXY Klinefelter syndrome. In the European survey mentioned above, Marteau et al. (2002) assessed responses to the prenatal diagnosis of XXY when counseling had been given by obstetricians, pediatricians, midwives, health visitors, or genetics specialists. Women counseled solely by genetics specialists were more than twice as likely (a relative risk of 2.4) to continue the pregnancy versus those counseled either by other professionals or by other professionals along with a geneticist. It seems probable that these differences may reflect the style of counseling. In a review of published reports, Jeon et al. (2012) noted the importance of these factors in influencing a decision: the specific type of sex chromosome abnormality, the gestational week at diagnosis, the parents’ age, the providers’ genetic expertise, and the number of children already had, or their desire for (more) children. They noted that among those parents choosing to continue a pregnancy, their socioeconomic status and ethnicity were particularly relevant. On the other hand, parents choosing termination were characterized by a fear or anxiety of having a child with a sex chromosome abnormality and also by having received directive counseling. The desirability for a consistent approach, with access to accurate information, is to be emphasized, as is—of course—the requirement to enable women’s choices to be well informed in the broadest sense, and for the counseling to be nondirective (Linden et al. 2002). Reimers et al. (2023) emphasize the importance of inclusive and non-stigmatizing language when discussing genetic diagnoses and of drawing upon the most up-to-date and unbiased SCA data. Beyond the clinic, there are support groups, public information resources, and communicating with other parents as means to become further informed about the implications of a SCA (in the short period of time during which a decision must be made), and Linden and colleagues note the pros and cons of taking these paths. As noted earlier with respect to trisomy 21, the views of those who had previously chosen

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  659 to terminate a pregnancy are less readily accessible. The prime responsibility for putting couples in the best position to make an appropriate decision lies with the counselor. As for subsequently informing the children from the pregnancies that are continued, Riggan et al. (2024) emphasize the importance of telling them of their sex chromosomal diagnosis and its implications in a timely, well informed, and sensitive manner. “Microarray-Level” Rearrangement Microarray analysis applied to prenatal samples can substantially increase the diagnostic pick-up (Figure 1–7). The other side of this two-edged sword is the fact that some micro-imbalances are not pathogenic, or may only be pathogenic in specific circumstances— for example when co-inherited with other variants from either parent. Indeed, one commentator wrote, somewhat provocatively, that prenatal array testing is likely “to produce a flood of information that is overwhelming, anxiety-producing, inconclusive, and misleading” (Shuster 2007); and Werner-Lin et al. (2016a) provide some testament, in documenting frustration and anxiety of couples receiving uncertain results, that this prediction may not have been entirely inaccurate. A practical response is to limit reporting to those variants that reach a certain threshold in terms of evidence of pathogenicity, so that results of unclear clinical significance or with very low penetrance2 are less often encountered (Armour et al. 2018; Maya et al. 2020). This approach is supported by data from McCoy et al. (2025), who followed up 46 children who had received VUS results in fetal life and found similar intellectual functioning and social functioning to those without a VUS. The counterpoint is that this approach will occasionally miss a clinically significant CNV. Parental testing may shed further light, but will not necessarily resolve the question, particularly for CNVs of incomplete penetrance. Goh et al. (2025) pose the following: “Consider a family with an affected child who has a CNV with 20% penetrance that was inherited from a parent with a mild phenotype. In contrast, a different family may have the same CNV with no affected family members. Some may wonder if the likelihood of a relevant phenotype occurring in a newborn child identified prenatally with this CNV is 20% regardless of the scenario, while others might wonder if there are known mechanisms that could impact the chance of penetrance for this CNV.” Counselors need to be quite au fait with the interpretations and to maintain close liaison with expert scientists in the field (the science); they must also be aware of the particular subtleties that counseling in this circumstance will demand (the art) (Werner-Lin et al. 2016b). The uncertainty of a prenatal result, for those continuing the pregnancy, can flow over into the child’s life: “They Can’t Find Anything Wrong with Him, Yet,” as Werner-Lin et al. (2017) title their paper apropos. There are other ways in which the use of microarray is changing the face of prenatal chromosome analysis. The two notable categories are the identification of de novo balanced chromosome rearrangements, and very small marker chromosomes. Although typically benign, but not always so, these abnormalities have historically necessitated detailed counseling and cytogenetic follow-up, and they have been a cause of anxiety for patients. The fact that these abnormalities may be invisible on microarray is a feature of 2 There is currently no consensus definition for low penetrance.

660  REPRODUCTIVE CYTOGENETICS the methodology that is, perhaps, almost an advantage in the prenatal setting. The capacity for next-generation sequencing (NGS) methodologies—“post-microarray” testing, one could almost say—to clarify the nature of apparently balanced rearrangements, is mentioned on p. 24. MOSAICISM: CONFINED, CONSTITUTIONAL, AND PSEUDO Mosaicism is the bane of cytogenetic prenatal diagnosis. Most times, it turns out to have been a false alarm, and the mosaicism in villus tissue or amniocytes does not reflect a true constitutional mosaicism of the fetus. This is a problem for the laboratory to resolve, inasmuch as they are able. We may list these two major categories: confined placental mosaicism, and true constitutional fetal mosaicism. A third category, pseudomosaicism, refers to an abnormality that arose during tissue culture in vitro (“cultural artifact”), but with the embryonic and extra-embryonic tissues being chromosomally normal. Pseudomosaicism can be eliminated by analyzing uncultured cells, using DNA-based techniques. A chromosomally abnormal cell line may exist only in the extra-embryonic tissues of the placenta (chorion, amnion), and the embryo is 46,N. This is confined placental mosaicism (CPM). CPM is encountered at CVS rather than at amniocentesis. It is uncommon that an observation of apparent CPM at CVS reflects a true constitutional mosaicism of the fetus. Grati (2014) offers an exhaustive treatment of the question of mosaicism at CVS, based upon an experience of more than 50,000 procedures, in which mosaicism was seen in 2.2%; of these, just 13% (0.3% of the total) proved to have a true fetal mosaicism (and see Table 22–2); she provides further authoritative review, with reference to the issues raised by noninvasive prenatal testing (Grati 2016). Comparable fractions are seen on prenatal diagnosis by microarray: 1.8% of CVS, and 0.5% on amniocentesis, in the analysis of Carey et al. (2014). Mosaicism for copy number variants can also be detected by microarray: Gu et al. (2018) showed that 3.0% Table 22–2.  Different Types of Mosaicism Identified After Chorionic Villous Sampling* TYPE NATURE TROPHOBLAST (DIRECT CVS, NIPT) MESENCHYME (CULTURE CVS) AMNIOCYTES RELATIVE FREQUENCY I CPM Abnormal Normal Normal 34.8% (308/886) II CPM Normal Abnormal Normal 42.3% (375/886) III CPM Abnormal Abnormal Normal 10.2% (90/886) IV TFM Abnormal Normal Abnormal 1.6% (14/886) V TFM Normal Abnormal Abnormal 5.8% (51/886) VI TFM Abnormal Abnormal Abnormal 5.4% (48/886) *Based upon a series of 52,673 CVS procedures, in which 1,136 (2.2%) showed mosaicism, and of which 886 had follow-up amniocentesis. CPM = confined placental mosaicism; CVS = chorionic villous sampling; TFM = true fetal mosaicism. Source: From Grati FR. Chromosomal mosaicism in human feto-placental development: Implications for prenatal diagnosis. J Clin Med 3:809–837, 2014. CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  661 of CVS3 prenatal arrays samples were indicative of mosaicism, including submicroscopic CNVs. Likewise, the Danish study of Lund et al. (2020) identified mosaicism for a microscopic or submicroscopic CNV in 1.3% of uncultured CVS3 samples, as well as in 2.8% of samples having a mosaic autosomal or sex chromosome aneuploidy.4 True fetal mosaicism was confirmed in 18% of mosaic whole chromosome aneuplodies, and in 25% of mosaic CNVs. A long-term follow-up study (Amor et al. 2006) is noted below. Considerable discussion follows, but at the outset we emphasize that true mosaicism of the fetus is infrequently observed, and that the majority of mosaicism identified at prenatal diagnosis, more especially at CVS, does not presage an abnormal baby. It is important to keep this perspective in talking with parents (according to the particular attributes of the mosaicism, as we go on to discuss) and to avoid causing any more anxiety than that which an “abnormal” result inevitably brings. Applied Embryology Interpreting mosaicism obliges an understanding of the earliest events of development of the conceptus, as we now outline (Figure 22–1). The zygote undergoes successive mitoses to produce a ball of cells (the morula, at Day 4). The morula then cavitates to produce an inner cyst, and it becomes the blastocyst (Days 6 to 9 in Figure 22–1). The outermost layer of the blastocyst is composed of trophoblast, and this tissue becomes the outer investment of the chorionic villi. The inner cell mass protrudes into the blastocystic cavity, and this will give origin to the embryo. It comprises two different cellular layers, the epiblast and the hypoblast. In a 64-cell blastocyst, most cells are trophoblasts, the inner cell mass comprises about 16 cells, within which only about four (epiblast) cells will give rise to the embryo itself. The hypoblast forms the spherical primary yolk sac (whose roof is, transiently, the ventral surface of the embryo). The primary yolk sac gives rise to the extra-embryonic mesoderm, sandwiched between itself and the outer cytotrophoblast and thus producing a three-layered sphere (Figure 22–1, Day 11). The mesodermal cells now invade the blastocystic cavity, and this mesodermal mass is in turn cavitated to produce the extra-embryonic celom such that there are outer and inner layers of extra-embryonic mesoderm (Figure 22–1, Day 13). The outer layer, underlying the trophoblast, gives rise to the mesenchymal core of the chorionic villus, and the inner layer becomes the outer (mesodermal) surface of the amniotic membrane. The amniotic cavity enlarges at the expense of the extra-embryonic celom (Figure 22–1, Week 4) and eventually obliterates it (by the end of the first trimester), with the mesodermal layer of the amnion fusing with the mesodermal layer of the chorion (Figure 22–1, Week 8). The epiblast gives rise to the amniotic cavity, the floor of which is the “dorsal” (ectodermal) surface of the embryo, and its roof is the amnion, these being continuous at their margins. Thus, the embryonic integument and the inner surface of the amniotic membrane—which are the source of the embryonic and amniotic epithelial cells present in amniotic fluid—have the same lineage. At the beginning of the third week, 3 Both cytotrophoblastic and mesenchymal samples. 4 The prevalence of CPM in this study was higher than most other studies, an observation that may be due at least in part to the indication for CVS being an increased-risk first trimester combined screen, a test which may itself enrich for CPM cases due to an association between low PAPP-A levels and CPM.

662  REPRODUCTIVE CYTOGENETICS Figure 22–1.  The timeline of embryogenesis, and the development of the different tissues. Note: “Embryoblast” (day 6) is synonymous with Inner Cell Mass, and these cells include those which will give rise to the actual embryo. The unlabeled black and white cells shown at days 8 and 9 (upper right in each panel) represent decidua placentalis. Source: From R Essers et al., Prevalence of chromosomal alterations in first-trimester spontaneous pregnancy loss, Nat Med 29:3233–3242, 2023. Courtesy A Salumets and MZ Esteki, and with the permission of Nature Medicine. the primitive streak arises from the epiblast, and this in turn gives origin to both endoderm and intra-embryonic mesoderm. Endoderm gives origin, among other tissues, to urinary tract and lung epithelia–desquamated cells which contribute to the cellular population of amniotic fluid. Albeit that the extra- and intra-embryonic mesoderms have different origins, there may be migration of some intra-embryonic mesodermal cells into the (extra-embryonic) amniotic mesoderm. Cells from the latter add a minor fraction to the population of amniocytes, but have a proliferative advantage and may come to comprise most of the cells present following in vitro culture. Amniocentesis is a procedure that samples cells having origin from the epiblast of the inner cell mass, and these cells rather closely reflect the true constitution of the embryo. Chorionic villus sampling, on the other hand, samples more distantly related cells: trophoblast cells (direct and short-term culture), which were the first lineage to differentiate from totipotent cells of the morula, and villus core cells (long-term culture), which reflect the more recently separated lineage of the extra-embryonic mesoderm. The differing origins of tissues sampled by different means are set out in Figure 22–3. Noninvasive prenatal testing (NIPT) samples cell-free DNA that originated from apoptotic trophoblast cells, and therefore is more closely related to CVS than to amniocentesis (indeed, it has been called a “liquid early CVS”). CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  663 Figure 22–1.  Continued. Figure 22–2.  Ultrasound of the Embryo. Notes: Ultrasound picture of embryo at 10–11 weeks gestational age, very close to actual size (note centimeter markers at right). Note amnion (A), amniotic cavity (AC), extra-embryonic celom (EC), umbilical cord (U), “physiological omphalocele” (O), yolk sac (Y), and placenta (P). The relative positions of embryo and other structures are similar to the depiction in Figure 22–1, Week 8. Source: Courtesy HP Robinson. 664  REPRODUCTIVE CYTOGENETICS Mechanisms of Mosaicism Mosaicism may involve aneuploidy for an intact chromosome or for an abnormal chromosome, along with a normal cell line. Two broad formats may apply, whereby the mosaicism arose: first, a mitotic error in an initially normal conceptus which gives rise Figure 22–3.  Cell Lineages vis a vis Prenatal Diagnostic Approaches. Notes: The cell lineages arising from differentiation in the very early conceptus are shown, with respect to their relevance at prenatal diagnosis. The fertilized egg (1) produces a trophoblast precursor (1b) and a totipotent stem cell (2), which in turn forms another trophoblast precursor (2b) and a stem cell (3) that produces the inner cell mass. The inner cell mass divides into stem cells for hypoblast (3b) and epiblast (4). The epiblast cell(s) (5) produces embryonic ectoderm and primitive streak, and the latter is the source of embryonic mesoderm and endoderm. The cell lineages sampled at various prenatal diagnostic procedures are indicated at right. E, epiblast; H, hypoblast; P, primitive streak; Y, yolk sac.a a  This construction is to be compared with that of Kennerknecht et al. (1993b), in which three postzygotic mitoses occur, producing eight totipotent cells, before the cells begin to take on their tissue identities. Robinson et al. (2002) propose a further variation, with some cells of the embryonic mesoderm migrating into the (otherwise extra-embryonic) mesodermal layer of the amnion. Source: From DW Bianchi et al., Origin of extra-embryonic mesoderm in experimental animals: Relevance to chorionic mosaicism in humans, Am J Med Genet 46: 542–550, 1993. Courtesy DW Bianchi, and with the permission of Wiley-Liss. CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  665 to an abnormal cell line; or, second, an initially abnormal conceptus, typically due to a meiotic error, with a subsequent mitotic event generating a normal cell line (Figure 3– 15). An “incomplete” trisomy rescue may generate a de novo sSMC (see below). The distribution of the normal and the abnormal cell lines in the fetus and the placenta depends upon the time and the place of the abnormal mitotic event. If, for example, a trisomic conceptus is “rescued” by the generation of a normal cell line, at a very early stage, in a cell that is going to give rise to the inner cell mass and to some of the extrafetal tissues, then the embryo may be 46,N, and the placenta will show mosaic trisomy. If rescue occurred at a later stage, the placenta might be entirely trisomic, with a mosaic trisomy of the fetus. These and other possible combinations are depicted in Figure 22–4. The eventual phenotype will be influenced by the tissue distribution of the cell Figure 22–4.  Types of Mosaicism of the Fetal-Placental Unit. Notes: In each image, the fetus is depicted enclosed in its sac at right, with the chorionic villi comprising the placenta to left. Gray areas indicate an aneuploid cell line; white areas indicate karyotypic normality. In reality, the distributions of the two cell lines are unlikely to be as clear-cut as is shown here. In the examples showing placental mosaicism, the path taken by the sampling needle will determine whether the abnormality is detected or missed at chorionic villus sampling. The cartoon of the fetus, sac, and placenta is close to the form and about two-thirds of the size that actually exists at 10 weeks 0 days (gestational age as measured clinically, dated from the last menstrual period), when crown-rump length is approximately 30 mm.

666  REPRODUCTIVE CYTOGENETICS lineages that contain the trisomic chromosome, and the normal:trisomic proportions in various tissues. The important distinction between confined placental mosaicism and true fetal mosaicism, as identified at CVS and follow-up amniocentesis, is outlined in Tables 22–2 and 22–3. Laboratory Assessment of Mosaicism The resolution of mosaicism in the cytogenetics laboratory and in its clinical interpretation can differ for CVS and amniocentesis, and we will consider them separately. Regarding the cytogenetic analysis of cultured cells, we can apply to both CVS and amniocentesis the concept of different levels of in vitro mosaicism, originally developed for amniocentesis by Worton and Stern (1984) and refined by Hsu et al. (1992) and Hsu and Benn (1999), as follows: Level I: A single abnormal cell is seen. With near certainty this is cultural artifact, and it is thus pseudomosaicism. The laboratory would not usually report the single-cell observation if the analysis of additional cells failed to confirm the abnormality.5 Level II: Two or more cells with the same chromosomal abnormality in a dispersed culture from a single flask, or in a single abnormal colony from an in situ culture (i.e., possibly or probably just a single clone). Some would also include the observation of two or more colonies from the same in situ culture. The abnormality is not observed in multiple colonies from other independent cultures. This form of mosaicism is almost always pseudomosaicism. It would not usually be reported to the physician if additional work-up failed to confirm the trisomy, but it may be reported if additional studies were inadequate, if fetal anomalies had been identified, or in the case of certain chromosome abnormalities which are well recognized as existing in the mosaic state (e.g., trisomy 16). A course of action to resolve the issue cytogenetically, in the case of amniocentesis, is given in Table 22–5. Level III: Two or more cells with the same chromosome abnormality, distributed over two or more independent cultures. Level III is likely to reflect a true mosaicism, and the cytogeneticist will report this finding immediately. (Some allow level III to include more than one colony in only a single flask, although this could be an “overinterpreted level II” if two colonies in the one flask had arisen from a single cell whose progeny migrated and established separated clones.) The distinction may not be quite as clear as this in practice, but this is a useful working definition. The mathematics of sampling comes into the picture: How many cells need to be looked at in order to establish what level of confidence that the possibility of mosaicism, of what extent, can safely be disregarded? Tables have been derived to assist in answering this question (Hook 1977; Sikkema-Raddatz et al. 1997a). Inevitably, 5 An exception may be mosaicism for an isochromosome, as a handful of reports have demonstrated true mosaicism in the context of a single abnormal cell at prenatal diagnosis (see below). CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  667 low-level mosaicism will on rare occasions be missed. Given the reality that only a limited number of cells can be karyotyped on the classical approach, the statistics will sometimes conspire against the cytogeneticist, and only normal cells will be examined. This has to be accepted: The test is not perfect. Molecular Methodologies and Mosaicism Microarray. Applying CGH arrays, Ballif et al. (2006) used experimental dilutions of a 46,XY sample with a 47,XY,+21 sample, in order to mimic trisomy 21 mosaicism. They demonstrated that mosaicism of 20% or greater could confidently be identified. In terms of actual experience, the threshold of detection in practice may be as low as 9% (Carey et al. 2014). When SNP arrays are used, the B allele frequency is more sensitive to the subtle loss or gain of a haplotype than the log R ratio (Conlin et al. 2010). Cross et al. (2007) set up mock samples from normal and trisomy 8 fibroblasts, and by analyzing the extracted DNA with a 50K SNP array, they established that down to a 20% level, mosaicism was readily recognized but fading out at approximately 10%. Conlin and colleagues were able to push the technology further and detect mosaicism as low as 5%. Quantitative Fluorescent Polymerase Chain Reaction. Mosaicism may be detected with reasonable efficiency on qualitative fluorescent polymerase chain reaction (QF-PCR). In one large retrospective study, Donaghue et al. (2005) reviewed 8,983 amniocentesis and CVS samples, from which 18 cases with mosaicism were identified. More (12) were detected by QF-PCR than by karyotyping (8), although neither approach picked up all. By their reckoning, a tissue load of 15% or more abnormal cells would allow detection of mosaicism by QF-PCR. Noninvasive Prenatal Testing. Although NIPT is frequently referred to as the testing of cell-free fetal DNA, the primary source is of apoptotic placental cells of the cytotrophoblast, and so it is more accurate to speak of cell-free placental DNA. Thus, the question of mosaicism, as with CVS, applies similarly to NIPT, and it is an important cause of a false-positive result (Grati 2016). This is particularly relevant when considering the choice of an invasive procedure for the confirmation of a high-risk NIPT result. Given that NIPT is typically performed at 10 weeks gestation, a check CVS can be offered without delay, but in the awareness that the same false-positive (i.e., due to CPM) result might be forthcoming.6 Consideration of the likelihood of CPM being present for a given abnormality, along with the presence of ultrasound abnormalities, can be used to guide selection of the test for confirmation. Confirmation rates after NIPT screening for trisomy 21 and trisomy 18 are high, due to the low frequencies of CPM, and CVS is usually recommended for these regardless of ultrasound findings (Scarff et al. 2023). For trisomy 13 the rate of CPM is higher, and therefore if ultrasound is normal, testing may be deferred until amniocentesis. NIPT is less effective at picking up mosaicism. Lund et al. (2024) found that only 44% of clinically significant mosaic cases were detected by NIPT, more likely so in higher-level mosaicism. 6 This point can have a real clinical relevance, as Srebniak et al. (2014) describe in a diagnosis of 45,X at NIPT. A confirmatory uncultured CVS showed 45,X/46,X,idic(Y) mosaicism, and termination of the pregnancy was initially chosen. But a check amniocentesis showed non-mosaic 46,X,idic(Y), and a decision was then taken to continue the pregnancy, accepting that the male child would likely be infertile. In retrospect, it was concluded that the 45,X line was confined to chorionic villus.

668  REPRODUCTIVE CYTOGENETICS Prediction of Phenotype in an Individual Mosaic Case Elegant theorizing notwithstanding, the pragmatic observations from published cases in the literature provide the mainstay of the advice that the counselor may offer the parents in an individual case. (Large series are better than single case reports, which are better than anecdote.) Are mosaicisms for some particular chromosomes, or types of aberration, of more concern than others? What is a low enough level of mosaicism, if any such exists, to have a degree of confidence that the child will be physically and intellectually normal? We set out below summaries of the recorded examples from the literature, none of which necessarily provide a firm answer, but which may serve as the basis for discussion and counseling. The numbers in some are very small. Another difficulty with these observational data is that, for the most part, the window of assessment of the child’s phenotype was confined to the neonatal period. Of course, many children who are eventually diagnosed with significant disability may have been well grown and morphologically normal at birth, with normal functional neurology (inasmuch as this may be assessed in a baby). On the other hand, it is possible to overdiagnose problems in babyhood, as a child who subsequently develops normally may prove (Warburton 1991; Joyce et al. 2001). An important concern in mosaicism is that a cell line inaccessible to analysis—specifically, in the brain—might contain the abnormal chromosome, notwithstanding a normal karyotype in the post-natal tissues which are normally examined, namely blood or buccal mucosa, and possibly skin. If so, cognitive functioning could be compromised. Those few reports that include follow-up data for some years into childhood (Baty et al. 2001; Amor et al. 2006) are therefore most valuable. Nevertheless, no certainty can be offered, recognizing that every case of mosaicism will be unique, in terms of the extent and qualitative tissue distribution of the abnormal lineage. CHORIONIC VILLUS CULTURE AND MOSAICISM, INCLUDING CONFINED PLACENTAL MOSAICISIM CVS mosaicism is detected in about 2% of procedures at the 10- to 11-week mark. Mosaicism from an early mitotic error in a single cell can give rise to confined mosaicism (confined to placenta, or to fetus) or to generalized mosaicism (present in both fetus and placenta), according to the destined lineage of that cell; the broad range of possibilities is shown in Figure 22–4. Depending upon the timing and site of the event producing the mosaic state, the karyotypes observed at CVS will vary. The extreme form is complete discordance, with a non-mosaic 46,N karyotype in fetus and non-mosaic aneuploidy in CVS, or vice versa.7 Clearly, an important distinction to make, inasmuch as it is possible to do so, is between a mosaicism confined to the placenta (CPM) and causing little or no compromise of its function, and the presence of an aneuploid cell line extending into the fetus, plus or minus an important effect upon the ability of the placenta to support fetal development. Follow-up amniocentesis is certainly advisable: A normal result, which is very 7 CPM is the main, but not the only, cause of discrepancy between the CVS and fetal/child karyotypes. One very rare explanation is that there was a resorbed co-twin with a different karyotype, with the sampling instrument having traversed its placental remnant (Tharapel et al. 1989). CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  669 often what eventuates, will substantially provide reassurance that the aneuploidy did not involve fetal tissue. In the large series of Grati (2014) as mentioned above, the great majority of mosaicism turned out to be CPM, with only a small fraction proving potentially to be true fetal mosaicism on follow-up amniocentesis. As shown in Table 22–2, Grati proposes the subclassification of CPM as follows: type I, aneuploidy confined to cytotrophoblast (recognized only at direct/short-term analysis); type II, aneuploidy confined to villous stroma; and type III, an aneuploid cell line in both cytotrophoblast and stroma. Since NIPT investigates the cytotrophoblast, it will detect only types I and III. A proviso is given in Daniel et al. (2004), who assess that on the order of 10% of CVS mosaicism for certain “rare chromosomes”8 interpreted as CPM, may in fact reflect a cryptic fetal mosaicism that would not be detected at follow-up amniocentesis and which might or might not have important phenotypic consequence. Origin of Trisomy in Confined Placental Mosaicism Robinson et al. (1997) studied 101 cases in which CPM had been identified at CVS, seeking to establish correlates of the origin of the trisomy. Some CPM trisomies are usually of mitotic (somatic) origin, the zygote having been 46,N. Others typically arise meiotically, and the zygote was trisomic ab initio. They determined that the meiotic or mitotic origins of the trisomy are substantially chromosome-specific. For example, trisomy 8 CPM is characteristically the consequence of a mitotic event, while in contrast, almost all cases of CPM for trisomy 16 have arisen at maternal meiosis I. From a meiotic origin, “correction” may generate a 46,N karyotype in the fetus, but there is a risk for this to be associated with uniparental disomy. Thus, of the trisomy 16 CPM cases, about half displayed UPD(16) in the fetus. A meiotic origin of the CPM typically implies a more guarded prognosis than if the error had arisen somatically. Trisomy 2 at CVS (see below) is an example of a mosaicism that conveys quite different implications according the meiotic or mitotic mechanism of its generation. Level III Mosaicism in CVS Level III mosaicism in CVS raises an immediate concern. Management at this point (which will usually be around 12–13 weeks) is aimed at demonstrating, as much as possible, fetal normality—or, if it so transpires, confirming a true fetal mosaicism. Amniocentesis with rapid FISH analysis of a large number of cells, along with detailed ultrasonographic assessment of fetal morphology, is usually the next plan of action; or a microarray analysis might be performed on the uncultured amniocytes. In fact, the majority of cases will return normal results after this additional work-up, since the mosaicism is likely confined to the placenta. A large body of data on level III mosaicism for autosomal trisomy was gathered by the European Collaborative Research Group on Mosaicism in CVS (EUCROMIC) (Hahnemann and Vejerslev 1997) comprising information on just over 92,000 CVS procedures, from 79 laboratories, during 1986–1994. Mosaicism (or non-mosaic fetoplacental discrepancy) was seen in 650 (1.5%) cases. Of these, 192 were followed up 8 In this study, chromosomes 5, 8, 9, 10, 11, 12, 14, 15, and 16.

670  REPRODUCTIVE CYTOGENETICS in detail, with karyotyping of fetal fibroblasts, fetal blood, amniocytes, or neonatal tissues. Most (84% of the 192) represented CPM. The abnormal cell line was present in either trophoblast (type I CPM; in 50%), villus mesenchyme (type II CPM; in 30%), or both (type III CPM; in 20%). Comparable numbers are due to a single Italian laboratory based upon a total of 60,347 samples, of which 2.2% showed mosaicism (Malvestiti et al. 2015). Of the 1,001 of these going on to amniocentesis, in only 131 (13%; 0.2% of the whole) was a true fetal mosaicism actually demonstrated. A greater risk (18.6%) applied when the abnormality had been detected on villus culture. The likelihood of confirmation at amniocentesis varies considerably according to the specific chromosome abnormality, being greatest when the chromosome concerned was a sex chromosome, marker, or one of those involved in the common trisomies (Table 22–3). Malvestiti and colleagues emphasize the value and validity of follow-up amniocentesis. Randomness of Sampling The vagaries of sampling may influence the interpretation, as the following examples show. We followed to term a woman in whom first-trimester CVS had shown trisomy 7 mosaicism with 47,XY,+7 in three out of eight clones; and yet three out of four placental samples (one from each quadrant), and peripheral blood from the (normal) baby, karyotyped 46,XY. Just one placental sample, which was not histologically distinguishable from the others, was 47,+7 (Watt et al. 1991). Presumably, the CVS sampling catheter had traversed this unrepresentative region of the placenta, and most of the sample that was eventually analyzed came from here. Table 22–3.  Follow-Up Amniocentesis Likelihood of Confirmation of a Mosaic Chorionic Villous Sampling Result at Follow-Up Amniocentesis, According to Type of Chromosome Aberration ABERRATION PHILLIPS ET AL. (1996) HAHNEMANN AND VEJERSLEV (1997) MALVESTITI ET AL. (2015) COMBINED LIKELIHOOD OF CONFIRMATION 47,+mar 8/30 22/65 30/95 (31.6%) Sex chromosome aneuploidies 17/109 51/153 68/262 (26.0%) Common trisomies (13, 18, 21) 15/79 15/66 29/151 59/296 (20.0%) Rearrangements 3/35 18/178 21/213 (9.9%) Polyploidy 1/28 2/63 3/91 (3.3%) Other autosomal trisomies 6/188 5/126 9/391 20/705 (2.8%) Notes. These data include cases in which the abnormal cell line was detected only in trophoblast (type I CPM). A greater risk is expected when the abnormality has been detected on villus culture. CPM = confined placental mosaicism. Sources: Data of Phillips et al. (1996), Hahnemann and Vejerslev (1997), and Malvestiti et al. (2015). CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  671 Similarly, in a case of i(5p) diagnosis at CVS following the birth of the (normal) baby, we identified a region of placental mosaicism (Clement Wilson et al. 2002). De Pater et al. (1997) did a CVS in a pregnancy of 37 weeks gestation in which severe growth retardation and a heart defect had been identified, and this showed non-mosaic trisomy 22. However, from a simultaneous amniocentesis only two out of 10 clones were 47,XX,+22, the other eight being normal, and a cord blood from the (abnormal) baby gave a non-mosaic 46,XX karyotype. Skin fibroblasts demonstrated mosaicism, 47,XX,+ 22[7]‌/46,XX[25]. Of 14 placental biopsies studied by interphase FISH, only one showed trisomy 22 cells, and at a low (~20%) percentage. Again, it may be that a small focus of trisomic tissue happened to be in the path of the CVS sampling needle, and the sample was aspirated while the needle was at this very spot. (This case is an example of “fetal-placental mosaicism,” as illustrated in Figure 22–4.) NIPT circumvents this question, in that effectively the whole placenta is sampled (Van Opstal et al. 2020) but with the counterpoint that low-level mosaicism detectable at CVS may be missed at NIPT (Lund et al. 2024). Discordance between results of CVS and post-natal sampling of the placenta can extend to “micro-level” imbalances. Lund et al. (2021) followed up five cases of mosaicism for a copy number variant detected at CVS. Abnormalities were confirmed in post-natal placental biopsies in four of the five cases, but in two of these, the placental biopsy findings were more complex than in the original CVS. Thus, CVS mosaicism for CNVs may be unreliable in predicting fetal CNVs, and in this setting, counseling should focus on the option of amniocentesis, rather than the CNV region involved. Different Trisomies Certain CVS trisomies are more or less likely to reflect the same trisomy in the fetus, and the pattern and distribution of the cell lines are also indicative. Trisomy 21 mosaicism on CVS is the most likely to represent a true fetal trisomy 21, whether in the non-mosaic or mosaic state. A risk applies also with trisomies 8, 9, 12, 13, 15, 18, and 20. On the other hand, CPM or fetoplacental discrepancy for trisomies 2, 3, 5, 7, 10, 11, 14, 16, 17, and 22 was never, in the EUCROMIC series, confirmed at fetal or post-natal studies. In some trisomies, a true fetal mosaicism may exist, but at such a low level that there might be no discernible effect upon the phenotype. Klein et al. (1994) reported such a case, a child born of a pregnancy in which trisomy 8 was observed in 81% of CVS cultured cells, 0% of amniotic fluid cells, and in 60% of a placental biopsy at delivery: The child had 4% and 1% mosaicism in blood at 2 and 7 months of age and 0% on a skin fibroblast study, and was normal in appearance, growth, and developmental progress at age 30 months. Of course, any fetal morphologic defect shown on ultrasonography would indicate the very substantial probability of a major degree of true fetal mosaicism, and in that case the choice of termination is appropriately offered. Prognosis The child subsequently born provides the direct evidence of a harm, or not, due to CPM. Amor et al. (2006) undertook a detailed post-natal follow-up, from ages 4 to 11 years, of 36 children from a “CPM pregnancy” and compared their outcomes according to a

672  REPRODUCTIVE CYTOGENETICS number of criteria with a control group of 195 children having had a normal chromosome result from prenatal diagnosis. The mosaicisms included trisomies 2, 7, 8, 12, 13, 17, 18, 20, X aneuploidies, markers, and one translocation. The children from the CPM pregnancies did just as well in terms of general health, development, behavior, and intrauterine growth, as did the control group. Only in respect of post-natal growth was there a small difference in favor of the control group; their mean percentiles for height and weight were 64.0 and 66.4, and those of the CPM children were 51.6 and 56.8 (this might have been an effect of subtly compromised placental function, but equally may have been random). These authors did note a statistically significant increase—which does not necessarily equate to biological significance—in CPM children being perceived by their mothers as “more active,” and they were suitably cautious about this observation. There have been a very few examples of suspected CPM at prenatal diagnosis, with a normal follow-up amniocentesis, that have been followed by the birth of a child with the same mosaicism (Stetten et al. 2004). In these, the mosaicism was in retrospect clearly not confined to the placenta but was in fact a true fetal-placental mosaicism. The direct application of DNA-based technologies without the need for culture provides an alternative window through which mosaic results can be interpreted. We identified a girl with a mosaic isodicentric chromosome 9 (idic(9)) involving the region 9p24.3q21.32, originally ascertained through NIPT but with diagnosis not confirmed until after birth (Lee et al. in press 2025). Amniocentesis and SNP microarray analysis on uncultured amniotic fluid confirmed the same relative copy number gain in ~20% of cells, but conventional G-banded analysis on cultured amniocytes showed a normal 46,XX karyotype in all 73 cells examined from 22 colonies. Post-natal SNP array on saliva, buccal and blood samples confirmed the 9p24.3q22.32 copy number gain, yet G-banded karyotype identified just one cell out of 105 examined carrying a supernumerary isodicentric chromosome 9. Presumably there was selection bias against the unstable isodicentric chromosome in culture. At age two years the girl had reduced head circumference and mildly delayed speech, but was otherwise healthy. Notwithstanding these rare examples, we return to the point: a normal amniocentesis is almost always followed by the birth of a chromosomally normal baby. Uniparental Disomy and Confined Placental Mosaicism A specific concern when CPM for a trisomy is diagnosed relates to uniparental disomy (UPD) (Kotzot 2008b). This is an issue in those trisomies involving an imprintable chromosome (namely 6, 7, 11,9 14, 15, 16, and 20). The embryo may “correct” by post-zygotic loss of the additional chromosome, while the placenta remains partly or wholly trisomic. The incidence of UPD in the setting of a mosaic CVS result is ~2% (Malvestiti et al. 2015). In the particular case of trisomy 15 on CVS followed by 46,N at amniocentesis, the Prader-Willi/Angelman methylation test can be applied. UPDs of the other chromosomes seem mostly to be without phenotypic effect per se, excepting the unlikely 9 Note that the mosaic paternal UPD(11p) causing Beckwith-Wiedemann syndrome is always mitotic in origin and non-mosaic paternal UPD11, which might arise rarely via trisomy rescue, is embryonic lethal.

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  673 possibility of isozygosity for a recessive gene. Where the UPD concerned is associated, or possibly associated, with a major clinical phenotype (and see Chapter 19), prenatal testing for these UPDs (UPD(6)pat, UPD(7)mat, UPD(11)pat, UPD(14)mat and pat, UPD(15)mat and pat, UPD(16)mat, and UPD(20)mat) is justified.10 Irrespective of imprinting, there remains also the question of a small residual trisomic cell line in the fetus, potentially contributing to an abnormal phenotype. Effect Upon Placental Function If a cytogenetically abnormal cell line is confined to the placenta, does this have any implication for placental function? A global statement cannot be made: Some trisomies may matter, and others not, and the fraction of placenta carrying trisomic tissue is likely an important variable. But for several trisomies at least, a placenta that is in part trisomic apparently retains a sufficient, or nearly sufficient, level of function, and mostly the (46,N) fetus is satisfactorily supported. Grati et al. (2020) analyzed 181 pregnancies and 757 controls and concluded that, other than for trisomy 16, the risk of an adverse outcome, for pregnancies carrying a CPM, is low. A contrary view comes from Eggenhuizen et al. (2021), who reviewed 328 cases of CPM from the literature, and while noting the existing of publication bias, judged that pregnancies with CPM involving chromosomes 2, 3, 7, 13, 15, 16 or 22 are at higher risk for fetal growth restriction, premature birth, and possibly structural fetal abnormalities, and particularly so for CPM trisomy 16. But they saw no such risk for CPM of chromosomes 9, 10, 12, 18 and 20. Thomsen et al. (2024) reviewed data from more than 90,000 CVS studies from Denmark covering the years 1983-2021, and identified 646 cases of mosaicism for an autosomal trisomy, of which 13% were true fetal mosaicism (TFM), and 87% CPM. Logically enough, a higher level of mosaicism was associated with a higher risk of TFM. With CPM for trisomies 2, 7, 8, 10, 12, 16, 18 and 20, there was an increased risk for low birth weight and preterm birth. Since the introduction of combined first-trimester screening (cFTS) in 2004, placental mosaicism has been more frequently detected, and fetal involvement with adverse outcome more often observed, suggesting that cFTS selects11 for a population at higher risk for both CPM and TFM. A quite different question is mosaicism with “placental mesenchymal dysplasia,” in which there is a normal and a uniparental cell line (p. 567). FALSE-NEGATIVE RESULTS FROM CHORIONIC VILLUS SAMPLING False-negative results are very rare, and more so since many laboratories no longer use direct or short-term CVS culture. False negatives are presumed to have arisen due to an early post-zygotic event, such that a normal cell line is generated in the extra-embryonic 10 If an SNP microarray has been used for karyotyping, evidence for UPD can be sought from the SNP profile; but it is important to note that some instances of UPD (specifically, meiosis I error where there has not been recombination) are not associated with stretches of homozygosity on SNP array. 11 Mosaicism may affect placental function, and cFTS is based upon placental function, as measured by HCG and PAPP-A levels.

674  REPRODUCTIVE CYTOGENETICS tissue from a basically abnormal conceptus; or, an abnormal cell line can arise from a normal conception, and this cell line then contributing to formation of the embryo (this latter scenario documented especially in the acrocentric isochromosome; Riegel et al. 2006). The largest formal series to address this question is due to van den Berg et al. (2006). These workers reviewed nearly 2,500 prenatal diagnoses from their own service, and comprehensively assessed the literature. In their own material, they had no false negatives. From the literature, most false negatives have been seen in the setting of a normal short-term culture, and then either an abnormal long-term result12 or, if no further testing done, an abnormal pregnancy outcome. This highlights a relative instability of the cytotrophoblast karyotype, with a tendency, as the most usual scenario in this context, to lose the additional chromosome from an initially trisomic conception. From long-term CVS culture, true false negatives numbered only in single figures, and several of these were likely due to maternal-cell contamination. Thus, practically all of the time, a normal long-term CVS result means that the baby will be chromosomally normal. Fluorescence in situ hybridization (FISH), applied to direct uncultured CVS, may be chosen to enable a more timely diagnosis (a faster “turnaround time,” in the laboratory jargon), and particularly in the circumstance of an ultrasound anomaly having been seen. This can target the common aneuploidies, which account for ~65% of all chromosome abnormalities. In one large series (Feldman et al. 2000), 115 direct CVS were analyzed by interphase FISH, from pregnancies in which 100 had a minor fetal anomaly by ultrasound, and 15 had a major anomaly. All of the FISH results were confirmed by routine cytogenetics, with no false positives or false negatives compared to the results after culturing. Although the authors did not separate the chromosome abnormalities found in CVS versus amniotic fluid, overall, they found aneuploidies by FISH in 11% of samples, with another 4% of cases having chromosome abnormalities by analysis of cultured cells that had shown a normal FISH result. Thus, the common aneuploidies are highly likely to be identified by uncultured, interphase FISH, but when the result is normal, routine karyotyping or microarray is still necessary to detect other abnormalities. RARE AUTOSOMAL TRISOMIES DETECTED AT NIPT NIPT has been widely implemented for the detection of trisomies 21, 18 and 13, as well as sex chromosome aneuploidies. Using a genome-wide approach, NIPT can also detect trisomies for other autosomes, collectively referred to as “Rare Autosomal Trisomies” (RATs), but this screening is controversial due to uncertainty about clinical relevance (Dungan et al. 2023). RATs are found in about 1 in 600 pregnancies in the general population, but more commonly in the setting of ultrasound abnormalities or advanced maternal age. The frequency and distribution of RATs detected by NIPT follows a similar patter to the detection of the same trisomies at CVS (Figure 22–5), consistent with cfDNA and direct-preparation CVS being of the same cell lineage. Comparing detection of RATs through NIPT and CVS, Benn et al. (2019) observed that RATs were seen slightly more frequently in CVS trophoblast samples (0.41%) compared to NIPT samples (0.32%), but that confirmation of trisomy at amniocentesis was 12 Or an abnormal result from a simultaneous amniocentesis, typically done in the context of abnormal fetal ultrasonography.

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  675 more likely when the RAT was ascertained through NIPT13 (9.3%) compared to through CVS (3.0%). The authors note that some of these differences are likely due to differences in referral criteria, and emphasize the need for caution when extrapolating outcome data from CVS to NIPT, and vice versa. Follow-up amniocentesis will confirm a RAT in ~10% of cases (Acreman et al. 2023; Lannoo et al. 2024), but similarly to CVS, the risk of fetal trisomy varies greatly according to the chromosome involved. Higher risks of fetal mosaic trisomy are observed for chromosomes 2, 9, 12, 15, 16 and 22, whereas for other chromosomes the risk is low, and approaching zero (Table 22–4). Lower risks also apply to pregnancies ascertained by population-based screening, compared to those with ultrasound abnormalities or other high-risk features. Even for RATs that are confirmed at amniocentesis, a wide range of clinical outcomes are possible. As with detection of CPM at CVS, the detection of a RAT at NIPT is associated with a range of pregnancy complications, including pre-eclampsia, preterm birth, intrauterine growth retardation, congenital anomalies, and perinatal death (Lannoo et al. 2024; Van Prooyen Schuurman et al. 2022). These results are well established for CPM of trisomy 16, but less clearly so for other chromosomes, given the smaller numbers and biases in ascertainment and reporting. Nonetheless, after excluding trisomy 16 from their analysis, Van Prooyen Schuurman et al. found that an increased risk of pre-eclampsia, preterm birth, and low birth weight persisted. A risk for UPD applies for trisomies of meiotic origin, with the highest risk applying to trisomy 15. AMNIOTIC FLUID CELL CULTURE AND MOSAICISM A mitotic error in the epiblast may produce mosaicism of both embryonic and amniotic tissue. A mitotic error in extra-embryonic epithelium causes mosaicism confined 13 Bearing in mind that NIPT samples the whole placenta. Figure 22–5.  Frequency of RATs detected by CVS and by NIPT. Source: Drawn from data in L Lannoo et al., Rare autosomal trisomies detected by non-invasive prenatal testing: an overview of current knowledge, Eur J Hum Genet 30:1323–1330, 2022 676  REPRODUCTIVE CYTOGENETICS to the amniotic membrane. An in vitro cell division defect causes pseudomosaicism. Separating confined placental mosaicism and pseudomosaicism from true mosaicism is critical, but not necessarily straightforward. Amniocentesis may be a first-tier investigation; or it may follow an abnormal NIPT result, in order to confirm or exclude the finding. In the latter case, almost always (96%), SNP analysis based upon uncultured amniocytes is concordant with the interpretation on cultured cells. Thus, SNP array may, of itself, suffice. The distinction is, in the first instance, based upon the number of abnormal cells seen, and whether one or more than one presumptive abnormal clone exists, according to the three levels I–III set out above. Level I mosaicism is seen in 2.5%–7% of amniocenteses, level II in 0.7%–1.1%, and level III in ~2 per 1,000 amniotic fluids (Wilson et al. 1989). Once the laboratory studies are completed, the cytogeneticist will provide an opinion about the level of mosaicism, taking into account technical aspects of the cultures. There is generally no point, and indeed it could be counterproductive, to report level I mosaicism. The only exception would be a single cell of a clinically relevant trisomy, and if the laboratory could not perform sufficient analysis, because of limited sample, to exclude substantial mosaicism. Some level II mosaicism and all level III mosaicism do, however, require to be conveyed to the patient, carefully and clearly interpreted. Level II Mosaicism at Amniocentesis Level II mosaicism reflects a true fetal chromosomal abnormality in only 1% or less of cases (Worton and Stern 1984; Ledbetter et al. 1992; Fryburg et al. 1993; Liou et al. 1993). The nature of the “mosaic chromosome” is important. If it is one that has been recorded, in life, in the non-mosaic trisomic state, or in the mosaic state, the level of concern is higher. This includes, for example, mosaic trisomies 8, 9, 13, 14, 15, 18, and Table 22–4.  Aspects of Rare Autosomal Trisomies detected at NIPT CHROMOSOME FREQUENCY (%) MAIN ORIGIN RISK OF FETAL TRISOMY UPD TESTING ALL STUDIES (%) POPULATION-BASED STUDIES (%) 2 0.006 mitotic 36 17 not indicated 3 0.012 mitotic 0 0 not indicated 7 0.073 mitotic 1 0 low risk 8 0.022 mitotic 6 8 not indicated 9 0.011 variable 39 20 not indicated 12 0.002 variable 25 17 not indicated 14 0.007 meiotic 6 0 low risk 15 0.020 meiotic 35 6 high risk 16 0.028 meiotic 17 16 not indicated 20 0.018 variable 0 0 low risk 22 0.019 meiotic 40 13 not indicated Source: L Lannoo et al., Rare autosomal trisomies detected by noninvasive prenatal testing: an overview of current knowledge, Eur J Hum Genet 30:1323–1330. 2022.

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  677 21 and mosaic isochromosomes 5p, 9p, 12p, and 18p. While true mosaicism for many of the other trisomies has been observed in the malformed fetus in a pregnancy advancing well into the third trimester or in an abnormal liveborn child, these cases are so rare that a level II amniotic fluid mosaicism is still more likely due to artifact rather than a true significant fetal mosaicism. High-resolution ultrasonography provides helpful information in this context. If further cytogenetic investigation is judged desirable—and it often is—repeat amniocentesis for interphase FISH analysis is the procedure of choice, with probe choice according to the chromosome in question. A large number of cells can be analyzed, and quickly. Fetal blood sampling, formerly the mainstay, is rarely used nowadays. (It is to be noted that not all mosaicism is necessarily present in blood, and for example fetal blood sampling only infrequently, if ever, detected a mosaic cell line in trisomy 5, 12, or 20, or i(12p); Berghella et al. 1998; Chiesa et al. 1998.) Strictly speaking, no amount of investigation could ever completely exclude the possibility of a true mosaicism of the fetus, albeit the distribution of the abnormal cell line may be rather limited and quite possibly of unimportant phenotypic consequence. We have seen, for example, a case of level III 47,XX,+13/46,XX mosaicism at CVS, followed by the demonstration of very low-level mosaicism at amniocentesis (1/28 colonies trisomic) and fetal blood sampling (1/400 cells trisomic). At birth, a cord blood sample from the baby showed 47,XX,+13 in 1 out of 150 cells; 2/32 cells were trisomic in amnion and 1/30 and 3/30 in two placental villus biopsies (Figure 22–6). It only needed the colony from one amniocyte not to have been analyzable, or one lymphocyte to have been passed over at each blood sampling, for the true state in the baby to have gone unrecognized. The child was reviewed at age 13 years: She was an above-average student and unremarkable on clinical examination; on analysis of 400 cells (blood and buccal cell), none showed trisomy 13 (MB Delatycki, personal communication, 2009). Rare Figure 22–6.  Level II Mosaicism for Trisomy 13. Notes: Repeat CVS showed a high level (66%) of trisomy 13 mosaicism. In the setting of normal ultrasonography, subsequent studies were done, and showed a very low level mosaicism, and a normal child was born. At age 13, no trisomy cells were seen (see text). Source: From the case in M Delatycki et al., Trisomy 13 mosaicism at prenatal diagnosis: dilemmas in interpretation, Prenat Diagn 18:45–50, 1998. Courtesy MB Delatycki and MD Pertile. 678  REPRODUCTIVE CYTOGENETICS similar examples exist to disquiet the counselor (Terzoli et al. 1990; Vockley et al. 1991), but a sense of perspective is to be kept: For each autosome, only the tiniest number of level II mosaicisms (zero for most chromosomes) have turned out to reflect, in fact, a recognized true mosaicism of the fetus. Level III Mosaicism at Amniocentesis Hsu and Benn re-evaluated the issues in 1999, and they have set forth useful guidelines. These are presented in detail in Table 22–5. While every autosome has now had Table 22–5.  Amniocyte Pseudomosaicism/Mosaicism Guidelines for Work-up for the Elucidation of Possible Amniocyte Pseudomosaicism/Mosaicism FLASK METHOD IN SITU METHOD A.  Indications for Extensive Work-up (1)  Autosomal trisomy involving a chromosome 21, 18, 13; or 2, 5, 8, 9, 21, 18, 13; or 2, 5, 8, 9, 12, 14, 15, 16, 20, 22 (SC, MC) (1)  Autosomal trisomy involving a chromosome 12, 14, 15, 16, 20, 22 (SCo, MCo) (2)  Unbalanced structural rearrangement (MC) (2)  Unbalanced structural rearrangement (MCo) (3)  Marker chromosome (MC) (3)  Marker chromosome (MCo) B.  Indications for Moderate Work-up (4)  Extra sex chromosome (SC, MC) (4)  Extra sex chromosome (SCo, MCo) (5)  Autosomal trisomy involving a chromosome 1, 3, 4, 6, 7, 10, 11, 17, 19 (SC, MC) (5)  Autosomal trisomy involving a chromosome 1, 3, 4, 6, 7, 10, 11, 17, 19 (SCo, MCo) (6)  45,X (MC) (6)  45,X (SCo, MCo) (7)  Monosomy (other than 45,X) (MC) (7)  Monosomy (other than 45,X) (SCo, MCo) (8)  Marker chromosome (SC) (8)  Marker chromosome (SCo) (9)  Balanced structural rea (MC) (9)  Balanced structural rea (MCo) C.  Standard, No Additional Work-up (10)  45,X (SC) (10)  Unbalanced structural rea (SCo) (11)  Unbalanced structural rea (SC) (11)  Balanced structural rea (SCo) (12)  Balanced structural rea (SC) (12)  Break at centromere with loss of one arm (SCo) (13)  Break at centromere with loss of one arm (SC) (13)  All single-cell abnormalities Notes: Criteria for extensive (A), moderate (B), and standard (C) work-up: A. Forty cells (20 cells from each of two flasks, excluding those cells analyzed from the culture with the initial observation of abnormality), or 24 colonies (excluding those colonies analyzed from the vessel with the initial observation). B. Twenty cells (from the flask without the initial observation), or 12 colonies (from vessels without the initial observation). C. Twenty cells (10 from each of two independent cultures), or 15 colonies (from at least two independent vessels). MC = multiple cells (single flask); MCo = multiple colonies (single dish); Rea = rearrangement; SC = single cell (single flask); SCo = single colony (single dish). Source: From LY Hsu and PA Benn, Revised guidelines for the diagnosis of mosaicism in amniocytes, Prenat Diagn 19:1081–1082, 1999.

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  679 a mention as a mosaic trisomy at prenatal or post-natal diagnosis, some are very rare, and others are of questionable significance. Some reported associations may not necessarily have been causal. Hsu and Benn propose the stringent requirement that, before embarking upon an extensive work-up, there be in the literature, for the particular chromosome, “two, or more, well-documented independent reports of confirmed amniocyte mosaicism with abnormal pregnancy outcomes.” The most extensive data treating the question are published in two reports from a collaboration of a number of American and Canadian laboratories: Hsu et al. (1997) with respect to the rare trisomies, and Wallerstein et al. (2000) on trisomies 13, 18, 20, and 21. We make much use of this material in the commentaries later, and every prenatal diagnosis laboratory will want to have a copy of these papers readily at hand. Ultrasonography provides useful adjunctive evidence, but apparent normality cannot be taken as a guarantee. Studies for uniparental disomy may need to be considered in the case of mosaicism for chromosomes known to be subject to imprinting. Further modifications to these guidelines can be anticipated, as new data come to hand. It is wise to attempt confirmation of a diagnosis of mosaicism, either on multiple fetal samples following pregnancy termination, or on blood and placenta in an infant. A post-termination chromosome study that did not confirm the abnormality could cause parents great distress, particularly when no malformations are seen on fetal examination; and it needs deciding with them beforehand whether they wish to have the results. An unconfirmed abnormality could be misleading in a twin pregnancy in which the diagnostic sample had come from a vanishing abnormal twin, but the post-termination tissue had come from the normal co-twin (Griffiths et al. 1996). Fejgin et al. (1997) refer to the “hopeful possibility” of mosaicism as a comfort to parents, with the post-termination tissue having sampled only the normal cell line. It is true that even multiple tissue sampling cannot be taken as having ruled out mosaicism, and a diagnosis of “apparent phenotypic normality” in a fetus still leaves open that a functional brain defect could have come to pass. Microarray Analysis of Uncultured Amniocytes Microarray analysis of uncultured amniocentesis samples detects mosaicism about as often as does karyotyping; however, differences between the two methodologies mean that results are not always concordant. First, the detection of abnormal cell lines by microarray is limited to those present at fractions above 5%–10%, whereas karyotyping allows, at least in principle, the analysis of tens or even hundreds of cells. Second, selection pressure in culture may result in either an underestimate or overestimate of the true proportion of aneuploid cells. Third, karyotyping is superior to microarray for resolving some sex chromosome mosaics, particularly where there was a mixture of monosomic and trisomic cells (e.g. 45,X/47,XXX mosaics). For most autosomal abnormalities, microarray on uncultured cells appears to be the optimal technique. Hao et al. (2020) compared microarray (on uncultured cells) and karyotype on 2,091 amniocentesis samples, and detected mosaicism in 13 cases (0.6%). Mosaic trisomy for an autosome was seen in seven samples, all of which were detected by microarray, but only four by karyotype. The percentage of autosomal aneuploid mosaicism on microarray was consistently higher than seen on karyotype, suggesting selection against the aneuploid cells in culture. In our experience, when the level of mosaicism in uncultured amniocytes exceeds 20%, the result is likely to be confirmed post-natally, at

680  REPRODUCTIVE CYTOGENETICS a similar percentage, in a saliva sample of the baby. In contrast, when the mosaicism in amniocytes is less than 20%, the result more likely represents CPM. In the specific setting of follow-up of increased-risk NIPT results, Donze et al. (2024) analyzed concordance between microarray (uncultured cells) and karyotype in 204 amniocentesis samples, and found eight cases discordant for mosaicism. Five of these had mosaicism only in cultured cells, comprising low-level mosaicism for trisomies 8, 13, 16 and 22, and an isodicentric chromosome 9. By the authors’ assessment, only two of these findings were clinically relevant (i.e., 1% of the 204 samples), and they argue that testing of amniocentesis samples after abnormal NIPT should be limited to SNP array on uncultured cells. Three cases had mosaicism detected only in uncultured cells, comprising mosaicism for trisomies 2, 9, and 22, of which at least one (trisomy 2) was clinically relevant. To conclude from these observations: following an abnormal NIPT result, SNP analysis based upon uncultured amniocytes, in order to confirm or exclude the finding, will almost always (96%) be concordant with the interpretation on cultured cells. Thus, SNP array may, of itself, suffice. TWIN PREGNANCY, DISCORDANT KARYOTYPES Discordant karyotypes may be observed in the setting of either dizygous (DZ) or monozygous (MZ) twinning (Figure 20–24). Selective termination of the abnormal twin is an option, although one that cannot ensure that the normal twin will be unharmed. Because twins may share circulations, the process of termination of the affected twin may lead to exsanguination of the normal one (Lewi et al. 2006). In MZ twins in which “trisomic rescue” has been the basis of one twin being karyotypically normal, a risk for UPD applies, and this would be a concern in the case of a “UPD-vulnerable” chromosome. SPECIFIC ABNORMALITIES In this section, we outline the risks for phenotypic abnormality of specific chromosomal abnormalities detected at prenatal diagnosis. Since the available data often derive from terminated pregnancies in which only major anomalies are recognized, many of these risk figures may be underestimates. For example, a trisomy 21 fetus may appear normal to the inexpert eye on external observation, but we naturally assume intellectual disability would have resulted; and the same may apply to several other chromosomal imbalances. New knowledge will continue to accumulate, and what appears here is printed on paper, not in stone. The small number of aneuploidies that may exist in the true non-mosaic state are noted first. In the mosaic list, almost every chromosome is represented, although in the CVS section we do include also a few instances of non-mosaicism. Autosomal Trisomy, Non-Mosaic Trisomies 13 and 18 (and extremely rarely 8, 9, 14, and 22) are practically the only non-mosaic autosomal trisomies besides + 21 that are detected at amniocentesis. Others occur, but virtually all miscarry before the usual time of amniocentesis. CVS, on the CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  681 other hand, is done at a gestational stage when a number of trisomies destined to abort have not yet done so. Trisomies 13 and 18 There is a high likelihood of spontaneous abortion after amniocentesis, and it is somewhat higher if detection is at CVS. Won et al. (2005) determined a rate of fetal death in utero following amniocentesis-proven trisomy 18 of 32%, while Yamanaka et al. (2006) arrived at a figure of 27%; from a somewhat different viewpoint, Cavadino and Morris (2017) showed a 30% survival rate to term of prenatally diagnosed trisomy 18. Data for survival of a liveborn child are due to Vendola et al. (2010) and are set out in Table 22–6. Further data from Lakovschek et al. (2011), following prenatal diagnosis and a continued pregnancy, give live birth rates of 33% for trisomy 13 and 13% for trisomy 18, with deaths typically within days. The outlook for a liveborn child is so bleak, with inevitable profound mental deficiency, barely a vestige of social response in those few who survive beyond early infancy (a capacity for communication limited to “body movement,” vocalization, and facial expression; Liang et al. 2015), and typically a requirement for full nursing care, that termination is sought by the majority of couples. Nevertheless, views can vary, and some infants are having heart surgery or other intervention (Weaver et al. 2021; Kosiv et al. 2023; Song et al. 2023). Others may elect to continue the pregnancy but chose post-natal comfort care provided by trained palliative care teams (Bierer et al. 2024). Carey (2012) emphasizes the need to bring the parents fully into the making of any decisions, and Chung et al. (2017) describe a family’s positive and rewarding experience of looking after their infant with trisomy 13, with the help of pediatric palliative care staff; a view to be acknowledged, in the spirit of recognizing the plurality of opinion that is seen in the world. Trisomy 21 We expect most readers will have an expert appreciation of the predicted DS phenotype, but we do recommend Bull’s (2020) review as a full and balanced account. Guidelines exist for the medical care of children and adolescents (Bull et al. 2022) and adults (Tsou et al. 2020). Marteau et al. (1994) appraised the views of obstetricians, geneticists, and genetic nurses to the prenatal diagnosis of DS, and recorded some remarkable differences. Table 22–6.  Trisomy Survival Probabilities of Survival to 1 Week, 1 Month, and 1 Year, for Liveborn Infants with Trisomies 13, 18, and 21 1 WEEK 1 MONTH 1 YEAR Trisomy 13 0.42 0.20 0.03 Trisomy 18 0.52 0.30 0.03 Trisomy 21 0.98 0.95 Source: From C Vendola et al., Survival of Texas infants born with trisomies 21, 18, and 13, Am J Med Genet 152A:360–366, 2010.

682  REPRODUCTIVE CYTOGENETICS The respective proportions who would counsel non-directively (see definitions above) were 32%, 57%, and 94%, and the respective proportions counseling directively in favor of termination were 62%, 40%, and 7%. Approximately 6% of obstetricians would counsel directively in favor of continuing the pregnancy, but practically no geneticists or genetic nurses would do so. Having received a positive 47,+21 result, what personal factors influence the parental decision? A 7½ year study, over 1989–1997, reports the views of 145 women in Michigan (Kramer et al. 1998). Most (87%) elected to terminate the pregnancy. The decision did not differ according to parity, race, religion, nor, perhaps surprisingly, with the presence or absence of ultrasonographic abnormality. Older mothers, those who had already had children, and those whose prenatal procedure was done at an earlier gestation were more likely to choose termination. With modern management, the survival of DS individuals approaches that of the general population (95% surviving at 1 year; Table 22–6), but comorbidities become prevalent with age, raising questions of practicalities of care as the parents themselves age (Glasson et al. 2002). On the other hand, if fetal ultrasonography shows a heart malformation and/or growth retardation, there is a risk of prenatal demise or post-natal death (Wessels et al. 2003). In the study of Won et al. (2005), the rate of fetal death in utero after an amniocentesis diagnosis of trisomy 21, in 392 women who decided to continue the pregnancy, was 10%. Other Autosomal Trisomy (In Particular 9, 10, 20, 22) Never (almost) do other non-mosaic true fetal trisomies survive through to a stage of extrauterine viability. Schinzel (2001) catalogs no more than about two dozen each of trisomy 9 and trisomy 22, and barely one or two of possible trisomies 7, 8, and 14, with survival through to the third trimester. Miscarriage is nigh on inevitable, usually within the 8- to 14-week gestation range. An example is trisomy 10, of which very rare examples as non-mosaics at prenatal diagnosis are known, but survival to term is seen only in mosaic forms, and these infants are very abnormal (Hahnemann et al. 2005). If natural abortion has not already occurred by the time the chromosomal result is received, and if there is supportive evidence otherwise, such as ultrasonographic defect, for there being a true fetal involvement, termination is appropriately offered. Schwendemann et al. (2009) reviewed the sonographic findings of fetuses with non-mosaic trisomy 9; heart defects and central nervous system malformations were the most frequent anomalies seen. Concerning non-mosaic trisomy 20, Stein et al. (2008) record five cases at prenatal diagnosis, the indication in each being the discovery of an anatomical abnormality on ultrasound, with early deaths in all except their own case, a child who in fact turned out to be mosaic on analysis of post-natal tissues; Morales et al. (2010) and Maeda et al. (2015) publish similar cases. Of all the other non-mosaic trisomies, it is only with trisomy 22 that there might be, very rarely, the possibility of a term pregnancy, and in some, limited post-natal survival (Phung et al. 2023). Autosomal Trisomy, Mosaic DETECTION OF MOSAICISM AT CHORIONIC VILLUS CULTURE The substantial majority of mosaic trisomies for a single autosome are followed by a normal result at amniocentesis and at karyotyping of the child (or of the aborted fetus) CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  683 (Table 22–7).14 In the EUCROMIC study, there were 192 gestations with mosaic or non-mosaic fetoplacental discrepancy for an autosomal trisomy, and in 84% CPM was confirmed. For mosaic trisomy 8, 9, 12, 15, and 20, only a single case of each was subsequently identified with aneuploidy in the fetus/child, compared with two each for chromosomes 13 and 18, and as many as seven for trisomy 21 (Hahnemann and Vejerslev 1997). With respect to mosaicism for multiple (>1) autosomal trisomies, the presence 14 As noted above, a few instances of apparent non-mosaic trisomy at CVS are also included here, on the assumption that—in the circumstance of a semblance of normal fetal development—a true fetal non-mosaic trisomy for that chromosome would in fact be improbable. We assume in these cases, rather, that this would be either “fetal mosaicism, non-mosaic placenta” or “fetal-placental mosaicism” with the sampling needle missing the karyotypically normal tissue, each of these scenarios being demonstrated in Figure 22–4. Table 22–7.  Mosaicism Outcomes Fetoplacental Mosaicism for Different Chromosomes and Percentage Confirmed at Amniocentesis or at Birth, Based on Data from 243,566 CVS Samples CHROMOSOME ALL FETOPLACENTAL MOSAICS (N) TRUE FETAL MOSAICS (N) TRUE FETAL MOSAICS (%) 1 1 0 0 2 164 2 1.2 3 51 0 0 4 8 1 13 5 11 1 9.1 6 7 0 0 7 161 0 0 8 102 13 13 9 51 4 7.8 10 37 1 2.7 11 10 0 0 12 34 4 12 13 80 6 7.5 14 20 1 5.0 15 48 2 4.2 16 90 10 11 17 6 0 0 18 140 21 15 19 2 0 0 20 62 8 13 21 119 47 39 22 32 12 19 Notes: Within these datasets there is considerable heterogeneity in terms of method of ascertainment, analysis techniques, percentage of mosaicism, and clinical follow-up, and these factors may influence interpretation of results. For rarer trisomies (those with fewer than 20 cases at CVS), confidence intervals for percentage with true fetal mosaicism are wide, and these percentages should be interpreted with caution. Source: Based on data from Hahnemann and Vejerslev (1997) (92,246 samples), Malvestiti et al. (2015) (60,347 samples) and Thomsen et al. (2024) (90,973 samples)

684  REPRODUCTIVE CYTOGENETICS or absence of a normal cell line is the key point: A fetal involvement is practically never seen if there is a normal cell line, and practically always seen if there is no normal cell line (MD Pertile, personal communication, 2002). The general rule that Robinson et al. (1997) advance is this: CVS mosaicism due to a preconceptual (meiotic) error conveys a significant risk for fetal trisomy/UPD, whereas a postconceptual (somatic) error is usually innocuous. Mosaic trisomies 15, 16, and 22 are mostly in the former category, for example, whereas trisomies 3 and 7 are typically of mitotic origin, and mosaic trisomy 2 can be either. The possibility remains for a residual effect due to (1) undetected (and presumably low-level) mosaic trisomy of the fetus, (2) uniparental disomy of the fetus, and (3) placental dysfunction as a consequence of a regional placental trisomy. The risks for these scenarios differ for different chromosomes, and we provide specific commentaries following. Mosaic Trisomy 1 at Chorionic Villus Sampling. One case was recorded by Malvestiti et al. (2015), and it was not confirmed at amniocentesis. Mosaic Trisomy 2 at Chorionic Villus Sampling and at NIPT. Two broad groups of trisomy 2 mosaicism are recognized (Robinson et al. 1997; Albrecht et al. 2001; Wolstenholme et al. 2001b). In the first, a majority (~90% of the total) are characterized by a small fraction of trisomic cells, and usually seen only in cultured mesenchymal cells. The pregnancy outcome is typically normal; in the series of Sago et al. (1997), 11/ 11 pregnancies had a normal outcome; and of the 77 cases reported by Malvestiti et al. (2015), none was confirmed at amniocentesis. It may be that these cases reflect a post-zygotic generation of the trisomic lineage in a restricted region of chorionic tissue in an otherwise normal conceptus, and this small trisomic region has no discernible effect upon placental function. The second, minority group is presumed due to trisomy “correction” in a 47,+2 conceptus, from either a maternal or a paternal error. The level of trisomic cells in the CVS is typically high, up to 100%, with the involvement of both trophoblast and the mesenchymal core. In the series of Thomsen et al. (2024), 2/76 cases were confirmed at amniocentesis, and these had levels of mosaicism at CVS, as detected by microarray, of 80% and 69%. The placenta being substantially trisomic apparently Figure 22–7.  Mosaicism Outcomes at Prenatal Diagnosis. Notes: This graph is derived from the data in Table 22-7. The increased risks associated with a prenatal diagnosis of mosaicism for trisomies 18 and 21, in particular, are evident. CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  685 compromises its function, and intrauterine growth retardation (IUGR) is a frequent observation, with a poor outcome (Roberts et al. 2003). One case is known of the very severe defect of “body stalk syndrome” associated with mosaic trisomy 2 at CVS (Smrcek et al. 2003). An atypical case in Chen et al. (2023c) attests to the variable levels of mosaicism between different tissues and the complexity of their interpretation. Following an initial trisomy 2 result at NIPT, the percentage of trisomic cells were, in chronological order: 11 week CVS, 100%; 20 week amniocentesis, 37% in cultured cells; 24 week amniocentesis, 28% FISH, 40% uncultured microarray, 0% cultured cells; post-natal cord blood, 0%; umbilical cord, 0%; placenta, 23%; and buccal swab at 6 weeks, 9%. The child, who was born with growth retardation and also had maternal UPD(2), had normal development at one year. Mosaic Trisomy 3 at Chorionic Villus Sampling. In the EUCROMIC study, of 10 cases with trisomy 3 at either short- or long-term culture, none proved to have fetal involvement, apart from one child with a normal karyotype at amniocentesis, and a very low 1/100 trisomy 3 count on blood as a newborn (Hahnemann and Vejerslev 1997). Similarly, none of the 27 cases reported by Malvestiti et al. (2015) or 14 cases of Thomsen et al. (2024) had fetal involvement. Zaslav et al. (2004) identified a case of trisomy 3, in which the initial amniocentesis showed 47,XX,+3[8]‌/46,XX[27], and a repeat procedure 47,XX,+3[1]/46,XX[18]. Fetal blood was normal in 100 cells. The baby was apparently normal at birth, except for IUGR. FISH of placenta demonstrated the trisomy 3; thus, it would likely have been found by CVS, had this been done. Mosaic Trisomy 4 at Chorionic Villus Sampling. This is very rare: there were none in the EUCROMIC study and none of the four cases reported by Thomsen et al. (2024) were substantiated at amniocentesis. Malvestiti et al. (2015) did confirm one of their four cases by amniocentesis, and the pregnancy was terminated, the fetus having abnormalities of the heart, colon, and spleen. Two cases are noted in Kuchinka et al. (2001). In one case, subsequent amniocentesis gave a 46,XX karyotype, but fetal demise occurred at 30 weeks, associated with considerable growth retardation (although no externally observable malformations). UPD(4)mat was demonstrated. It remains open whether the unfortunate outcome was the consequence of the UPD or due to placental trisomy. The second case did not proceed to amniocentesis; biparental disomy 4 was demonstrated in the child. Follow-up at 1 year raised some reservation: Although development was judged to be normal, growth indices were low, including a head circumference at about the third centile (in other words, borderline microcephaly). To complicate the story, mother and child carried a balanced t(10;15). The case in Marion et al. (1990), and followed up several years later (Brady et al. 2005), was actually an amniocentesis diagnosis, but since post-natal studies showed trisomy 4 mosaicism in the placenta, it is not unreasonable to consider this is a potential CVS example. The child, at age 14 years, had a low-normal intellect and some physical body asymmetries (of hand, ear, and breast). Blood was 46,XX; skin biopsy confirmed constitutional +4 mosaicism. In another case, Gentile et al. (2005)

686  REPRODUCTIVE CYTOGENETICS identified mosaic trisomy 4 by amniocentesis (22% of cells). The pregnancy presented at 22 weeks gestation with micrognathia, abnormal brain development, and spinal and cardiac defects. At termination, trisomy 4 mosaicism was confirmed in placental and fetal skin cultured cells; the cord blood karyotype was normal. Molecular analysis excluded uniparental disomy of chromosome 4, but showed that the trisomy 4 was of maternal meiotic origin. An extraordinary example of mosaic trisomy 4 at CVS with double mosaicism for trisomies 4 and 6 at amniocentesis, 47,XY +4/47,XY,+6/46,XY, is described in Wieczorek et al. (2003). The double trisomy mosaicism was confirmed on skin (but not blood) karyotyping in the child, whose phenotype, while certainly abnormal, was less so than might have been anticipated. Mosaic Trisomy 5 at Chorionic Villus Sampling. Thomsen et al. (2024) recorded five cases, of which one was confirmed at amniocentesis. That pregnancy was investigated because of an increased-risk first-trimester screen, and trisomy 5 was identified in 2/25 cells at CVS and 2/100 cells at follow-up amniocentesis. The pregnancy continued to livebirth, but further follow-up was not provided. Only three cases are recorded in the EUCROMIC study; in none was a fetal trisomy subsequently shown (Hahnemann and Vejerslev 1997). Another three cases are listed in Malvestiti et al. (2015), again without fetal involvement. Chen et al. (2020a) report the diagnosis of CPM for trisomy 5 at CVS with follow-up amniocentesis showing maternal uniparental disomy for chromosome 5. The baby, born to a 45-year-old mother, also had Down syndrome, with no additional phenotypic features attributable to the CPM or UPD(5)mat. Mosaic Trisomy 6 at Chorionic Villus Sampling. Very few examples are known. Malvestiti et al. (2015) record three cases, and Thomsen et al. (2024) four cases, none confirmed on amniocentesis. A detailed case report is given in Miller et al. (2001). A young mother had a 12-week CVS because of ultrasonographic anomalies (crown-rump length at 11-week size, nuchal translucency), with 60% of cells in short-term culture and 22% of long-term cells showing 47,XX,+6. Amniocentesis was declined. An abnormal heart rate at 25 weeks led to emergency delivery, and a growth-retarded infant with numerous anomalies was born. Her blood karyotype was normal, but trisomy 6 cells were found in placenta and umbilical cord samples. Growth indices remained below the third centile. On follow-up at age 2¾ years, neurodevelopmental progress was “near normal.” Skin taken at the time of surgery showed 3% (hand) and 20% (inguinal area) mosaicism. The case in Kitamura et al. (2024) was ascertained via the detection of 47,XY,+6 in 20% of cells at amniocentesis, but post-natal analysis of the placenta showed trisomy 6 in 4% of cells. The baby was healthy at age 1 year, and the post-natal karyotype normal. Mosaic Trisomy 7 at Chorionic Villus Sampling and at NIPT. This is typically a mitotically arising mosaicism. Kalousek et al. (1996) looked at 14 cases of trisomy 7 CVS mosaicism and feto-placental discordance, the fraction of trisomy ranging from 7% to 88% in 11, and with three showing 100%. Twelve infants were judged normal, and in the eight of these tested, all proved to have biparental inheritance. Two infants were of low birth weight, and the one of these tested was the only of the series with UPD and a meiotic origin; the cultured CVS in this case was 100% trisomic. In a case we studied, mentioned also above, three post-natal placental samples karyotyped normal, and one with trisomy 7; the baby was normal (Watt et al. 1991).

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  687 In the EUCROMIC study, of 32 cases with trisomy at either or both short- and long-term culture (including three with non-mosaic trisomy), none proved to have fetal involvement (Hahnemann and Vejerslev 1997). Similarly, of the 74 cases recorded by Malvestiti et al. (2015) and 55 cases by Thomsen et al. (2024), none were confirmed at amniocentesis. The conclusion is that the great majority of trisomy 7 mosaicism detected at CVS arises mitotically, does not imply a risk for UPD, is confined to the placenta, does not obviously compromise intrauterine growth, and is associated with the birth of a normal baby. Following detection of CPM for trisomy 7 at NIPT, however, an increased risk of low birth weight has been noted (Van Prooyen Schuurman et al. 2022). Mosaic Trisomy 8 at Chorionic Villus Sampling and at NIPT. A well-recognized but highly variable post-natal phenotype (Warkany syndrome) accompanies trisomy 8 mosaicism, which may also include an increased risk for cancer (Altiner et al. 2016). Fetal defects are recorded on pathology examination (Jay et al. 1999). Typically, the mosaicism is the consequence of a post-zygotic nondisjunction in an initially 46,N conceptus (Danesino et al. 1998).15 In the three-case series of Hahnemann and Vejerslev, Thomsen et al. and Malvestiti et al., there were a total of 102 cases, of which 13 were confirmed at amniocentesis. Thomsen et al. (2021) reviewed 60 cases of trisomy 8 mosaicism at CVS from the literature, and added 37 of their own. Fetal involvement was confirmed in 18 out of 97 pregnancies (19%), including eight pregnancies initially interpreted to be CPM following a normal amniocentesis, but subsequently seen as true fetal mosaicism. This circumstance calls to mind the scenario proposed by Wolstenholme (1996): True fetal mosaicism is typically associated with low levels of trisomy 8 in trophoblast cells (short-term CVS culture), high levels in extra-embryonic mesoderm (long-term CVS culture), and low levels in amniocytes and fetal blood cells. Hu et al. (2022) report a case of trisomy 8 mosaicism initially detected at NIPT, and confirmed by the detection of 10% trisomy 8 cells at amniocentesis. Hydronephrosis and spinal irregularity were detected antenatally, and at three-year follow-up the child had language delay and facial asymmetry. A high risk NIPT result for trisomy 8 led to a diagnosis of paternal UPD(8) at amniocentesis in the case of Yu et al. (2022), with the child reported to be developing normally at age 3 years. In this case, the initial abnormality was likely a paternal meiotic error. Mosaic Trisomy 9 at Chorionic Villus Sampling and at NIPT. Malvestiti et al. (2015) record 18 cases of mosaic trisomy 9 at CVS, none of which were confirmed at amniocentesis. In contrast, Thomsen et al. (2024) confirmed three of their 24 cases at amniocentesis. Saura et al. (1995) presented seven cases of trisomy 9, five of which gave a non-mosaic result, with the outcomes being abnormal in most. In the EUCROMIC study, of nine cases with trisomy 9 at either or both short- and long-term culture (including three with non-mosaic trisomy in one or both cultures), one proved to have fetal involvement (Hahnemann and Vejerslev 1997). This single case had non-mosaic trisomy at both short- and long-term culture. Slater et al. (2000) report a case of trisomy 9 non-mosaic at CVS, but with level II mosaicism found at amniocentesis, with only two cells 47,XX,+9. At fetal blood sampling, all 85 cells analyzed were 46,XX. Molecular studies revealed UPD(9)mat. A blood sample from the newborn infant had 15 Trisomy 8 mosaicism of meiotic origin is sometimes found in miscarriage samples, but when detected post-natally, appears to cause a phenotype distinct from Warkany syndrome. Valind et al. (2014) detected trisomy 8 cells of meiotic origin in a Wilms tumor sample of an otherwise healthy 2 year old boy, and Baidas et al. (2014) made a similar finding in the bone marrow of an otherwise healthy 32 year old woman who presented with recurrent aphthous ulcers that were refractory to treatment.

688  REPRODUCTIVE CYTOGENETICS the karyotype 47,XX,+9[4]‌/46,XX[50]; upon retrospective review of the fetal blood, 3 of 102 cells were trisomic 9. Minor anomalies were noted in the child, who had been followed up to age 1 year. It is probable that this phenotype reflected a minor degree of residual trisomy in the child’s soma. Li et al. (2022a) reported 16 NIPT cases in which trisomy 9 was suspected. Follow-up amniocentesis identified two cases of non-mosaic trisomy 9, three cases of mosaic trisomy 9, two cases with UPD(9), and nine with normal karyotype (presumed CPM). Mosaic Trisomy 10 at Chorionic Villus Sampling. From the three-case series of Hahnemann and Vejerslev, Thomsen et al., and Malvestiti et al., only 1/37 cases proved to be of true fetal mosaicism. In this case of Thomsen et al. (2024), ascertained via first-trimester combined screening, trisomy 10 was identified at CVS in 4/15 cells but follow-up amniocentesis was normal. However, after pregnancy termination, there was trisomy 10 in 20/30 fetal cells. Jones et al. (1995) presented one case in which direct culture showed trisomy 10 mosaicism, while long-term culture and amniocentesis were 46,XY, but with UPD(10)mat. The child subsequently born was apparently normal. Mosaic Trisomy 11 at Chorionic Villus Sampling. From the three-case series of Hahnemann and Vejerslev, Thomsen et al., and Malvestiti et al., there were a total of 10 cases, none confirmed at amniocentesis. Mosaic Trisomy 12 at Chorionic Villus Sampling. Malvestiti et al. (2015) report 12 cases, one of which was confirmed at amniocentesis. Hahnemann and Vejerslev (1997) and Sikkema-Raddatz et al. (1999) describe three cases, two of which involved a true fetal mosaicism. Of these latter, one fetus appeared grossly normal post-termination, and one infant was abnormal. Thomsen et al. (2024) reported 2/21 cases in their series as representing true fetal mosaicism, although both were atypical. In the first, trisomy 12 was detected in 54/57 cells at CVS, was absent from amniocentesis sample, but was confirmed in fetal tissue following stillbirth. In the second case, CVS showed non-mosaic trisomy 12, but follow-up amniocentesis was normal apart from FISH studies, which found one cell with trisomy 12 and one with monosomy 12 out of 100 cells examined. The pregnancy continued to live birth, but further follow-up was not given. Mosaic Trisomy 13 at Chorionic Villus Sampling and at NIPT. A high level of trisomy 13 cells may well reflect significant mosaicism of the fetus. Ultrasonography and amniocentesis may clarify the picture. Mosaic trisomy 13 may present a very abnormal post-natal phenotype (Delatycki and Gardner 1997). A difficulty arises in the case of very low-level (a percent or so) mosaicism, in which case it is possible the child could be normal (Delatycki et al. 1998). In the EUCROMIC study, of 15 cases with trisomy 13 at either or both short- and long-term culture (including four with non-mosaic trisomy in one culture), two (14%) proved to have fetal involvement (Hahnemann and Vejerslev 1997). Malvestiti et al. (2015) record 42 cases of trisomy 13 at CVS, of which only one was confirmed at amniocentesis. Thomsen et al. (2024) confirmed three out of 23 cases at amniocentesis, however it is notable that in one of these, post-termination karyotype on fetal tissues was normal, despite amniocentesis having identified trisomy 13 in 36/ 60 cells. If trisomy 13 is indicated at NIPT, and in the setting of normal ultrasonography, amniocentesis may be a better option than CVS for confirmation/exclusion, given how often it is that that CPM will have been the cause (Rogers and Mardy 2023). Mosaic Trisomy 14 at Chorionic Villus Sampling. Malvestiti et al. (2015) note 14 cases, none of which were confirmed at amniocentesis. Similarly, three examples of 47,+14/46,N were recorded in the EUCROMIC study, none showing fetal trisomy

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  689 (Hahnemann and Vejerslev 1997). Thomsen et al. report five cases, with one being confirmed at amniocentesis and in a fetal sample post pregnancy termination. The case in Ralph et al. (1999) proceeded to follow-up amniocentesis, which also showed the mosaicism, and in addition maternal uniparental isodisomy 14. Fetal death in utero supervened; no morphological abnormality was identified. Other prenatal cases (or retrospectively diagnosed, on post-natal placental biopsy) with the syndrome of maternal UPD(14), following “correction” of trisomy, are known (Morichon- Towner et al. 2001; Engel and Antonarakis 2002). Growth restriction, and possibly dysmorphism and minor anomalies, may be associated. Mosaic Trisomy 15 at Chorionic Villus Sampling. Malvestiti et al. (2015) report 26 cases, 16 of which were confined to the cytotrophoblast, and none of which were confirmed at amniocentesis. Thomsen et al. (2024) identified true fetal mosaicism in one out of 11 cases of trisomy 15 mosaicism at CVS. In two EUCROMIC studies, cases of trisomy 15 CPM were examined, in which direct and long-term cultures had been done (EUCROMIC 1999; Hahnemann and Vejerslev 1997). Few of these cases demonstrated true fetal mosaicism. Most often, the trisomy 15, mosaic or non-mosaic, was found in trophoblast and villus mesenchyme, and rarely in the fetus. The authors theorize that chromosome 15 (and 16) participates more often in trisomy rescue. This would increase the potential risk for UPD 15, and more often than not, the trisomy 15 would be meiotic in origin. The recommendation is that amniocentesis be offered to all patients with a CVS diagnosis of mosaic or full trisomy 15, prudently to check for the possibilities of UPD, and true fetal mosaicism. Mosaic Trisomy 16 at Chorionic Villus Sampling and at NIPT. Almost all CPM for trisomy 16 (which may present as mosaic or non-mosaic trisomy 16 on CVS) is due to a maternal meiosis I nondisjunction. The important follow-up investigation is an amniocentesis. If this gives a normal karyotype, CPM is very probable. IUGR with a low birth weight is common, but catch-up growth is typically observed. Malformation may be present, but usually these are minor or surgically reparable birth defects. Normal intellectual capacity is well recorded (Langlois et al. 2006; Neiswanger et al. 2006). However, a more severe phenotype may result, and ultrasonography may indicate this likelihood, the complications including major malformation, and fetal death in utero. The degree of severity may relate to the presence or absence of fetal trisomy (which may not be revealed until post-natal tissue sampling), or, in the case of CPM, to the existence of uniparental or biparental disomy of the fetus, although this latter point is controversial (Scheuvens et al. 2017). From the combined data of Hahnemann and Vejerslev, Thomsen et al., and Malvestiti et al., 90 cases of trisomy 16 at CVS were recorded, 10 of which (11%) were confirmed at amniocentesis. In a prospective cohort study Grau Madsen et al. (2018) confirmed a diagnosis of true trisomy 16 mosaicism in 2/28 (7%) of pregnancies. Of those with CPM, pregnancy complications were seen in two-thirds, ranging from small for gestational age (58%), malformations (20%), preterm delivery (32%), to intrauterine demise or fetal death (12%), and these were associated with a higher fraction of trisomic cells. One-third were born at term with normal birth weight and without malformations. Following detection of CPM trisomy 16 by NIPT, van Prooyen Schuurman et al. (2022) identified risks of 27% for pre-eclampsia, 26% for premature birth, and 53% for birth weight <3rd centile. If the mosaicism is seen at subsequent amniocentesis, the prognosis is less favorable (see below); and yet we have seen a normal post-natal outcome on 2½-year follow-up in this setting (Coman et al. 2010a).

690  REPRODUCTIVE CYTOGENETICS Mosaic Trisomy 18 at Chorionic Villus Sampling and at NIPT. In the combined datasets of Hahnemann and Vejerslev, Thomsen et al., and Malvestiti et al., there were 140 cases of non-mosaic trisomy 18 at CVS, and 21 (15%) of these proved to have fetal involvement. Harrison et al. (1993) studied placental karyotypes from pregnancies in which trisomy 18 had been diagnosed, whether at pre- or post-natal diagnosis, and mosaicism was detected in seven of 12, involving the trophoblast. This supports the view that mosaic trisomy 18 at CVS may on occasion reflect a full trisomy of the fetus (and also leads to the conclusion that fetal survival may, in the context of this particular trisomy, be enhanced if there is a diploid placental fraction). At NIPT, a quarter of trisomy 18 results are in the mosaic range, with the majority of these confirmed at amniocentesis (Rafalko et al. 2021). Mosaic Trisomy 19 at Chorionic Villus Sampling. Malvestiti et al. (2015) report two cases, one confined to cytotrophoblast and one to mesenchyme; neither was confirmed at amniocentesis. Mosaic Trisomy 20 at Chorionic Villus Sampling and at NIPT. Mosaic trisomy 20 is one of the most common mosaicisms detected at amniocentesis (see below), but observation at CVS is less frequent. Malvestiti et al. (2015) list 29 cases, of which three (10%) were confirmed at amniocentesis. In the EUCROMIC study, of 12 cases with trisomy 20 at either short-term or at both short- and long-term culture (including four with non-mosaic trisomy in short-term culture), one (8%) proved to have fetal involvement (Hahnemann and Vejerslev 1997). Thomsen et al. identified 21 cases in their series, with four (19%) confirmed at amniocentesis. Six cases were reported by Robinson et al. (2005), two of which had compromised outcomes: developmental delay in one, and growth retardation and stillbirth in the other; follow-up amniocentesis had shown trisomy at levels of 11% and 59%, respectively. Steinberg Warren et al. (2001) described a child, followed to age 8¾ years, normal other than hypomelanosis of Ito, from a pregnancy with a non-mosaic trisomy 20 diagnosed at CVS; culture from a subsequent amniocentesis failed. As the pigmentary skin sign in the child indicated, he was in fact mosaic, and proven so to be on skin culture; and this mosaicism would probably have been revealed had the amniocentesis been successful. We may presume the likely circumstance as depicted in “Fetal mosaicism, non-mosaic placenta” in Figure 22–4. If there is CPM trisomy 20 at NIPT, an increased risk of pre-eclampsia (10%) applies (Van Prooyen Schuurman et al. 2022). Mosaic Trisomy 21 at Chorionic Villus Sampling and at NIPT. Chromosome 21 naturally commands special attention. In the EUCROMIC study, of 22 cases with trisomy 21 at either or both short- and long-term culture (including eight with non-mosaic trisomy in one culture), in nine (40%) there transpired to be a fetal involvement (Hahnemann and Vejerslev 1997). Similar proportions were confirmed at amniocentesis by Malvestiti et al. (2015), with 19/56 (34%), and Thomsen et al. (2024), with 19/41 (46%). Beverstock et al. (1998) report a “near false-negative” finding of mosaic trisomy 21, in which trisomic cells were observed in long-term CVS culture, and then, at follow-up amniocentesis, in only one culture. True mosaic trisomy was proven at fetal blood sampling, and at tissue culture post abortion. At NIPT, results in the mosaic range are unusual, and most of these are confirmed as trisomy 21 at CVS or amniocentesis (Rafalko et al. 2021). Mosaic Trisomy 22 at Chorionic Villus Sampling. Malvestiti et al. (2015) report 11 cases of mosaic trisomy in cytotrophoblast and/or mesenchyme, none of which were seen at amniocentesis. In contrast, Thomsen et al. (2024) describe 18 cases, six of which

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  691 were confirmed at amniocentesis. When trisomy 22 is confirmed at amniocentesis, fetal defect is typically associated, but the degree may vary considerably. Wolstenholme et al. (2001a) described their own case of non-mosaic trisomy 22 diagnosed at direct and cultured CVS, with 47,XX,+22/46,XX mosaicism subsequently shown at amniocentesis (3/ 60 cells +22) and fetal skin biopsy (6/170 cells +22). Fairly subtle fetal dysmorphism was noted post-termination, and multiple tissue samplings showed mostly low but consistent trisomy mosaicism: 1% trisomic cells in skin, muscle, blood, and kidney; 3% in lung; 5% in liver; and 21% in spinal cord. It is probable that neurological compromise would have eventuated, quite likely of severe degree, had a child been born. Wolstenholme et al. reviewed 11 other cases of mosaic and non-mosaic trisomy 22, the mosaicisms mostly being of high percentages at CVS, and (in the six cases proceeding to amniocentesis) low percentages at amniocentesis. Of nine cases in which post-termination samplings were done, six showed mosaicism in at least some tissues (see also the case of De Pater et al. 1997, mentioned above in the section on “Level III Mosaicism”). In the three cases with 0% trisomy at fetal sampling, all had manifested severe IUGR. This may have been the consequence of functional insufficiency of the trisomic 22 placenta; there is also the point that occult fetal trisomy can never be excluded. Bryan et al. (2002) studied a child born of a pregnancy with a non-mosaic 47,XY,+ 22 karyotype having been shown at CVS. There was IUGR, but the child apparently showed post-natal catch-up. He typed 46,XY on peripheral blood (with biparental disomy) and was phenotypically normal, except for hypospadias. Detection of Mosaicism at Amniotic Fluid Cell Culture Considering the three major trisomies, Hsu et al. (1992) have determined that mosaicism for chromosomes 13, 18, and 21 very frequently predicts fetal abnormality, in half or more of cases. As for rare trisomies, Hsu et al. (1997) undertook a wide survey, based on the experiences of a number of American and Canadian laboratories and drawing on previous reports in the literature; the reader wishing full detail will need to refer to the original document. Some mosaic trisomies are associated with a high risk for phenotypic abnormality in the fetus or term infant, with figures of >60% for mosaic trisomies 2, 16, and 22, whereas trisomies 7, 8, and 17 are toward the lower end of the scale (<20%). Ultrasonography has a role in the assessment: Most cases in which the mosaicism involves the fetus to a substantial degree will display morphologic/growth abnormality. Nevertheless, normal ultrasonography cannot allow firm reassurance. Some mosaic states might cause structural defects too subtle to be discerned at fetal imaging, and yet be associated in the child with considerable, possibly severe, functional neurological compromise. In chromosomes known to be subject to parent-of-origin imprinting, uniparental disomy needs also to be factored in to the assessment. Comments on individual trisomies follow (for the most part, we are here considering only abnormalities seen on amniocentesis as the first invasive prenatal procedure, and not follow-up amniocenteses done to clarify an abnormal CVS result). The best estimate of the percentage of cells with an autosomal aneuploidy is achieved by microarray analysis of DNA extracted from uncultured amniocytes. These are rare observations, and in the survey of Forabosco et al. (2009), the most frequent mosaic autosomal trisomies recognized at amniocentesis were, in descending order: trisomies 21 (1 in 4,000 amniocenteses), 20 (1 in 5,000), 13 and 18 (1 in 22,000), 9 (1 in 30,000), and, each at 1 in 90,000, trisomies 2, 6, 7, 8, 15, and 17. 692  REPRODUCTIVE CYTOGENETICS Mosaic Trisomy 2 at Amniocentesis. In Hsu et al.’s (1997) survey, trisomy 2 conveyed the highest risk of any of the “rare trisomic” autosomes for an abnormal outcome, namely 90%, with a variable pattern of major defects. Wang et al. (2020b) provide a clear demonstration (Figure 22–8). It is probable that mosaic trisomy 2 detected at amniocentesis would be in the same group as the high-level mosaic CVS case (see above). Tuğ et al. (2017) report a case with severe fetal malformation initially identified on serum screening and ultrasonography, with amniocyte fractions 47,XX,+2[12]/46,XX[73]. A trisomic line in the fetus/child may take some diligence to find. Sago et al. (1997) reported a case in which there was level III mosaicism with trisomy 2 cells present in 27% of amniocytes (and biparental disomy). The child was severely abnormal, and while blood and skin karyotyped as 46,XY, 4% of liver cells were 47,+2. Similarly, Chen et al. (2013c) report a fetus with cardiac defect, polydactyly, and dysmorphism, and who had mosaic trisomy 2 at amniocentesis. Trisomy 2 cells had arisen due to maternal meiosis I error and comprised 100% of the placenta, 50% of the amniotic membrane, and 10% of the fetal liver, but fetal lung, skin, and blood showed a normal karyotype. Mosaic Trisomy 3 at Amniocentesis. Only two cases were identified in Hsu et al.’s (1997) review, in one of which the child had multiple malformations, with the mosaicism confirmed on skin fibroblast culture. The child in the other case was normal. Marked intrauterine, and subsequently post-natal, growth restriction was the prominent feature in the cases in Zaslav et al. (2004) and Sheath et al. (2010); in both, development in early infancy was judged to be normal. Tang et al. (2017) provide further evidence that low-level mosaicism for trisomy 3 can be associated with a normal outcome. In their case, ascertained in the context of an increased-risk maternal serum screen and normal ultrasound, the prenatal detection of trisomy 3 in 10% of amniocytes and 3/200 cord blood lymphocytes was followed by the birth of a healthy baby girl, with a normal blood karyotype. Mosaic Trisomy 4 at Amniocentesis. Hoyas et al. (2021) review ten cases, with normal and abnormal outcomes both, approximately proportional to the fraction of trisomic cells. Four cases had developed normally by age 1 year, with the percentage of + 4 cells between 8% and 31%; normal ultrasonography was a positive pointer. The case Figure 22–8.  Trisomy 2 Mosaicism at Amniocentesis. Notes: These images are of interphase FISH on uncultured amniocytes, two cells shown here. The red spots reflect a probe recognizing chromosome 2 centromere; the green is for chromosome 22, as a control. The cell at left, with three red spots, is trisomic; the cell at right, with two red spots, disomic. Second trimester NIPT results had indicated the chromosome 2 quantum to be increased. Follow-up amniocentesis showed mos 47,XY,+2[8]‌/46,XY[19]; ultrasonography at 24 weeks revealed fetal IUGR. Source: From T Wang et al., Prenatal diagnosis of mosaic trisomy 2 and literature review, Mol Cytogenet 13:36, 2020. Courtesy L Guo, and with the permission of Springer Nature.

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  693 in Hoyas et al. is unique, of monozygous twins with discordant karyotypes: Twin A had multiple major structural anomalies and karyotyped non-mosaic trisomy 4 at amniocentesis, whereas the karyotype of Twin B was normal 46,XX. Trisomy 4 cells have never been detected in the peripheral blood of affected individuals. Mosaic Trisomy 5 at Amniocentesis. Hsu et al. (1997) recorded five cases. In one, a high level of trisomic cells (80%) was associated nevertheless with a phenotypically and karyotypically normal infant. In two, the child was abnormal, both showing the mosaicism on post-natal study. Brown et al. (2009) and Bonasoni et al. (2022) describe cases in which a fetal heart defect had been the factor leading to an amniocentesis, with respective proportions of +5 cells of 50% and 19%. In both, autopsy analysis showed trisomy cells in several tissues, but not blood. A further case in Reittinger et al. (2017) may have reflected a “correction” of trisomy, as the severely malformed infant eventually born had, on peripheral blood analysis, uniparental disomy 5 (and no trisomy). Mosaic Trisomy 6 at Amniocentesis. Hsu et al. (1997) recorded three cases, each with the same low-level (6%) trisomy in amniocytes, and each with a normal outcome. Reports are on record of the diagnosis following recognition of fetal defects at ultrasonography, the defects ranging from minor to severe (Wallerstein et al. 2002; Wegner et al. 2004; Destree et al. 2005). One case of fetal death in utero at 23 weeks was associated with 48% trisomy cells on fetal skin analysis (Cockwell et al. 2006). Chen et al. (2006b) report a case with low-level (3%–10%) mosaicism, with normal fetal blood karyotype, biparental inheritance, in which the parents chose termination, and the trisomy was absent on cultured fetal tissue. They suggest low-level trisomy 6 mosaicism may be a benign finding. A similar case, but with 20% mosaicism at amniocentesis, and where the parents chose to continue, is provided by Kitamura et al. (2024): the baby appeared healthy at birth, and post-natal cytogenetic investigations were normal. Mosaic Trisomy 7 at Amniocentesis. Hsu et al. (1997) recorded eight cases, with fractions of trisomic cells ranging from 5% to 48%. Only one resulted in the birth of a phenotypically abnormal child, but low-level 47,XY,+7/46,XY mosaicism was confirmed in two phenotypically normal children on foreskin analysis. Warburton (2002) emphasized that this relatively low-risk assessment is the appropriate one to offer, and she notes also that UPD(7), while unlikely, may be worth testing for. A low-level mosaic case, taken to termination with pathology study and multiple fetal karyotyping, with entirely normal findings, led Chen et al. (2005a) to agree with the view that optimistic advice may usually be appropriate. A microarray-era example is that of Chen et al. (2024e) who identified 24% mosaicism for trisomy 7 in DNA from uncultured amniocytes, but a 46,XY karyotype in cultured amniocytes. At birth, the baby was assessed to be phenotypically normal, with no evidence of trisomy 7 in peripheral blood. Other cases with a post-amniocentesis abnormal outcome (and in which ascertainment was necessarily biased) include the following: Mosaicism was verified post-natally on skin fibroblast analysis in the child reported in Kivirikko et al. (2002), in whom fetal blood sampling and mid-trimester ultrasonography had been normal; there was facial asymmetry and mild dysmorphism along with rather impressive hypomelanosis of Ito, while mental development was “considered to be within normal limits,” although no detailed assessment had been done. The fraction of trisomic colonies in the 47,XX,+7/ 46,XX case of Bilimoria and Rothenberg (2003) was rather high, at 41%, and in addition uniparental heterodisomy was shown in the 46,XX line; the pregnancy had come to attention because of an increased-risk maternal serum screen. On a neonatal blood sample, all cells were 46,XX, while on the contrary, all placental cells analyzed were

694  REPRODUCTIVE CYTOGENETICS trisomic. The child was small for dates and had some minor anomalies. These authors mention an anecdote of a trisomy 7 mosaic woman “graduating from college and getting married.” Petit et al. (2012) describe a case of IUGR leading to amniocentesis, which was interpreted at the time as normal. The child proved to be delayed in growth and development, and displayed hypomelanosis of Ito. Blood analysis showed 46,XY with maternal UPD(7). The skin fibroblast karyotype, however, was 47,XY,+7/46,XY; and restudy of stored amniocytes from long-term culture showed mosaic trisomy 7. Mosaic Trisomy 8 at Amniocentesis. Counseling is difficult, and advice must be cautious. An observation of trisomy 8 in amniocytes predicts a distinct probability, but by no means a certainty, of the clinical syndrome, namely Warkany syndrome. It is not possible to put a good figure on the level of risk. Vice versa, a true fetal mosaicism may not necessarily be detected at amniocentesis. Lund et al. (2020) estimate that true fetal mosaicism is missed by amniocentesis in 11% of cases, and emphasize the importance of careful ultrasound, and the possible role of cordocentesis, as well as noting that microarray analysis of DNA from uncultured cells may inform the interpretation. A finding of apparently normal morphology at fetal examination following termination in some 47,+ 8/46,N pregnancies might be misleading, since the physical component of the clinical syndrome is relatively minor (Hsu et al. 1997). In the series of van Haelst et al. (2001), the two cases of trisomy 8 mosaicism detected at amniocentesis both turned out to be pseudomosaicism. Mosaic Trisomy 9 at Amniocentesis. The risk is high (Chen et al. 2003a). Hsu et al. (1997) recorded data on 25 cases, with pregnancy termination being done in 21. Abnormality was identified in most of the 21, and mosaicism was confirmed in the seven having skin fibroblast studies. In the four pregnancies continuing, one abnormal child was born, with 47,+9/46,N mosaicism on blood karyotyping, the other three pregnancies resulting in apparently normal newborns. An overall figure of 56% applies for the risk that the fetus is abnormal. This high percentage figure is not surprising, and indeed it may well be an underestimate of the risk for functional abnormality in the child (intellect not being assessable in the newborn), considering the well-recorded phenotype of mosaic trisomy 9 in older individuals. A review of the outcomes in surviving individuals is given in Bruns and Campbell (2015). Mosaic Trisomy 10 at Amniocentesis. In one case listed in Daniel et al. (2004), a 47,XX,+10[27]/46,XX[83] karyotype was associated with severe fetal defects, this observation being the basis of the referral for prenatal diagnosis. They were able to ascertain that the cause was a post-zygotic duplication of the maternal homolog. Mosaic Trisomy 11 at Amniocentesis. Of the four reported examples, all have had a normal outcome. One child came from a pregnancy with a 26% fraction of trisomic cells, with 46,N findings on post-natal tissues, and followed through to 1 year of age. Basel-Vanagaite et al. (2006) raise the question that this mosaicism may typically be a benign finding. Mosaic Trisomy 12 at Amniocentesis. This is one of the more frequently described mosaicisms, and often implies a high risk. Hsu et al. (1997) accumulated 23 cases, comprising 12 continuing pregnancies and 11 terminations. In most of those proceeding to fetal or neonatal fibroblast karyotyping, the mosaicism was subsequently confirmed, though most of the fetuses appeared to be normal. It is possible, however, that some subtle physical features, and possibly unsubtle neurological deficit, might have eventuated had these “normal” fetuses been born. The clinical range in the few recorded

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  695 liveborn patients with true trisomy 12 mosaicism is quite variable, from lethality in the newborn period through to an otherwise normal man with Kartagener syndrome being investigated for infertility (DeLozier-Blanchet et al. 2000). Of the 12 continuing pregnancies in Hsu et al. (1997) the outcomes were abnormal in five and grossly normal newborns in seven. Three of these normal infants followed for 5 months to 5 years were all judged to be continuing along normally, and Staals et al. (2003) add another 3-year-old to this list. The proportion of trisomic cells at amniocentesis apparently is not a very helpful guide in prognosis. In one case in Daniel et al. (2004), associated with fetal defect at 18-week termination, the trisomy had resulted from a post-zygotic duplication of one homolog. Mosaic Trisomy 13 at Amniocentesis. The risk for abnormality is very high. A collaboration of 23 American and Canadian laboratories provided data on the outcomes of 25 prenatal diagnoses of 47,+13/46 mosaicism (Wallerstein et al. 2000). Care was taken to exclude cases in which ascertainment had been biased by abnormal ultrasonography. In 21, the pregnancies were terminated. Various abnormalities were identified in 10 of these; the range of percentages of abnormal amniocytes was very wide, 6%–94%, average 58%. No defect was detectable in the remaining 11 aborted fetuses, although the assessment was limited to simple inspection. Four pregnancies proceeded to apparently normal live birth; the percentages of abnormal amniocytes in these were lower, ranging from 5% to 13%. We mentioned above the very low-level mosaicism at a post-CVS follow-up amniocentesis in Delatycki et al. (1998), with a normal outcome. Mosaic Trisomy 14 at Amniocentesis. Chen et al. (2013e) summarized 10 cases from the literature. In three of the 10 cases, the trisomy 14 resulted from an isochromosome 14. Four of the 10 pregnancies continued to term, from which three infants appeared normal, and one had multiple abnormalities resulting in neonatal death. Five pregnancies were terminated, in which four had multiple abnormalities, and one had micrognathia only. One pregnancy ended in intrauterine fetal death at 18 weeks gestation, in the absence of anatomical abnormality. A risk exists for UPD(14) over and above any defect due to the mosaic trisomy per se, and this should be checked. Mosaic Trisomy 15 at Amniocentesis. Trisomy 15 is usually the consequence of a maternal meiosis I nondisjunction. Amniotic fluid mosaicism may well reflect a true mosaicism of the fetus. In Hsu et al. (1997), six of the 11 cases recorded had an abnormal outcome, the risk being greater when the trisomy level was higher (>40%). Zaslav et al. (1998) review seven cases of low-level mosaic trisomy 15 detected at prenatal diagnosis, in each the amniocentesis having been done for advanced maternal age. All seven chose to terminate, and a variety of defects were documented in most but not all. In their own case, the trisomic cell line in the initial amniocyte analysis was at a low level: 47,XX,+ 15[2]‌/46,XX[37]. Fetal tissues were also at low levels (lung 2%–5%, heart 8%–15%, skin 6%–10% on metaphase and interphase analysis, respectively), but the placenta showed 100% trisomy on metaphase analysis and 95% using FISH on interphase cells. These authors also document from the literature four cases of abnormal liveborns with trisomy 15 mosaicism. There is the additional question of UPD(15)mat, the considerable phenotypic consequence of which—that is, Prader-Willi syndrome—may be superadded upon that of a trisomy 15 mosaicism. Chen et al. (2020c) identified discrepancy between a normal karyotype from cultured amniocytes, and a microarray result from uncultured amniocytes that detected 30% trisomy 15. The amniocentesis was repeated eight weeks later with a similar result, but with

696  REPRODUCTIVE CYTOGENETICS percentage of trisomy 15 reduced to 15%.16 Biparental inheritance of chromosome 15 was confirmed, and the pregnancy proceeded to the delivery of a phenotypically normal but growth-retarded infant at 32 weeks of gestation. Trisomy 15 was absent from blood but detected in 2% of buccal mucosa cells, this confirming a low-level mosaicism in the child. Mosaic Trisomy 16 at Amniocentesis. Neiswanger et al. (2006) reviewed the question of trisomy 16 mosaicism diagnosed prenatally, including 36 cases from amniocentesis; and they reported their own findings in three cases in which no prior CVS had been undertaken. Of these three, all had abnormal outcomes: IUGR but with normal cognitive development as judged at 14 months; IUGR and major malformations including cardiac dextroposition; and IUGR with hypoplastic left heart, leading to neonatal death. In their literature review, the figures for complication were as follows: infant death, 33%; prematurity, 64%; IUGR, 69%; physical anomalies, 75%; and just one assessed as a normal outcome, 3% (these figures being considerably worse than for CVS diagnosis). They note that level II mosaicism, in this context, may well reflect a true fetal mosaicism. Thus, the earlier opinion of Hsu et al. (1998) is supported: “Mosaic trisomy 16 detected through amniocentesis is not a benign finding, but associated with a high risk of abnormal outcome, most commonly intrauterine growth retardation, congenital heart defect, developmental delay, and minor anomalies.” Rieubland et al. (2009) diagnosed two cases post-natally, noting a considerable phenotypic difference between the two, one normally grown and developing at age 11 months, but with a severe hypospadias; the other with IUGR, body asymmetry, numerous physical anomalies, and dying at 7 months: yet further illustrating the challenge in offering advice at prenatal detection. Notwithstanding, we have seen an eventual apparently normal outcome, the child assessed at 2½ years of age, albeit that delivery by cesarean section at 36 weeks had been necessitated due to fetal distress with IUGR. Trisomy 16 had been detected at high level on CVS and at amniocentesis, and low-level (8%) post-natally on buccal mucosal cells (Coman et al. 2010a). The question of an influence of UPD is open to argument. Scheuvens et al. (2017) proposed that UPD(16)mat is not of itself pathogenic, and that any associated phenotypic abnormality is actually the consequence of a cryptic trisomy 16 mosaicism; nevertheless, UPD was more frequent in those pregnancies in Neiswanger et al. (2006) with IUGR, and in the infants with anomalies. Yong et al. (2003) tested for UPD in a series of infants from mosaic trisomy 16 pregnancies, and the fraction with UPD(16) mat at 40% was close enough to the one-third expectation from random loss of one chromosome. These infants were more severely affected than those with biparental inheritance of 16. Mosaic trisomy 16 has a particular association with very low levels of pregnancy-associated plasma protein-A (PAPP-A) on first-trimester serum screening. Mosaic Trisomy 17 at Amniocentesis. de Vries et al. (2013) summarized 28 cases from the literature, a proportion at least of which had arisen mitotically. The most common outcome (19 of the 28, 68%) was the birth of a healthy infant without evidence 16 A reduction in trisomic cells over time is consistent with origin being primarily from the amnion, and with a progressive decline in the proportion of cells from the amnion compared to the proportion from the fetus.

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  697 of trisomy 17 in blood and/or fibroblasts. In nine cases the trisomy was confirmed on fibroblasts—six after birth, and three following termination of pregnancy. In seven of the nine cases of true mosaicism, trisomy was not detected in blood, suggesting selection against trisomic cells in this tissue and possibly explaining why so few post-natal cases are reported. In cases of true fetal mosaicism, the clinical phenotype includes cerebellar hypoplasia, ventricular septal defect, scoliosis, growth and intellectual retardation, and body asymmetry (Baltensperger et al. 2016). The longest follow-up is reported in Witters and Fryns (2008), a child at age 36 months who was significantly delayed, with a developmental age of 26 months. And yet a number of normal outcomes are on record, as Abrams et al. (2005) document in their own case, with the child reportedly normal as a 2-year-old, and as they note similarly in a handful of other cases from the literature. They advise that an optimistic view is warranted, if the ultrasonography is normal. This view is supported in Chen et al. (2016b), whose patient had a prenatal diagnosis of 47,XX,+17[4]‌/46,XX[17], confirmed on repeat study with interphase FISH (5/105 cells); a normal child was in due course born, 46,XX on cord blood.17 Chen et al. (2020b) report monozygous twins with discordant karyotypes at amniocentesis: Twin A had the mosaic karyotype 47,XX,+ 17[3]/46,XX[23], whereas Twin B was normal. At 2 year follow-up, the only discernable phenotype was preaxial polydactyly of the right hand in Twin A. Mosaic Trisomy 18 at Amniocentesis. The risk is very high. In the collaboration of Wallerstein et al. (2000), 31 prenatal diagnoses of trisomy 18 mosaicism were available for review. In just over half of these, the abortuses (induced termination or natural abortion) were abnormal. In 11, no defects were discerned at fetal examination. Just three pregnancies came to live birth, and these babies were apparently normal. The percentages of trisomic amniocytes in these three cases ranged from 2% to 20% (mean 9%), compared with 2%–95% (mean 37%) in those with abnormal outcome. A very rare abnormality is 45,X/47,XX,+18 mosaicism, in which the phenotype can vary from fairly mild to severe (Schluth-Bolard et al. 2009; Tyler et al. 2009). Mosaic Trisomy 19 at Amniocentesis. A single case is recorded in Hsu et al. (1997), and in which there was a normal outcome at live birth. Mosaic Trisomy 20 at Amniocentesis. This is one of the most commonly observed mosaic aneuploidies. Trisomy 20 may exist in three forms: as confined placental mosaicism, as placental-fetal mosaicism with an apparently normal phenotype in the child that is subsequently born, or as a fetal mosaicism with phenotypic consequence (Hsu et al. 1991). There may be no dysmorphic features, only some “soft” signs, or rarely an unambiguous facial dysmorphism; a characteristic if subtle syndrome is proposed (Willis et al. 2008). In certain fetal regions in which the trisomy may exist, in particular kidney and gut, the imbalance apparently has no discernible untoward effect, and in fact aneuploid cells may normally be cultured from urinary sediment. (Recognizing that amniotic fluid has a substantial contribution from fetal urine production, presumably some of the “amniotic fluid cells” from which the diagnosis of trisomy 20 had been made may have actually had origin from the fetal urinary tract.) In the collaboration of Wallerstein et al. (2000) comprising 152 diagnoses, 10 (7%) were recorded with an abnormal outcome (six liveborns, four abortuses). There was 17 The observation of 5/90 cells trisomic 17 on uncultured urinary cells from the child may well have reflected an occult constitutional mosaicism but not necessarily so, since a low level of trisomic cells can be a normal finding in this tissue.

698  REPRODUCTIVE CYTOGENETICS correlation with the level of mosaicism: Abnormality was observed in 20% of infants where there had been >50% trisomic cells at amniocentesis, and in 5% of those with <50%. Baty et al. (2001) reviewed 17 cases in which follow-up of the children extended beyond 1 year, of whom 12 (71%) had developed normally. The remaining five had various degrees of speech and motor delay. A more optimistic interpretation comes from James et al. (2002), who tracked down all cases diagnosed at amniocentesis in New Zealand from 1991 to 2001, numbering 13, with follow-up well into childhood for nine of these (the longest to age 10 years). The range of the trisomic fraction of amniocytes was 8%–50%. All were essentially normal, except for one child who had minor anomalies at birth, resolving by 6 months of age, and deformation due to breech delivery may have been the cause, although weight was below the third centile; and in the only case in which termination was chosen, rather subtle (indeed borderline) external fetal anomalies were noted, and cultured tissue showed low-level (skin 2%, kidney 7%) trisomy mosaicism. Baty et al. (2001) followed up two cases with higher fractions of trisomic cells at amniocentesis, 83% and 57% in one, and of 90% in the other, and the children, at ages 9 and 8 years, respectively, were of normal intelligence and of essentially normal morphological appearance. Each did, however, display quite prominent hypomelanosis of Ito, presumably reflecting a fairly widespread distribution of a trisomic 20 lineage, at least in skin. Nevertheless, reservation must remain. Reish et al. (1998) offer the sobering example of a 15-month-old child with considerably delayed gross and fine motor skills and poor language acquisition, who had 54% trisomic cells from a skin biopsy (a normal karyotype on peripheral blood). In the pregnancy, amniocentesis had shown a 45% mosaicism, fetal ultrasonography was normal, and the parents had been “cautiously counseled.” Likewise, Wallerstein et al. (2005) report a child who had seemed normal at birth, and 46,XX on blood, but who went on to manifest a “pervasive developmental disorder.” Trisomy 20 had been present in only 4/63 colonies at amniocentesis; trisomy was further documented in urinary sediment at age 4 years. They comment that “optimism regarding developmental outcome should be tempered with some caution.” Bianca et al. (2008) summarize the issues and advise along these lines: A second CVS or amniocentesis would add little value; fetal blood sampling is not useful, and neither is UPD analysis; the level of mosaicism does not predict outcome. This agrees with the views of some and contradicts others, as noted above; notwithstanding, Chen et al. (2023d) advise that low-level mosaic trisomy 20 at amniocentesis can be a transient and benign condition. Mosaic Trisomy 21 at Amniocentesis. The risk for DS is very high. The collaborative study of Wallerstein et al. (2000) accumulated 96 cases for review. Half had an observably abnormal outcome, with confirmatory cytogenetic study performed in a minority. Most of these were fetuses post-termination with various abnormalities; six were liveborns, five of these with a clinical diagnosis of DS and one with an isolated heart defect. An apparently normal appearance (assessment limited to inspection in 39, autopsy in two) was recorded in 41 aborted fetuses. Among these, 20 were submitted to further cytogenetic analysis (repeat amniocentesis, fetal tissue, fetal blood, placenta), with eight showing 8%–90% trisomic cells, and 12 with 0%. Seven liveborns were normal, two being followed up beyond the newborn period; none had confirmatory karyotyping. The mean amniotic fluid proportion of trisomic cells was 17%, range 6%–31%, in these CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  699 normal children. This compares with a mean of 35% in those with a demonstrably abnormal outcome. But even in the group with the lowest level of amniotic fluid trisomy (3%–10%) half had an abnormal outcome. From the whole material, a risk for phenotypic abnormality of 50% should be seen as a minimum estimate, since subtler defects at fetal or neonatal assessment would have escaped notice, and a potential compromise of intellectual function was of course not assessable. However, the report in Chen et al. (2016c) illustrates that a confirmed true mosaicism in the child, if of low level (2/38 in this case, on post-natal cord blood), can be associated with phenotypic normality, at least as judged at age 7 months; two amniocenteses had returned trisomy 21 fractions of 5/53 and 6/26. Mosaic Trisomy 22 at Amniocentesis. Chen et al. (2019) determined a very high risk for abnormality for 47,+22/46, with 14 out of 20 outcomes being abnormal. The mosaic trisomy 22 levels of the six normal cases ranged from 5% to 28%, which overlapped with the 3.6%–50% levels from the 14 abnormal cases. Growth retardation is the most common ultrasound finding, but a range of congenital anomalies are also seen. Berghella et al. (1998) described a normal fetal blood result following trisomy 22 mosaicism diagnosis at amniocentesis, but fetal skin biopsy showed 47,+22/46, and structural abnormalities were subsequently identified in the aborted fetus. Four cases are noted in the review of Wolstenholme et al. (2001a), these all having followed an initial detection at CVS. Three out of the four showed some degree of normal/trisomy mosaicism at fetal samplings post-termination. Leclercq et al. (2010) record a normal phenotypic outcome in a single case, followed up to age 4 years, albeit that the child showed the mosaicism on skin, in 6% of cells. Three other cases were abnormal at autopsy study (two following fetal death in utero, and one a medical termination). Mosaic Partial Trisomy at Amniocentesis. It is not feasible to list here recorded cases, and each must be judged on its merits. One specific example is worth noting, in that it may represent simply cultural artifact associated with a fragile site. This is mosaicism for a del(10)(q23). Zaslav et al. (2002) document a case of 46,XY,del(10)(q23)[9]‌/46,XY[45] detected at amniocentesis. The phenotypically normal child had the del(10q) in only 3/ 100 blood cells, this culture having been stressed by growth in a low-folate medium. We are aware of a handful of essentially similar case, all involving 10q23, and none resulting in a documented abnormal child. Polyploidy TRIPLOIDY Close to 100% of the time, triploidy aborts spontaneously, but in some cases not until the pregnancy is well advanced; those surviving through to the second or third trimester are usually digynic. In the prenatal diagnosis series of Lakovschek et al. (2011), in which no intervention was taken, no triploid infant was born alive. This being so, the offer of termination is appropriate when triploidy is diagnosed, particularly given the increased risk of maternal complications (Massalska et al. 2020). Cassidy et al. (1977) described the emotional turmoil suffered by the family when a triploid infant, predicted to die immediately, survived for the extraordinary period of 5 months.

700  REPRODUCTIVE CYTOGENETICS The ultrasonographic observations are typically quite obvious, and include severe abnormalities of the brain, heart, and kidneys, limb defects, and oligohydramnios (Massalska et al. 2021). The ultrasonographic appearance of cystic placenta points to diandric triploidy, whereas digynic triploidy is characterized by asymmetric IUGR (p. 642). In general, cfDNA testing does not detect triploidy, but a very low fetal fraction may raise suspicion. Sarno et al. (1993) reported a unique case of complete placental/ fetal discordance with triploidy on CVS and a normal diploid karyotype on amniocentesis and fetal blood sampling, with the birth of a normal baby; such a possibility warrants consideration where triploidy on CVS accompanies an ultrasonographically normal fetus. True triploidy mosaicism is very rare (p. 345). Wegner et al. (2009) report a prenatal diagnosis, the pregnancy ending in fetal death in utero at 25 weeks, with the remarkable mixed-sex karyotype of 46,XX/69,XXY. Numerous abnormalities were revealed at anatomical pathological examination. They were able to demonstrate that the initial conception had been dispermic (one X- and one Y-bearing sperm) and that the 69,XXY lineage had arisen by the delayed incorporation of the Y-bearing male pronucleus into a cell with a 46,XX nucleus; they preferred the expression “mixoploidy” to describe this scenario. In the mosaic case of Mahon et al. (2022), amniocentesis showed 69,XXX[6]‌/ 46,XX[24], the procedure done on the basis of growth restriction. A growth-retarded and dysmorphic female infant was born at term; at age 15 months, she was clinically stable but developmentally delayed. A very rare case is “hypotriploidy” with 68 chromosomes. One case of 68,XX hypotriploidy was diagnosed prenatally, following an ultrasound picture which was similar to that of classic digynic triploidy (Pasquini et al. 2010). TETRAPLOIDY Tetraploidy seen at prenatal diagnosis, in the context of normal ultrasonography, is usually an in vitro cultural artifact, or possibly a vestige from the blastocystic stage of normally occurring trophoblastic tetraploidy (Krieg et al. 2009). In their series of 76,104 CVSs, Grati et al. (2020) identified 72 cases (0.1%) of mosaic tetraploidy: All were confined to the placenta (32 CPM type I, 15 CPM type II, and 25 CPM type III), and there were no adverse perinatal outcomes. Balkan et al. (2012) tell a salutary story, concerning a pregnancy with an increased-risk screen, mild fetal pyelectasis and hyperechogenic bowel on ultrasound, and the amniocentesis showing non-mosaic 92,XXYY; but cordocentesis then demonstrating 46,XY, and a normal child subsequently born. True tetraploidy is very rare, and Teyssier et al. (1997) recorded only 10 cases, two of which had been discovered at amniocentesis; further cases are listed in Stefanova et al. (2010), who describe their own case of a newborn who died at age 30 hours. Ultrasonographic demonstration of growth retardation and enlarged cerebral ventricles may be typical but rather non-specific signs. While tetraploid/diploid mosaicism is almost always a cultural artifact, Edwards et al. (1994), having observed true normal/tetraploid mosaicism in two severely disabled individuals, nevertheless caution that a tetraploid cell line is not absolutely certain to be an innocuous finding. The 2½ year old child in Stefanova et al. (2010) with mosaic tetraploidy was quite abnormal. In a prenatal case, Meiner et al. (1998) showed 92,XXYY/46,XY mosaicism on fetal blood sampling following the diagnosis of non-mosaic 92,XXYY at amniocentesis, in the setting of growth retardation discovered at ultrasonography, and confirmed at subsequent fetal pathology study.

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  701 Structural Rearrangement Structural rearrangements are seen in about 1 in 1,000 cytogenetic prenatal diagnoses (Warburton 1991). It is typically a matter of urgency to do parental chromosome studies, in order to distinguish between a familial or a de novo rearrangement in the fetus. If one parent is discovered to have the same apparently balanced autosomal rearrangement, and in the context of normal ultrasonographic anatomy, there is no firm evidence for an increased risk of fetal abnormality, and many would counsel to the effect of no discernibly increased risk. Sex chromosome rearrangements require separate attention. De Novo Balanced Chromosome Rearrangement A major difficulty posed by the de novo balanced chromosome rearrangement (BCR) seen at prenatal diagnosis is that, at the level of classical cytogenetic analysis, is “apparently balanced,” and when the interpretation at ultrasonography is normal. But even with the highest resolution banding, a submicroscopic abnormality (deletion or duplication, or gene disruption) may still be present (Figure 5–25). However, now that microarray is often a first-line genetic test, the issue may be by-passed: De novo rearrangements with no copy number gain or loss will not be detected, and the test will be reported as normal. When the prenatal diagnosis is made on classical methodology, the starting point is an acknowledgment that precedents are recorded for a de novo BCR having disrupted or compromised a locus, and therefore that the discovery of such a rearrangement at prenatal diagnosis could potentially herald an abnormal child. Of course, these translocations are to be taken seriously. Equally, the truly balanced translocation carrier state (every one of which in the world must have been de novo at some point in the near or distant past) is very familiar, as Chapter 5 attests at length. Very many translocations are indeed balanced, in terms of their functional genetic consequences. Thus, a normal child is very possible, and as the observations have shown, this is the more likely outcome. Following a first-line test on classical cytogenetic analysis, microarray analysis should be offered as a second-tier test. In the study of De Gregori et al. (2007) of 14 prenatal diagnoses of de novo simple translocation, the ultrasonography being normal in 12, all proved to be balanced at the level of array-CGH. However, genomic sequencing around the breakpoints would be needed to detect a true disruption18 of a gene (Ordulu et al. 2016). Nevertheless, we should emphasize the pragmatic observation that most pregnancies with prenatal diagnosis of a de novo inversion or simple reciprocal translocation go on to produce a normal baby. Presumably, these normal cases reflect breakpoints in DNA that does not code for a gene or for a control element (or if a gene is disrupted, its haplo-state is sufficient), and in which there is no concomitant microdeletion. Of course, abnormal ultrasonography dictates a different perspective. Thus, for example, when Price et al. (2005) identified growth and anatomical abnormalities suggestive of Cornelia de Lange syndrome (CdLS), the subsequent finding of a presumed de novo translocation (father not available for testing) 46,XX,t(3;5)(q21;p13) enabled a clear interpretation, the CdLS gene being located at 5p13, and presumably disrupted by the rearrangement. 18 In the post-natal setting, Redin et al. (2017) performed whole genome sequencing on 273 patients with both a BCR and a congenital abnormality, and identified gene disruption in 35%, pathogenic genomic imbalances in 5% and disrupted topologically associated domains (TADs) in 7%.

702  REPRODUCTIVE CYTOGENETICS On post-natal observation, one can be wise after the event. If a child with a particular phenotype has a rearrangement involving a breakpoint known to be in the region of a Mendelian locus, or of other recorded rearrangements producing the similar phenotype, the conclusion could reasonably be drawn that the cytogenetic abnormality was the cause of that abnormal phenotype. As just one example, a child with a de novo inv(7) (p22q21.3) having a particular split hand/foot malformation would invite the inference of a causal link, given the similarity of the limb defect with other 7q21.3q22 rearrangements (Cobben et al. 1995). Empiric Risk Estimation. Warburton (1991) conducted a review of major laboratories in the United States and Canada over a 10-year period and collected data based on more than a third of a million procedures. We make frequent reference to this work. A de novo translocation was identified in about 1 in 2,000 amniocenteses, a Robertsonian translocation in about 1 in 9,000, and an inversion in 1 in 10,000 (all on classical methodology). In Warburton’s study, serious malformations were identified in 6% of pregnancies with a de novo simple reciprocal translocation, either at elective termination or at live birth. This is some 3% above the background risk of approximately 3% for malformation and/or serious functional defect that applies to all pregnancies. Thus, we may draw the inference that in approximately 3% of these de novo translocations, the chromosomal defect was causative. For a de novo inversion, peri- or paracentric, the risk from Warburton (1991) for phenotypic abnormality is 9.4%, which is 6%–7% over and above the background risk. The numerator is small, however, and the 95% confidence limits span 2%–25%. Since, in theory, a two-breakpoint inversion should not imply a greater risk than the two-break reciprocal translocation, the figure for this latter category as noted above, namely 3% (or a little above), might reasonably be seen as appropriate also for the inversion. Although if one breakpoint is in an acrocentric short arm, the risk might be that much less (Leach et al. 2005). Warburton emphasized that the outcome data are imperfect, given the lack of long-term follow-up, and the questionable accuracy of phenotypic assessment in terminated pregnancies. Having made that point, she did then go on to say “there was no case in which a live birth originally reported as normal was later classified as abnormal after longer follow-up. In fact, the opposite tended to occur: Several cases described as having neonatal problems were later described as completely normal.” Follow-up Studies Small studies with follow-up into childhood have been undertaken (Gyejye et al. 2001), and these suggested that the figures presently offered are in the vicinity of the truth. We undertook detailed follow-up, to mean age 6 years, in 16 children with prenatally detected de novo balanced chromosome rearrangements (Sinnerbrink et al. 2013). One congenital anomaly (congenital hip dysplasia) was reported; but compared to population norms, no significant differences were observed with respect to health care needs, intelligence, or mental health. A large collaborative exercise involving 29 Italian prenatal laboratories, covering the period 1983–2006, brought together the findings on a total of somewhat more than a quarter of a million diagnoses (amniocentesis, CVS, and fetal blood) (Giardino et al. 2009). From these, 246 de novo balanced rearrangements were identified: 177 reciprocal translocations, 45 Robertsonian translocations, 17 inversions, and seven complex chromosome rearrangements. But follow-up data, in the 80% of cases in which the pregnancy was continued, were insufficient to derive risk figures for clinical outcomes, due to logistic and legal considerations, albeit that the authors comment that “none of the newborns have been reported to display visible malformations.”

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  703 Longer-term follow-up (mean 17 years) comes from a Danish registry study, based upon 41 individuals who had had a prenatal diagnosis of a de novo BCR in the absence of prenatal ultrasound anomalies (Halgren et al. 2018). Congenital malformations were neither more common nor more severe in BCR carriers than in matched controls, but the frequency of neurodevelopmental and/or neuropsychiatric disorder was significantly increased, 19.5% in BCR carriers compared with 8.3% in controls. This figure increased further, to 26.8% with longer clinical follow-up. These disconcerting data, somewhat at odds with the foregoing, suggest that neurodevelopmental compromise, rather than congenital abnormalities, are the major morbidity associated with de novo BCRs, and that their impact may have been underestimated by previous studies, due to lack of follow-up. Microarray was normal in all 32 subjects of this series where it was performed, indicating a low prognostic value from this test. On the other hand, sequencing of breakpoints, performed retrospectively, was informative in 7/27 cases, detecting disrupted neurodevelopmental genes (ARID1B, NPAS3, CELF4), disruption of regulatory domains of known developmental genes (ZEB2, HOXC), and two complex BCRs with multiple breakpoints, all in fetuses with adverse outcomes. The authors recommend the sequencing of breakpoints in all pregnancies with de novo BCRs, but warn that this approach still has relatively low sensitivity (~50%) for the prediction of abnormal outcomes. The “Carrier Fetus” Who Will Become a Carrier Adult. We discuss in Chapter 27 the issue of the genetic testing of children. In the case of prenatal diagnosis in which a de novo apparently balanced state is discovered, of course the child has already been tested, and “untesting” is not a practical matter. Consider the example of the mosaic test result mentioned below, the whole-arm translocation 46,XY,t(1;5)(p10;q10)/46,XY. Naturally, parents may want to know what reproductive implications this may have for their as-yet-unborn child. In this example, the genetic risk for the child will be, as the reader can readily determine, essentially that of a likely propensity to miscarriage, should the translocation cell line involve the gonad. It is the counselor’s responsibility to communicate this sort of information in outline form to the parents, along with the advice that the child could, in the fullness of time, attend the clinic on his or her own behalf. The information must be clearly conveyed. It could be seen as a failure of the counselor’s duty of care if in the next generation an affected child were born, the parents being unaware of the genetic risk (Burn et al. 1983; and p. 809). DE NOVO BALANCED HETEROLOGOUS ROBERTSONIAN TRANSLOCATION Heterologous Robertsonian translocations are present in 1 in 1000 persons, with similar proportions inherited and de novo (Shaffer et al. 1992). The great majority of de novo cases will be disomic, non-mosaic, and of biparental inheritance, and a normal phenotype is to be expected. The risk for phenotypic defect over and above the baseline is due to UPD and, theoretically, to occult mosaic trisomy. For inherited balanced Robertsonian translocations, Moradkhani et al. (2019) surveyed 28 diagnostic laboratories, and arrived at a risk of UPD of 0.06% when a parent is a known carrier of a rob translocation involving chromosome 14 or 15, and the fetus carries the same translocation. Comparable data are not available for de novo robs, because they will seldom be ascertained, in the absence of a relevant family history. However, on post-natal observation, UPD is twice as common in de novo robs as in inherited robs (Moradkhani et al.). Given that ~half of Robertsonian translocations arise de novo, the risk of UPD, in the setting 704  REPRODUCTIVE CYTOGENETICS of a de novo Robertsonian translocation involving an imprinted chromosome, can be estimated to be around 0.12%. With Robertsonian translocations forming predominantly in maternal meiosis (Page and Shaffer 1997), the risk applies mainly to maternal UPD. DE NOVO BALANCED HOMOLOGOUS ROBERTSONIAN TRANSLOCATION A chromosome comprising two long arms of the same acrocentric chromosome may be either a homologous Robertsonian translocation or an isochromosome: for example, rob(13q13q),19 or i(13q). If the formation of a homologous rob had been through the fusion of the maternal and paternal homologs, which of course must have occurred as a post-fertilization event, then the rearrangement manifestly has to be a true Robertsonian translocation, and the inheritance is biparental. In that case, a phenotypically normal child is the expectation, other things being equal (Abrams et al. 2001); infertility would, however, be anticipated (Chapter 7). All Robertsonian isochromosomes, and some homologous translocations, will display uniparental inheritance. The importance of uniparental disomy depends upon the chromosome involved. In Berend et al.’s (2000) Robertsonian series, there were six identified with an homologous translocation, all de novo, and four of these had UPD, two UPD(13)pat, and two UPD(14)pat. Barring isozygosity for a single gene mutation, normal outcomes are to be expected following prenatal diagnosis of a Robertsonian translocation (isochromosome) comprising a chromosome not subject to imprinting (chromosomes 13, 21, 22). This is actually recorded for the i(13q) UPD (Berend et al. 1999). No prenatal diagnosis reports exist for i(21q) UPD or i(22q) UPD, but the post-natal state of normality in each of these is known (Engel and Antonarakis 2002). Isodisomy for at least part of the chromosome will exist in the i(13q) UPD, i(21q) UPD, and i(22q) UPD states, and this raises the question of a risk, not readily quantifiable but likely very small, for a Mendelian autosomal recessive disorder due to isozygosity, the parent being heterozygous for the mutation in question. On the other hand, for the imprintable chromosomes 14 and 15, the risk for clinical defect is absolute following prenatal diagnosis of the rea(14) UPD and the rea(15) UPD, and the clinical syndromes of UPD(14) or UPD(15), maternal or paternal, would inevitably ensue. DE NOVO BALANCED INSERTION Only two cases are recorded, to our knowledge, of a de novo apparently balanced autosomal interchromosomal insertion detected prenatally (Hashish et al. 1992; Chen et al. 2020d). In both cases the child proved to be phenotypically normal. Van Hemel and Eussen (2000), in their review of nearly 90 families with an interchromosomal insertion, note that of the nine probands with congenital anomalies having a balanced insertion, seven were de novo and only two familial. It might reasonably be suggested that the risk for the interchromosomal insertion (three breakpoints) would be similar or possibly a little greater than the de novo apparently balanced reciprocal translocation (two breakpoints). Submicroscopic balanced insertions are at least six times as common but will escape detection at prenatal diagnosis, coming to light only in future generations if they generate apparent de novo CNVs (Nowakowske et al. 2012). 19 The formally correct nomenclature is actually der(13;13)(q10;q10).

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  705 DE NOVO BALANCED AUTOSOMAL RING CHROMOSOME The 46,r(A) ring chromosome is discussed in Chapter 11, and the reader is referred to specific instances listed therein. Rings that are truly balanced, reflecting a tip-to-tip telomere fusion, are nevertheless likely to cause growth retardation (or, in the case of r(20), epilepsy). Microarray analysis can reveal a very subtle deletion, which even targeted multiplex ligation-dependent probe analysis (MLPA) and FISH could not detect, as Manolakos et al. (2009) show in a case of ring 15 chromosome prenatal diagnosis. The prenatally diagnosed r(4) in Akbas et al. (2013) had 4p and 4q deletions of only 130 kb and 2.4 Mb, and this was associated with IUGR. The outlook for the child might be similar to that of the man with a r(4) whose case we mention on p. 296. Equally, microarray may demonstrate no apparent loss of material, as Papoulidis et al. (2010) report with a ring 21, the baby subsequently born being assessed as normal at 10 months. DE NOVO BALANCED WHOLE-ARM TRANSLOCATION Very few de novo whole-arm translocations are recorded, “although the existing examples suggest an optimistic prognosis can be given” (Farrell and Fan 1995). A whole-arm X-autosome translocation is mentioned below. DE NOVO BALANCED COMPLEX REARRANGEMENT A de novo apparently balanced complex chromosome rearrangement (CCR) has a high risk for intellectual impairment and physical malformation, but equally, normal children have been born. Chen et al. (2006a) and Giardino et al. (2006) reviewed the published cases, in some of which amniocentesis had been triggered by an increased-risk maternal serum screen, or the observation of fetal anomaly on ultrasound. The outcomes were abnormal in about half, the abnormalities ranging from developmental delay to single or multiple malformation. Intuitively, the risk would be greater with a higher number of breakpoints, and Madan et al. (1997) provide support for this view. Microarray analysis may clarify whether a true quantitative imbalance exists; however, a CCR with a breakpoint occurring within a gene might not (as with any such rearrangement) be detected, as exemplified in the t(2;12;18)(q22.3;12q22;q21.33) reported in Engenheiro et al. (2008), in which the 2q22.3 breakpoint disrupted the ZEB2 gene, causing Mowat-Wilson syndrome. In a report of three CCRs diagnosed prenatally, all proved to be unbalanced upon array-CGH analysis (De Gregori et al. 2007). As noted above, in the registry study of Halgren et al. (2018), the two individuals who were revealed by whole genome sequencing to have complex rearrangements that were cryptic to karyotyping (with four and seven breakpoints respectively) both had neurodevelopmental phenotypes at follow-up. This is consistent with Redin et al. (2017), who identified cryptic complexity in 21% of post-natally ascertained BCRs, and confirms that, at prenatal diagnosis, complexity of BCRs is a risk factor. DE NOVO BALANCED X-AUTOSOME TRANSLOCATION In the case of a de novo apparently balanced X-autosome translocation, there are the additional possible complications of (1) gonadal dysfunction if the breakpoint is within one of the critical regions of the X chromosome, and (2) the unpredictability of the patterns of inactivation with the possibility of severe abnormality. On theoretical grounds, the risk may be about twice that for the simple autosomal translocation given earlier (Waters et al. 2001), although Abrams and Cotter (2004), reviewing the literature, arrived at a risk figure as high as 50% (and disregarding a possible risk for reproductive 706  REPRODUCTIVE CYTOGENETICS health). Nevertheless, in the case they report, a normal daughter, with follow-up to age 17 months, was born after amniocentesis (for advanced maternal age) had shown a de novo 46,X,t(X;6)(q26;q23) karyotype, with the normal X late replicating. They, and we, hope that further such cases will be reported. In the series of Halgren et al. (2018) there were two prenatal diagnoses of X-autosome translocations. One pregnancy was terminated, but in the other, with karyotype 46,X,t(X;15)(q24;p11)dn, the pregnancy ended in livebirth, and at follow-up at 6 years, the girl was assessed as having an IQ of 67, facial dysmorphism and obesity. Microarray was normal, and breakpoints could not be determined by sequencing. On the specific issue of an Xp21 breakpoint, the question of Duchenne muscular dystrophy arises. Evans et al. (1993) actually showed normal dystrophin on the invasive procedure of fetal muscle biopsy following detection at amniocentesis of an apparently balanced rcp(X;1) with the X breakpoint at p21, and so predicted the child would not have Duchenne/Becker muscular dystrophy; and their prediction proved to be correct. In a case of de novo 46,X,t(X;9)(p21.3;q22) diagnosed at amniocentesis, Feldman et al. (1999) showed apparent integrity of the dystrophin locus on FISH. Methylation analysis indicated preferential inactivation of the normal X. On these two observations, the couple decided to continue the pregnancy; but fetal demise occurred at 34 weeks, probably due to chorioamnionitis following premature rupture of membranes at 33 weeks. No fetal defects were seen; dystrophin staining of muscle was normal. These days, molecular study would be a simpler approach. DE NOVO BALANCED YQ-AUTOSOME TRANSLOCATION The balanced Yq-autosome reciprocal rearrangement, with a 46-chromosome count, has the gonosomal breakpoint in proximal Yq (the breakpoints usually given as q11, q11.2, or q12). Given the common association with male infertility, most Yq-autosome translocations arise de novo, but antenatal detection is rare. Hsu (1994) reviewed 23 reports, in which the usual ascertainment was through infertility (oligospermia/azoöspermia) in the adult male, with a few being found incidentally, and including one at prenatal diagnosis. Only three, including two from the early 1970s in which the detail of the rearrangement was less certain, were identified through a malformed child. It may be that such translocations should be regarded as conveying no greater risk for an abnormal intellectual phenotype than do reciprocal autosomal translocations, but acknowledging a frequent, perhaps inevitable, compromise of fertility (p. 175). In the particular case of a de novo translocation with Yqh material on the short arm of an acrocentric (which is, to be precise, an unbalanced rearrangement), this is unlikely to be the basis of any phenotypic defect (p. 174). MOSAICISM FOR A DE NOVO BALANCED STRUCTURAL REARRANGEMENT Reciprocal Translocation Mosaicism. True mosaicism for a balanced reciprocal translocation, 46,rcp/46, is very rarely recognized (Leegte et al. 1998). The great majority of this type of mosaicism seen at prenatal diagnosis is level I or II, and this is pseudomosaicism due to in vitro change. Some breakpoints (6p21,13q14) are preferentially involved (Benn and Hsu 1986). In terms of implications for fetal phenotype, it can usually be disregarded. True mosaicism for a reciprocal translocation has been seen at prenatal diagnosis, and Hsu et al. (1996) accumulated 11 examples showing

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  707 one normal cell line and one with a balanced autosomal translocation. In no instance in which the pregnancy proceeded (nine of the 11) had phenotypic abnormality been observed. Chen (2024a) summarized another four cases, all with normal phenotype at follow-up. Regarding post-natal (and therefore biased) ascertainment, mosaicism for a balanced translocation has been reported with congenital abnormalities and dysmorphism (Saura et al. 1987; Aughton et al. 1993), suggesting it is not always benign. Concerning a possible risk for unbalanced progeny in the next generation if the gonad were involved, each such case would need to be individually assessed; the parents would need to know to give their child access to the information in the fullness of time. Robertsonian Translocation Mosaicism. In four cases in Hsu et al. (1996) of diagnosis at amniocentesis of mosaicism for a balanced heterologous translocation, 45,rob/ 46, the outcome was normal in all (the mosaicism confirmed post-natally in the two infants studied). The specific translocations were 13q14q, 13q22q, and 14q21q. Chen et al. (2024b) reported prenatal diagnosis of mosaic rob 13q21q with normal post-natal follow-up at 4 years. Whole-Arm Translocation Mosaicism. The mother reported in Wang et al. (1998) with 46,XX,t(10q;16q)/46,XX mosaicism was normal (although her child was abnormal). We know of one case of level III mosaicism for a balanced whole-arm translocation at amniocentesis, 46,XY,t(1;5)(p10;q10)/46,XY, with 30% of cells in three separate cultures showing the translocation, and confirmed on a cord blood sample at delivery (10 cells out of 50 with the translocation). On follow-up at age four years, the child was normal and healthy (D Grimaldi and B Richards, personal communication, 2001). Complex Rearrangement Mosaicism. The only known example to our awareness of de novo mosaicism with an apparently balanced CCR and a normal cell line detected prenatally is that described in Hastings et al. (1999b), 46,XX,t(3;10)(p13;q21.1), inv(6) (p23q12)/46,XX, and this case was associated with fetal abnormality. Inversion Mosaicism. In four cases in Hsu et al. (1996) of diagnosis at amniocentesis of mosaicism for a balanced inversion (pericentric or paracentric), 46,inv/46, the outcome was normal in all (all four were studied post-natally, with the mosaicism found in only one). DE NOVO UNBALANCED STRUCTURAL REARRANGEMENT, MODAL NUMBER 46 Autosomal. For any de novo autosomal structural rearrangement in which imbalance is cytogenetically visible, serious phenotypic abnormality is highly likely. Microarray analysis may be used to identify the breakpoints of the rearrangement and aid in the prediction of phenotype. Many cases, indeed most, are unlikely to be exactly the same as those in the literature or on the databases, and the counselor will need to make an informed evaluation. Ultrasonography may clarify the question if abnormalities are seen, but an apparently normal sonogram does not guarantee that the child would be normal (Al-Kouatly et al. 2002). If a “jumping translocation” (p. 319) leads to imbalance, fetal defect is very probable (Annable et al. 2008). In the mosaic state, the risk may be high if pseudomosaicism is judged to be unlikely. Hsu et al. (1992) record 34 cases with at least one cell line having an unbalanced rearrangement (thus, presumed to be a true mosaicism). In follow-up studies, phenotypic abnormality was noted in half, and cytogenetic confirmation obtained in 65%. Each rearrangement needs to be considered on its merits. The dilemma of deciding how

708  REPRODUCTIVE CYTOGENETICS best to advise couples is illustrated in Cotter et al. (1998). They describe the karyotype 46,XX,der(4)t(4;5)(q34;q12)/46,XX detected at amniocentesis, imparting, in the abnormal cell line, trisomy for most of 5q. This was confirmed on two subsequent amniocenteses, with an average overall of 17% of amniocytes abnormal, but with a 46,XX result on fetal blood sampling, and normal ultrasonography. The parents were advised that “few data were available” to determine risk; they made a decision to continue the pregnancy. In the event, the child appeared normal at birth and at 2-year follow-up; 100 cells at cord blood karyotyping were normal. Chen (2024a) report two similar cases, including one in which a mosaic 13q duplication was confirmed in post-natal blood and buccal mucosa; both infants appeared normal, but longer-term follow-up was lacking. In contrast, 46,XX,add(15)(p10), t(2;15)(p10;q10)/46,XX mosaicism detected at 30-week prenatal diagnosis (performed due to IUGR), and shown on both amniocentesis and fetal blood sampling, was associated post-termination with fetal anomalies consistent with a partial trisomy 2p (Pipiras et al. 2004). Cotter et al. rightly call for others’ experience in similar cases to be published. X-Autosomal. Prediction with respect to the unbalanced X-autosome translocation is precarious (and see Chapter 6). While the pattern of inactivation may lessen the effect, and indeed convert an invariably lethal imbalance to a survivable state, the degree to which selective inactivation may occur in fetal tissues is not knowable, and a significant defect remains very probable, the risk as high as 50% (Abrams and Cotter 2004). Had the child with an unbalanced der(X)t(Xp;22q) described on p. 171 (Figure 6–17) been identified at amniocentesis, and with the DiGeorge critical region intact and no inactivation on the 22q segment, a prediction of typical Turner syndrome might have been reasonable. In the event, this child proved to have a significant intellectual disability. Contrary examples in which a prediction of major abnormality would have been mistaken are rare. Sund et al. (2020) highlight how an increased risk for Turner syndrome at NIPT might, on rare occasions, lead to a diagnosis of an unbalanced X-autosome translocation. In their case, the pregnancy resulted in the birth of a female infant with the unbalanced karyotype 46,X,der(X)t(X;15)(q27.3;q15.1) and a phenotype that included complex congenital heart disease, dysmorphism and global developmental delay. Chen et al. (2023b) reported an unbalanced X;21 translocation detected by amniocentesis with tertiary monosomy (concomitant deletion of both X chromosome and autosome) and significant fetal abnormality. Y-Autosomal. Autosomal material attached to the heterochromatin of a Y chromosome is to be seen in essentially the same light, as if it had been a translocation to an autosome (and see Chapter 6). A rare but recurrent unbalanced karyotype seen at prenatal diagnosis is the t(Y;1)(q12;q21) translocation in mosaic state, which endows essentially a 1q trisomy in the tissue with the translocation (Li 2010). The phenotype is lethal. Vice versa, if Y material is attached to an autosome, and if autosomal material is lost at that site, the autosomal monosomy of itself determines phenotypic defect (Klein et al. 2005). A somewhat different and very rare category is that in which a near-intact Y, missing only part of the pseudoautosomal region, combines with an acrocentric chromosome. Borie et al. (2004) describe the prenatal diagnosis of 45,X,dic(Y;22)(p11.3;p11). Had this dicentric chromosome included all the Yp material, the child might have been normal. But in fact the SHOX locus, at Yp11.3, was deleted, and the otherwise normal male child had short stature. Yq;15p Variant. In the population there is a common variant whereby the heterochromatin of Yq becomes translocated to the short arm of chromosome 15; this occurs

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  709 in about 1 in 2,000 individuals. Occasionally, translocations with other breakpoints will occur between these two chromosomes, such as the case reported by Chen et al. (2007), in which the father’s karyotype was 46,X,t(Y;15)(q12;p13) and the female fetus inherited the abnormal chromosome 15. Because the derivative chromosome has deleted the repetitive 15 short arm and replaced it with Yq heterochromatin, no phenotypic effect would be expected. The authors suggest that methylation analysis for chromosome 15 should be considered, although in fact no cases of UPD(15) due to this common variant have been reported. DE NOVO UNBALANCED STRUCTURAL REARRANGEMENT, MODAL NUMBER 47: SUPERNUMERARY CHROMOSOME A supernumerary chromosome may be of substantial size, and identifiable as to its makeup; or it may be smaller, and its origin uncertain. The latter are referred to as supernumerary marker chromosomes (SMCs), and these have also been described as markers, extra structurally abnormal chromosomes (ESACs), and accessory chromosomes. Some as rings and isochromosomes are discussed separately below. The SMCs we mostly consider here are the small SMCs (sSMC); these are defined as structurally abnormal chromosomes that cannot be identified or characterized unambiguously by conventional banding cytogenetics alone, and which are generally equal in size or smaller than a chromosome 20 in the same metaphase spread. Some are quite harmless, and associated with phenotypic normality, and others are not: They are a very heterogeneous group. Small SMCs are encountered in about 1 in 1,000 prenatal diagnoses analyzed cytogenetically, frequently in the mosaic state with a normal cell line. Upon the discovery of an sSMC at prenatal diagnosis, an urgent parental chromosome analysis is required. The majority will prove to be de novo,20 and Liehr et al. (2009) emphasize the point: De novo sSMCs ascertained at prenatal diagnosis are without phenotypic consequence in about three-fourths of cases. These questions are to be asked: From which chromosome is it derived, and does it comprise euchromatin or heterochromatin? Is it a recognized type of sSMC, for which precedents are recorded? Precise characterization is necessary, and this requires the use of FISH or microarray (Liehr et al. 2009; Domaradzka et al. 2021). On FISH, ~80% are shown to derive from one of the acrocentric chromosomes, most commonly chromosome 15 or chromosome 22, and often involving only the pericentromeric region and/ or the satellites (Crolla et al. 1998; Lin et al. 2006). Use of microarray as the first-tier prenatal cytogenetic analysis circumvents some of these issues: Small benign SMCs that comprise only heterochromatin will not be detected at all. Larger SMCs will be detected on microarray as copy number gains of the relevant chromosomal segment. According to Kurtas et al. (2019), most de novo sSMCs are the legacy of a partial trisomy rescue. This multistep process begins with maternal meiotic nondisjunction followed by partial trisomy rescue of the supernumerary chromosome. This chromosome then lags at anaphase and becomes trapped within a micronucleus, where it is shattered via the process of chromothripsis (Chapter 12). The micronuclear material is later re-embedded into the main nucleus, where DNA repair “stitches together” only some parts of the original chromosome to form the sSMC. 20 Familial SMCs are noted, according to their chromosomal provenance, in Chapter 14; see also Brøndum-Nielsen and Mikkelsen (1995) and Hastings et al. (1999a). 710  REPRODUCTIVE CYTOGENETICS De Novo Identifiable Supernumerary Chromosome of Substantial Size. An additional chromosome which is of sufficient size that it can be characterized on initial routine analysis as a deleted or rearranged form of a specific autosome will imply a very high risk of abnormality, approaching 100%, due to partial trisomy of that chromosome. Once a supernumerary chromosome has been identified, it is no longer referred to as an SMC; it is now described as a ring or derivative—for example, r(7), der(22), or neo(13q31), or whatever may be the precise description. De Novo Small Supernumerary Marker Chromosome. De novo sSMCs have been described for most chromosomes (Liehr 2025b); two-thirds of sSMCs are acrocentric-derived (Dalprà et al. 2005). Huang et al. (2019) identified 40 de novo SMCs from 68,087 amniocentesis samples (1 in 1700), of which seven had a copy number gain detected on microarray, and could therefore be interpreted as pathogenic. De Novo Minute Marker. The very small SMC (minSMC) may comprise only centromeric material and be harmless. We discuss a prenatal case on p. 626, a minSMC apparently comprising no more than chromosome 18 centromere; the child turned out to be normal. The tiny bisatellited microchromosome can be thought of as the reciprocal product of the Robertsonian rearrangement; these microchromosomes also are typically harmless (Liehr et al. 2021) (Figure 22–9). Figure 22–9.  A De Novo Supernumerary Marker Chromosome at Amniocentesis. Notes: Amniocentesis had been done on the basis of advanced maternal age. The very small marker chromosome (mar) was seen in all cells. FISH on cultured amniocytes with probes recognizing the centromeric region of acrocentric chromosomes 13 and 21 (green), and 14 and 22 (red), showed the marker (arrowed) to be derived from either 14 or 22. Having proven an acrocentric basis, the advice given could be optimistic. A healthy child was born. Source: From CP Chen et al., Prenatal diagnosis and molecular cytogenetic characterization of a familial small supernumerary marker chromosome derived from the acrocentric chromosome 14/22, Taiwan J Obstet Gynecol 61:364–367, 2022. Courtesy CP Chen, and with the permission of Elsevier.

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  711 21 When microarray is used as the first-line test, a mosaic isochromosome may be mistaken for a duplication. For example, an isochromosome (4 copies), if present in 50% of cells, may be interpreted as a non-mosaic duplication (3 copies). Supernumerary Marker Chromosome Outcomes. With precise cytogenetic characterization of an sSMC identified prenatally, using FISH and microarray (Jang et al. 2016), and with concomitant ultrasound examination, it should be possible precisely to categorize the genetic risk. Earlier published reports of liveborn children with various types of SMCs—a notably heterogeneous material—are mostly biased by ascertainment in favor of phenotypic abnormality. Series of prenatally diagnosed fetuses are deficient in that there is usually only a short-term follow-up of liveborn children, while pathological assessments following termination can only show major structural malformations (Warburton 1991). Brøndum-Nielsen and Mikkelsen (1995) report a 10-year experience in Glostrup, Denmark, during which nine de novo SMCs were identified. In seven cases, termination of pregnancy was chosen, with some of these showing defects at pathological examination; and in the two pregnancies continuing, one infant with a minute acrocentric-derived SMC was normal at birth, while one with a ring-like 17 was mildly delayed at age two years. In the similar survey of Hastings et al. (1999a), data were presented on 31 prenatally diagnosed SMCs, of which 21 were de novo. In 10 of these 21 proceeding to FISH analysis, six being mosaic, five were shown to be 15-derived and three 14- or 22-derived; the remaining two included a r(8) and a der(16). Of the six in which the pregnancies continued, only the r(8) child was physically and developmentally abnormal. Repeating the point: With FISH and microarray, most sSMCs should admit of precise cytogenomic analysis, and the prenatal advice based upon knowledge of the specific involved segment. Supernumerary Autosomal Ring Chromosomes. Autosomal ring chromosomes, as a supernumerary 47th chromosome, imply a high risk of phenotypic abnormality. They originate from a variety of chromosomes and contain euchromatin. Certain of these, in which only one arm of the chromosome is represented in the ring, are specifically recorded in association with phenotypic abnormality: r(1p), r(5p), r(7q), r(8q), r(9p), r(10p), r(20p), and r(20q) (Hastings et al. 1999a; Anderlid et al. 2001). Uniparental disomy may complicate the picture: James et al. (1995) and Anderlid et al. report supernumerary rings, from chromosomes 6 and 9, associated with UPD(6) and UPD(9), respectively. Very small rings, that might also have been categorized as sSMCs, might not necessarily cause an abnormal phenotype: For example, two infants in Kitsiou-Tzeli et al. (2009) born following prenatal diagnosis of 47,+r(20)/46,N mosaicism were judged normal in early infancy. Autosomal Isochromosomes. Autosomal isochromosomes are typically seen in the mosaic state as a supernumerary isochromosome (or as isodicentric isochromosome), and thus the discovery of 47,+i/46,N (or 47,+idic/46,N) is always a concern, whether at a level II or even level I mosaicism.21 Such a karyotype raises the prospect of an effective mosaic tetrasomy for the chromosomal arm concerned. A 46-chromosome karyotype in which one homolog is replaced by an isochromosome typically implies a trisomy for one arm of that chromosome, and monosomy for the other. These are certainly rare observations: In an amniocentesis-based survey from Italy, based on slightly less than 90,000 diagnoses, the most frequent were, in order, isochromosomes of 20q, 9p, 18p, and 12p, at approximately 1 in 30,000, 45,000, 45,000, and 90,000, respectively (Forabosco et al. 2009). Brief commentaries on these, and on certain other isochromosomes, follow.

712  REPRODUCTIVE CYTOGENETICS 47,+i(5p). Blakey-Cheung et al. (2020) reviewed the literature and identified six cases where mosaic tetrasomy 5p had been diagnosed prenatally, but disconcertingly, an additional two cases where prenatal testing (one CVS and one amniocentesis) had been interpreted as normal, and a diagnosis of tetrasomy 5p was only made following the birth of a clinically affected child. Sijmons et al. (1993) assessed a dysmorphic and neurologically compromised child with a 5p isochromosome in 3/31 lymphocytes and 12/14 skin fibroblasts, and yet upon retrospective checking, only one of 217 cells from a stored short-term CVS culture was 47,XY,+i(5p). We contrast this unfortunate experience with ours of seven cases of i(5p) mosaicism identified at CVS, six of which went on to follow-up amniocentesis (Clement Wilson et al. 2002). Three children were followed up to 2½, 3¼, and 4 years, and their normality was quite apparent. In one of these children, a circumscribed area of the placenta following delivery karyotyped 47,+i(5p), adjacent parts karyotyped 47,+i(5p)/46,N, and most of the placenta (and the child himself ) had a normal karyotype. The CVS sampling had presumably needled this small region of confined placental i(5p) mosaicism. One pregnancy tested 100% i(5p) at CVS, and the parents chose termination; no i(5p) cells were detected from fetal skin culture. In another with a 65% load at CVS, a follow-up amniocentesis showed 45% of cells with the isochromosome, and post-termination tissues showed 15%–30%. From the foregoing, we may conclude that a CVS diagnosis with a normal follow-up amniocentesis and with normal ultrasonography suggests, but cannot confirm, a normal child. As for the primary detection of i(5p) mosaicism at amniocentesis, only five cases are recorded, with all having an abnormal outcome (Blakey-Cheung et al. 2020). Grams et al. (2011) reported monozygotic twins who were discordant for i(5p) at CVS and amniocentesis; multiple abnormalities were seen on ultrasound in the affected twin, but the pregnancy ended in spontaneous loss of both twins at 18 weeks gestation. 47,+i(8p). López-Pajares et al. (2003) review the small number of reported cases. Two examples are given of discordance between amniocentesis (normal) and post-natal blood (tetrasomy 8p), an unusual pattern for isochromosomes (but cf. the i(9p) below). A disconcerting story is told in Nucaro et al. (2006): i(8p) mosaicism was seen at long-term cultured (but not short-term) CVS, with a normal result after amniocentesis, but resulting in a child severely disabled and epileptic, and with a 5% level of the i(8p) on blood. 47,+i(9p). The highly variable clinical picture and the subtleties of different breakpoints are discussed in Dhandha et al. (2002). Isochromosome mosaicism can be the basis of a false-negative test result at prenatal diagnosis. Thus, Eggermann et al. (1998) reported an abnormal baby born to a 39-year-old mother, in whom amniocentesis at 14 weeks gestation had returned a normal karyotype. On blood analysis, the child had an i(9p) in 32% of cells. From one skin biopsy, 50 cells had a normal karyotype, but on a second biopsy, five out of eight cells showed the i(9p) chromosome. Similarly, we have seen a girl with mosaicism for an isodicentric chromosome 9, in which microarray on DNA from uncultured amniocytes showed a copy number gain at 9p24.3q22.32, but karyotype on cultured cells was normal (Lee et al. 2025). The particular attribute of the i(9p) is for blood, but not skin, to show the abnormality, and this is likely the explanation for its non-detection at amniocentesis. Pertile et al. (1996) support this interpretation, in their follow-up of a (non-mosaic) CVS diagnosis of idic(9)(q13). An extensive search at amniocentesis revealed a single abnormal colony, which might well otherwise have been missed. Finally, fetal blood sampling showed the

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  713 idic(9) in 8% of cells. A more severe case is recorded in Tang et al. (2004), which showed the isochromosome in all amniocytes at 24 weeks, and in most blood and fibroblast cells from the very malformed infant (who died at 1 month of age). 47,+i(10p). A single case is on record, the diagnosis having been made following the recognition of fetal defects on ultrasonography (Wu et al. 2003). 47,+i(12p), Pallister-Killian Syndrome. The 12p isochromosome is the basis of the well-known Pallister-Killian syndrome. Salzano et al. (2018) review the prenatal profile. The fractions of abnormal cells detected at prenatal diagnosis can vary greatly. Bernert et al. (1992) showed in one example 100% of short-term CVS cells and 10% of amniotic fluid cells having the 47,+i(12p) karyotype, whereas in Kunz et al. (2009), at CVS the isochromosome was seen only in long-term culture; in both cases, the pregnancies were terminated. Horn et al. (1995) reported a pregnancy in which CVS gave a 46,XY result on direct (17 cells) and cultured (eight cells) analysis (and 28 further cells on a retrospective study), and the abnormal newborn baby was 46,XY on a peripheral blood study (100 cells counted); at 18 months, a clinical diagnosis of Pallister-Killian syndrome was made, and the karyotype on skin fibroblast culture was 47,XY,+i(12p)/ 46,XY, with 85% of cells having the isochromosome. (Had it been an amniocentesis rather than CVS that had been done, abnormal cells would probably have been seen.) Classical karyotyping typically returns a normal result because the i(12p) is lost in stimulated lymphocyte cultures; thus, microarray on uncultured cells is the preferred prenatal test. Iso(12p) can also be detected by genome-wide cell-free DNA testing (Chau et al. 2020). 46,i(13q). A de novo “Robertsonian” translocation, leading to trisomy 13, is, in the majority of cases, actually an isochromosome, as discussed above. Isodicentric 15. About half of all SMCs are an idic(15) (also referred to as pseudodicentric 15, or inverted duplication 15; and see p. 428). These are typically dicentric and bisatellited, although one of the centromeres may be suppressed. The smallest ones (smaller than chromosome 21q) appear to be harmless, but larger ones result in the “idic(15) syndrome,” characterized by intellectual disability, epilepsy, and autistic features. The boundary between smaller and larger is in 15q12. The use of D15S10 or SNRPN FISH probes, which recognize sequences in 15q12q13, enables distinction of harmless and pathogenic chromosomes (Eggermann et al. 2002), a distinction that can also be made using microarray (Wang et al. 2004). Rare idic(15)s have been associated with UPD(15), and it may be warranted to check for this possibility (Hastings et al. 1999a). 47,+i(18p). Schinzel (2001) notes that more than 75 cases of 47,+i(18p) have been recorded. Multiple physical anomalies and a moderate to severe degree of intellectual disability characterize the clinical picture. Boyle et al. (2001) emphasize the plausibility of a premeiotic origin, and the caution therefore that gonadal mosaicism may exist in a parent, as they illustrate in their report of affected half-sisters. Iso(18p) has been diagnosed following a cfDNA result showing increased risk for trisomy 18 (Tamaki et al. 2020). 46,i(18q). The karyotype produces a combination of monosomy 18p and trisomy 18q. Chen et al. (1998) record that many 18q isochromosomes diagnosed prenatally are associated with very severe malformation, such as holoprosencephaly and cloacal dysgenesis. Levy-Mozziconacci et al. (1996) describe a case presenting at 22 weeks gestation with abnormal ultrasonography, and although the direct CVS was 46,XX in all cells, amniocentesis and fetal blood sampling showed the isochromosome (an isodicentric, in this instance) in all cells: an example of complete CVS-amniocentesis discordance.

714  REPRODUCTIVE CYTOGENETICS 46,i(20q). An i(20q) identified at amniocentesis in mosaic form appears most often to be a benign finding; a rather surprising conclusion. It may be an unusual sort of mosaicism in being confined, or largely so, to amniocytes, the abnormal cell line having arisen as a post-zygotic event, and its growth perhaps favored in vitro (Robinson et al. 2007). This is consistent with the pregnancy reported by Chen et al. (2023a), in which two separate amniocenteses, at 17 and 18 weeks, detected i(20) in 12% of cultured cells, but microarray on uncultured cells was normal; the baby was healthy with normal blood karyotype. The few reported cases with fetal defect would presumably reflect a tissue distribution which included the fetal anatomy. Goumy et al. (2005) counsel caution, and point to the advisability of careful ultrasonography, targeted in particular to the brain and vertebrae. If there is a theoretical risk of UPD(20), this has not been observed in practice. 46,i(21q). This rearrangement is an isochromosome, not a Robertsonian translocation (Shaffer et al. 1991). The phenotype is that of Down syndrome. Gilardi et al. (2002) report a case in which the isochromosome probably arose post-zygotically in an early cell destined to form the lineage of the inner cell mass and the extra-embryonic mesoderm, such that a direct CVS gave a non-mosaic 46,XX result, while long-term CVS and post-termination fetal studies showed non-mosaic 46,XX,i(21q); a similar story comes from Brisset et al. (2003). Post-zygotic formation and placental mosaicism explain why isochromosome 21q is a common cause of false-negative cell-free DNA screening results for Down syndrome (Huijsdens-van Amsterdam et al. 2018). The i(21q) can also exist in a 47-chromosome karyotype. Nagarsheth and Mootabar (1997) showed a 47,XY,+i(21q)[6]‌/46,XY[19] karyotype at amniocentesis; the parents elected to continue the pregnancy, and the abnormal child had only one out of 120 peripheral blood lymphocytes with the i(21q), the other 119 being normal. These authors suggest that some previously reported cases of supposed i(12p) mosaicism may have been, in fact, i(21q). 47,+i(22q). A single case of an isochromosome for 22q being detected at amniocentesis is recorded in Guzé et al. (2004). The isochromosome was probably generated post-zygotically, with the subsequent production of additional abnormal cell lines. The pregnancy continued to full term: The child had several defects and died on the second day of life. Isodicentric 22. The bisatellited idic(22) typically, but not invariably, causes cat eye syndrome (p. 456). If the idic(22) lacks proximal 22q euchromatin, normality is very probable, whereas those containing euchromatin can lead to a phenotype anywhere between full cat eye syndrome and normality (Crolla et al. 1997). Normal Variants Chen et al. (2006c) review the question of variants detected at prenatal diagnosis. They identified 16 variants of euchromatin or heterochromatin in 21,832 amniocenteses. Eight of nine euchromatic variants were proven inherited, and seven were C-band positive. The remaining C-band-positive, heterochromatic variants were all inherited from a carrier parent. Concerning the specific case of the nucleolar organizing region translocation or interstitially inserted satellite, and as noted in Chapter 17, “genetic counseling should be reassuring” if this is discovered at prenatal diagnosis (Faivre et al. 1999, 2000; Chen et al. 2004c). The Y;15 variant is noted above. Microarray-level normal variants CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  715 are not usually reported by the laboratory at prenatal diagnosis. If further interpretation is required, the resources outline in Chapter 18 can be applied to the specific variant. Sex Chromosome Abnormalities THE CLASSIC FULL ANEUPLOIDIES A sex chromosome abnormality is not an uncommon discovery at prenatal diagnosis, with an overall incidence of 1 in 250–300, and with detection increasing with the uptake of NIPT (Linden et al. 2002; Steffensen et al. 2023). The three most common are the sex chromosome trisomies (SCTs), XXY, XXX, XYY, which share many neurodevelopmental characteristics and are sometimes considered together, whereas 45,X has a more distinctive clinical presentation. As Boyd et al. (2011) write, “The importance of providing parents with accurate information about the frequency of the diagnosis, and the variability of the condition on the basis of outcomes from unbiased population-based follow-up studies on the specific chromosome abnormality, cannot be overemphasized.” Two of these aneuploidies (XXY and 45,X) may be firmly predicted in terms of an abnormality of development of the reproductive system: Children with Klinefelter and 45,X Turner syndrome will, with near certainty, be infertile. Some will choose pregnancy termination, although it is of interest that in France, coincident with multidisciplinary centers for prenatal diagnosis being put in place in 1997, the termination rates during the period 1976–2012 fell (from 41% to 12% for XXX, and from 26% to 7% for XYY) (Gruchy et al. 2016), and similar observations are made in some other, but not all, jurisdictions. Liao and Li (2014) wonder if the question can be side-stepped, in the setting of pregnancies tested for some other reason (e.g., thalassemia), by not interrogating the sex chromosomes. Similar considerations apply to the inclusion of sex chromosome abnormalities in NIPT. For those couples deciding to continue a pregnancy, Robinson et al. (1986) offer a useful commentary. Parents of children predicted to be infertile might feel a sense of loss—a “sadness and regret about their child’s anticipated loss and about their own loss of grandchildren” and “concern about their children’s wholeness and, by extension, their own.” Parents may take some comfort from knowing that infertility is by no means an uncommon problem in the general population, and further comfort from the advice that recent advances in artificial reproductive technology may now enable the infertility to be overcome, in some individuals. The picture for intellectual and psychological functioning is less predictable. Earlier adult studies defining a strong association with learning difficulties and psychological disturbance were contaminated by ascertainment bias (and counselors’ personal experience may have been more with those children whose problems were sufficiently severe that they had come to medical attention). Children identified in newborn populations screened for cytogenetic abnormalities and subsequently followed up constitute a group unbiased in their ascertainment, although perhaps subject to other but less important biases (Puck 1981). Data from the study of such children in several American and European cities, followed from infancy through childhood, adolescence, and young adulthood, have since given a reasonably clear picture of the natural history of the more common sex chromosome aneuploidies (Linden et al. 2002). In general, the IQ averages 10–15 points below that of the siblings. Hook’s (1979) early proposition has held up: Some sex chromosome

716  REPRODUCTIVE CYTOGENETICS aneuploidies influence brain function in such a way that the development of intellectual capacity, emotional maturity, and speech and language skills are affected to some extent; but none of these effects necessarily occurs, none is specific to sex chromosome aneuploidy, and some may be amenable to corrective intervention. There is considerable overlap with the XX and XY population. Hong et al. (2014) reviewed the cognitive and neurological aspects of sex chromosome aneuploidies and noted shared features across the sex chromosome aneuploidies, comprising impairments in executive functioning, motor skills, and higher-order social cognitive ability. In their comprehensive review, Tartaglia et al. (2020) note that although “the list of neurodevelopmental and psychological risks associated with SCT that can manifest from infancy into adulthood is long, marked variability in the presence and severity of these features is a consistent research finding.” Ratcliffe (1999) and Bender et al. (2001) provide long-term follow-up data, well into adulthood. Bender et al. followed eight 45,X, 10 47,XXX, and 11 47,XXY individuals through to an age range of 26–36 years, using siblings as controls, and noted the IQs of the aneuploid groups to be considerably less compared with the sibs. Nevertheless, the variation is wide, and these authors emphasize the point that “sex chromosome aneuploidy does not exert its influence in a vacuum, but rather interacts with the host of other genetic and environmental influences that collectively guide human development.” As Le Gall et al. (2017) show, an independent, concomitant CNV can exacerbate the clinical, and especially the neurocognitive phenotype, and may indeed be the more significant factor. Children with sex chromosome aneuploidies seem more susceptible to either the good or the bad effects of a stable or of a dysfunctional family setting, than do their 46,XX and 46,XY siblings (Stewart et al. 1990; Bender et al. 1995). Children identified at prenatal diagnosis, a group biased toward higher socioeconomic status, may do better academically and socially than the cohorts followed from birth, although it was nevertheless true in the study of Linden and Bender (2002) that these children had “a strong risk for developmental problems, particularly for learning disabilities . . . [albeit that] these problems were not often severe.” There may, however, be an increased risk for psychosis in childhood and adulthood (Kumra et al. 1998). More prenatally ascertained cases are being found through noninvasive prenatal screenings, and this has provided both a need and an opportunity for follow-up of the children so diagnosed. A pioneering clinic in Colorado, the eXtraordinarY Kids Clinic, providing a multidisciplinary management for children and adolescents with a sex chromosome aneuploidy, has been well received by parents, and it may offer a useful model (Tartaglia et al. 2015). Figure 22–10 shows developmental data on the children in the eXtraordinarY Babies Study. As well as the unbiased prenatally ascertained data, we now have (essentially) unbiased adult-derived information, with the genomic study of very large population-based cohorts giving insight into the long-term health and wellbeing of people with sex chromosome aneuploidy. Davis et al. (2025) utilized genomic and clinical data from the Million Veteran’s Program in the USA to identify more than 1500 men with 47,XXY or 47,XYY. Of these, only 26% of XXY men and remarkable low of just 1% of XYY men had been diagnosed clinically. The study results showed that these men had successfully served in the military, with similar performance metrics to their XY peers, but were more likely to have a history of a broad range of medical and psychological conditions. Similar data come from 356 men with XXY or XYY identified from the UK Biobank, which CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  717 identified surprisingly similar disease risks shared across XYY and XXY men, including increased risk of type 2 diabetes, atherosclerosis, and venous thrombosis (Zhao et al. 2022). Compared to XY men, men with sex chromosome trisomy were more likely to be taller, less likely to have a university or college degree, and more likely to live without a partner. With the foregoing narrative in mind, what should parents say to others, if they decide to continue the pregnancy, after the prenatal diagnosis of an SCT? Should the family know, should they tell friends, and should school personnel be aware? And when should the child learn about his or her chromosomal condition? Linden et al. (2002) have considered these questions, and in general make a case for openness within the family, but see no need, indeed potential disadvantage, for those outside to be told. We next outline the predicted outlook for the more commonly encountered sex chromosome aneuploidies. Attention is paid mostly to gonadal function and to intellectual and social development. Figure 22–10.  Developmental Progress in Children with a Sex Chromosome Trisomy. Notes: Data were collected as part of the eXtraordinarY Babies Study, a prospective natural history of developmental and health trajectories in a prenatally identified sample of infants with sex chromosome trisomy (SCT). In all SCT conditions, compared with normative data, there was increased variability and a later median age of skill development across multiple gross motor and expressive language milestones. However, results also show significant overlap with the general pediatric population, suggesting that for many children with prenatally identified SCT, early milestones present within, or close to, the expected timeline. Source: From T Thompson et al., Quantifying the spectrum of early motor and language milestones in sex chromosome trisomy, MedRxiv in press 2025. Courtesy N Tartaglia, and with the permission of the Cold Spring Harbor Laboratory.

718  REPRODUCTIVE CYTOGENETICS XXY (KLINEFELTER SYNDROME) Almost certainly, the child becomes an infertile adult, although in recent times testicular sperm extraction with in vitro fertilization (IVF) has enabled a small number of men to become fathers (p. 475). Some have used gamete donation from a father or brother. Penile size is usually normal; the testes will be small. Androgen deficiency can be managed by replacement therapy with testosterone.22 It may be that treatment induces a more masculine body habitus, improved self-esteem, vitality, ability to concentrate, and sexual interest (Nielsen 1990; Winter 1990). Gynecomastia may be present, transiently, in some 50%; if it persists, it can be treated surgically. As an average statement, verbal IQ is reduced by some 18 points, and performance IQ by 11 points (Leggett et al. 2010). Learning difficulty at school is to be expected. Of 13 XXY boys studied by Walzer et al. (1990), 11 had persistent reading and spelling problems. Bender et al. (1993) note that a deficit in verbal fluency and reading is “the most homogeneous and consistent cognitive impairment found in any sex chromosome abnormality group,” and this may reflect a specific dysfunction of the left cerebral hemisphere. Specific characteristics included a lowered level of motor activity, a pliant disposition, and a cautious approach to new situations; thus, in the classroom setting, they are perceived as “low-key children, well liked by their teachers, and presenting few behavioral management problems.” Leggett et al. conclude that these boys “do not usually have major problems with social interaction and adaptation, although they may be timid and unassertive.” Speculatively, the neural substrate of this passivity may reside in an underdevelopment of the amygdala, a brain nucleus that underpins aspects of social processing (Patwardhan et al. 2002). Six Danish XXY boys were followed from birth to age 15–19 years by Nielsen and Wohlert (1991), and all but one needed remedial teaching. Their career plans were carpenter, draftsman, gardener, unskilled laborer, mechanic, and undecided. Stewart et al. (1990) comment that “XXY boys are unlikely to reach a level of personal and social development that is consistent with their family background.” Ratcliffe (1999) commented upon a rate of psychiatric referral being above that of male controls (26% cf. 9%), with the neurotic score (not the antisocial score) being higher. (She also notes anecdotal mention of men from a Klinefelter clinic with professions including physician, engineer, minister, and accountant.) In a summary of psychosocial adaptation from several studies, recurring adjectives to describe the XXY personality were shy, immature, restrained, and reserved. In the Denver study, 11 young adults with XXY “appeared to have met the demands of early adulthood with fair success, although slightly less well than did their siblings”; they appeared to have a diminished insight into their own psychology (Bender et al. 1999). Their mean IQ of 91 compared with 109 in normal male sibling controls. We have noted above the ameliorative effect of growing up in a stable and supportive family. In a cohort of 934 Danish XXY males, the incidence of criminal convictions was 1.4 times that of controls, but this difference disappeared after adjusting for socioeconomic indices (Stochholm et al. 2012). The origin of the extra X chromosome in XXY males may be maternal (60%) or paternal (40%), and Larsen et al. (2024) addressed the question of whether parent-of- origin influences might be a source of phenotypic variability within KS. In fact, deep 22 A role for testosterone therapy in infancy to mimic the normal “mini-puberty” is controversial, and not routinely recommended (Aksglaede et al. 2020) CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  719 phenotyping in 58 XXY males, while identifying minor neuroanatomical differences, found no evidence of a substantial parent-of-origin effect on cognition, psychopathology or behavior. White et al. (2023) addressed the question “What should we tell prospective parents?” and identified six key counseling points (Table 22–8). XXX Physical development of the XXX female is generally unremarkable, although there is a tendency toward tallness. Gross and fine motor skills are likely to be somewhat impaired, and children are awkward and poorly coordinated. Pubertal development and fertility appear for the most part uncompromised. In a very few, genitourinary malformations (ovarian, uterine, renal, bladder) are recorded, of which the karyotype may or may not have been causal (Haverty et al. 2004; Linden and Bender 2002). It is the neural substrate in which the important vulnerability applies (Otter et al. 2010). Thus, major concerns in childhood relate to intelligence and language development and poor self-confidence, and, in adulthood, psychosocial maladjustment and, occasionally, frank psychiatric disease. Full scale and verbal IQ is, on average, reduced by some 10–20 points. Language comprehension and use of speech are impaired in over half the cases. Learning difficulty is likely, and many will benefit from additional remedial teaching, but few require education outside the mainstream. In one small study of 11 girls, nine needed special education intervention, and one was placed in a class for disabled children (Bender et al. 1993). While girls who had been diagnosed prenatally do better than those ascertained post-natally (as naturally is to be expected), it remains true that their neurocognitive capacity is somewhat compromised (Wigby et al. 2016). They are likely to be more vulnerable if there is a concomitant CNV of reduced penetrance (González-Del Angel et al. 2023). Harmon et al. (1998) and Bender et al. (1999) reported a longer follow-up in these young women, into adolescence and young adulthood, and documented difficult Table 22–8.  Key Counseling Points Following Prenatal Diagnosis of Klinefelter Syndrome 1.  Genetic counseling by a clinician familiar with the issues and complexities of KS should be offered. 2.  A review by a pediatric endocrinologist during the mini‐puberty stage is recommended to assess the clinical +/− biochemical profile of each infant. 3.  Testosterone replacement therapy during adolescence is the standard treatment for prevention of the sequalae of hypogonadatrophic hypogonadism in individuals with KS. 4.  Discussions on fertility-preserving options should be introduced in late adolescence, depending on the individual maturity and wishes of the young person. 5.  KS is associated with cognitive, language, and learning disabilities, attention and executive functioning difficulties, as well as internalizing and externalizing behavioral and psychological disorders; however, the rate of these neurodevelopmental outcomes may be over‐reported due to ascertainment bias. Early diagnosis and intervention may improve outcomes. 6.  The focus of antenatal counseling with parents emphasizes the role of appropriate screening and early management of evolving issues as the key to improved clinical outcomes. Source: M White et al., Klinefelter Syndrome: What should we tell prospective parents? Prenat Diagn 43:240–249, 2023.

720  REPRODUCTIVE CYTOGENETICS adaptation to the stresses of life. On a measure of social adjustment (in work, leisure, family, marital, parental), the XXX women scored significantly less well than their sisters. Their mean IQ was 82 (cf. sisters, 103). However, Ratcliffe (1999) described most XXX young women in the Edinburgh survey as “physically attractive, and displaying a common-sense attitude that counterbalanced their low educational achievements” (and relieved to be free of the pressure they had felt while at school). The observations in the similar study of Rovet et al. (1995) were more promising, although, as Harmon et al. point out, this was a group from a higher socioeconomic stratum, and presumably both genetic and environmental factors would have been more favorable. An XXX girl who might otherwise have had an IQ of 130 can yet do well despite a reduction to 110; to the contrary, a drop from 90 to 70 would be a considerable handicap. Many counselors will know from their own experience how variable can be the phenotype. Davis et al. (2025) identified 61 XXX women enrolled in the Million Veterans Program, of whom 28% had been diagnosed clinically. The authors report a “more reassuring perspective to the trisomy X literature”: Compared to their XX peers, women with XXX were 6 cm taller and had lower income and employment status, but there was little minimal difference in terms of health23 and psychological concerns. In the UK Biobank, 110 women were identified with trisomy X (Tuke et al. 2019). They had had a normal menarche age, but went through menopause on average five years earlier than XX women; this did not seem to affect their fertility, as they had had a similar number of pregnancies and pregnancy losses as XX women. Nonetheless, 9% had experienced premature ovarian insufficiency (POI). They were found to have lower income, to be of below average cognitive ability, and more likely to live alone. XYY The multicenter prospective study documented in Evans et al. (1990) reviewed progress in 39 boys and young men. The particular physical attribute of the XYY male is increased stature. Sexual activity is normal, and fertility is apparently uncompromised. Motor proficiency may be impaired. While the IQ is in the normal range, it is usually lower than those of sibs or controls, and about half of XYY boys have a mild learning difficulty, and may display poor attentiveness and impulsivity in the classroom. There is an overlap in the cognitive profiles between individuals with Klinefelter syndrome and those with XYY syndrome, mainly characterized by deficits in executive function and language-related skills. The vignettes from the series of Ratcliffe et al. (1990) of 10 Scottish subjects who had left school give an idea of what XYY young men are capable of: One ran a market stall, two were chefs, and the others were a private in the army, a waiter, a supermarket assistant, a video shop assistant, a technician, a laborer, and one was training as a painter and decorator. In a cohort of children aged 8–16 years selected for the XYY karyotype having been diagnosed prenatally, and of higher socioeconomic status, a considerable range in academic ability was observed, with most coping satisfactorily, and IQs ranging from 100 to 147 (Linden and Bender 2002). Perhaps the major concern is in psychosocial adaptation. These boys can have a low frustration tolerance, and some are prone to temper tantrums in childhood, progressing to aggressive behavior in teenage, and may need help to learn to cope with this. They may find it difficult to “read” social situations, and antisocial behavior is more common 23 An increased risk was identified for kidney disease, congestive cardiac failure and glaucoma in women with 47,XXX. CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  721 (Ratcliffe 1999). The functioning of the family may be as much an ingredient as the karyotype in psychosocial development. Fryns et al. (1995) identified 50 XYY males among 98,725 patients referred for chromosomal analysis, and they note that this fraction of 50/98,725, approximately 0.05%, is very close to the newborn incidence; they thus drew a conclusion that the XYY phenotype differs little from the norm. They do, however, acknowledge a high (86%) risk for psychosocial pathology in those XYY males with concomitant borderline intelligence. In Ratcliffe’s follow-up report into adulthood, some disconcerting data are noted, not incongruent with the conclusions of Fryns et al. Psychiatric referrals were fivefold compared with male controls (47% cf. 9%), and the rate of criminal conviction was fourfold, the mean IQ of those convicted being lower than those who were not (although most offenses were minor and against property rather than persons). More reassuring data come from a cohort of 161 Danish XYY males, in which the incidence of criminal convictions was just 1.42 times that of controls, a difference which disappeared after adjusting for socioeconomic parameters (Stochholm et al. 2012). In the data of the UK Biobank, only one of 143 men with XYY had been diagnosed clinically, emphasizing the bias in clinically ascertained cases (Zhao et al. 2022). Men with XYY were, on average, 7.9 cm taller than their XY counterparts, were less likely to have completed tertiary education, had below average cognitive abilities, and were more likely to live alone. As noted above, increased risk was observed for a range of common and potentially preventable diseases, including type 2 diabetes and cardiovascular diseases. In the Million Veteran’s Study, Davis et al. (2024) also identified a surprisingly similar profile for common disease risk shared by XXY men and XYY men. The basis of these risks is unknown. 45,X (TURNER SYNDROME) Unlike the foregoing three aneuploidies, monosomy X has a very high in utero lethality, peaking at around 12–15 weeks gestation. Spontaneous abortion follows amniocentesis-detected 45,X in three-fourths of cases (Hook 1983). But some survive pregnancy and are born as infants with Turner syndrome. Robinson et al. (1990) note that “variability among 45,X girls is considerable; and precise predictions about any child’s prognosis are not possible.” They also emphasized that “a supportive environment that provides stimulation and encouragement is of considerable importance.” The traits following comprise the core phenotype: • Gonadal failure with infertility is almost certain (Lippe 1991). In the survey of Sutton et al. (2006), infertility was seen, by the women with Turner syndrome themselves, as the most concerning component of the phenotype. Classically, a spontaneous onset of puberty, with breast development and onset of menses, has been regarded as being very infrequent, although Pasquino et al. (1997) propose that the fraction who enter a spontaneous puberty may be as high as 9%, and they suggest that earlier figures may have been biased downward by a policy, previously, of not karyotyping short girls who had had an onset of menstruation. Most girls and women with Turner syndrome will require hormone replacement therapy for initiation of puberty, starting at 11-12 years old, continuing until the usual age of menopause (Gravholt et al. 2024). Spontaneous pregnancies are more likely in women who have a history of spontaneous menarche and a 45,X/46,XX karyotype.

722  REPRODUCTIVE CYTOGENETICS Childbearing via ovum donation may be successful in some cases. The utility of oöcyte cryopreservation in girls and women with Turner syndrome is unknown. • Stature will be short, and this is mostly due to haploinsufficiency of SHOX. In a study of adult Danish women with Turner syndrome, never having had growth hormone therapy, the average height (with standard deviation) was 147 cm ± 7 cm (4 feet 10 inches ± 2½ inches) (Gravholt and Weis Naeraa 1997), which may be slightly taller than in some other ethnic populations. Gravholt et al. report a deficit of about 20 cm compared with population mean. Growth hormone treatment results in a an approximately 1 cm gain in height for each year of treatment, such that most treated individuals will have a final height in the lower normal range. • Neuropsychological functioning is impaired. The average IQ is reduced compared to siblings. At long-term follow-up in the Denver cohort (Bender et al. 1999), nine young women with 45,X had a mean lower IQ (85) compared with normal female sibling controls (104). Their educational achievements were, however, better than those of the XXX women from the same study: Eight were high school graduates, and five had college degrees. In one notable case, Reiss et al. (1993) report monozygous twins, one non-mosaic 45,X and the other 46,XX, the former’s performance IQ being 18 points less than her sister, but the verbal IQs practically the same. In fact, girls with Turner syndrome appear possibly to have superior skills in some language domains compared to their 46,XX peers (Temple and Shephard 2012). Across the lifespan, development in the preschool years is largely normal, other than some delay in motor skills. At school age, a cognitive profile emerges characterized by relative strength in verbal abilities compared to non-verbal abilities. Vulnerabilities include handwriting, drawing, mathematics, attention, executive function (the ability to plan, organize, monitor, and execute multistep problem-solving processes), and visual-spatial appreciation with a lesser volume of the right parietal cerebral cortex (Gravholt et al. 2024). These cognitive differences continue into adolescence and adulthood. Difficulties initiating and maintaining social and intimate relationships are more common. • Certain physical differences are associated, of which the major are neck webbing and cardiovascular malformation, of which coarctation (narrowing) of the aorta is the classic one (Birjiniuk et al. 2023). In the setting of prenatal diagnosis, fetal echocardiography should be performed, with post-natal follow-up within the first three days of life. Renal ultrasound should be performed at diagnosis to identify congenital anomalies of the kidney and urinary tract. • Morbidity in adult life is increased and comprehensive management guidelines are available (Gravholt et al. 2024). Certain common diseases are more frequently seen: obesity, both insulin-dependent and insulin-resistant diabetes, hypothyroidism, heart disease, hypertension, stroke, and liver cirrhosis. Even in the absence of structural heart disease, lifelong cardiac surveillance is recommended due to the increased risk of aortic dissection. Weakness of the bones (osteoporosis) implies a risk for fracture. The 45,X karyotype accounts for less than half of Turner syndrome diagnoses, and there is a range of variant forms (Table 15–2). The delineation of karyotype-phenotype correlations has been challenging due to small numbers, clinical heterogeneity, and the role of mosaicism, but some general karyotype-phenotype associations have been established, and these are listed in Table 15–3.

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  723 There is a possibility that Y-chromosome material may be present, even if the karyotype is apparently non-mosaic 45,X. Huang et al. (2002) reviewed 74 cases of 45,X diagnosed prenatally, most having been ascertained via, or discovered with, abnormal fetal ultrasonography. Of six with normal ultrasonography, three showed a male genital phenotype. The explanations, upon more detailed analysis, were as follows: In one, a segment of Yp was translocated to a chromosome 14, shown on FISH with an SRY probe; and in the other two, there was low-level mosaicism for an idic(Y) marker. Apparently normal male children were born. Some women with Turner syndrome who are 45,X on karyotyping may actually show Y sequences on molecular study, and these women do have a greater risk for gonadoblastoma (Mendes et al. 1999). SEX CHROMOSOME POLYSOMY The clinical pictures of the karyotypes 48,XXXX, 48,XXXY, 48,XXYY, 48,XYYY, 48,XXYY, 48,XXXY, 49,XXXXX, 49,XXXXY, 49,XXXYY, 49,XXYYY, and 49,XYYYY are outlined in Linden et al. (1995), Visootsak and Graham (2006), and Demily et al. (2017). The phenotypes may resemble commonly identified traits in the sex chromosome trisomies, such as impairment of language skills, executive function, and social adaptation, but with increased severity of these core phenotypic features, and with the addition of delayed developmental milestones and variable intellectual compromise (Hong and Reiss 2014). While the authors’ comment is well taken that the current perception of the seriousness of phenotypic abnormality may have been overstated due to ascertainment bias, and indeed they describe normal (but low) IQs in some of the 2n = 48 karyotypes, it remains true that most have intellectual impairment and abnormal behavior (Cammarata et al. 1999). Nuchal thickening is a frequent prenatal observation in 49,XXXXY (Peitsidis et al. 2009). Behavior in the 49,XXXXY male may modestly improve with testosterone therapy (Samango-Sprouse et al. 2023). X AND Y CHROMOSOME MOSAICISM True mosaicism involving the sex chromosomes seen at prenatal diagnosis presents a challenge in interpretation, and skilled ultrasonography, with respect to external genital anatomy, is central in determining the fetal gender. The problem is that the tissue analyzed at prenatal diagnosis may or may not reflect the distribution in the gonad. The presence of a Y chromosome in at least some gonadal tissue—or to be precise, the presence of the Y-borne SRY gene—will promote testicular development, which might or might not be complete, and which might or might not secrete male-inducing hormones. Thus, we may observe gender states from normal (although possibly infertile) female, through Turner-like female, genital ambiguity, mixed gonadal dysgenesis, even ovotesticular disorder of sex development (p. 763), to male with incomplete pubertal development, and to normal (although often infertile) male. XX/XY Mosaicism. At prenatal karyotyping, this is usually pseudomosaicism resulting from the growth of maternal cells in a 46,XY pregnancy (Worton and Stern 1984). (Obviously, such pseudomosaicism would normally be undetected if the fetus is female.) Level III XX/XY mosaicism, curiously enough, is most likely to indicate a phenotypically normal female fetus in which the XY source is unknown, particularly when the XX cells predominate. A male “vanished twin” is a theoretical possibility (Worton and Stern), and indeed a quite plausible explanation given the frequency with which a twin pregnancy prior to seven weeks is followed some months later by the birth of a singleton baby (Sampson and de Crespigny 1992). Analysis of placental membranes after delivery

724  REPRODUCTIVE CYTOGENETICS in a pregnancy from which one twin has “vanished” can reveal a fetus papyraceous or a remnant empty sac (Nerlich et al. 1992). One can imagine tissue of the (male) twin remnant having been, by chance, in the path of the amniocentesis needle that sampled cells from the remaining (female) fetus. A girl born following such a prenatal diagnosis (Hunter et al. 1982) was followed through to mid-adolescence, and her development was entirely normal (AGW Hunter, personal communication, 2002). A similar case was studied by I Hayes and A George (personal communication, 2009) with an XX:XY ratio of 90:10 on FISH of uncultured amniotic fluid, although non-mosaic 46,XX on cultured cells; ultrasonography indicated female external genital morphology. Following the birth of a normal girl, examination of seven sites from the placenta, and one site each from the cord and sac, all revealed a non-mosaic 46,XX complement, as did the cord blood sample. A true fetal XX/XY karyotype is rare indeed, and it is more likely due to the fusion of two conceptuses—that is, XX//XY chimerism (but other mechanisms exist; see Chapter 24). Presumably depending upon the gonadal distribution of XX and XY cells, the genital anatomy will be male, female, or in between. Malan et al. (2007) report XX// XY chimerism at prenatal diagnosis, the child (subjected to pelvic ultrasonography) proving to be an apparently normal girl. Ovotesticular DSD, with imperfect or ambiguous genital anatomy, has been recorded from an XX/XY amniocentesis result with the same karyotype demonstrated in the child (Amor et al. 1999; Chen et al. 2005b, 2006e; Malan et al. 2007). Yaron et al. (1999) had a case presenting at amniocentesis with normal male morphology on ultrasound. The XX/XY mosaicism was confirmed on a second amniocentesis, and in due course on the normal male newborn infant (including on genital skin). Amor and colleagues note the point that intellectual compromise is not to be anticipated. Hughes et al. (2006) provide guidelines on management for children with intersex conditions. Infertility is predicted; but remarkably enough, one XX//XY man has fathered a child following IVF with retrieved sperm (Sugawara et al. 2005). X/XY Mosaicism. Patients coming to medical attention with 45,X/46,XY mosaicism can be divided into three distinct groups based on their genitalia: typical female genitalia (variant of Turner syndrome), ambiguous genitalia (mixed gonadal dysgenesis), and normal male genitalia (Guzewicz et al. 2021). A risk for gonadal tumor applies (Müller et al. 1999). While post-natal cases often present with genital abnormality, at prenatal diagnosis a phenotypic male infant is the outcome in the considerable majority (90%–95%) of X/XY gestations—in other words, cases whose ascertainment was unbiased—and going through to birth (Hsu 1994; Huang et al. 2002). Fertility may be compromised, however, and other medical and neurodevelopmental manifestations of Turner syndrome such as short stature and cardiovascular defects may occasionally be present. We lack long-term follow-up data on prenatally ascertained cases (Alkhunaizi et al. 2024). Of 14 pathology studies on fetuses post-termination in Chang et al. (1990), two had ovotestes and one a “precancerous” gonadal lesion. In a registry study that included mostly post-natally diagnosed cases, Stochholm et al. (2024) found that mortality rates were doubled in males and quadrupled in females with X/XY, and called for further research into the long-term health of these patients. X/XX Mosaicism. Huang et al. (2002) reported their experience with 17 cases of X/ XX mosaicism at amniocentesis. The ratios of X to XX cells ranged from 2:23 to 12:3. One case with IUGR (ratio 6:12) terminated in stillbirth, while the remaining 16 had

CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  725 normal ultrasonography. Of the eight cases continuing to term and for which information was available, two liveborn babies had the features of Turner syndrome (ratios 7:10 and 3:14), with the mosaicism confirmed post-natally in one of these. The remaining six (ratios ranging from 3:15 to 12:8) “reportedly had a normal female phenotype.” To quote Huang et al., “The percentage of 45,X cells in amniocytes does not seem to be an indicator of pregnancy outcome, as there was considerable overlap between cases with normal and abnormal outcome.” In a unique case of a monozygous twin pregnancy, one fetus showed nuchal swelling and the other appeared normal (Gilbert et al. 2002). Fetal blood sampling showed low-grade 45,X[2]‌/46,XX[23] mosaicism in the former and a normal 46,XX karyotype in the latter, in contrast to post-natal skin fibroblast karyotyping results of non-mosaic 45,X and 45,X[2]/46,XX[78], respectively. Tokita and Sybert (2016) followed up 23 females with prenatally diagnosed 45,X/ 46,XX mosaicism, and noted the importance of accurate counseling in the context of increased detection of this karyotype by noninvasive prenatal testing. Follow-up was until mean age 11 years (range 0.1–27 years), and the mean percent aneuploidy was 42% at prenatal diagnosis and 23% in post-natal blood. Structural heart defects were documented in six females (26%), renal pathology in four (17%), and thyroid dysfunction in three (13%). No 45,X/46,XX female had formal IQ assessment, but persistent learning difficulties were reported in three (13%). Growth was comparable to population norms. Of the six patients older than 16 years, all had completed secondary school, and all had undergone spontaneous puberty. Results were compared with a cohort of 59 females with post-natally ascertained 45,X/46,XX mosaicism. Compared to the prenatally ascertained cohort, post-natally ascertained 45,X/46,XX females had shorter stature and a higher percentage aneuploidy in peripheral blood (40%); they also had a higher frequency of heart defects (39%), renal pathology (43%), and primary amenorrhea (50%). Combined data from both prenatally and post-natally ascertained cohorts suggest that the higher levels of percentage aneuploidy (on blood karyotype) were associated with an increased risk of congenital heart disease and a decreased chance of spontaneous menses, but not with the presence of other complications. Insight into the long-term outlook well into adult age is gleaned from the UK Biobank data, and this offers some reassurance. 186 women with 45,X/46,XX mosaicism were included (Tuke et al. 2019). Most had not been diagnosed clinically, but presumably some would have been detected at prenatal diagnosis, had it been performed. This group of women had normal intelligence and household income, went through menarche and menopause at an average age, had an average number of children, were not at increased risk of pregnancy loss, and there was no evidence of increased risk of cardiac complications. The only difference noted compared to XX women was a 4 cm reduction in average height. Similarly, a number of healthy pregnant women have only been diagnosed themselves with X/XY mosaicism after receiving an abnormal cfDNA result (Steinfort et al. 2024). X/XX/XXX, X/XXX, and XXX/XX Mosaicism. One reported case of X/XX/XXX mosaicism illustrates the difficulty in extrapolating the distribution of cell types from one tissue to another (Schwartz and Raffel 1992). Amniocentesis gave the proportions 16:64:20, respectively. Cord blood gave similar findings, although in placental tissue (chorion), the percentages were 2:57:41. The baby appeared normal. Huang et al. (2002) reported a case each of X/XXX and X/XX/XXX mosaicism diagnosed at amniocentesis, the former pregnancy producing a newborn with features of Turner syndrome, and the other a normal female.

726  REPRODUCTIVE CYTOGENETICS As might be expected, the cardiovascular, renal and skeletal features of X/XXX are less severe than in non-mosaic monosomy X, and spontaneous puberty and pregnancy are more common (Tang et al. 2019). Sybert (2002) reviewed hers and others’ data and concluded that ~60% of girls with X/XX/XXX and X/XXX could be predicted to have short stature and that “it is fair to suggest that residual ovarian function is possible, and to caution that premature ovarian failure is common.” Klamut et al. (2024) noted that some features of Turner syndrome were more common in with X/XXX than with X/XX, and attributed this to the higher proportion of 45,X cells seen in the X/XXX girls. XXY/XY Mosaicism. Up to 20% of Klinefelter syndrome are mosaic, and these men are, on average, more androgenized and have higher rates of fertility compared to their non-mosaic counterparts (Samplaski et al. 2014). X/XYY and X/XY/XYY Mosaicism. The X/XYY and X/XY/XYY mosaic states are (necessarily) abnormal in post-natally ascertained cases, but prenatally diagnosed cases have consistently manifested an apparently normal male genital phenotype, albeit that the mosaicism may be confirmed in the child subsequently born (Pettenati et al. 1991; Hsu 1994). Presumably according to the distribution of X and XYY tissues, the gender in X/XYY mosaicism can be of either sex, or there can be ambiguity, these three states documented in one of the earliest reviews (Mulcahy et al. 1977). There is a tumor risk, and gonadoblastoma was identified at gonadectomy in a virilized female with mixed gonadal dysgenesis (Gibbons et al. 1999). Infertility is likely, but it may be treatable (Dale et al. 2002). XXY/XX Mosaicism. XXY/XX is a rare karyotype that shows extreme clinical heterogeneity ranging from no abnormal phenotype to multiple physical and behavioral symptoms, and ambiguous genitalia with ovotestes (Guess et al. 2024). The phenotype presumably reflects the proportion and distribution of the XXY and XX cell lines. STRUCTURALLY ABNORMAL SEX CHROMOSOME X Chromosome Deletion. The possibility of an inherited X-autosome translocation should be checked by doing the mother’s karyotype; it may transpire that she has the same karyotype.24 Cytogenetically visible X chromosome deletions in the female, 46,X,del(Xp) or 46,X,del(Xq), predict the possibility, but not the certainty, of an incomplete form of Turner syndrome and/or premature ovarian failure (Chapter 15). Brown et al. (2001) describe a mother, of tall stature (5 feet 10 inches), having a prenatal diagnosis of del(X)(q22q26); she herself had the same karyotype, and “the parents took comfort in the observation that in the mother the deletion had no apparent phenotypic effect.” A normal baby girl was born. Mother and daughter showed completely skewed X-inactivation, the abnormal X being consistently inactive. In the male, the 46,Y,del(X) state would be nonviable for all but the very smallest cytogenetically visible deletions, and major abnormality would be probable for those pregnancies that might be viable. X Chromosome Duplication. De novo X chromosome duplications in the female, 46,X,dup(X), may determine a nil, minor, or major phenotypic impairment accordingly, as the pattern of X-inactivation may or may not be protective and if a functional disomy is not prevented (Chapter 15). Zhang et al. (1997) provide detail according to the extent and site of the duplication in a review of post-natally diagnosed cases. Women with an 24 X-autosome and Y-autosome rearrangements are discussed above, under De Novo Rearrangements, Apparently Balanced, and Unbalanced, respectively. CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  727 isodicentric X chromosome with Xq deletion are monosomic for genetic material distal to the Xq breakpoint and trisomic for the Xp arm. The phenotype typically includes primary or secondary amenorrhea, and other Turner syndrome features are lacking unless there is mosaicism for a 45,X cell line (Tsai et al. 2006). Normality has been reported with respect to an isodicentric X, idic(X)(q27) comprising practically a double copy of the X, identified prenatally with the abnormal chromosome being late replicating, and indeed one such child was “academically advanced and enrolled in a gifted and talented program.” In contrast, isodicentric X chromosomes with Xp deletion are associated with a Turner syndrome phenotype. In the male, functional disomy for the duplicated segment would likely cause severe defects, often lethal in utero; this is likely the scenario in the case in the family described in Figure 22–11 with a concomitant del Xp and dup Xq. X-Y Translocation. The most common form of the t(X;Y) has the X breakpoint in the distal short arm, at Xp22.3 or distal to Xp22, and the Y breakpoint at Yq11 or Yq12. The t(X;Y) arises during spermatogenesis by recombination due to highly similar sequences at these loci. The other, intact sex chromosome may be an X or a Y chromosome, and the two states differ as follows. 46,X,der(X)t(X;Y). A de novo X-Y translocation would be expected to herald a female child, who will likely have short stature due to SHOX haploinsufficiency but otherwise normal pubertal development and fertility, with a 50% risk to pass the t(X;Y) to male or female offspring (Daghsni et al. 2025). Intelligence is typically normal, but Figure 22–11.  A Del/Dup X Chromosome in a Mother and her Fetus. Notes: The mother was of short stature (150 cm), but otherwise normal. She had had four miscarriages, in one of which, a male, major fetal malformation (holoprosencephaly) had been identified. Ultrasonography in the current pregnancy had discovered multiple malformation. Amniocentesis gave a male result with an add(Xp22.2). Microarray of amniocyte DNA discovered 46,X,delXp22.33p22.32 and dupXp27.1q28. She herself had the same del/dup X chromosome. The fetal malformation likely reflected a functional distal Xq disomy including the MECP2 locus, and a distal Xp monosomy, whereas the mother had been protected by favorable inactivation. Source: From Q Lin et al., Prenatal detection and molecular cytogenetic characterization of Xp deletion and Xq duplication: a case report and literature review, BMC Med Genomics 17:57, 2024. Courtesy Q Lin, and with the permission of Springer Nature.

728  REPRODUCTIVE CYTOGENETICS learning difficulties have been described in some affected females. The Yq segment present of the der(X) is primarily composed of heterochromatin with little or no effect. 46,Y,der(X)t(X;Y). If the intact sex chromosome is the Y, the child is expected to be male. An Xp deletion with size between 1.5 Mb and 8.5 Mb results in a contiguous gene deletion syndrome. Severity depends on the genes deleted, which can be assessed by microarray. For small Xp deletions, the phenotype is confined to short stature and infertility. A more extensive loss of loci might determine a nullisomy that would cause physical and intellectual disability, or be lethal in utero. Other rare types include dicentric X;Y translocations, and der(X) and der(Y) chromosomes with a range of p and q arm breakpoints on X and Y (Hsu 1994). The phenotypes are male if SRY is present, and otherwise female. Infertility is typical, and in the male, short stature. In the der(Y) case, in which there may be an effect of functional X disomy, genital anomaly and other malformation is common, as is intellectual disability. A detailed case is described in Ghosh et al. (2008) in which the recognition of an ultrasound brain anomaly at 21 weeks led to amniocentesis with the discovery of a de novo 46,X,der(Y)t(X;Y)(p22.13;q11.23). The Yqh region was replaced by Xp material, which thus existed in the functionally disomic state. Other Abnormal X Chromosomes. X chromosome abnormalities are characteristically seen in the mosaic state, the other cell line typically being 45,X (and see Chapter 15). Mosaicism with a large ring X or an Xq isochromosome, 45,X/46,r(X) and 45,X/46,X,i(Xq) respectively, would lead to variant Turner syndrome. An Xp isochromosome, i(Xp), would probably always be lethal because there would be a functional Xp trisomy (Lebo et al. 1999). In an X inversion, there may be gonadal insufficiency in the otherwise normal female; and gonadal insufficiency may likewise accompany the de novo intrachromosomal insertion X, ins(X) (Dar et al. 1988; Dahoun et al. 1990). The “tiny ring X” syndrome is discussed on p. 480; a severe phenotype would mostly be the prediction from prenatal diagnosis, but exceptions exist, with a Turner-like picture or, in one extraordinary case, a normal male outcome (Turner et al. 2000; Chen et al. (2006d). Y Isochromosome. The least rare Y isochromosome (or isodicentric Y) is the 46,X,i(Yq), in which the essential imbalance is a double dose of Yp material and absence of some or most of Yq.25 As reviewed in Chapter 15 (p. 481), the condition may be seen in both non-mosaic and (more usually) mosaic form, the latter with a 45,X cell line. The phenotype in post-natally identified cases has ranged from sterile but otherwise normal male, through female with gonadal dysgenesis, to actual genital ambiguity (Bruyère et al. 2006; DesGroseilliers et al. 2006). In contrast, the outlook from unbiased (i.e., not following an abnormal ultrasound) prenatal diagnosis is markedly in favor of normal male physical development, albeit that infertility will be very probable, and indeed practically certain. If ultrasonography indicates male genitalia, a normal male phenotype is to be anticipated. Willis et al. (2006) reviewed 15 cases, with follow-up from four months to nine years: All but one had presented as normal males, and “development has been normal in all cases where follow-up was reported.” A similarly optimistic interpretation comes 25 An interesting question, not entirely theoretical in the present context, is what extrapolation, if any, can be made from the XYY syndrome, in which there is a double dose of Yp but of course also of Yq material. Neas et al. (2005) suggest that trisomy for the pseudoautosomal region PAR1 might lie behind aspects of the cognitive phenotype in the XXX and in some i(Y) karyotypes; and the same might apply to XYY. CHROMOSOME ABNORMALITIES DETECTED AT PRENATAL DIAGNOSIS  729 from Bruyère et al. (2006): In a series of 12 cases from these authors, all nine in which diagnosis had been unbiased and the pregnancies continued, led to births of normal males and normal development in those who were further followed up. Although a question about cognitive development is not entirely settled (Tuck-Muller et al. 1995; Neas et al. 2005) and few reports give follow-up into adolescence or adulthood, at least anecdotally many do well. Y Ring Chromosome. Layman et al. (2009) report their own cases and review the 45,X/46,X,r(Y) karyotype, as identified in males in whom testes were descended. Variable short stature and gonadal failure were typical (and see p. 483). These authors note the confounding factors, in terms of predicting phenotype at prenatal diagnosis, of the bias toward genital abnormality in post-natally identified infants, versus the frequent lack of follow-up in apparently normal males following a prenatal diagnosis, leading to a bias in the other direction. As for normal Yqh variation identified at prenatal diagnosis, this is reviewed in Cotter and Norton (2005). Microarray will not recognize this. The Y;15 variant is mentioned on p. 174.

23 PREIMPLANTATION GENETIC TESTING CHROMOSOMAL PREIMPLANTATION genetic testing is done in the setting of in vitro fertilization, and in principle it enables a chromosomally normal embryo to be transferred to the uterus at either a few days post-conception or, more usually, following frozen transfer at a subsequent menstrual cycle. Thus, for those facing a high genetic risk, that risk can be bypassed, and the prospect of pregnancy termination for the reason of genetic abnormality can be minimized. Advances in the late 20th century in the fields of in vitro fertilization (IVF), human embryo culture, manipulation and cryopreservation, and molecular genetics, set the stage for the development of preimplantation genetic testing (PGT). From an essentially research-based exercise in a very few laboratories in the early 1990s, it has progressed to being, in the 2020s, a diagnostic procedure routinely available through most IVF clinics. PGT is applied in two main settings: for the diagnosis of chromosome abnormalities, and for the detection of a Mendelian condition.1 Initially, the two categories were distinguished by the methodology applied—FISH in the former, DNA testing in the latter—but DNA-based methodologies are now the mainstay in both categories. Indeed, techniques have converged such that it is possible to accomplish chromosomal and Mendelian PGT with a single test. CIRCUMSTANCES IN WHICH CHROMOSOMAL PREIMPLANTATION GENETIC TESTING MAY BE APPROPRIATE Carriers of Balanced Structural Rearrangements (PGT-SR) The carrier of a balanced chromosomal rearrangement typically has a high risk to produce unbalanced embryos, as discussed at length in previous chapters. Particularly in the context of an unfortunate reproductive history, often with several miscarriages or with one or more terminations following prenatal diagnosis of an unbalanced fetal karyotype, the attraction of PGT-SR is obvious: Only an embryo with a normal or balanced chromosomal constitution is transferred, with the expectation that, for each embryo transferred, there is a good chance of ongoing normal pregnancy. A related benefit for those otherwise undergoing IVF is that they avoid the time, effort, cost, and disappointment of transferring embryos which would not have been viable. Where there exists a risk of a liveborn child with an unbalanced karyotype, this risk is, in principle, almost eliminated. In practice, the two main categories are reciprocal (rcp) and Robertsonian 1 PGT can also be used to detect and quantify mutations in the mitochondrial genome.

732  REPRODUCTIVE CYTOGENETICS (rob) translocations, but chromosomal PGT-SR can likewise be applied to pericentric inversions and complex chromosome rearrangements. Preimplantation Genetic Testing for Aneuploidy Screening (PGT-A) PGT-A is the practice of evaluating embryos for chromosome aneuploidy in chromosomally normal parents. The rationale behind PGT-A is that a substantial number of human embryos are aneuploid, and certain patient populations, including couples of advanced maternal age or with histories of recurrent miscarriage and repeated implantation failure, may be predisposed to producing aneuploid embryos. Thus, aneuploidy is assumed to be an important reason that many IVF pregnancies fail. Logically, therefore, assessing the chromosome status of embryos, and discarding screened embryos that are aneuploid, should improve the proportion of transferred embryos that result in pregnancy. PGT-A is also used in the setting of a couple who have had a previously chromosomally abnormal pregnancy, or a previous liveborn child with Down syndrome or other aneuploidy. Here, the risk of recurrence of the specific aneuploidy is usually low, and the benefit of PGT-A, at least in terms of risk reduction, is small. For couples who are naturally fertile, it is questionable whether such a small risk reduction justifies the use of IVF (with its associated small risks to both mother and baby); but for couples who are already using IVF because of infertility, the addition of PGT-A may be considered appropriate, and it may give the couple more confidence of a successful outcome. Gender Selection Gender diagnosis at PGT can be achieved via PGT-A, and may be appropriate in the context of a sex-related genetic risk, whether Mendelian or non-Mendelian—an example of the latter being autism (Amor and Cameron 2008). In practice, the use of sex-selection PGT has declined, as direct gene testing has become feasible for many sex-linked disorders. A notable exception is for males affected by an X-linked disease, who wish to avoid transmitting the faulty gene to their (female) offspring, a circumstance in which sex selection can substitute perfectly for a direct gene test. EMBRYOLOGY PROCEDURES Those who make the decision to embark upon chromosomal PGT will need to enroll (if not already) in an IVF program. In IVF, hormone treatment is given to stimulate the ovaries to produce a large number2 of oöcytes in a single stimulated IVF cycle. Ovum “pick-up” is conducted by transvaginal endoscopy under ultrasound guidance. The ova are collected, and fertilization is achieved using either conventional IVF or intracytoplasmic sperm injection (ICSI). 2 Preferably double figures; but if only a few, this does not per se betoken an increased risk for trisomy (Honorato et al. 2017). PREIMPLANTATION GENETIC TESTING  733 On day 1, about 18 hours after exposure to sperm, the oöcytes are checked for the presence of two pronuclei and two polar bodies3 as evidence that fertilization4 has occurred. Shortly thereafter, syngamy will occur, and during the next 48 hours, the first few mitoses will have produced cleavage-stage embryos of six to eight cells. In the next 72 hours, these few cells give rise to 200–300 daughter cells, and this is the blastocyst stage (Figure 20–2). In the early days of IVF, embryos were transferred at the cleavage stage, on the assumption that the uterus would provide the best environment for the survival of the embryo, while also allowing assessment of embryo morphology, in order to select the best quality embryos for transfer. But during the 2010s, there was a shift in practice, with transfer now taking place on day 5 or 6, by which time the embryos have become blastocysts. This shift was facilitated by the improvement in embryo culture techniques, and supported by the observation that blastocysts appeared to have better implantation potential compared with cleavage-stage embryos; an observation subsequently confirmed in a Cochrane review, which showed that embryos transferred at the blastocyst stage were about 25% more likely to result in a live birth (Glujovsky et al. 2022). This picture likely reflects a “self-selection” whereby only the most viable embryos develop into blastocysts. There are three types of PGT biopsy, done at three sequential stages of gametic and embryonic development: polar body biopsy, blastomere biopsy, and blastocyst (trophectoderm) biopsy. In practice, the great majority of PGT biopsies are done at the blastocyst stage. Polar Body Analysis Polar body (PB) genetic analysis (satisfyingly requiring recall of some elementary facts of biology) has been used for PGT (or “preconception diagnosis”) in a few laboratories, and legal or logistic constraints against PGT in some jurisdictions have propelled interest (Verpoest et al. 2018; Oberle et al. 2024). The process of biopsy is illustrated in Figure 23–1, with the first and second polar bodies (PB1 and PB2) aspirated from beneath the “shell” of the zona pellucida. Both polar bodies are required for diagnosis, and they can be removed simultaneously or sequentially (Harper 2018). PB analysis allows a focus on the more vulnerable gamete—that is, the ovum—since the great majority of 3 IVF laboratories use morphological assessment for the presence of two pronuclei as a “pseudo-genetic test” for ploidy. The rationale is that embryos with one pronucleus (2% of all embryos) are at high risk of haploidy, which is incompatible with embryo development; and embryos with three pronuclei (5% of all embryos) are at high risk of triploidy and its incumbent risk of miscarriage. However, Capalbo et al. (2024) argue that pronuclei number is a poor predictor of embryo ploidy, and that many livebirths have followed the transfer of embryos with 0, 1, or 3 pronuclei. They suggest that rather than discarding abnormally pronucleated embryos, PGT-A be applied to them as a more accurate ploidy assay, allowing those embryos that are euploid to be “rescued” for clinical use. 4 Since fertilization in vitro can be observed as it actually happens, the fine detail of the process can be appreciated. The first act is penetration of the ovum by the sperm. To the embryologist, this is only the prelude to conception; the true moment of conception is the point at which the male and female pronuclei fuse, their chromosomes aligning on a common metaphase plate (“syngamy”). Once that event has taken place, the zygote has come into existence. At the first mitosis, it loses that name and becomes, in IVF parlance, a “cleavage-stage embryo,” or simply an embryo.

734  REPRODUCTIVE CYTOGENETICS segregation errors leading to aneuploidy occur here, due either to nondisjunction, or to premature pre-division of sister chromatids. Disomic or nullisomic gametes could be identified and thus excluded from fertilization. By way of example, imagine that the asterisked gametocyte at meiosis I in Figure 23–3 is PB1, and that the two chromosomes shown within it are chromosome 22s—that is, PB1 is disomic 22. The “empty” gamete to the right, therefore, would be a nullisomic 22 oöcyte, and thus of course to be discarded. Figure 23–2 shows an actual example, and while FISH is no longer used for this purpose, the colored probes provide an Figure 23–1.  The process of polar body biopsy. The egg (left) is held in place by a suction pipette, which is applied directly to the zona pellucida (the “shell” that invests the egg itself, seen in faint darker circular outline). The first and second polar bodies are located in the space between the zona pellucida and the egg’s cell membrane. The egg is manipulated so that the polar bodies are at the 1 to 2 o’clock position. The pipette (right) has entered a laser-cut hole in the zona pellucida, and the two polar bodies have been aspirated into its lumen. Source: From M Montag et al., Polar body biopsy: A viable alternative to preimplantation genetic testing and screening, Reprod Biomed Online 18 Suppl 1:6–11, 2009. Courtesy M Montag, and with the permission of Elsevier. Figure 23–2.  FISH Analysis of the First Polar Body. Normal Haploidy for Five Chromosomes (left), and Disomy 22 (right). Notes: The 5-color FISH probes recognize chromosomes 13 (red), 16 (light blue), 18 (dark blue), 21 (green), and 22 (yellow). With each chromosome being two-chromatid, each has two colored spots. The arrows in the polar body at right indicate two no. 22 chromosomes, enabling an inference that the associated ovum is nullisomic 22. Source: From F Vialard et al., Evidence of a high proportion of premature unbalanced separation of sister chromatids in the first polar bodies of women of advanced age, Hum Reprod 21:1172–1178, 2006. Courtesy F Vialard, and with the permission of Oxford University Press. PREIMPLANTATION GENETIC TESTING  735 intuitive illustration. Preferably, the interpretation requires analysis of both PB1 and PB2 (Verpoest et al. 2018). The reader may determine, on study of Figure 23–4 with respect to pre-division of sister chromatids at meiosis, why analysis of just one PB could in some instances mislead. The asterisked cell at meiosis I could be, say, PB1, and tested as disomic, suggesting nullisomy of the egg; and yet the ovum alongside is actually monosomic normal. At meiosis II, the nullisomic PB2 (the fourth gamete of this row) allows the full picture to be interpreted: the opposite aneuploid results from PB1 and PB2, taken together, indicate a euploid egg. Potential advantages of PB biopsy are that specimens are obtained earlier, and that the biopsy is less disruptive to the embryo. On the other hand, PB is expensive, given the need to test both polar bodies in every embryo, although new technologies may be more cost-effective (Oberle et al. 2024). It is time-consuming, and temporal constraints will stretch laboratory resources. The windows of opportunity for performing biopsy are narrow. Abnormalities arising in paternal meiosis, estimated to represent less than 10% of aneuploidies (Templado et al. 2013), will not be detected, nor will unbalanced structural rearrangements of paternal origin. PB biopsy will not detect mosaic post-zygotic aneuploidies. While PGT-A by microarray analysis of polar bodies may not increase live birth rate, there are less embryo transfers, with less requirement for cryopreservation and fewer miscarriages (Verpoest et al. 2018). The procedure may yet come to have a place (Madritsch et al. 2024). Cleavage-Stage Biopsy (Day 3 Biopsy of a Blastomere) In the early days of PGT, the usual approach was to remove one, or at most two,5 cells (blastomeres) from each embryo on day 3. This required a hole to be made in the zona Figure 23–3.  Nondisjunction Detected via Polar Body Analysis (Classical Scenario). 5 The live-birth rate is less when two cells are removed (De Vos et al. 2009). 736  REPRODUCTIVE CYTOGENETICS pellucida (which has not yet been cast off ), the cells being extracted by very gentle suction. These cells were then subject to genetic analysis to determine whether or not they had a normal/balanced chromosome constitution, or an aneuploidy, unbalanced rearrangement, or other genetic abnormality. One6 embryo shown to be chromosomally normal/balanced was then transferred to the uterus, on day 4 or 5, and with good fortune would develop into a normal infant. Blastocyst-Stage Biopsy (Days 5–7 Biopsy of Trophectoderm) This is now the usual PGT timeframe, and numerous studies have demonstrated its superior performance compared to day-3 biopsy. The embryo is incubated through the morula (~day 4) and then the blastocyst (days 5–7) stages. By now, cell number has increased, and differentiation between inner cell mass (which develops into the embryo Figure 23–4.  Nondisjunction Detected via Polar Body Analysis (Predivision Scenario). 6 Elective single embryo transfer (eSET) is currently the IVF preference (ESHRE Guideline Group on the Number of Embryos to Transfer, 2024). The attraction of PGT, according to this criterion, is obvious. PREIMPLANTATION GENETIC TESTING  737 proper) and trophoblast (which develops into the placenta) has begun, allowing sampling to be focused on the trophectoderm (in a tissue-origin sense, a very early chorionic villus sampling). The biopsy procedure involves making a hole in the zona pellucida and allowing a small part of the trophoblastic lining of the blastocele cavity to herniate through (“assisted hatching”); part of this tiny bulge can be excised by laser or teased away by manipulation (Figures 23–5, 23–6). Blastocysts are more robust than early-stage Figure 23–5.  The process of in vitro fertilization (IVF) (with or without intracytoplasmic sperm injection, ICSI) and preimplantation genetic testing (PGT) at the blastocyst stage. (a) Oöcytes are obtained from the female, and sperm from the male (by testicular aspiration, if necessary). (b) Oöcytes and sperm are mixed in vitro; or, single sperm are injected into an oöcyte (ICSI). (c) Syngamy, the fusion of male and female pronuclei, occurs. (d) The day-3 cleavage embryo. At the blastocyst stage, days 5 – 7. (e), trophectoderm cells are removed (Figure 23-6), and these cells are then subject to cytogenomic analysis. (f) Normal (or balanced) embryos are chosen for later transfer to the uterus, normally following cryopreservation.

738  REPRODUCTIVE CYTOGENETICS embryos and have a greater tolerance for biopsy (although one consequence of the blastocyst approach is that little time is left to obtain genetic results; hence embryos usually need to be frozen after biopsy, forgoing the option of fresh transfer). There is considerable attrition over the timeframe of blastocyst development, with only about half of cleavage-stage embryos surviving to become a blastocyst (Clouston et al. 2002), and this attrition appears preferentially to target aneuploid embryos. Adler et al. (2014), comparing aneuploidy rates from the day-3 cleavage embryo to the day 5 blastocyst, found that the overall proportion of euploid embryos increased from 23% to 32%, indicating loss of at least some aneuploidies by day 5; nevertheless, among the aneuploid embryos that do actually reach day 5 or 6, a wide range of trisomy and monosomy is observed (Rodriguez-Purata et al. 2015) (Figure 20–12). The appearance of a blastocyst may allow an inference concerning ploidy status. Thus, “good-looking” blastocysts may have a better chance than those of “poor” appearance; but while those of better morphology may have an advantage over the “lesser lookers,” the latter still have a fair chance of succeeding (Forman 2017; Irani et al. 2017). Improved predictive models have been developed using artificial intelligence to assess embryo morphology, combined with maternal age (Sfakianoudis et al. 2022). Although this approach has an impressive accuracy of about 70% for the prediction of ploidy, this is not enough to replace PGT-A (Barnes et al. 2023). Figure 23–6.  Blastocyst Biopsy. Notes: The process of blastocyst biopsy, at days 5 – 7. The blastocyst is held in place by the suction pipette on the left, which is applied directly to the zona pellucida (the “shell” that invests the blastocyst), faintly visible surrounding the blastocyst. Trophectoderm has herniated through a laser-generated hole, in the zona pellucida, visible in this view at 3 o’clock; the inner cell mass is seen within the blastocyst, at 10 o’clock. Suction will be applied through the biopsy pipette (right), and typically five or six cells from the trophectoderm gently teased off.

PREIMPLANTATION GENETIC TESTING  739 Noninvasive PGT Cell-free DNA, secreted by the blastocyst into the culture fluid medium as the cells grow, reflects the genome of the embryo and is analyzable: this is “noninvasive PGT” (NI-PGT) (Xu et al. 2016). A similar, but slightly more invasive, approach is to use an ICSI needle to aspirate a sample of blastocele fluid, which includes both embryonic cells and cell-free DNA (Campos and Nel-Thelmaat 2024). While NI-PGT methodologies have proposed advantages of being easier, cheaper, and lower-risk compared to embryo biopsy, concerns about accuracy and interpretation of results have thus far precluded widespread adoption (Moustakli et al. 2024). GENETIC ANALYSIS OF BIOPSIED CELLS Various cytogenetic and molecular methods have been used to analyze the biopsied material, with molecular approaches in principle superior, in that copy-number analysis of all 24 chromosomes is achievable and with higher resolution. Cost of analysis is particularly relevant in PGT (in contrast to prenatal testing or postnatal chromosome analysis), because of the need to test multiple embryos for each IVF cycle. Metaphase Chromosome Analysis. The first, and simplest, analysis of chromosomes in human embryos was done by spreading and counting stained metaphase chromosomes on glass slides (Angell et al. 1983). This method provided the earliest evidence of the extent of aneuploidy in human embryos; but the metaphase spreads were too few in number, and too poor in quality, ever to be of use clinically. Fluorescence In Situ Hybridization. FISH was the methodology that launched PGT for chromosome abnormalities as a practicable possibility. Chromosome-specific FISH probes applied to the interphase nuclei of human embryos allowed a rapid and targeted assessment of a subset of chromosomes, or of chromosome regions. Although only about five multicolor FISH probes could be applied at one time, the first set of probes could be washed off and a second round of probes used, allowing analysis of about 10 chromosome loci per cell. For the detection of structural rearrangements, the choice of FISH probes for a particular PGT had to consider all possible segregation outcomes. By observing the number of colored spots in the nucleus of a blastomere removed from the IVF embryo, the chromosome complement could be deduced. Although PGT-A using 5–9 probe FISH was initially adopted with enthusiasm by many IVF centers, results were disappointing, and a meta-analysis compiling nine published randomized controlled trials concluded that PGT-A using FISH should not be recommended (Mastenbroek et al. 2011). With the benefit of hindsight, the major limitation of PGT-A using FISH could be seen as the technology itself, and the validity of much of the substantial body of literature on FISH PGT is now questioned. First, only a minority of chromosomes could be tested by FISH. Second, even for the chromosomes that were tested, it was inevitable that a proportion of results would be inaccurate (Wells et al. 2008). Single-cell FISH cannot achieve 100% of resolvable signal on every chromosomal target. With 10 or more signals per cell to be interpreted, the error rate compounded. This is not normally a problem for other FISH applications, where many nuclei or metaphases are available for study. In PGT, it is a critical limitation. 740  REPRODUCTIVE CYTOGENETICS The disconcerting adverse outcomes were a wastage of normal embryos (diagnosed as monosomies), and the misdiagnosis of embryos with a trisomy as normal. Comparative Genomic Hybridization. The first analysis for the full 24-chromosome copy number was achieved with comparative genomic hybridization (CGH), a technique using whole genome amplified DNA from the embryo and DNA from a karyotypically normal individual, the two samples labeled with different-colored fluorochromes. The two labeled DNAs are then co-hybridized to normal metaphase spreads, and the relative intensity of the two fluorochromes is analyzed. Although this technique could lead to the birth of a healthy child (Wilton et al. 2001), it was too complex and time-consuming for routine application in the clinic. Array Comparative Genomic Hybridization. The more practicable use of CGH was enabled by the hybridization of DNA samples to microarrays rather than to metaphase spreads (hence, array-CGH). Array-CGH analysis is performed by scanning and imaging the array, and measuring the relative intensity of the two hybridization signals (Chapter 2). Single Nucleotide Polymorphism Microarray. Single nucleotide polymorphism (SNP) arrays provide both copy-number information and genotyping (Chapter 2). Compared to array-CGH, more complex bioinformatic analysis can be undertaken: this may include the incorporation of parental genotypes, enabling the detection of smaller copy-number variants, and copy-number neutral abnormalities such as uniparental disomy. Data from SNP-arrays can also be used for linkage analysis in single-gene disorders, allowing aneuploidy and single-gene PGT detection to be undertaken in the same test (Handyside et al. 2010; Natesan et al. 2014). Another potential advantage of SNP-array is that, for translocation carriers, it is possible to distinguish embryos with a normal karyotype from those that have the balanced translocation, via the analysis of family haplotypes (although many translocation patients will not have enough embryos available to allow the luxury of discarding those carrying the translocation). SNP arrays may also be preferred for PGT-SR when the size of the segments predicted to be unbalanced is below the resolution of next-generation sequencing. Next-Generation Sequencing (NGS). In next-generation sequencing (short-read sequencing), DNA samples from an embryo biopsy, following whole genome amplification, are fragmented, and the nucleotide sequence is determined (Yin et al. 2013). The number of fragments sequenced from each chromosome (or part of each chromosome) should be proportional to copy number, thus enabling inference of monosomy/disomy/ trisomy of the embryo. Diagnostic accuracy is high (Kung et al. 2015). A comparable methodology using short-read sequencing as a molecular counter, in order to detect copy-number changes (rather than sequence changes), is utilized in NIPT (Chapter 21). Although low-pass NGS can be used for both PGT-A and PGT-SR, for translocations in which both translocated segments are very small—these being difficult to detect by PGT-NGS in their unbalanced form and also carrying the greatest risk of a viable abnormal pregnancy—SNP microarray may be the preferred technology. Nanopore Sequencing. Nanopore-based single-molecule short-read sequencing has been applied to trophectoderm samples for aneuploidy screening. The results are comparable in accuracy compared with NGS-based PGT, and with the advantage of a faster result (Wei et al. 2018; Tan et al. 2023). This method, termed short-read transpore rapid karyotyping (STORK), has also been applied to CVS and amniocentesis samples, and products of conception (Wei et al. 2022). Nanopore sequencing can also generate long

PREIMPLANTATION GENETIC TESTING  741 reads, and this can be used in PGT-SR for translocation couples in order to generate sequence reads across translocation breakpoints, providing another means to identify balanced carrier embryos (Chow et al. 2020; Madjunkova et al. 2020). As noted above, nanopore sequencing has been applied to polar body biopsies for PGT-A, with results comparable to microarray analysis (Oberle et al. 2024). The Problem of Mosaicism Multi-cell trophectoderm biopsy for PGT has led to the discovery that mosaic aneuploidy—wherein samples contain a mixture of euploid and aneuploid cells—occurs much more frequently than in prenatal or postnatal cytogenetic analyses. PGT-A using NGS can detect mosaicism by identifying an intermediate chromosome copy number,7 down to a level of approximately 20%, which corresponds to one aneuploid cell in a 5-cell trophectoderm biopsy. The rate of mosaic diagnosis in clinical testing is estimated to be 2%–20% (American Society for Reproductive Medicine 2023), a surprisingly wide range that is attributed to different test methodologies, assays, reporting practices, and philosophies of individual laboratories. In some instances, detected mosaicism will be apparent rather than real, resulting from technical artifact. When real, in most cases the aneuploid cells will have arisen in the first mitotic cell divisions of embryos that were euploid at conception, the first few cell divisions being particularly vulnerable to segregation error (Currie et al. 2022). The embryonic genome is largely inactive at this stage, with the initial cell divisions being under the control of maternal RNAs and proteins (Braude et al. 1988). The mitotic checkpoints in cleavage-stage embryos are weak, and while this might be useful in facilitating the progress of these early mitoses, a consequence is that cells may divide before the chromosomes have properly replicated (McCoy et al. 2017). Further evidence for the underlying mechanisms comes from mouse studies, showing that the 4-to-8-cell division is particularly vulnerable to replication stress, leading to high rates of chromosome breakage and segmental aneuploidy (Takahashi et al. 2024). The six- to eight-cell embryo contains probably only one or two cells whose descendants will go on to form the inner cell mass and thus, eventually, the embryo proper and the fetus. It is assumed that, at least in surviving embryos, the mosaic aneuploid cells are excluded due to apoptosis or to being outcompeted—or, alternatively, are sequestered away from the inner cell mass and into the placenta, where they may present later as confined placental mosaicism (Chapter 22). Although occasionally some aneuploid cells will become incorporated into the embryo, this appears to be an uncommon occurrence; and in fact, the great majority of mosaic aneuploid embryos that become established as a pregnancy lead to the birth of a chromosomally normal child. In embryos that do not survive, it may be that the presence of aneuploid cells is the cause of embryo arrest (McCoy et al. 2023). 7 Note that the presence of mosaicism is not observed directly, but inferred from the presence of an intermediate copy number on NGS profile. Other possible explanations for an intermediate copy number are artifact, DNA amplification bias, contamination, mitotic state, and variation in embryo and culture and biopsy techniques. 742  REPRODUCTIVE CYTOGENETICS A landmark study in 2024 shed new light on the true extent of mosaicism, looking at every cell of blastocyst-stage embryos (Chavli et al. 2024). Remarkably, 82% of embryos contained at least some cells with numerical and/or structural abnormalities indicative of mosaic aneuploidy (Figure 23–7). Most of these embryos were euploid-aneuploid mosaic, with a range of numerical and structural abnormalities present. These findings represent a challenge of PGT-A, which relies on the genetics of the biopsied cell(s) being representative of the embryo as a whole (Figures 23–8); as such, the presence of mosaicism could mislead. Figure 23–7.  Chromosomal Status of Blastocysts. Notes: Percentages of human blastocysts that have normal, abnormal and mosaic results, as detected by single cell sequencing in 55 embryos. Source: Redrawn from EA Chavli et al., Single-cell DNA sequencing reveals a high incidence of chromosomal abnormalities in human blastocysts. J Clin Invest 2024:e174483, 2024. Courtesy EB Baart, and with the permission of the American Society for Clinical Investigation. Figure 23–8.  The Sensitivity of Blastocyst Biopsy, in the Detection of Mosaicism. Notes: On trophectoderm biopsy at PGT-A (PGT for aneuploidy) using standard DNA methodology (above), mosaicism of less than 20% would likely be missed. Research using single-cell analyses of the whole embryo (below) could reveal otherwise undetectable mosaicism. Source: From EA Chavli et al., Single-cell DNA sequencing reveals a high incidence of chromosomal abnormalities in human blastocysts. J Clin Invest 2024:e174483, 2024. Courtesy EB Baart, and with the permission of the American Society for Clinical Investigation.

PREIMPLANTATION GENETIC TESTING  743 PREIMPLANTATION GENETIC TESTING OUTCOMES PGT for Aneuploidy Every patient undergoing IVF has a risk to generate an aneuploid embryo. The risk varies according to individual factors, and these several factors may influence a decision in favor of PGT for Aneuploidy (PGT-A): advanced maternal age, recurrent implantation failure, recurrent pregnancy loss, unexplained infertility, and prior aneuploid conception. The aim of PGT-A is to improve the chance of a successful pregnancy by choosing only those embryos known to be euploid. Logically, it would seem that surely PGT-A would improve the odds. But the question is this: is PGT-A actually beneficial in practice? The ability of NGS-based PGT-A accurately to predict a successful delivery was assessed in a large prospective, blinded, nonselection study in which 484 blastocysts were biopsied and transferred, but with PGT-A being performed only after clinical outcome was known (Figure 23–9). Out of 102 PGT-A aneuploid embryo transfers, 40% resulted in a biochemical pregnancy, but none ended in a live birth. On the other hand, from 312 PGT-A euploid transfers, 65% resulted in sustained implantation or delivery. The comparison is stark. While this study demonstrates the prognostic ability of NGS-based PGT-A, we yet await large randomized controlled trials to arrive at a definitive answer regarding the clinical benefit of PGT-A. Early research suggests an age-related difference. For older women, aged ≥35 years, PGT-A indeed appears to be effective in reducing the risk of miscarriage and results in a significantly higher birth rate. But no such benefit is observed in women younger than 35 (Simopoulou et al. 2021). Figure 23–9.  Outcomes Following Euploid and Aneuploid Embryo Transfer. Notes: The two timelines reflect outcomes after the implantation at IVF of a euploid cf. an aneuploid embryo. The distinction in terms of outcome is very evident. Source: From AW Tiegs et al., A multicenter, prospective, blinded, nonselection study evaluating the predictive value of an aneuploid diagnosis using a targeted next-generation sequencing-based preimplantation genetic testing for aneuploidy assay and impact of biopsy, Fertil Steril 115:627–637, 2021. Courtesy AW Tiegs, and with the permission of Elsevier. 744  REPRODUCTIVE CYTOGENETICS There are several reasons why PGT-A might not be as effective as logically expected. PGT-A remains a screening test, and as such is subject to a risk of false-positive and false-negative results. False-negative PGT-A would lead to embryos screened as euploid resulting in miscarriages, and to ongoing pregnancies in which aneuploidy is detected. Such observations, while rare (<1% of embryos transferred), are on record (Tiegs et al. 2016; Greco et al. 2023). There are two main sources of false-positive and false-negative PGT-A results. First, PGT provides many opportunities for technical errors arising through the processes of embryo biopsy, whole genome amplification, and genetic analysis. Second, the phenomenon of mosaicism appears to be common, and potentially undermines the strategy of PGT-A by generating multiple pathways to misdiagnosis. False-positive PGT-A results have been proven upon rebiopsy studies, but mostly are limited to mosaic and segmental aneuploidies. Non-mosaic full aneuploidies, which are typically of meiotic origin, are almost always confirmed at rebiopsy of the embryo. Similarly, if blastocysts testing non-mosaic aneuploid are transferred, only a very small fraction go through to successful normal live birth. At the request of parents, Barad et al. (2022) transferred 76 embryos that had been diagnosed at PGT-A with full chromosome aneuploidies. Five pregnancies resulted, but only one continued to live birth.8 For fully euploid embryos, rebiopsy studies have found little evidence of false-negative results, with euploid PGT-A results confirmed in >99% of cases (Kim et al. 2022). OUTCOMES FOLLOWING THE TRANSFER OF MOSAIC EMBRYOS While a conservative approach, based on prenatal and pediatric experience with mosaicism, might have been to discard mosaic embryos, in fact the PGT-A literature allows otherwise, and two key facts are now understood. First, many healthy babies have been born following the transfer of mosaic aneuploid blastocysts; and second, pregnancies following transfer of mosaic embryos do not have an increased risk of untoward fetal or neonatal outcome. Indeed, if a mosaic embryo is transferred and a pregnancy results, in <1% is the mosaicism ever confirmed in the fetus or neonate (American Society for Reproductive Medicine 2023). In the case of mosaic aneuploidies and non-mosaic segmental aneuploidies, between 40% and 80% are confirmed at rebiopsy (Victor et al. 2019; Girardi et al. 2020; Wu et al. 2021; Cascante et al. 2023). Aneuploidies that are both mosaic and segmental are particularly unlikely to be confirmed at rebiopsy, and in fact when these embryos are transferred, the livebirth rate is indistinguishable from that of euploid embryos (Grkovic et al. 2022).9 Healthy livebirths have also been observed following the transfer of embryos diagnosed as non-mosaic segmental aneuploidy, but with a lower livebirth rate of about 25% (Besser et al. 2024), indicating that some of these embryos were actually mosaic. Mosaicism does, however, have an impact upon the reproductive potential of the embryo: implantation rates are less, and miscarriage more so, compared to euploid embryos (Viotti et al. 2021). Risks vary with the type and extent of mosaicism (Figure 23–10). As noted above, embryos with low-level mosaicism of a single chromosome 8 This embryo had been diagnosed at PGT-A with 46,XX,+14,-18. 9 Yet, mosaic segmental aneuploidies at PGT-A are not always a benign finding. Barad et al. (2022) transferred an embryo diagnosed at PGT with mosaic dup(10)(q11.21-q21.1), and the outcome was the birth of a male infant non-mosaic for the dup(10), and with coarctation of the aorta.

PREIMPLANTATION GENETIC TESTING  745 segment appear to have similar outcomes to euploid embryos. Conversely, these four factors are associated with worse clinical outcomes (Ma et al. 2022): • Mosaicism level >50% • Whole chromosomes involved • Mosaicism for three or more chromosomes • Mosaicism for both gain and loss In light of this knowledge, couples can have the option of transferring mosaic embryos, but with an awareness of possible adverse outcomes including risks to the pregnancy and to the child to be born. The American Society for Reproductive Medicine (2023) provides guidance in this matter and has put forth these advisories that can guide decision-making: • A lower percentage of mosaicism is associated with better rates of implantation and ongoing pregnancy. • Higher-level mosaicism may predict a risk of adverse outcome, due to misdiagnosis of a true aneuploid embryo. • Segmental mosaic aneuploidies have better ongoing pregnancy rates compared to whole chromosome mosaics. • Although intuitively, mosaicism for one of the classic survivable aneuploidies (trisomies 13, 18, 21) might seem to pose a higher risk, in fact this appears not to be so. • Monosomic and trisomic mosaicisms have similar pregnancy and miscarriage rates. • Embryos with mosaic aneuploidies of three or more chromosomes do less well than those with one or two. • Whether embryos with segmental mosaic aneuploidy versus those with whole chromosome mosaics differ in terms of risk, is not yet clarified. • A theoretical risk of uniparental disomy (UPD) or confirmed placental mosaicism (CPM) in pregnancies from a mosaic embryo, in fact, seems not to apply. Figure 23–10.  Outcomes after Transfer of Euploid and Mosaic Embryos. Notes: Clinical outcomes of the euploid embryos are compared with mosaic groups, sorted according to mosaic level, and type (segmental cf. full aneuploidies). segm = segmental; chr = chromosome. Source: Adapted from M Viotti et al., Using outcome data from one thousand mosaic embryo transfers to formulate an embryo ranking system for clinical use, Fertil Steril 115:1212–1224, 2021. Courtesy M Viotti, and with the permission of Elsevier 746  REPRODUCTIVE CYTOGENETICS In practice, mosaic embryos, as do the euploid, upon transfer follow a path to one of three outcomes: a failure of implantation, a miscarriage, or a live birth with no apparent abnormal phenotype. There does not seem to be an increased risk, vis-à-vis the euploid embryo, of pregnancy complications or abnormal prenatal or postnatal karyotypes. Viotti et al. (2023) collated data for more than 2,000 mosaic embryo transfers, including birth data for 488 infants, and results were largely reassuring. Birth outcomes showed no difference in birth weight or gestational age compared to euploid controls, and of the 488 livebirths, only one had a serious birth defect (a congenital heart defect). In three out of 250 cases (1.2%) where testing was performed in the pregnancy, there were fetal findings consistent with the previous PGT mosaic result (Viotti et al. 2023; Greco et al. 2023). The first was an embryo with low-level mosaicism for deletion of 1p36.33-p31.1, a finding that was confirmed at amniocentesis and in brain tissue from the products of conception after pregnancy termination. The second was an embryo with low-level mosaicism for trisomy 21 that was confirmed at CVS and amniocentesis, and the pregnancy terminated after the additional detection of ultrasound abnormalities. The third was an embryo with low-level mosaicism for duplication of 4q32.2q34.3 that was confirmed at CVS, with the pregnancy continuing to term, resulting in an infant with no birth defect. In addition to these three cases, Viotti et al. (2023) also identified two cases with prenatal mosaic results that were different from the original one detected with PGT-A, and assumed to be unrelated. Notably, this overall rate of 2% mosaic results at prenatal testing is not substantially different from the 1% to 2% mosaic results detected at CVS in the general population (Chapter 22). Thus, in summary, the somewhat surprising conclusion is this: Mosaicism of the embryo, rather often, seems not to matter. PGT for Structural Rearrangement (PGT-SR) We have data from many thousands of reported PGT-SR cycles to call upon. Naturally, the risk that an embryo will be abnormal at PGT, not yet having been subject to survival pressure, is substantially—and often very substantially—higher than during the pregnancy or at birth. Reciprocal Translocations. Xie et al. (2022) analyzed segregation fractions from the study of 10,846 blastocysts, from 2,253 couples, and tested by NGS as follows: normal (alternate) 46%, adjacent-1 31%, adjacent-2 12%, and 3:1, 4:0, and others all ~5%. Thus, slightly more than half were chromosomally imbalanced. These figures are averages derived from aggregated data, but of course each rcp translocation will have its own unique segregation pattern (as discussed at length in Chapter 5). Factors that reduce the proportion of alternate segregation are female sex of rcp carrier and involvement of an acrocentric chromosome (Figure 23–11). Alternate segregation is also favored if the translocated segment is of relatively large size, as a proportion of the chromosome arm (Xie et al. 2022). Robertsonian Translocations. Segregation outcomes for rob translocations are more favorable than with rcp. Dang et al. (2023) tested 3,423 blastocysts with NGS, from 763 couples, and found the following: normal (alternate) 70%, adjacent 29%, and 3:0 in 0.7%. Alternate segregation was more frequent in male (82%) than in female (60%) carriers, the rates not varying with carrier age. Outcomes were similar across all rob translocations. One or more transferrable embryos were identified in 82% of cycles, and 57% of transferred embryos resulted in a live birth.

PREIMPLANTATION GENETIC TESTING  747 Insertional Translocations. Zhang et al. (2023c) evaluated PGT-SR outcomes in 187 embryos from 23 couples, one of whom was an insertional translocation carrier, and found 30% of embryos were suitable for transfer (normal/balanced). Predictors of higher percentage of normal/balanced embryos were interchromosomal insertions (compared to intrachromosomal insertions) and male sex of carrier parent. Pericentric Inversions. Xie et al. (2019) analyzed PGT-SR results for 379 embryos from couples, one of whom carried a pericentric inversion. Overall, 17% of blastocysts were unbalanced for the inversion, but unsurprisingly (Chapter 9), this varied with the size of the inversion segment. When the inverted segment was <57% of the chromosome length, only 8% were unbalanced, compared to 34% unbalanced when the inverted segment was >57% of the chromosome length. Overall, 51% of transferrable blastocysts resulted in a live birth. Paracentric Inversions. As noted in Chapter 9, for paracentric inversions, unbalanced recombinant chromosomes are either acentric or dicentric, and typically do not lead to viable offspring. PGT-SR is rarely performed for paracentric inversions, but Xie et al. (2019) were able to study 197 blastocysts from couples, one of whom carried a paracentric inversion, and found unbalanced segregants in just 3.6%, representing a small fraction of the 30% of embryos that were aneuploid. After transfer, 63% of normal/balanced blastocysts resulted in a live birth. Complex Balanced Chromosome Rearrangements (BCRs). When there is more than one balanced rearrangement, or a rearrangement involving more than two chromosomes, normal/balanced embryos are predicted to be scarce. Yet, there is still cause for some modest optimism that suitable embryos might be found. Li et al. (2020) collated PGT results for more than 200 embryos from couples with complex chromosome rearrangements, and found 12% were euploid (see also Chapter 10, Figure 10–13). Both partners of a couple carrying different balanced rearrangements is another challenging circumstance. Liu et al. (2021) report a couple where each partner carried a different reciprocal translocation: 46, XX,t(10;16)(q25.2;q12.1) and 46,XY,t(9;14)(p21.1;q12). They were advised of a low chance of success, but against the odds, two out of three Figure 23–11.  Alternate Segregation in Blastocysts. Notes: These data show the impact of sex of reciprocal translocation carrier, and involvement of an acrocentric chromosome, on the proportion of embryos with alternate (balanced) segregation. Source: From data in P Xie et al., Risk factors affecting alternate segregation in blastocysts from preimplantation genetic testing cycles of autosomal reciprocal translocations, Front Genet 13:880208, 2022.

748  REPRODUCTIVE CYTOGENETICS blastocysts tested were found to be balanced/normal, and transfer of one of these embryos led to the birth of a healthy boy. GENETIC COUNSELING PGT is sufficiently complicated, not to mention expensive, that it will not usually be the first option for fertile couples wishing to avoid the birth of a child with a chromosomal disorder. High-risk scenarios, such as the person being a translocation carrier, might, however, warrant consideration sooner rather than later. Carriers of structural rearrangements may view access to the procedure as empowering, but equally may find the process stressful; discarding an embryo with an unbalanced rearrangement, having had a child with that condition, may raise uncomfortable ambiguities (Karatas et al. 2010). For infertile couples (whether or not there is a chromosomal basis of the infertility) who require an IVF procedure to conceive, advice about a place for PGT-A will need to be tempered by a continuing understanding of the biology of early human embryo development, and the fact that PGT is, as yet, an imperfect tool for predicting the health and viability of an embryo. For couples presenting for PGT on the basis that one of them carries a chromosomal rearrangement, a number of points need to be raised. The Reasons for Choosing PGT Some couples may have had conventional prenatal diagnosis with successive terminations of pregnancies due to a high-risk translocation, and be unwilling to face this prospect again. It may be difficult to distinguish between two possibilities. Either, this may have been a run of bad luck, with an optimistic outlook for the next pregnancy as a realistic possibility and therefore allowing the counselor to suggest a further natural attempt; or, the series of abnormal pregnancies may reflect a strong predisposition of that translocation to generate unbalanced gametes. Avoiding the possibility of termination following conventional prenatal diagnosis is, for those who have had that experience, a strong motivation (Lavery et al. 2002). The Limited Success Rate As discussed above, many IVF/PGT procedures do not produce the desired end result of a “take-home baby,” and the figures for PGT pregnancies are fairly similar to those applying to all IVF. For couples who require IVF in order to conceive (e.g., when a male translocation carrier also has oligospermia), the benefits of PGT—avoidance of transfer of nonviable embryos and reduced risk of miscarriage—are likely to outweigh the disadvantages, of which financial cost may be prominent. In contrast, couples who would otherwise have no difficulty conceiving may struggle weighing the pros and cons of IVF with PGT, compared to natural conception and conventional prenatal diagnosis (Kanavakis and Traeger-Synodinos 2002). In this comparison, the chances of ultimately having a healthy child may be similar between the two pathways, but the challenges of the two choices are very different. One study that compared livebirth rates in translocation PREIMPLANTATION GENETIC TESTING  749 couples who chose PGT-SR, against those choosing natural conception, found no difference in the livebirth rate, although the risk of miscarriage was less with PGT-SR (Ikuma et al. 2015). The counselor can help the couple in their decision-making by encouraging them to consider the relative merits and challenges of PGT versus natural conception. As noted below, the reality is, for those who face the dual challenge of advanced maternal age and a BCR, the chance of success may be low. The Outlook at PGT-A For PGT-A, a realistic consideration has to be that possibly not a single euploid embryo will be obtained from an IVF stimulation cycle. The risk for this unfortunate circumstance is at its lowest for women in the age range later 20s through early 30s, but rises steeply from the late 30s—and, as logically expected, this risk is mirrored by a fall, according to age, in the chance of retrieving at least one euploid embryo (Figure 23–12). Maternal Age Considerations in PGT-SR Given that PGT-SR typically also incorporates PGT-A, the additional impact of de novo aneuploidy will influence the proportion of embryos available for transfer. Whereas the proportion of embryos unbalanced for the parental rearrangement is constant, independent of parental age, the proportion of de novo aneuploid embryos will naturally increase with advancing maternal age.10 In a population with median maternal age of 30 years and de novo aneuploidy rate of 35%, Zhou et al. (2024) found that, after considering both unbalanced segregants of the balanced rearrangement and de novo aneuploidy, the proportion of embryos suitable for transfer was 30% for rcp translocations, 46% for rob translocations, and 57% for inversions (Table 23–1). By comparison, in a slightly older cohort with median maternal age 33 years, the de novo aneuploidy rate was 45%, and the proportion of embryos available for transfer was reduced to 19% for rcp translocations, 33% for rob translocations and 36% for inversions (Oğur et al. 2023). In the same study, the proportion of embryos with de novo aneuploidy was 59% in women ≥35 years, compared to 41% in women <35 years. Hence, for women above age 35 years, the proportion of tested embryos that are suitable for transfer may be less than 20%, and a probable outcome of a PGT cycle may be that no embryo will be suitable for transfer. When suitable embryos are obtained, the outlook is considerably brighter regardless of maternal age, with around half of transferred euploid embryos leading to the birth of a healthy child (Oğur et al. 2023). The Question of Mosaicism The complexity and inherent imprecision of the interpretation of mosaic aneuploid findings are discussed above, and Moran et al. (2023) rehearse the challenges facing 10 Embryos from translocation carriers appear to have a similar rate of de novo aneuploidy as the general population; that is, there is not a significant interchromosomal effect (ICE).

750  REPRODUCTIVE CYTOGENETICS counselors who have the task of conveying these uncertain findings to their patients. It is important that patients should be counseled prior to their IVF cycle about the possibility of mosaic results, including their frequency, difficulties in interpretation, and options regarding transfer. Considering this information, some couples might elect to decline PGT-A. If one or more mosaic embryos are available for transfer, post-test counseling should discuss the question of lower implantation rates and higher miscarriage rates compared to euploid embryos, and the small (but poorly understood) risk of liveborn with aneuploidy (American Society for Reproductive Medicine 2023). In the event of an ongoing pregnancy, prenatal testing should be considered. Amniocentesis provides the Figure 23–12.  Maternal Age and Chance for Euploidy at PGT. Notes: The relationship is shown between maternal age and the probability (upper) that no euploid blastocysts will be available from a single IVF cycle, or (lower) that at least one euploid embryo (day 3 or day 5) will be retrieved. These shapes of these two curves reflect each other, and show that the best chances lie between the maternal ages of the later twenties to the early thirties. The odds are a little less favorable in younger women, and a lot less in older. Sources: From JM Franasiak et al., The nature of aneuploidy with increasing age of the female partner: a review of 15,169 consecutive trophectoderm biopsies evaluated with comprehensive chromosomal screening, Fertil Steril 101:656-663.e1, 2014, courtesy JM Franasiak, and with the permission of Elsevier; and from ZP Demko et al., Effects of maternal age on euploidy rates in a large cohort of embryos analyzed with 24-chromosome single-nucleotide polymorphism-based preimplantation genetic screening, Fertil Steril 105:1307–1313, 2016, courtesy ZP Demko, and with the permission of Elsevier. Table 23–1.  Carriers of Balanced Rearrangements, and Blastocyst Results from PGT-SR RCP TRANSLOCATION (N=2943) ROB TRANSLOCATION (N=927) INVERSION (n=220) COMPLEX BCR (n=39) INSERT TRANSLOCATION (n=187) Euploid 30% 46% 57% 46% 30% Unbalanced BCR alone 36% 19% 7% 36% 44% Unbalanced BCR and de novo aneuploid 16% 10% 6% 10% 9% De novo aneuploid alone 18% 24% 30% 8% 16% Total 100% 100% 100% 100% 100% Total unbalanced BCR 52% 29% 13% 46% 53% Total de novo aneuploid 34% 34% 36% 18% 25% Notes: Biopsied trophectoderm cells were amplified using whole genome amplification and next-generation sequencing in order to detect numerical and segmental chromosome abnormalities. Complex BCR is defined as a BCR involving three or more chromosomes or with three or more breakpoints. Data on type of inversion was not provided. Mosaic embryos are included in the de novo aneuploid fraction. Sources: Data of F. Zhou et al. Preimplantation genetic testing in couples with balanced chromosome rearrangement: a four-year period real-world retrospective cohort study. BMC Pregnancy Childbirth 24:86, 2024; and Z Zhang et al. Clinical outcomes in carriers of insertional translocation: a retrospective analysis of comprehensive chromosome screening results. F S Rep 5:55– 62, 2023. 752  REPRODUCTIVE CYTOGENETICS best representation of chromosome complement within fetal tissues, and therefore the best chance of detecting true fetal mosaicism, although being aware that even with amniocentesis, mosaicism may escape detection. CVS and NIPT may provide additional information, but it is possible that on very rare occasions confined trophoblast mosaicism might evolve into a confined placental mosaicism at CVS, and leave the question unanswered. UPD studies may be considered when the embryo mosaicism involves an imprintable chromosome. Follow-up in the Pregnancy Understandably, some couples will be unenthusiastic about an invasive procedure that could possibly put at risk the pregnancy in which there has been so much investment (Meschede et al. 1998). Nevertheless, couples need to be aware that chromosomal PGT cannot provide a “guarantee,” although the misdiagnosis rate, for whatever reason, is very low when good-quality embryos are transferred. Prenatal testing should be offered. Ultrasonography may be an acceptable, if imperfect, compromise, only proceeding to CVS or amniocentesis if anomalies are detected. Noninvasive prenatal testing (NIPT) can target full aneuploidies, but consultation with the laboratory will be required to determine if unbalanced segregants of a translocation are detectable (Flowers et al. 2020). Nature May Intervene A natural pregnancy may be achieved while the couple waits for the IVF/PGT preparations to be made. For example, the adjacent-2 karyotype shown in Figure 5–15 came from culture of the products of conception of this couple’s third miscarriage and no normal pregnancies, the woman being a t(13;16) carrier. The outlook did not seem very promising, and plans were being put in place for IVF; but the couple then reported a naturally conceived pregnancy in which amniocentesis showed a 46,XY karyotype. The Children Resulting Does PGT carry any risks to the embryo, pregnancy, or to the child? We may consider three categories of potential risk. First, there is a small risk of an embryo not surviving the biopsy process. With improvements in biopsy technique, this risk is now less than ½%; and although it is undoubtedly disappointing for a couple to lose an embryo in this way, the small risk is unlikely to deter couples from PGT. Second, might the biopsy procedure itself compromise the chance of success? In fact, this seems not to be so. In a review of nearly 500 PGT blastocyst transfers, Tiegs et al. (2021) found no difference in sustained implantation between the study group and an age-matched non-biopsied control group (48% vs. 46%). The third and most important question relates to the child born from a PGT pregnancy. It is well recognized that there are differences, albeit small, in health between children conceived by IVF (without PGT) and naturally conceived children. IVF-conceived babies are more likely to be of low birth weight and to be born pre-term, and perinatal mortality and birth defects are more frequent (Halliday 2007; Halliday et al. 2010).

PREIMPLANTATION GENETIC TESTING  753 Nevertheless, it is reassuring that, in the longer term, the health of adult “IVFlings” appears to be similar to that of their in vivo–conceived counterparts (Halliday et al. 2019). For IVF in general, the usual counseling is for an absolute increase of approximately 1% in the risk of congenital abnormalities, compared to background risk.11 The risk for Beckwith-Wiedemann syndrome in IVF babies generally is noted in Chapter 19, p. 551. And, it is encouraging that IVF babies from PGT have no greater risk for pregnancy complication, neonatal outcomes, and birth defects, than IVFlings generally (Alteri et al. 2023; Van Heertum et al. 2024). Thus we may conclude that, in fact, there is likely no detrimental effect of embryo biopsy per se. Further, their health and development as children lie within a normal range (Ginström Ernstad et al. 2023). Perhaps unsurprisingly, for a child in whom so much has been invested, PGT infants score well on a scale of “warmth/affection” (measured by observing how infants may be cuddled and kissed, and how positively and kindly spoken to) (Banerjee et al. 2008). 11 There is conflicting evidence regarding whether IVF conception increases the risk of chromosome abnormalities in offspring. If such a risk really exists, it is likely to be very low, and therefore IVF conception is not, in itself, an indication for chromosomal PGT.