SMFM consult: Chromosomal microarray for prenatal diagnosis

By Society for Maternal-Fetal Medicine (SMFM); Lorraine Dugoff, MD; Mary E Norton, MD; and Jeffrey A Kuller, MD

Chromosomal microarray analysis (CMA) is a high-resolution whole-genome screening technique that can identify most of the chromosomal imbalances detected by conventional cytogenetic analysis, as well as smaller submicroscopic deletions and duplications known as copy-number variants (CNVs). CNVs may cause a wide range of disorders, including neurodevelopmental disorders and congenital anomalies such as cardiac defects. CMA is recommended as the first-tier test in postnatal evaluation of congenital abnormalities and neurodevelopmental disorders. With accumulating experience over the last decade, and data demonstrating improved detection of chromosomal abnormalities compared to conventional karyotyping, CMA is proving to be a valuable diagnostic tool in the prenatal setting. CMA can be performed on uncultured DNA samples, including those obtained from chorionic villus sampling (CVS) and amniocentesis, which may lead to quicker turnaround than karyotyping.

Q: What are the different types of microarray?

Two major microarray platforms are used in prenatal diagnosis: single nucleotide polymorphism (SNP) arrays and comparative genomic hybridization (CGH) arrays. With SNP and CGH arrays, DNA from a fetal sample, such as CVS or amniocentesis, is hybridized to an array platform consisting of DNA probes on a solid surface, such as a microscope slide or a silicon chip.

CGH compares the fetal DNA sample with a normal reference DNA sample. An SNP is a variation at a single position in a DNA sequence among individuals. With SNP arrays, only the DNA test sample is hybridized to the array platform. While CGH arrays are able to detect only copy number variants, SNP arrays can also detect triploidy and regions on the 2 homologous chromosomes that are identical to each other, as occurs with uniparental disomy (UPD) and consanguinity. In the case of UPD, both copies of a chromosome are inherited from the same parent, instead of one from each parent. UPD has been associated with genetic disorders such as Prader-Willi syndrome, which can occur when both copies of chromosomes 15 are maternally inherited. SNP arrays can also detect some cases of maternal cell contamination and mosaicism.

Arrays may include probes that cover the whole genome, or may be targeted, with concentrated coverage in known disease-causing regions of the genome and more limited coverage of the rest of the genome.

Q: What can CMA detect? How does it differ from a karyotype?

A standard karyotype can detect aneuploidies (abnormalities in chromosome number), relatively large structural abnormalities such as deletions or duplications that are microscopically visible to a resolution as low as approximately 5–10 Mb, and balanced or unbalanced translocations and inversions. CMA has a higher resolution than conventional karyotyping, allowing for detection of much smaller submicroscopic deletions and duplications, typically down to a 50–100 kb level. CMA can also detect some copy number changes near the chromosomal breakpoint sites in rearrangements that appear to be balanced on a conventional karyotype.

Another advantage of CMA is that this technique does not require dividing cells, in contrast to conventional karyotyping, which requires cell culture. This difference can allow for quicker turnaround. In addition, CMA can be performed on macerated tissue obtained from stillbirth specimens that may not grow in tissue culture, and thus, may be more likely to provide a result as long as sufficient good-quality DNA can be obtained.

Q: What are the limitations of CMA?

Because CMA looks for genomic imbalance, it cannot detect totally balanced chromosomal rearrangements, such as translocations or inversions. The large majority of balanced rearrangements, however, result in a normal outcome. In addition, CMA does not provide information about the chromosomal mechanism of a genetic imbalance. Therefore, a karyotype should be performed in such cases to rule out a translocation that may have been inherited. Low-level mosaicism may not be detected by CMA and some arrays do not detect triploidy. SNP arrays, however, are generally able to detect lower levels of mosaicism, as well as triploidy. CMA will not detect all CNVs, such as those that are in regions not represented on the array platform and very small CNVs that are below the level of detection. In some cases, a postnatal CMA may identify a CNV that was not identified prenatally due to the higher resolution of postnatal arrays. In addition, CMA will not detect point mutations within single genes, including those that cause disorders such as sickle cell anemia, cystic fibrosis, and many skeletal dysplasias.

Q: What are the risks and disadvantages of CMA?

A disadvantage of CMA is the inability to precisely interpret the clinical significance of a previously unreported CNV or to accurately predict the phenotype of some CNVs that are associated with variable outcomes. CNVs are characterized as benign, clinically significant (ie, pathogenic), and VUS. The overall prevalence of VUS is approximately 1% to 2%. Fortunately additional information on the classification of CNVs is rapidly accumulating, which should lead to a decrease in incidence of VUS reported over time.

Patients in whom a fetal VUS is detected by prenatal diagnosis should receive counseling from experts who have access to databases that provide updated information concerning genotype-phenotype correlations. Patients should be educated regarding the significance of the finding including the potential range of outcomes, and should be provided with resources and support. Further testing should be offered if indicated. One of the initial steps in the evaluation is to determine if either parent has the same CNV as was detected in the fetus. Although de novo CNVs are more likely to be pathogenic, an abnormal fetal outcome cannot always be excluded even if a parent with the same CNV as a fetus is normal, as some have a variable outcome. When interpreting VUS, it may be helpful to evaluate the specific genes that are contained in the deleted or duplicated regions. In general, small duplications are less likely to be clinically significant than are small deletions.

Q: When should array be offered, and what are the indications?

The American College of Obstetricians and Gynecologists (ACOG) and SMFM recommend that “[A]ll pregnant women should be offered prenatal assessment for aneuploidy by screening or diagnostic testing regardless of maternal age or other risk factors …The differences between screening and diagnostic testing also should be discussed.” CMA in particular is recommended when genetic analysis is performed in cases with fetal structural anomalies and/or fetal demise. CMA replaces the need for fetal karyotype in these cases. ACOG and SMFM recommend that either fetal karyotype or CMA be performed when invasive prenatal diagnosis is performed in cases with structurally normal fetuses regardless of maternal age.

Some clinical providers now recommend CMA as a first-line test whenever fetal chromosomal analysis is planned, while other clinical providers reserve CMA for cases in which there are fetal structural abnormalities, to avoid the possibility of discovering a VUS. The prevalence of significant abnormalities identified by CMA in cases with a normal karyotype and normal ultrasound was 1/60 (1.7%) in the study by the Eunice Kennedy Shriver National Institute of Child Health and Human Development. This prevalence is high enough that providers should discuss the benefits and limitations associated with CMA and conventional karyotype with their patients who are considering amniocentesis and CVS.

Chromosomal microarray analysis should be considered as further evaluation when an apparently balanced de novo rearrangement is detected by karyotyping in order to exclude an imbalance at one or both of the translocation breakpoints. CMAs may also prove helpful in identifying the chromosomal origin and gene content of marker or ring chromosomes identified with conventional karyotype.

Q: When is it appropriate to perform just a karyotype?

Conventional karyotype and/or rapid FISH testing may be more appropriate when a common aneuploidy such as trisomy 21, 18, 13 or monosomy X is strongly suspected based on prenatal ultrasound findings. In these circumstances, conventional karyotype and FISH analysis may provide a more rapid turnaround, allow for more sensitive detection of low-level mosaicism, and rule out a translocation-associated trisomy. A CMA can be performed in the event that the FISH or karyotype is normal. Conventional karyotype to identify potential balanced translocations is the most appropriate first-line test for couples with a history of recurrent miscarriage. Due to limited data, CMA is not currently recommended as a first-line test to evaluate first-trimester pregnancy losses.

In some situations a karyotype should be performed following an abnormal microarray result. When trisomy of an acrocentric chromosome (13, 14, 15, 21, or 22) is identified by CMA, a karyotype should be performed to rule out an unbalanced Robertsonian translocation that might have been inherited. Depending on the size, either FISH or a karyotype is sometimes recommended to rule out inherited rearrangements in some cases involving smaller copy number gains. This information is needed to provide accurate information regarding future recurrence risks.

Q: How should patients be counseled prior to CMA?

Trained genetic counselors, geneticists, or other providers with expertise in the complexities of interpreting CMA results should perform pre- and post-test counseling. Providers should be familiar with the microarray platform used by their laboratory, including the rate of VUS. Patients should be informed that compared with conventional karyotype, CMA will detect a potentially pathogenic CNV in an additional 6%–7% of cases with fetal structural abnormalities on ultrasound and in 1%–1.7% of cases with a structurally normal fetus. Patients should also be informed of the 1.4%–2.1% chance that a VUS will be detected. Pretest counseling should include a discussion of the spectrum of disorders that can be detected with CMA, including disorders with severe neurologic phenotypes as well as those with more mild or adult-onset phenotypes. Patients should also be informed that CMA does not detect every genetic disease or syndrome, including autosomal recessive disorders associated with single-gene point mutations. Patients should also be informed that CMA can detect consanguinity and non-paternity in some cases.

Q: What samples can be used?

CMA may be performed on DNA obtained from amniocentesis, CVS, fetal cord blood, and stillbirth specimens. DNA obtained from the mesenchymal core cells of the chorionic villi and uncultured amniocytes is preferable to DNA from cultured cells to allow for quicker turnaround and to avoid the possibility of culture artifacts. Some labs require that a maternal blood specimen be sent with the original CMA specimen, while other labs request parental samples only when a CNV is detected, in order to distinguish between an inherited and a de novo CNV.

Q: Are there differences between prenatal and postnatal microarray?

In the postnatal setting, CMAs are utilized to explain existing abnormalities, while in the prenatal setting CMAs are obtained to predict fetal outcomes. In many prenatal cases, patients opt to have CMA for reassurance that a significant finding is absent. CMAs are recommended as the first-tier diagnostic test for postnatal evaluation of children with multiple congenital anomalies, developmental delay/intellectual disability, and/or autism spectrum disorders, with clinically significant findings reported in approximately 15% of cases with normal conventional karyotypes. In postnatal cases, identification of a diagnosis is important to parents for many reasons, including ending the search for a diagnosis, obtaining resources, planning future care for the child, and planning future pregnancies. The benefit of finding a clinically significant abnormality with CMA may offset the downside of finding a VUS. In the prenatal setting, particularly in cases with a structurally normal fetus on ultrasound, a VUS may cause considerable stress and anxiety as the parents may consider the option of pregnancy termination. It may be difficult to interpret the significance of a CNV prenatally due to the limitations of fetal imaging and the limited information currently available correlating prenatal CNV findings with postnatal phenotypes.

In order to decrease the likelihood of identifying a VUS, many specialists advocate using a targeted rather than a whole-genome approach in prenatal cases. Targeted arrays use platforms that primarily identify CNVs in which clinical interpretation is non-equivocal, including trisomies, or well-documented microdeletion/duplication syndromes. However, targeted arrays also may result in a lower diagnostic yield. The probe density of targeted arrays has increased over the last several years, and as knowledge regarding classification of CNVs continues to expand, the difference between whole genome arrays and targeted arrays is narrowing. As above, postnatal CMA does generally have a higher resolution than those used in a prenatal setting, and therefore may identify a CNV that was not identified prenatally in some cases.

What about the cost of CMA and insurance coverage for it?

Chromosomal microarray currently is more expensive than conventional karyotyping but the cost is expected to decrease with increasing volumes and technical advances. Insurance coverage in the United States largely conforms to the recommendations of the joint ACOG and SMFM Committee Opinion.