In prenatal diagnosis, chromosomal microarray (CMA) has not yet fully replaced conventional karyotyping but has rapidly become the recommended test in pregnancies with ultrasound abnormalities. In this review, we provide an overview of the published data concerning this technology and the controversies concerning its use in the prenatal setting. There is abundant evidence indicating the added detection of pathogenic abnormalities with CMA in comparison to the traditional karyotyping, especially in fetuses with multiple or isolated ultrasound abnormalities such as congenital heart disease, increased nuchal translucency, or oral cleft. On the other hand, there is also a risk to detect variants of unknown significance, late-onset disorders, and variants in susceptibility loci. However, it has been shown that pregnant couples tend to prefer a maximum of information about the health of their unborn child. Taken together, CMA has considerable diagnostic and prognostic values during pregnancy and should therefore be the test of choice.
Introduction
Chromosomal microarray analysis (CMA) refers to molecular karyotyping using different array platforms such as bacterial artificial chromosome (BAC) arrays, oligonucleotide arrays, or single-nucleotide polymorphism (SNP) arrays . BAC and oligonucleotide arrays use DNA from both the patient and a control sample. These DNA samples are labelled with different fluorochromes and hybridized to complementary probes on a chip, and the intensity of the colors is then compared. Abnormal ratios between the patient and the control DNA reveal copy number variants (CNVs). This approach is therefore also referred to as the “array-based comparative genomic hybridization” (aCGH). SNP arrays on the other hand use SNPs to detect CNVs. The A and B alleles of a given SNP are labelled with different fluorochromes, and the comparison of their intensities with the internal references reveals the CNVs . In contrast to aCGH, SNP arrays have the advantage to detect the allelic imbalances and therefore also triploidies, low-level mosaics, and regions of homozygosity, which might indicate parental consanguinity or uniparental disomy, observed prenatally in 0.4–5% of the cases . This technology is widely used for the clinical evaluation of individuals with congenital anomalies, cognitive deficits, developmental delay, growth abnormalities, or behavioral problems . Although conventional karyotyping reveals a disease-causing aberration in about 4% with a maximum resolution of 5–10 Mb, CMA is capable of detecting clinically significant submicroscopic aberrations up to a few kb in an additional 14–18% of such cases .
Conventional karyotyping has been the gold standard in prenatal diagnosis. In the last few years, however, several centers have introduced CMA as a diagnostic tool in the prenatal setting . Although CMA has not yet replaced karyotyping for all indications, it has rapidly become the recommended test in pregnancies with ultrasound abnormalities . Fetal materials suitable for CMA are chorionic villi, amniotic fluid, fetal blood, fetal pleural effusion, or fetal urine. Typically, uncultured material is used. Pathogenic CNVs found in chorionic villi in fetuses without ultrasound abnormalities or which do not fit the clinical indication should be confirmed in the long-term cultured chorionic villi or even in amniotic fluid to exclude confined placental mosaics. Such mosaics have been reported not only for the entire chromosomes but also for the submicroscopic aberrations . An alternative approach might be separating the mesenchymal core cells to be used in CMA from the trophectoderm in native chorionic villi .
Despite the advantages of CMA, such as its superior sensitivity, the faster turn-around time thanks to the possibility of using uncultured or even nondividing material (e.g. intrauterine fetal demise), there are also some disadvantages compared to conventional karyotyping. CMA is unable to detect balanced chromosomal aberrations and cannot reveal the chromosomal location of extra chromosomal material. Microscopic karyotyping remains necessary in such cases. In addition, the higher resolution scan of the genome results in a greater chance of revealing variants of unknown significance (VOUS), which are CNVs containing no genes or genes with no known function, variants in susceptibility loci (VISLs) that are CNVs containing genes with incomplete penetrance , CNVs indicating a predisposition to late-onset diseases , and CNVs that are relevant for future pregnancies only (e.g. X-linked CNVs in a female fetus) . All these concerns may cause counseling problems, parental distress, and unwarranted terminations of pregnancies . Accordingly, a pretest counselling may be implemented as a part of prenatal CMA, as reported in recent publications and clinical guidelines . Srebniak et al. recently suggested a simplified pretest counselling providing information about the higher detection potential, the need of parental testing in some cases, and the chances of detecting unclear results.
Several professional societies (American College of Medical Genetics [ACMG], American College of Obstetricians and Gynecologists [ACOG], Canadian College of Medical Genetics, and Italian Society of Human Genetics) do not encourage replacing prenatal karyotyping with CMA but recommend it as an adjunct test in specific cases only .
Classification of results
A) Pathogenic/causative findings
Three categories of pathogenic findings have been suggested: causative array findings, VISLs, and unexpected diagnoses . CNVs that overlap critical regions of known syndromes (or that overlap other genomic regions that are defined as clinically significant) are likely to be pathogenic in nature . According to Riggs et al., a CNV should be considered pathogenic if there are at least three independent loss-of-function mutations or duplications in unrelated individuals with a similar phenotype published in the literature .
VISLs are CNVs characterized by extreme phenotypic heterogeneity and variable expressivity, such as 16p11.2 microdeletion . Because the incidence of VISLs among affected individuals is increased in comparison with the general population, it has been suggested to group VISLs within the pathogenic CNVs . Several reports showed a statistically higher incidence of VISLs in fetuses with ultrasound abnormalities compared to those without such abnormalities (2.6% vs 1.35%) . VISLs are often inherited from “apparently” unaffected parents . The term “apparently” is justified by the fact that the carrier controls are phenotypically intermediate between the affected carriers and the noncarriers . Most likely, in addition to the presence of a VISL, a second hit such as other genetic or nongenetic factors might affect the phenotypic expression . Currently, there is no information available about the development of children, in whom a VISL was detected prenatally. Consequently, it is important to offer a long-term follow-up to collect such information in the future, particularly in an uneventful pregnancy.
Unexpected diagnoses are those that do not fit the test indications.
B) VOUS
Another category of findings is the variants of unknown significance (VOUS), which can be classified as “likely benign” or “likely pathogenic.” This distinction is made based on the fact whether the variant is inherited or de novo . If the CNV is inherited from a healthy parent, it is less likely to be pathogenic. Therefore, parental testing is important in such cases.
Reporting policy
The classification as VISL, VOUS, or unexpected diagnosis is controversial, and there is no widely accepted policy on this. Especially, the prenatal disclosure of VISLs is a topic of debate because of their unquantifiable risks for neurodevelopmental disorders such as epilepsy, autism, and psychiatric disorders and of the unpredictable phenotype after diagnosis, because a neurodevelopmental phenotype most often cannot be ascertained by ultrasound examination . Many healthcare professionals fear that their disclosure might pose too big a psychological burden to the couple. For this reason, some laboratories classify them as VOUS and report them only in case of ultrasound abnormalities.
In the recent UK guidelines, it has been recommended that VOUS that cannot be linked to a potential phenotype on the basis of the involved genes should not be reported.
Walser et al. performed a survey among genetic counselors in the United States, and the participants found it important to disclose all types of results. Concerning the patients, it seems that they are not influenced in their choice by concerns voiced by professionals, such as the right not to know. About half the women undergoing invasive testing chose CMA, and they were more likely to accept CMA if the indication was structural abnormality . Van Opstal et al. encountered 14 cases of VISLs in 1330 pregnancies without ultrasound abnormalities and Van der Steen at al. reported on 12 parental experiences after disclosure of unclear prenatal results . Importantly, none of the participants felt that a termination of pregnancy would be an option and they reported no congenital anomalies or dysmorphic features at birth, although the newborns were still too young to be assessed for neurological development. Notably, all couples were offered psychological support to cope with the situation after disclosure of the unclear prenatal finding, but none of them needed it. In another report, however, this was different. Bernhardt et al. observed that patients needed specialized support and that some women continued to worry even after delivery and regretted to have taken the test . Notably, in the study of Van der Steen et al., all couples were counselled by a senior geneticist, whereas in another study, such support was not offered in all cases. Therefore, a key factor in mitigating parental anxiety upon disclosing an unclear result appears to be proper post-test genetic counselling . The fact that all but one woman would choose to be informed of an unclear prenatal result again, if offered a choice, underlines this finding. This is in line with earlier reports showing that parents highly appreciate individualized choices .
In another study, van der Steen et al. showed that 79% of pregnant couples wished to decide themselves about the resolution of the test and the type of findings they wished to receive. They showed that 94% of the couples chose testing at high resolution, knowing that this would increase the likelihood of identifying unclear results. Moreover, 84% of them wished to be informed about VISLs. Surprisingly, there was no significant effect of educational level on the choice of array resolution. Interestingly, when asked about their decision 4 weeks after the results, 90% were satisfied with their choice and 19% were worried about the possible consequences of their decision. However, a limitation of this study was that no VISL was detected .
Very limited data exist on unexpected prenatal diagnoses . The few reports that exist mention adult onset conditions only . No consensus on what kind of unexpected diagnosis should be reported has been reached.
Srebniak et al. showed that about 55% of parents undergoing prenatal CMA wanted to be informed about adult-onset conditions . The ACOG recently recommended that all findings with and without pathogenicity and regardless of the age of onset should be reported . Other countries, especially in Europe, support a more restricted disclosure approach, while giving at least some choice to parents to decide what results they wish to receive . Nevertheless, particular caution should be used in potentially untreatable late-onset disorders . It is therefore important to set up a policy for how to deal with such findings, and it is recommended to routinely discuss abnormal prenatal CMA cases in a multidisciplinary team of clinical geneticists, laboratory specialists, gynecologists, and psychologists.
Deletions revealing carrier status for recessive disorders may also be found in array testing, and these are a separate category of findings. According to the ACMG, reporting heterozygous recessive mutations is beyond the scope of genomic array testing and is hence not recommended .
Classification of results
A) Pathogenic/causative findings
Three categories of pathogenic findings have been suggested: causative array findings, VISLs, and unexpected diagnoses . CNVs that overlap critical regions of known syndromes (or that overlap other genomic regions that are defined as clinically significant) are likely to be pathogenic in nature . According to Riggs et al., a CNV should be considered pathogenic if there are at least three independent loss-of-function mutations or duplications in unrelated individuals with a similar phenotype published in the literature .
VISLs are CNVs characterized by extreme phenotypic heterogeneity and variable expressivity, such as 16p11.2 microdeletion . Because the incidence of VISLs among affected individuals is increased in comparison with the general population, it has been suggested to group VISLs within the pathogenic CNVs . Several reports showed a statistically higher incidence of VISLs in fetuses with ultrasound abnormalities compared to those without such abnormalities (2.6% vs 1.35%) . VISLs are often inherited from “apparently” unaffected parents . The term “apparently” is justified by the fact that the carrier controls are phenotypically intermediate between the affected carriers and the noncarriers . Most likely, in addition to the presence of a VISL, a second hit such as other genetic or nongenetic factors might affect the phenotypic expression . Currently, there is no information available about the development of children, in whom a VISL was detected prenatally. Consequently, it is important to offer a long-term follow-up to collect such information in the future, particularly in an uneventful pregnancy.
Unexpected diagnoses are those that do not fit the test indications.
B) VOUS
Another category of findings is the variants of unknown significance (VOUS), which can be classified as “likely benign” or “likely pathogenic.” This distinction is made based on the fact whether the variant is inherited or de novo . If the CNV is inherited from a healthy parent, it is less likely to be pathogenic. Therefore, parental testing is important in such cases.
Reporting policy
The classification as VISL, VOUS, or unexpected diagnosis is controversial, and there is no widely accepted policy on this. Especially, the prenatal disclosure of VISLs is a topic of debate because of their unquantifiable risks for neurodevelopmental disorders such as epilepsy, autism, and psychiatric disorders and of the unpredictable phenotype after diagnosis, because a neurodevelopmental phenotype most often cannot be ascertained by ultrasound examination . Many healthcare professionals fear that their disclosure might pose too big a psychological burden to the couple. For this reason, some laboratories classify them as VOUS and report them only in case of ultrasound abnormalities.
In the recent UK guidelines, it has been recommended that VOUS that cannot be linked to a potential phenotype on the basis of the involved genes should not be reported.
Walser et al. performed a survey among genetic counselors in the United States, and the participants found it important to disclose all types of results. Concerning the patients, it seems that they are not influenced in their choice by concerns voiced by professionals, such as the right not to know. About half the women undergoing invasive testing chose CMA, and they were more likely to accept CMA if the indication was structural abnormality . Van Opstal et al. encountered 14 cases of VISLs in 1330 pregnancies without ultrasound abnormalities and Van der Steen at al. reported on 12 parental experiences after disclosure of unclear prenatal results . Importantly, none of the participants felt that a termination of pregnancy would be an option and they reported no congenital anomalies or dysmorphic features at birth, although the newborns were still too young to be assessed for neurological development. Notably, all couples were offered psychological support to cope with the situation after disclosure of the unclear prenatal finding, but none of them needed it. In another report, however, this was different. Bernhardt et al. observed that patients needed specialized support and that some women continued to worry even after delivery and regretted to have taken the test . Notably, in the study of Van der Steen et al., all couples were counselled by a senior geneticist, whereas in another study, such support was not offered in all cases. Therefore, a key factor in mitigating parental anxiety upon disclosing an unclear result appears to be proper post-test genetic counselling . The fact that all but one woman would choose to be informed of an unclear prenatal result again, if offered a choice, underlines this finding. This is in line with earlier reports showing that parents highly appreciate individualized choices .
In another study, van der Steen et al. showed that 79% of pregnant couples wished to decide themselves about the resolution of the test and the type of findings they wished to receive. They showed that 94% of the couples chose testing at high resolution, knowing that this would increase the likelihood of identifying unclear results. Moreover, 84% of them wished to be informed about VISLs. Surprisingly, there was no significant effect of educational level on the choice of array resolution. Interestingly, when asked about their decision 4 weeks after the results, 90% were satisfied with their choice and 19% were worried about the possible consequences of their decision. However, a limitation of this study was that no VISL was detected .
Very limited data exist on unexpected prenatal diagnoses . The few reports that exist mention adult onset conditions only . No consensus on what kind of unexpected diagnosis should be reported has been reached.
Srebniak et al. showed that about 55% of parents undergoing prenatal CMA wanted to be informed about adult-onset conditions . The ACOG recently recommended that all findings with and without pathogenicity and regardless of the age of onset should be reported . Other countries, especially in Europe, support a more restricted disclosure approach, while giving at least some choice to parents to decide what results they wish to receive . Nevertheless, particular caution should be used in potentially untreatable late-onset disorders . It is therefore important to set up a policy for how to deal with such findings, and it is recommended to routinely discuss abnormal prenatal CMA cases in a multidisciplinary team of clinical geneticists, laboratory specialists, gynecologists, and psychologists.
Deletions revealing carrier status for recessive disorders may also be found in array testing, and these are a separate category of findings. According to the ACMG, reporting heterozygous recessive mutations is beyond the scope of genomic array testing and is hence not recommended .
CMA in fetuses with ultrasound abnormalities
CMA has now become the technique of choice to follow up the structural anomalies identified by prenatal ultrasound . In fetuses with ultrasound anomalies and a normal karyotype, CMA can detect clinically significant submicroscopic CNVs in 5.2–10% of the cases ( Fig. 1 ) . The most common fetal anomalies associated with CNVs occur in the cardiac, skeletal, urogenital, renal, and central nervous system . The incidence of submicroscopic abnormalities depends on the CMA resolution that is applied . Before the introduction of CMA, only targeted testing based on specific ultrasound abnormalities could detect submicroscopic chromosomal aberrations. However, it is difficult to determine by fetal ultrasound which loci should be interrogated. Screening with fluorescence in situ hybridization (FISH) for the 22q11 microdeletion before array testing for example seems to be of limited value, as has also been reported by Tyreman et al. who missed one atypical deletion nested in the distal Di George region .
Srebniak et al. recently described the pathogenic findings in the biggest cohort of prenatal tests in fetal ultrasound abnormalities published so far. They analyzed 1033 prenatal cases after having excluded the most common trisomies by rapid aneuploidy test . They observed an increase in pathologic findings from 1.7% microscopically visible to 4.3% in total. They concluded that karyotyping alone is no longer an appropriate diagnostic approach to detect unbalanced chromosome aberrations in case of ultrasound anomalies.
A recent review reports the results on 3359 fetuses with structural ultrasound anomalies and a normal karyotype . They reported a causal submicroscopic CNV in 5.6% of the fetuses with structural anomalies restricted to one anatomical system. The lowest prevalence was found for the indication increased nuchal translucency (NT) (3.1%), whereas in the group with cardiac anomalies, the prevalence was 4.6%, in line with the data obtained in infants with nonsyndromic structural heart defects (0–4.3%) . In poly-malformed fetuses, however, they found pathogenic genomic microimbalances in 9.1% of the cases.
CMA in fetuses with congenital heart disease
Congenital heart disease (CHD) is a common birth defect affecting 0.5–0.7% of newborns . With the availability of advanced surgical techniques, normal or near-normal heart function can be restored after cardiovascular surgery for most types of CHD. In the prenatal setting, the incidence of chromosomal anomalies in fetuses with CHD is reported to be as high as 18–22%, mostly being trisomy 21, trisomy 18, and 22q11 microdeletions . Furthermore, fetuses with CHD carry a residual risk of additional genetic anomalies including microdeletion or microduplication syndromes such as Williams–Beuren syndrome or monogenic disorders such as Noonan syndrome . Recently, a systematic meta-analysis compared the detection rate of imbalances by CMA to that by conventional karyotyping and 22q11 microdeletion analysis by FISH in fetuses with cardiac malformations. Thirteen publications including a total of 1131 cases met the inclusion criteria. CMA yielded additional clinically valuable information in 7.0% of fetal CHD cases. Subgroup analysis showed an incremental yield of 3.4% and 9.3% in isolated and nonisolated CHD cases, respectively. A VOUS was detected in 3.4% of the cases . Ventriculum septum defect with extra cardiac malformation and left ventricular outflow tract malformations showed the highest rate of pathologic results . Even transpositions of the great arteries and heterotaxy, which are not considered to be associated with chromosomal anomalies detected by karyotyping, were found to have pathogenic CMA results. In another recent study on CHD, the authors showed an incremental yield of pathogenic CNVs smaller than 10 Mb of 5.2% in comparison to that of conventional karyotyping, in line with previous reports . In view of these results, several authors recommend CMA for all types of CHD, after the exclusion of the most common aneuploidies.
CMA in fetuses with increased NT
The prevalence of genetic disorders in presence of increased NT has been previously reported to be as low as 0.5% and as high as 6.6% . More than 50 genetic conditions have been described in association with an increased NT . In a recent meta-analysis, the incremental yield of CMA over karyotyping was assessed . Seventeen studies including a total of 1696 cases were analyzed, and the overall incremental yield was 4.0% in fetuses with isolated NT and 7.0% in presence of other malformation. VOUS were found in 1% of the cases. The incremental yield of this meta-analysis is in line with that of other studies and a systematic review, which reported a yield of 2.5–10% . Performing CMA rather than karyotyping could hence lead to a decrease in undiagnosed genetic disorders in fetuses with an increased NT. In a publication by Senat et al., in which a cohort of 162 fetuses with an apparently isolated increased NT at 11–14 weeks of gestation and normal karyotype were followed up until 2 years of age, an 11.1% rate of abnormalities at birth or postnatally (18/162) was observed . Moreover, in case of persistent unexplained NT at 20 weeks of gestation, the risk for adverse outcome was reported to be as high as 18% .
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
