10115510730290Birth Defects Res A Clin Mol TeratolBirth Defects Res. Part A Clin. Mol. Teratol.Birth defects research. Part A, Clinical and molecular teratology1542-07521542-076024265126451903410.1002/bdra.23195HHSPA709970ArticleExpanding Diagnostic Testing beyond Cytogenetics: Implications for Birth Defects Research and SurveillanceJacksonJodi M.Division of Birth Defects and Developmental Disabilities, National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention (CDC), Atlanta, GeorgiaDruschelCharlotte M.New York State Congenital Malformations Registry, New York State Department of Health, Troy, New YorkShapiraStuart K.Office of the Director, National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention (CDC), Atlanta, Georgia23720150611201311201329720159711726729

For decades, classical cytogenetic techniques that yielded a karyotype were the mainstay for identifying and characterizing the causes of certain genetic syndromes and birth defects. The capacity to identify chromosome anomalies expanded in the late 1980s and the 1990s with the development and maturation of fluorescent in situ hybridization (FISH) techniques to interrogate submicroscopic regions of the chromosomes for deletions or duplications. Beyond karyotypes and FISH, new technologies—chromosomal microarrays and next generation DNA sequencing—have markedly increased the number of birth defects and genetic syndromes that now have a known cause. These new testing techniques that can unambiguously confirm a diagnosis—as occurred previously, for example, with FISH for 22q11.2 deletions—will increase the specificity and sensitivity for classifying birth defects and improve prevalence estimates. Although these testing techniques improve the resolution of analysis of smaller and more complex chromosome and DNA anomalies, their interpretation can be problematic, particularly when the test results are of unknown significance. Broader usage of new prenatal screening technologies, such as testing for chromosome and DNA anomalies in cell-free fetal DNA, will likely impact prevalence estimates of certain birth defects included in surveillance systems. These new advancements in genetic testing can create challenges for birth defects surveillance and research programs in learning how to abstract, interpret, classify, store, and incorporate new findings into surveillance systems, as well as categorizing the data in epidemiological studies. Birth defects research and surveillance programs must be mindful of these new challenges and thoughtful in addressing them.

TRADITIONAL CYTOGENETICS: HISTORICAL BACKGROUND

Advances in the field of cytogenetics over the years have yielded an increased understanding of the causes of numerous syndromes, diseases, and structural birth defects. These advances have arisen based on the desire to understand human chromosomes and the information that they contain. Although the interest in chromosomes waxed and waned during the late 1800s and early 1900s, chromosomes have been studied in ever-increasingly detailed and revealing ways since the mid-1900s. The first recognizable drawings of human chromosomes were published in 1882 by Walther Flemming (Flemming, 1882), but it took until 1956 to realize that 46 was the correct number of chromosomes in human cells (Tjio and Levan, 1956). The subsequent development of banded karyotype analysis allowed for the identification of the aneuploidies responsible for Turner (monosomy X), Klinefelter (47,XXY), Down (trisomy 21), Edwards (trisomy 18), and Patau (trisomy 13) syndromes (Ford et al., 1959; Jacobs and Strongs, 1959; Lejeune et al., 1959; Edwards et al., 1960; Patau et al., 1960), as well as chromosomal translocations involved in cancer cytogenetics such as the Philadelphia chromosome (Nowell and Hungerford, 1960). Analysis of material from spontaneous abortions showed that over half of miscarriages are due to aneuploidies (Kajii et al., 1973; Boué et al., 1975; Carr and Gedeon, 1978). As cytogenetic analysis became commonplace, families of individuals affected by syndromes or birth defects could have cytogenetic testing whereby the results could refine the possibility of recurrence, thus informing future reproductive choices. Initially, all cytogenetic tests were done postnatally on cells derived from tissues such as blood, biopsies, bone marrow, and products of conception. Beginning around the late 1960s, the reproductive decision process could include an existing pregnancy because it became possible to diagnose chromosomal aberrations prenatally using the technique of amniocentesis and subsequent karyotype (Steele and Breg, 1966; Nadler, 1968).

GENETIC TESTING BEYOND TRADITIONAL CYTOGENETICS

Chromosome analysis expanded beyond banded karyotypes with the advent of fluorescent in situ hybridization (FISH) to diagnose submicroscopic chromosomal deletions and duplications (Langer et al., 1981; Langer-Safer et al., 1982), which together with polymerase chain reaction (PCR) (Saiki et al., 1985), became the mainstay technologies of modern day cytogenetics. However, recent genetic testing has begun to move beyond traditional cytogenetics. The American College of Medical Genetics and Genomics published guidelines in 2007 (Shaffer et al., 2007) regarding the clinical use of chromosomal microarrays (CMA) (an umbrella term encompassing both array-based comparative genomic hybridization [aCGH], which detects copy number variation, and single nucleotide polymorphism arrays), which has several advantages over FISH and traditional cytogenetic karyotyping. There is improved laboratory expediency with CMA as it allows for the simultaneous interrogation of hundreds of thousands of loci for potential duplication or deletion, which is not possible with FISH. aCGH has a resolution of 1 kilobase (Kb) while FISH can only detect changes approximately 100 Kb or larger and traditional karyotypes typically have a resolution of 5000 Kb or larger (Miller et al., 2010; Evangelidou et al., 2013). This improved resolution with aCGH facilitates the detection of many structural abnormalities that would have been missed using only a traditional karyotype; for example, in a recent study, aCGH detected losses and gains in approximately 20% of apparently balanced translocations previously analyzed by karyotype (Manning et al., 2010). Based on the advantages of aCGH over traditional karyotypes, Miller et al. (2010) recommended that CMA become the first-tier genetic test for patients with unexplained developmental disabilities, intellectual disabilities, autism spectrum disorder, or multiple congenital anomalies, while continuing to use karyotype and FISH first for individuals with suspected trisomies and for those with a history of recurrent miscarriage. A recent study concurred that CMA should be the first-tier test, but presented evidence that traditional cytogenetic analysis remains useful in some cases for detecting chromosomal mosaicism and characterizing structural chromosome rearrangements (Bi et al., 2013). Beginning in 2006, CMA has been performed prenatally with cells derived from amniocentesis or chorionic villus sampling (CVS) (reviewed in Evangelidou et al., 2013), although current recommendations are for it to be offered only as an optional screening tool with validation requiring karyotype and FISH analyses (American College of Obstetrics and Gynecology, 2009). In addition to CMA, another new technology changing the field of genetic diagnosis is next generation DNA sequencing (NGS), a massively parallel shotgun sequencing technique. NGS allows for the sequencing of whole genomes or whole exomes (the portion of the genome that codes for proteins) at a fraction of the cost with greatly increased coverage depth compared with first generation sequencing (Wetterstrand, 2013).

Perhaps the most significant recent development in genetic diagnosis, at least prenatally, is based on finding cell-free fetal DNA (cffDNA) in maternal blood in quantities substantial enough to be used to screen for aneuploidies. Although the discovery of cffDNA occurred in 1997 (Lo et al., 1997), it took the next ten years to develop the methods that would reliably differentiate between cffDNA and maternal cell-free DNA (Lo and Chiu, 2007, 2008), as fetal DNA represents just 10% of the cell-free DNA found in maternal plasma (Nygren et al., 2010). Numerous studies have used NGS to validate the use of cffDNA for prenatal detection of trisomy 13, 18, 21, and sex chromosome aneuploidies. The best results to date have shown a nearly 100% detection rate of trisomy 21 with a false-positive rate of less than 1%, and detection rates for trisomies 13 and 18 are now approaching those levels; however, it must be noted that these validation studies have been done in high risk pregnancies, not as population-based screenings (reviewed in Langlois et al., 2013). Because of this caveat and the varying detection rates, cffDNA analysis is currently being offered only as a screening test rather than as a diagnostic test, and results indicative of an aneuploidy require validation through karyotype analysis of cells collected by means of amniocentesis or CVS (Gregg et al., 2013). Further down the road is the prenatal use of NGS and cffDNA to obtain whole genome and exome fetal DNA sequences for prenatal diagnosis of a myriad of genetic conditions; however, this technology is still in the research phase and not yet clinically available (Kitzman et al., 2012).

INCORPORATING NEW GENETIC TESTING RESULTS INTO SURVEILLANCE ACTIVITIES: HOW MUCH INFORMATION IS TOO MUCH?

With each technological advancement in genetic testing, there have been challenges in learning how to interpret and incorporate the newly generated findings into clinical practice, as well as into surveillance activities and epidemiological studies. First and foremost, for those working in surveillance programs, it is not inherently obvious how to interpret the terminology used by laboratories to describe test results. Beyond standard karyotype nomenclature, which by itself has some degree of technical complexity, the terminology to describe FISH and CMA test results (ISCN Committee, 2013) might be even more perplexing to some who work in clinical medicine and population surveillance. Laboratory reports are often long and packed with detail describing the results, as well as the technology that was used to perform the tests. It can be challenging to determine what components of the report need to be abstracted and included in a surveillance record; surveillance programs will need to make decisions regarding the extent of data that will be collected and train those who report or abstract test results according to surveillance protocols. Another related issue is how to classify actual test results. For years, surveillance systems captured only two types of cytogenetic tests: prenatal and postnatal chromosome analysis. With the current assortment of new tests and even newer ones on the horizon, it is not clear to many programs how to classify FISH, CMA, and now NGS test results because they do not exactly fit under the previous rubric. Surveillance programs will likely need to expand their diagnostic test categories to capture data that are generated by new advances in testing.

INCORPORATING NEW GENETIC TESTING RESULTS INTO SURVEILLANCE ACTIVITIES: WHAT IS TRULY SIGNIFICANT?

Technology has now progressed to the point where a single test has the potential to produce much more data than are needed to answer the question originally posed, with some results being of unknown clinical significance. For example, in CMA testing, the reports of results may include CNVs (copy number variations—duplications or deletions in chromosomal segments) that are known to be pathologic or benign, as well as CNVs of unknown significance (not known whether they are benign and represent normal variation in the population, or they are truly pathologic and related to the person’s phenotype, i.e., birth defects and other features and medical problems). In some cases, the indication for performing the CMA test might have been for a phenotype that would have otherwise not been eligible for inclusion as a case in a birth defects surveillance system (e.g., seizures, low muscle tone); in such cases when the CMA test shows only CNVs of unknown significance, surveillance programs will need guidance whether or not to include these cases as true abnormalities. Even if a CNV is clearly known to be pathologic, the surveillance program might not be certain whether the presence of the CNV is related to the infant’s phenotype or is just a coincidental finding. This abundance of uncertain data makes it challenging, not only in deciding whether to include certain cases in a surveillance system, but also how to classify cases with uncertain data into categories for epidemiological studies. In addition, when pathologic CNVs and particular phenotypes are clearly related, the fact that most surveillance programs will encounter very few cases with each particular CNV makes it problematic to evaluate these small numbers of cases in epidemiological and outcomes studies; only by combining resources and expertise from multiple surveillance programs might some of these classification and analysis issues be resolved.

Although there are several complexities for surveillance programs in dealing with results from CMA tests, these are potentially orders of magnitude larger when it comes to NGS results. Sequencing a person’s exome or whole genome typically yields thousands of genetic sequence variants that need to be sorted through to determine if they are pathologic, benign, or of unknown significance, and among those potentially pathologic, which variants might or might not be responsible for the person’s phenotype. With both CMA and NGS tests, CNVs or variants that are of unknown significance today could be determined in the future to be either pathologic or benign; it will be a challenge for surveillance programs if they choose to update and reclassify cases based on new information or include cases that were initially excluded because of uncertain data. Finally, large amounts of data are generated by CMA or NGS tests, so there may be limitations on how much data could be recorded or stored long-term in a surveillance database. Learning how to incorporate data from these new tests within the confines of birth defect research and surveillance presents significant challenges to the field that will take time to resolve.

THE IMPACTS OF NEW GENETIC TESTING ON PREVALENCE

Could the newer diagnostic testing technologies impact the reported birth prevalence of certain birth defects that are ascertained by surveillance programs? In a word—possibly. As a result of CMA and NGS testing, there will definitely be a shift from cases of unknown etiology to those that have a known genetic cause. This was observed previously with the incorporation of FISH testing into clinical practice; for example, in population surveillance, cases with 22q11.2 deletion are now frequently identified (Tézenas Du Montcel et al., 1996; Devriendt et al., 1998; Goodship et al., 1998; Botto et al., 2003) when before the availability of FISH to diagnose this submicroscopic deletion, many such cases included in surveillance systems could have been of unknown etiology, clinically diagnosed as something else, or not ascertained at all (Emanuel et al., 1998; Katzman et al., 2005). Therefore, reported rates will likely increase for specific conditions that are diagnosed with expanded technologies. However, use of other technologies, such as cffDNA analysis might cause the reported birth prevalence of certain conditions to decrease. Several studies have reported that if surveillance programs fail to account for cases of trisomy 13, 18, or 21 that are stillborn or identified by prenatal diagnosis and electively terminated, there are significant underestimations of the prevalence (Stoll et al., 1998; Forrester and Merz, 1999, 2002a,b; Parker et al., 2003; Siffel et al., 2004; Crider et al., 2008). These under-ascertained cases typically occurred in women offered prenatal diagnosis because of their age, abnormal serum screening test results, or abnormal ultrasound findings. Because cffDNA analysis for prenatal cytogenetic abnormalities will likely be offered to the broader population, there is potential that this newer testing technology will further impact the estimated prevalence of aneuploidy disorders such as Down syndrome, as well as other conditions with cytogenetic abnormalities. Therefore, surveillance programs that do not account for cases identified by prenatal diagnosis may further underestimate population rates.

THE ROAD AHEAD: HOW TO AVOID FALLING OFF THE CLIFF

In summary, whether performed prenatally or postnatally, the myriad information available from CMA, NGS, and cffDNA is changing how genetic syndromes and birth defects are diagnosed and classified. The new genetic testing technologies are making significant contributions toward understanding the causes of birth defects and developmental disabilities. Birth defects surveillance programs need to keep abreast of these new diagnostic tests; develop the means to incorporate new genetic data for case ascertainment, classification, and etiological studies; and be cognizant of the potential effects on birth prevalence estimates. Finally, programs should also remain forward-thinking; if history is any indicator, newer and even more exciting genetic testing opportunities are probably just around the corner.

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

American College of Obstetrics and Gynecology2009Committee Opinion No. 446: array comparative genomic hybridization in prenatal diagnosisObstet Gynecol1141161116320168129BiWBorganCPursleyAN2013Comparison of chromosome analysis and chromosomal microarray analysis: what is the value of chromosome analysis in today’s genomic array era?Genet Med1545045723238528BottoLMayKFernhoffPM2003A population-based study of the 22q11.2 deletion: phenotype, incidence, and contribution to major birth defects in the populationPediatrics11210110712837874BouéJBouALazarP1975Retrospective and prospective epidemiological studies of 1500 karyotyped spontaneous human abortionsTeratology1211261162621CarrDHGedeonMM1978Q-banding of chromosomes in human spontaneous abortionsCan J Genet Cytol20415425570443CriderKSOlneyRSCraganJD2008Trisomies 13 and 18: population prevalences, characteristics, and prenatal diagnosis, metropolitan Atlanta, 1994–2003Am J Med Genet A14682082618348276DevriendtKFrynsJPMortierG1998The annual incidence of DiGeorge/velocardiofacial syndromeJ Med Genet357897909733045EdwardsJHHarndenDGCameronAH1960A new trisomic syndromeLancet178779013819419EmanuelBSBudarfMLScamblerPJ1998The genetic basis of conotruncal cardiac defects: the chromosome 22q11.2 deletionHarveyRPRosenthalNHeart developmentNew YorkAcademic Press463478EvangelidouPAlexandrouAMoutafiM2013Implementation of high resolution whole genome array CGH in the prenatal clinical setting: advantages, challenges, and review of the literatureBiomed Res Int 20132013346762FlemmingW1882Zellsubstanz, Kern und Zelltheilung LeipzigGermanyFCW VogelFordCEJonesKWPolaniPE1959A sex-chromosome anomaly in a case of gonadal dysgenesis (Turner’s syndrome)Lancet171171313642858ForresterMBMerzRD1999Trisomies 13 and 18: prenatal diagnosis and epidemiologic studies in Hawaii, 1986–1997Genet Test333534010627940ForresterMBMerzRD2002aEpidemiology of Down syndrome (Trisomy 21), Hawaii, 1986–97Teratology6520721211967919ForresterMBMerzRD2002bPregnancy outcome distribution and prenatal diagnosis of autosomal abnormalities, Hawaii, 1986–1999Teratology66S7S1112239737GoodshipJCrossILiLingJWrenC1998A population study of chromosome 22q11 deletions in infancyArch Dis Child793483519875047GreggARGrossSJBestRG2013ACMG statement on noninvasive prenatal screening for fetal aneuploidyGenet Med1539539823558255ISCN CommitteeISCN 2013: an international system for human cytogenetic nomenclature2013ShafferLGMcGowan-JordanJSchmidMBasel, SwitzerlandS Karger AGJacobsPAStrongsJA1959A case of human intersexuality having a possible XXY sex determining mechanismNature18330230313632697KajiiTOamaKNiikawaN1973Banding analysis of abnormal karyotypes in spontaneous abortionAm J Hum Genet255395474741848KatzmanPJWangBSawhneyMWangN2005Differential detection of deletion 22q11.2 syndrome by specialty and indicationPediatr Dev Pathol855756716222476KitzmanJSnyderMWVenturaM2012Noninvasive whole-genome sequencing of a human fetusSci Transl Med4137ra76Langer-SaferPRLevineMWardDC1982Immunological method for mapping genes on Drosophila polytene chromosomesProc Natl Acad Sci U S A79438143856812046LangerPRWaldropAAWardDC1981Enzymatic synthesis of biotin-labeled polynucleotides: novel nucleic acid affinity probesProc Natl Acad Sci U S A78663366376273878LangloisSBrockJAGenetics Committee2013Current status in noninvasive prenatal detection of Down syndrome, trisomy 18, and trisomy 13 using cell-free DNA in maternal plasmaJ Obstet Gynaecol Can3517718123470070LejeuneJTurpinRGautierM1959Chromosomic diagnosis of mongolismArch Fr Pediatr1696296314415503LoYMChiuRW2008Noninvasive prenatal diagnosis of fetal chromosomal aneuploidies by maternal plasma nucleic acid analysisClin Chem5446146618202154LoYMCorbettaNChamberlainPF1997Presence of fetal DNA in maternal plasma and serumLancet3504854879274585LoYMDChiuRWK2007Innovations: prenatal diagnosis: progress through plasma nucleic acidsNat Rev Genet8717717146468ManningMHudginsLProfessional Practice and Guidelines Committee2010Array-based technology and recommendations for utilization in medical genetics practice for detection of chromosomal abnormalitiesGenet Med1274274520962661MillerDTAdamMPAradhyaS2010Consensus statement: chromosomal microarray is a first-tier clinical diagnostic test for individuals with developmental disabilities or congenital anomaliesAm J Hum Genet8674976420466091NadlerHL1968Antenatal detection of hereditary disordersPediatrics429129184235769NowellPCHungerfordDA1960Chromosome studies on normal and leukemic human leukocytesJ Natl Cancer Inst258510914427847NygrenAODeanJJensenTJ2010Quantification of fetal DNA by use of methylation-based DNA discriminationClin Chem561627163520729299ParkerMJBuddJLDraperESYoungID2003Trisomy 13 and trisomy 18 in a defined population: epidemiological, genetic and prenatal observationsPrenat Diagn2385686014558033PatauKSmithDWThermanE1960Multiple congenital anomaly caused by an extra autosomeLancet179079314430807SaikiRKScharfSFaloonaF1985Enzymatic amplification of beta-globin genomic sequences and restriction site analysis for diagnosis of sickle cell anemiaScience230135013542999980ShafferLGBeaudetALBrothmanAR2007Microarray analysis for constitutional cytogenetic abnormalitiesGenet Med965466217873655SiffelCCorreaACraganJAlversonCJ2004Prenatal diagnosis, pregnancy terminations and prevalence of Down syndrome in AtlantaBirth Defects Res A Clin Mol Teratol7056557115368554SteeleMWBregWRJr1966Chromosome analysis of human amnioticfluid cellsLancet13833854159775StollCAlembikYDottBRothMP1998Study of Down syndrome in 238,942 consecutive birthsAnn Genet4144519599651Tézenas Du MontcelSMendizabaiHAyméS1996Prevalence of 22q11 microdeletionJ Med Genet337198863171TjioJHLevanA1956The chromosome number of manHereditas4216WetterstrandKADNA Sequencing costs: data from the NHGRI Genome Sequencing Program (GSP)Available at: http://www.genome.gov/sequencingcosts. Accessed August 3, 2013