Prenatal diagnosis: New study suggests exome sequencing could replace microarray as a first-tier test for severe central nervous system anomalies
Could NGS technologies replace cytogenetics as the gold standard for prenatal genetic diagnosis? With the increasing availability of prenatal-exome sequencing (pES) and prenatal-genome sequencing (pGS), together with a significant reduction in their associated costs in the past few years, we’re possibly on the verge of yet another revolution in the fetal diagnostic field.
Microarray vs NGS
Prenatal genetic testing is regularly performed in high-risk pregnancies to identify or confirm fetal abnormalities, and molecular cytogenetic methods have constituted the recommended practice for prenatal testing for several decades. Karyotyping first enabled the diagnosis of large-scale chromosomal abnormalities. Then, in the early 21st century, chromosomal microarray analysis (CMA, or just microarray) enabled the identification of submicroscopic copy number variants (CNVs), significantly increasing the prenatal diagnostic yield and consequently becoming the benchmark first-tier genetic test for fetal structural anomalies.
However, cytogenetic testing still leaves the majority of cases undiagnosed (ranging from 75 to 95%, depending on the system and severity of the fetal abnormalities). This can be partially explained by the fact that some disorders are caused by single nucleotide polymorphisms (SNPs) or small insertions or deletions (indels).
These types of shorter variants can be detected using higher resolution sequencing technologies, such as panels, exome sequencing, or genome sequencing. While NGS techniques have already become the gold-standard post-natal diagnostic tools for disorders of monogenic etiology, their utilization in prenatal cases still poses several technical and economic challenges. The identification of structural variants from exome samples shows great variability in results and still lacks consensus in the bioinformatics field. The issue of CNV detection could be solved with genome sequencing, which usually has better calling outcomes. However, this option would result in even higher costs per sample (especially compared to CMA), and more variants of unclear significance as well as secondary and incidental findings.
Accurate genetic diagnostic solutions are of significant importance in the prenatal stage, mainly given the challenging identification of phenotypes. Fetal physical assessment is limited to imaging techniques such as ultrasound, or fetal MRI at best. It has also been thoroughly reported that, even in the presence of pathogenic variants, prenatal cases may have variable expression and incomplete penetrance, which makes them harder to match to the phenotypic information available on most databases. Other issues that are very important to consider are the time constraints present in pregnancy, and the psychological impact that inconclusive results can have on the expectant families.
In particular, the central nervous system (CNS) is involved in less than 10% of all fetal malformations, but CNS disorders account for over 30% of termination of pregnancies due to fetal anomalies. Moreover, the diagnostic yield of CMA (combined with karyotyping or alone) is less than 20% for this system and, even when diagnosed, is usually done late in pregnancy.
The novelty of the study
To shed some light on the debate as to which should be the standard genetic test for prenatal cases involving the CNS, the journal Ultrasound in Obstetrics & Gynecology recently published an article that compared the diagnostic yield between CMA and pES in fetuses with major central nervous system (CNS) anomalies. The study aims to establish which type of testing has the highest diagnostic yield for fetuses with major CNS malformations. Establishing which is the best prenatal diagnosis tool is of special relevance taking into account the time-sensitive context of pregnancy, and the fact that the recurrence rate can have a great impact on future family planning choices.
With that objective in mind, researchers from Tel Aviv Sourasky Medical Center, led by Professor Yuval Yaron, studied 114 cases of pregnancy termination due to severe CNS anomalies. In accordance with the recommendations of the American College of Obstetricians and Gynecologists, all fetuses were analyzed using CMA. As shown in the image below, 11 cases with pathogenic or likely pathogenic variants were identified by microarray alone (10% diagnostic yield). Then 86 cases with normal CMA results were also analyzed by pES, as shown in the image below.
All cases were analyzed using hg19 as a reference genome. CMA analysis was performed with ChAS Software by Affymetrix, and pES cases were processed using the Franklin genetic analysis platform, end-to-end, from FASTQ to report. All variants were classified according to the latest American College of Medical Genetics and Genomics (ACMG) guidelines as Pathogenic, Likely Pathogenic, Variant of Uncertain Significance (VUS), Likely Benign, or Benign.
Using the Rainbow algorithm, Franklin’s bioinformatic pipeline allowed researchers to accurately call not only SNPs and small indels, but also CNVs. This proprietary detection tool mainly utilizes coverage-based predictors combined with a machine learning algorithm, and has a resolution of two exons for heterozygous structural variants and even one single exon when the CNV is present in a homozygous state. Variants were then prioritized by Franklin’s AI-based algorithm, taking into account their automatically-calculated ACMG classification and correlation to the fetal phenotypes, expressed in HPO terms.
The study reports that in 38 out of the 86 CMA-negative cases that agreed to further testing, pathogenic or likely pathogenic variants were uncovered by pES analysis. Using NGS after normal microarray results on cases with major CNS malformations resulted in a 44% added diagnostic yield, on top of the 10% provided by CMA testing. Interestingly, the researchers found that this percentage was even higher for specific CNS disorders, such as cases with multiple brain anomalies, reaching an impressive 58% added diagnostic yield.
Another 10% of the cases received a report with a VUS that was worth informing due to its potential clinical relevance. The authors suggest that the currently elusive clinical significance of these variants could be discerned in the near future, thanks to data-sharing tools that connect genetic cases from all around the world, such as GeneMatcher and the Franklin community.
It is also important to note that of the 38 cases resolved by pES analysis by the researchers, 20 (53%) were related to inherited variants. Consequently, these families were at risk for recurrence. That information, which could only be uncovered using NGS technology, was crucial in terms of their future reproductive options and decisions.
Finally, pES reanalysis was performed on 10 out of the 11 CMA-positive cases in order to assess the detection rate of Franklin’s CNV caller. The remaining case was dismissed due to low-quality metrics that reflected insufficient fetal DNA on the sample. Yaron et al discovered that the structural variants detected by microarray analysis could also be identified with pES, validating Franklin’s bioinformatic pipeline for CNV calling.
In summary, the research group concluded that in fetuses with major CNS anomalies, pES offers significant additional diagnostic capabilities as compared to using CMA alone. In addition, taking into account the fact that this NGS technology was found to be equally successful in detecting CNVs, the publication suggests that exome sequencing should be considered a first-tier diagnostic test in prenatal cases with severe malformations of the nervous system.
The future of prenatal testing
The question of prenatal diagnosis doesn’t end here. The role of next-generation sequencing in fetal genotyping is open for debate, and experts are still divided on which of the currently available technologies will be the golden standard for testing in the future.
As discussed in this article and many others, exome sequencing has shown to significantly increase prenatal diagnostic yield for a wide range of diseases, especially for diseases that compromise multiple organ systems. However, despite its undeniable clinical utility, pES is still limited to about 85% of genes, and it fails to identify mosaics and variants in intronic or other problematic regions.
Other candidates, such as pGS and Low Pass Sequencing (LPS), have their own advantages and disadvantages. While the former can achieve even higher clinical utility than pES and provide even higher diagnostic yield than all other techniques, there still exist several practical challenges, such as the high cost and the lack of interpretation data for most variants, resulting in a large number of VUSs that complicate the analysis. As for the former, this type of pGS with genome coverage of up to 10x, known as LPS or simply low-pass, could offer a more cost-effective alternative, with clinical results similar to exome sequencing.
In any case, most genetic experts agree that extensive interdisciplinary collaboration is needed to bring more clarity to the field. Data sharing initiatives like the Fetal Sequencing Consortium aim to increase our understanding of genotype-phenotype interactions in prenatal cases and identify new disease etiologies, in order to reduce the amount of VUSs and ultimately improve clinical diagnosis and pregnancy management, independently of the chosen sequencing technique.
In summary, there are several different views among genetic experts on what the future of prenatal diagnostics looks like. Old or new, invasive or non-invasive, SNPs or CNVs (or both) — there is a case to be made for all mentioned technologies. Yet, if there is anything most experts can agree on, it is that NGS offers exciting opportunities in the field of prenatal diagnosis.
To learn more about this topic, stay tuned for our next article, in which three different experts will present their current efforts for prenatal diagnosis! Leave your questions for the specialists here.