Article Text

PDF

Non-invasive prenatal diagnosis: progress and potential
  1. Rebecca Daley1,2,
  2. Melissa Hill1,
  3. Lyn S Chitty1,3
  1. 1North-East Thames Regional Genetics Service, Great Ormond Street Hospital for Children NHS Foundation Trust, London, UK
  2. 2Fetal Medicine Unit, University College London Hospital NHS Foundation Trust, London, UK
  3. 3Department of Genetics and Genomic Medicine, UCL Institute of Child Health, London, UK
  1. Correspondence to Rebecca Daley, North-East Thames Regional Genetics Service, Great Ormond Street Hospital for Children NHS Foundation Trust, Level 5, York House, 37 Queen Square, London WC1N 3BH, UK; rebecca.daley{at}nhs.net

Abstract

Non-invasive prenatal diagnosis and testing by analysis of cell-free DNA in the maternal circulation is a rapidly evolving field. Current clinical applications include fetal sex determination, fetal rhesus D determination, the diagnosis of some single gene disorders, and a highly accurate screening test for aneuploidies. In the future it is likely to be used for the diagnosis of an increasing range of monogenic disorders, and may even be used to profile entire fetal genomes. The introduction of these tests into clinical practice brings clear benefits but also poses several ethical, social and service delivery challenges. Here, we discuss the current clinical applications, discuss some of the technical and ethical challenges, and look to what the future might bring as technology continues to evolve.

  • Non-Invasive Prenatal Diagnosis (NIPD)
  • Non-Invasive Prenatal Testing (NIPT)
  • Cell-Free Fetal DNA (cffDNA)
  • Genetics
  • Fetal Medicine

Statistics from Altmetric.com

Introduction

For many decades prenatal diagnosis of genetic conditions has required an invasive test such as chorionic villus sampling or amniocentesis. As these tests carry associated risks of miscarriage of up to 1%, much research has been focused on identifying a non-invasive approach. Originally, work concentrated on the isolation of fetal cells in maternal blood during pregnancy. However, there were several problems with using this approach as the fetal cells were difficult to isolate and analyse, and may not be pregnancy-specific.1 Once cell-free fetal DNA (cffDNA) was identified in the maternal circulation,2 efforts were directed towards the analysis of this as a source of fetal genetic material. Analysis of cffDNA has led to the development of non-invasive prenatal diagnosis (NIPD) or testing (NIPT). Non-invasive prenatal techniques have rapidly evolved with current clinical uses including fetal sex determination in pregnancies at high genetic risk, fetal rhesus D (RHD) determination, the diagnosis of some single gene disorders, and a highly accurate screening test for aneuploidies, with scope for further applications. The introduction of cell-free DNA (cfDNA) testing into clinical practice brings the clear benefits of earlier testing, improved safety and ease of access. However, there are several ethical, social and service delivery challenges associated with both these benefits and the tests in general, including potential routinisation, erosion of informed consent, misuse and costs.3 Here, we give an overview of the current uses of cfDNA testing, discuss the future potential and explore some of the social and ethical issues which must be addressed before more widespread implementation into routine public sector care.

What is cffDNA?

Cell-free DNA (cfDNA) is comprised of small fragments of extracellular DNA that circulate freely in the maternal plasma. The cfDNA includes some from the fetus (cffDNA), shed from the placenta once formed at around 4 weeks’ gestation, but the majority is maternal in origin, with cffDNA only accounting for around 10% on average.4 The cffDNA fragments represent the entire fetal genome,5 and are pregnancy-specific as they are rapidly cleared from the maternal circulation after delivery. Thus, analysis of cfDNA offers significant potential for use in prenatal diagnosis. Analysis is less reliable when the proportion of cffDNA in the maternal circulation, known as the fetal fraction, is below 4%.6 The fetal fraction increases with advancing gestation and, while this often reaches levels sufficient for testing for some applications by around 7 weeks’ gestation, in the majority of women, to ensure high accuracy, many tests are not offered until later gestations.

Using cffDNA in clinical practice

The key barrier in the development of safer prenatal diagnosis based on the analysis of cfDNA has been the difficulty in identifying cffDNA against the background of high levels of maternal cfDNA. Initially, it was only possible to identify DNA sequences that are not maternal in origin, but are either paternal in origin or have arisen de novo in some autosomal dominant conditions, for example, Y-chromosome sequences used for fetal sex determination. NIPD can also be applied to autosomal recessive conditions if the parents carry different mutant alleles by excluding the presence of paternal mutant alleles in maternal plasma. More recently, advances in DNA technology, such as next-generation sequencing (NGS), have allowed accurate quantification of specific sequences in maternal plasma, and subsequently enabled non-invasive testing for conditions where the maternal cfDNA must also be accounted. This includes the common aneuploidies, autosomal recessive conditions, where the mother and father carry the same mutant alleles and some X-linked conditions. Testing for aneuploidies is currently considered an advanced screening test, hence NIPT, as there is a small chance of a discordant result, and invasive testing is recommended by many national bodies to confirm positive results.7

Fetal sex determination

Determination of fetal sex by identifying the presence or absence of Y-chromosome sequences was the first NIPD technique to be developed for clinical application, as reviewed by Devaney et al.8 Determination of fetal gender is important to inform the management of pregnancies with unexpected detection of fetal genital ambiguity, and those at risk of a serious X-linked condition, or congenital adrenal hyperplasia (CAH).9 NIPD for fetal sex determination is now well established in a number of countries, including The Netherlands, France, Spain and the UK, and has been shown to be reliable8 and cost effective10 when offered from 7 weeks’ gestation. In pregnancies at risk of X-linked conditions, where it is males that are primarily affected, the requirement for invasive testing for definitive diagnosis is reduced by around 50% as this is only required if a male fetus is identified. In pregnancies at risk of CAH, affected female fetuses may have abnormal development of their external genitalia (virilisation), which can be ameliorated by maternal dexamethasone administration before 9 weeks’ gestation. Fetal sex determination using cffDNA allows for targeting of dexamethasone administration to female-bearing pregnancies.9 The introduction of NIPD for fetal sex determination in the UK has been welcomed by service users who report practical benefits from safe early testing as well as psychological benefits, such as a feeling of having control over the pregnancy and peace of mind.11 However, there is some concern about its potential misuse for social fetal sex determination.

Haemolytic disease of the newborn

Rhesus negative (RhD−) mothers, who carry a Rhesus positive (RhD+) baby, are at risk of development of Rhesus D antibodies that can subsequently cause haemolytic disease of the fetus and newborn (HDFN). As an RhD− mother does not carry a copy of the RHD gene, NIPD can be used to determine whether the RHD gene has been inherited from the father. Worldwide, this approach has been widely used in pregnancies at known high risk of HDFN to inform pregnancy management.12 Women found to be carrying an RHD− baby can have standard antenatal care, while those with an RHD+ fetus require close monitoring in a fetal medicine unit as they may require intrauterine transfusions or early delivery. A non-invasive fetal RHD genotyping service is now available in many countries, with some also offering NIPD for c, E, C and Kell sensitised women.12 NIPD for routine fetal RHD genotyping has recently been shown to be highly accurate and has potential for use in maternity care to direct the use of antenatal prophylaxis using anti-D immunoglobulin.13 In Denmark and The Netherlands, this service is currently offered between 25 weeks’ and 28 weeks’ gestation,13 but earlier RHD genotyping could maximise the benefits by further reducing the costs and risks of unnecessary anti-D (a human blood product) administration for earlier sensitising events as well. A recent UK study has shown that fetal RHD genotyping using a high throughput methodology is robust enough to be used routinely from 11 weeks’ gestation to direct subsequent anti-D prophylaxis,14 and women and health professionals alike would welcome the introduction in order to avoid unnecessary anti-D administration.15 A review of NIPD for fetal RHD genotyping showed high sensitivity (99.5–99.8%) and specificity (94.0–99.5%),13 but unfortunately a small minority continue to receive potentially unnecessary anti-D because of inconclusive or false positive results, often due to the rhesus pseudogene. However, false negative results are of more concern as they place women at risk of sensitisation and potential HDFN in future pregnancies. The continuation of cord blood serology in all women predicted to be carrying a RHD− baby may help reduce these problems as sensitisation is predominantly associated with delivery and postnatal prophylaxis is highly effective.

Single gene disorders

Around 20% of diagnostic prenatal tests in the UK are undertaken for pregnancies at risk of single gene disorders, amounting to approximately 2600 tests for 146 different single gene disorders in 2011–2012.16 Although there is significant potential need, introduction of NIPD for monogenic disorders into clinical practice has been slow, not least because the number of families at risk of individual disorders is small and NIPD often requires development of bespoke assays. Nonetheless, NIPD has been reported for a variety of single gene disorders, including achondroplasia, thalassaemia and cystic fibrosis (table 1).17

Table 1

List of NIPD for single gene disorders publications

Current clinical use for the prenatal diagnosis of monogenic disorders is limited to cases where the mother does not carry the mutant allele. For example, NIPD for achondroplasia and thanatophoric dysplasia has successfully been translated into clinical practice in the UK with full approvals and funding for use in the National Health Service.18 NIPD, to exclude the inheritance of a paternal altered alleles, is also possible for autosomal recessive disorders when the mother and father carry different altered alleles (table 1). This technique is more suitable for those conditions with a large number of possible mutations, such as thalassaemia, but invasive testing is still required for definitive diagnosis when the paternal allele is detected in the maternal plasma.

The development of single molecule counting techniques using digital PCR (dPCR)19 or NGS20 means that it is now possible to determine whether a fetus has inherited a maternal mutant allele by comparing the ratio of mutant to wild-type, or normal, alleles (figure 1). If the fetus is affected, there will be relatively more copies of the mutant allele present in the maternal plasma. Initially, a dPCR-based method followed by relative mutation dosage (RMD) analysis to estimate the allelic ratios by taking account of the fetal fraction was reported.19 A number of RMD studies have been published that demonstrate the application of this technique in NIPD of autosomal recessive diseases19 and X-linked disorders,22 as shown in table 1. In 2010, a technique involving massively parallel sequencing (MPS) and relative haplotype dosage (RHDO) analysis was reported, which involved genome-wide sequencing.5 Since then, a more targeted approach of RHDO analysis has been demonstrated.21 The main challenge for most of these techniques is the need to accurately estimate the fetal fraction, which is straightforward in male fetuses as the number of Y-chromosome sequences present can be used to determine the amount of cffDNA. However, quantification of fetal fraction is particularly difficult in female fetuses as there is no universal, reliable fetal marker available for use in a routine clinical setting. Additionally, the costs of some of these approaches are currently very high. Consequently, a number of technical and translational challenges must be overcome before NIPD for single gene disorders becomes more widely applicable.

Figure 1

Inheritance of mutant or altered alleles.

The introduction of NIPD for single gene disorders brings specific issues in terms of ethical practice. Some apply to traditional prenatal diagnosis, but others are more specific to NIPD, such as the ease of access and safety of the procedure, and have the potential to undermine informed choice if NIPD is seen as routine and may engender feelings of pressure to have testing.23 Offering NIPD through specialist genetic services where health professionals have specialist knowledge about prenatal testing and the condition in question as well as training in prenatal counselling is supported by stakeholders.

NIPT for aneuploidies

In 2008, the first proof of principle studies demonstrated that NIPT for Down's syndrome (DS) was possible using massively parallel shotgun sequencing (MPSS).24 This approach involves extracting cfDNA and using MPSS to estimate the number of sequences mapping to individual chromosomes (whole genome MPS). The relative levels of uniquely mapped DNA sequences from the chromosome in question (eg, chromosome 21 for DS) are then compared with that of a reference euploid sample. Thus, women carrying a fetus with DS will have slightly more chromosome 21 cfDNA than those of the euploid sample. An alternative approach targets sequences from the chromosome in question only (targeted MPS). Several large studies have validated NIPT for DS using these techniques, and report high sensitivities (98.6–100%) and specificities (97.9–100%).25 This approach can also be used for the detection of the other major trisomies6 ,26 ,27 and for sex chromosome anomalies,26 ,27 but the accuracy reported varies and can be less than for DS.

The pace of development of NIPT for aneuploidy has been extremely fast and completely driven by the commercial sector, which has much to gain due to the huge potential market. There was only 3 years between the publication of the first proof of principle studies24 and the launch of clinically available NIPT tests for aneuploidy in Asia and the USA in 2011, and Europe and the UK in 2012. However, at present, these tests are only available through the private sector and we are not aware of any routine implementation in a public health service as yet. Developments have continued at speed with reports of the detection of other unbalanced chromosomal abnormalities, such as large duplications and deletions,28–30 with subsequent incorporation into NIPT by at least one company.31 This latter development is perhaps premature, as the limit of detection for these other chromosomal rearrangements and microdeletion syndromes is unknown, making pretest counselling difficult. Furthermore, it is as yet unclear whether or not this might increase the false positive rate.

NIPT for aneuploidy is not considered diagnostic at present due to the small but discernible risk of discordant results caused by a number of factors. The cffDNA originates from the placenta and, therefore, can result in discordant results due to confined placental mosaicism.29 The approaches used in NIPT analyse the total cfDNA, the majority of which is maternal in origin, and may therefore detect maternal chromosomal rearrangements.32 Finally, false negative and inconclusive results can result from a low fetal fraction associated with early gestation or an increased Body Mass Index, as maternal cfDNA levels may be higher in obese women due to increased shedding of maternal DNA from adipose tissue.4

Work done in the UK, Europe and USA indicates that women and health professionals welcome the potential for a test for aneuploidy that is highly accurate, safe and can be conducted early in pregnancy.33 ,34 However, these studies have consistently reported concerns that mimic those reported by stakeholders surveyed about NIPD for single gene disorders, namely, that ease of access and test safety may potentially lead to routinisation and erosion of informed choice. It is thought these challenges can be overcome with approaches to counselling that support parents to make informed choices by providing a clear explanation of NIPT, the possible advantages and disadvantages, a discussion on the implication of results and allowing time for reflection. Regulation and guidelines from professional bodies and strategies for the training and education of health professionals to facilitate best practice are essential.33

NIPT for aneuploidy is currently only available in the private sector with medical referral, where uptake is reported as being high. However, uptake in the USA has been reported to be related to the cost to the family,35 and in a Hong Kong-based study, uptake of NIPT was almost exclusively by women in high-income brackets.34 As such, this does raise concern around equity of access for NIPT for families on lower incomes. Introduction into the public sector requires further evaluation, not least because most studies report accuracy in high-risk pregnancies and further evaluation is required in those at lower risk, but also to develop the infrastructure, including patient and health professional educational packages, required for safe and efficient implementation.3 It is unclear where in the care pathway NIPT should be introduced. At present, it would be too costly to replace current DS screening tests with NIPT, but an alternative approach with NIPT being offered as a contingent test after current DS screening may be effective and more affordable.36 However, the exact risk in cut-offs to be used are yet to be determined. There are now studies in the public sector in progress in the UK,37 and about to start in The Netherlands and Canada which will hopefully address these issues, with the expectation that NIPT for aneuploidy will be introduced into routine maternity care in the not too distant future.

Entire fetal genome profiling

Several groups have used sequencing of parental genotypes and various approaches to compare their haplotypes and derive the entire fetal genome.5 This approach clearly has potential for the diagnosis of multiple genetic conditions with a single test, but routine application in the near future seems unlikely as the techniques and data analysis are time-consuming and extremely expensive. Furthermore, this approach has potential to detect a very wide range of conditions, and has raised a number of ethical and societal concerns,3 which need to be addressed in the near future to direct the path of future developments.

Conclusion

There has been rapid development of safer prenatal diagnosis based on analysis of cfDNA in maternal plasma for a range of clinical indications. With the technological advances offered by NGS, indications have expanded to include a highly accurate screening test for aneuploidy. The widespread introduction of NIPT for aneuploidy has been driven by the commercial sector, and the challenge now is to develop the infrastructure required for implementation into public sector maternity care. Further research is also needed to expand the range of tests available to families at high risk of monogenic disorders. However, the relative ease of access and safety of these tests, together with the potential of broadening the scope of testing in the future, bring significant ethical issues that need to be addressed in advance of clinical implementation.

References

View Abstract

Footnotes

  • Contributors All authors contributed to the conception and design of the work and the interpretation of the review findings. All authors contributed to the drafting of the article and the approval of the final version for publication.

  • Funding Any work by the authors’ described here is funded by the National Institute for Health Research (NIHR) Programmes Grants for Applied Research (RP-PG-0707-1017—the ‘RAPID’ project) and the NIHR Biomedical Research Centre at Great Ormond Street Hospital.

  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.