Can microarray analysis reveal additional abnormalities in prenatal diagnosis?

Keywords::

• amniocentesis
• chorionic villus sampling
• copy number variants
• karyotype
• microarray
• microdeletions
• microduplications
• prenatal array comparative genomic hybridization

Prenatal diagnosis took a seismic leap forward in the 1960s, with the universal introduction of metaphase karyotyping [1]. With chromosome banding introduced in the 1970s, not only whole chromosome aneuploidy, but deletions, duplications and other chromosomal rearrangements that were five to ten megabases in size or larger could now be detected. Karyotyping, however, has limitations. It remains based on the subjective assessment of gains and losses and is prone to considerable interpersonal and interlaboratory variation in detection rates. With the advent of chromosomal microarray analysis, we are now witnessing a further step forward in prenatal diagnosis. Chromosomal microarray is able to detect small genomic deletions and duplications (copy number variants [CNV]) that are 100-times smaller, in the kilobase rather than megabase range, than those routinely seen in standard karyotyping. This can be achieved in addition to nearly all the chromosomal imbalances detected in conventional cytogenetic analysis.

CNVs result from a variation from the expected number of copies of a segment of DNA, compared with a reference (normal) genome. These variants are usually larger than 1 kb. If pathogenic, CNVs are likely to have arisen as a result of sporadic chromosomal rearrangements that occurred during meiosis [2]. Approximately 12% of the human genome exhibits CNVs [3]. Many of these smaller aberrations occur frequently enough to have well characterized microduplication or deletion syndromes, such as DiGeorge syndrome. Others may occur less often, but have equally significant phenotypic consequences.

Chromosomal microarrays are now the first-tier diagnostic test for postnatal evaluation of individuals with developmental disabilities, congenital abnormalities or autism spectrum disorder [4]. In this setting, microarray offers a significantly higher yield (15%) compared with that of standard G-banding karyotyping. We now know that chromosomal microarrays in a prenatal settings provide clinically important information in addition to that provided by karyotyping. Chromosomal microarrays identify aneuploidies and unbalanced rearrangements with efficacy equal to that of conventional karyotyping [5].

Microarray construction

Comparative genomic microarray hybridization directly compares a test sample to a normal reference sample, identifying genomic regions that are either under- or over-represented in the study sample. Both samples are labeled with different fluorophores and hybridized to a glass slide printed with several thousand probes derived from most of the known genes, and select non-coding genes of the genome. The probes can be created using synthetic oligonucleotide probes (30–50 base pairs in length) or larger (150–175 base pairs) bacterial artificial chromosome probes, the latter used less frequently now. For single nucleotide polymorphism arrays, only a single test DNA is labeled and hybridized for comparison with previously validated controls. Compared with oligonucleotide or bacterial artificial chromosome microarrays, single nucleotide polymorphism arrays have the benefit of identifying consanguinity, triploidy and most uniparental disomy [6].

The ideal construction of a prenatal microarray remains uncertain at present. Array designs for prenatal diagnosis should be designed to minimize findings of uncertain clinical significance. As opposed to postnatal arrays used to evaluate infants with known abnormal phenotypes, the phenotype in the fetus is incomplete and unpredictable; therefore, array designs for prenatal testing must minimize findings of uncertain clinical significance.

Microarray in prenatal diagnosis

As with the postnatal use of microarray analysis, early prenatal experience in the use of microarrays involved evaluation of phenotypically abnormal cases. In 2011, a systematic review of cases published until the end of 2009 showed that pathogenic CNVs, or variants of unknown significance, were detected in 3.6% of cases for which conventional karyotyping was ‘normal’ [7]. An additional 5.2% (95% CI: 1.9–13.9) were detected when the indication was a fetal anomaly on ultrasound. This systematic review highlighted the varying resolution of the microarray assays used. Studies also varied concerning whether an attempt was made to determine the presence of a benign CNV in parental samples.

In December 2012, the NICHD study on chromosomal microarray versus karyotyping for prenatal diagnosis was published. This showed that microarray analysis was equivalent to standard karyotyping for prenatal diagnosis of common aneuploidies [5]. When the indication was advanced maternal age or a positive aneuploid screening result, additional clinically relevant information was provided in 1.7% of cases (95% CI: 1.2–2.4). When a fetal anomaly was detected in a fetus with a normal karyotype, microarray identified a further 6.0% (95% CI: 4.5–7.9) of genetic alterations.

One of the potential disadvantages that has often been discussed is unexpected identification of novel, previously unreported variants of unknown clinical significance. To facilitate a more accurate assessment and interpretation in these cases, a number of international CNV databases are continually being developed. The International Standards of Cytogenomic Array Consortium [101] and the University of California Santa Cruz Genome Browser [102] databases help stratify findings of uncertain clinical significance into benign or pathogenic variants. It is critical to develop large, prenatally identified databases of CNVs to avoid the ascertainment bias associated with discovery in postnatal cases studies, which are only studied because of existing abnormalities. Postnatal follow-up of prenatally diagnosed cases is also needed and indeed a follow-up of the children in the NICHD study is currently ongoing. Whether a given CNV is considered to be clinically relevant may also depend on ascertainment. Some CNVs not otherwise considered of clinical relevance may be considered if detected in a proband having a known deletion syndrome [8].

The clinical utility of extant CNV databases can nonetheless be demonstrated by reviewing the findings of the NICHD study. Uncertain findings occurred in 3.4% (130 out of 3822) of all karyotypically abnormal cases analyzed with the use of microarray [5]. The study commenced over 5 years ago, allowing a significant amount of time for databases to accrue new cases and published literature to report new clinical findings. When a reinterpretation of the initial categorization was carried out on the basis of the literature available towards the end of 2012, only 56 of the original 94 uncertain results requiring review by the NICHD study’s clinical advisory committee would have remained in an uncertain category. Thirty of the 94 cases would be classified as clearly pathological, and eight likely to be classified as benign. Given information newly available in 2012, only 1.5% of CNVs detected on karyotypically normal samples would now be interpreted as variants of unknown significance. Increasing complexity in the interpretation of clinical and laboratory findings requires close collaboration between laboratory directors, clinical geneticists, counselors and practitioners.

Introduction of new genetic technology requires a new approach to consent and counseling. Pre- and post-test counseling by trained medical geneticists and genetic counselors are central to the prenatal testing process. Discussion is necessary concerning the lack of precise correlation between genotype and phenotype, coupled with the genetic principles of variable expressivity and penetrance. Although the clinical information encountered in microarray is new, the question of uncertainty and how, as clinicians, we should approach uncertainty in prenatal diagnosis without patients is not new. The history of new applications of genetic techniques in prenatal diagnosis suggests that introduction always provokes robust debate [9,10]. Yet open discussion with patients and ongoing data collection reduce uncertainty and lead to better outcomes over time [11].

At present, the American College of Obstetricians and Gynecologists recommends that pregnant women are offered the opportunity to have prenatal testing performed, regardless of risk. If the 1.7% (one in 60) frequency of clinically relevant microdeletions and duplications observed in the NICHD study [5] is borne out in other populations, offering invasive testing and microarray analysis routinely in all pregnancies seems prudent. Indeed, other studies have been consistent with this [12,13]. Women should thus be informed of the more limited identification of common aneuplodies, which is possible using noninvasive screening approaches. They should balance their desire for information about the health of their fetus against the risk of procedure-induced loss of pregnancy following an invasive procedure. This analysis is a very personal one.

Conclusion

Chromosomal microarray is equally efficacious in identifying aneuploidies and unbalanced rearrangements to karyotyping. Furthermore, chromosomal microarrays in a prenatal setting provide additional clinically significant cytogenetic information. Microarrays should thus become the first-tier test for all invasive prenatal testing, including that for structural chromosomal abnormalities. Conventional chromosomal analysis (karyotyping) is still necessary because only this analysis can provide certain information (e.g., location of a microduplication). Only karyotyping can identify balanced rearrangements. Selection of the most appropriate array for use in prenatal diagnosis must balance the need to identify all pathologic imbalances against increased likelihood of many variants of uncertain significance being identified. Interdisciplinary co-operation is paramount in both the design of the arrays and the pre- and post-test counseling, ensuring clear and the most appropriate communication for each pregnant woman who undergoes the microarray testing.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.