NEW YORK (GenomeWeb) – A growing number of clinical laboratories is providing noninvasive prenatal trisomy testing using cell-free DNA from maternal plasma, commonly referred to as NIPT. In Europe, in particular, such testing is often not performed by large centralized laboratories, like in the US, but by a variety of commercial and non-commercial entities.
While professional societies, like the American College of Medical Genetics and Genomics, have put out position statements on the use of NIPT, no detailed guidelines exist for how test results should be reported, and according to a survey conducted by three UK-based external quality assessment (EQA) organizations last year, reporting practices differ widely between laboratories.
In an effort to establish standards for reporting NIPT results, the three EQA providers — CEQAS (Cytogenomic External Quality Assessment Service), EMQN (European Molecular Genetics Quality Network), and UK NEQAS (United Kingdom National External Quality Assessment Service) — and their clinical collaborators have developed a set of consensus minimum best practice guidelines, which were published in Prenatal Diagnosis last month.
At a workshop during the European Society of Human Genetics annual meeting in Copenhagen last week, the EQA providers presented those recommendations, along with initial results from the first pilot NIPT external quality assessment scheme they conducted, in which about 40 laboratories participated.
While there was consensus on many reporting requirements among the authors, which included Lyn Chitty, a genetics and fetal medicine expert at University College London, and Erik Sistermans of the VU University Medical Center in Amsterdam, who oversees a nationwide NIPT study in the Netherlands, they could not agree on whether the fetal DNA fraction should be stated on the report or whether positive and negative predictive values and individualized risks should be included.
The best practice guidelines came out of discussions at a workshop held at the International Society for Prenatal Diagnosis (ISPD) annual conference in Berlin in 2016 that was attended by stakeholders from 19 countries, including commercial NIPT providers, clinicians, laboratory scientists, and policy makers.
Prior to the ISPD meeting, in an attempt to gain information about the status quo of noninvasive prenatal testing and reporting practices, the three EQA organizations had sent out a survey to almost 3,000 laboratories in their database.
The survey results revealed that labs differ markedly not only in their testing methodologies for NIPT but also in how they report the results, despite the fact that ISO 15189 standards, to which many clinical laboratories in Europe adhere, already mandate certain reporting elements.
Of the 121 laboratories who responded to the survey, 75 said they were currently performing NIPT for fetal aneuploidy, and another 25 said they were in the process of validating an NIPT assay.
Among those who already offered NIPT, testing volumes differed significantly, ranging from up to 50 to more than 1,000 tests per month.
Almost three quarters of labs said they use a counting-based next-generation sequencing approach that measures dosage of chromosomal DNA, and a little more than 10 percent reported using a SNP-based sequencing approach. The remaining 15 percent made use of arrays, Sanger sequencing, MLPA, or PCR.
About 70 percent of NGS labs reported using a whole-genome sequencing approach, with the remaining 30 percent applying targeted sequencing. Also, almost 80 percent of NGS labs had an Illumina platform and almost 40 percent a Thermo Fisher Scientific sequencer.
Turnaround times varied from 3 days to 21 days, with the largest number of labs saying they return results within eight to 14 days.
Almost half the laboratories said they only report fetal sex when requested, and about 23 percent said they always report fetal sex, while 28 percent don't report it.
In addition to the three most common trisomies, almost half the labs said they test for sex chromosomal abnormalities, and almost 40 percent said they report other trisomies or subchromosomal abnormality.
Forty-three laboratories submitted sample NIPT reports along with the survey — an example of a report with a low-risk finding and one with a high-risk result — which served as the basis for the discussion that resulted in the guidelines.
All reports included the patient's name, and more than 80 percent her date of birth, while almost 20 percent only provided her age. ISO 15189 standards already recommend using at least two unique patient identifiers, including name and date of birth, and the authors suggested following those existing guidelines.
They also agreed that a NIPT report should state the sample type, date of sample collection, and date of sample receipt by the laboratory, a practice that not all labs followed — only half, for example, mentioned the sample type.
Reports should also state the gestational age at the time the sample was taken, according to the authors, a practice that about three quarters of the labs followed. In addition, labs should date the report and include contact details of the recipient, which several labs failed to provide.
The authors further recommended that the reports should state the test methodology, for example, next-generation sequencing, which nine laboratories did not include, and should mention the limitations of the test, which 11 labs did not provide.
Approximately 60 percent of the labs surveyed said they measure fetal DNA fraction, but the authors could not agree on whether fetal fraction should be measured or reported. Not all testing platforms measure fetal fraction, they wrote, and while its inclusion may reduce the number of false negative results, it may also result in an increase in test failures. The authors also did not reach consensus on whether negative and positive predictive values, along with an individualized risk, should be included.
Labs differed a lot in how they reported a positive test result, though they all said such a result must be confirmed by other tests. Most commonly, they referred to a positive result as "high risk", though some called it an "increase in the number of copies of chromosome #", "suspicious of trisomy", or "abnormal result for chromosome #".
Since NIPT is considered a screening test, not a diagnostic test, the authors said results should ideally be reported as "high risk/low risk of aneuploidy." Also, reports should clearly state that a high-risk result needs confirmation by invasive testing.
In addition to developing the minimum guidelines, CEQAs, EMQN, and UK NEQAS conducted a pilot EQA scheme last year to assess laboratory performance for NIPT. Sandi Deans, director of UK NEQAS for Molecular Genetics, reported initial results at last week's workshop, noting that participating labs will receive their individual results this month.
Establishing a quality assessment scheme for NIPT was technically challenging, she said, because cell-free DNA tends to be unstable in plasma and because patient samples are difficult to obtain, especially trisomy-positive samples.
For this reason, the organizers decided to work with artificial samples, provided by SeraCare Life Sciences. Those samples contained fragmented fetal DNA derived from trophoblast cells, including from confirmed trisomies, and fragmented normal female DNA, which were added to a plasma-like matrix. After extensive validation, samples were shipped to laboratories at ambient temperature, Deans said, as the DNA had proven to be quite stable.
Labs were supplied with samples from three artificial patient cases: a sample with low trisomy risk, a sample with a high risk of trisomy 21, and a sample with a high risk of trisomy 18.
Using artificial samples restricted the scheme to labs that rely on counting-based methods, which are employed by tests from Sequenom, Illumina, BGI, Berry Genomics, and Roche's Ariosa, for example. Labs using SNP-based methods, such as Natera's Panorama test, could not participate because the fetal DNA did not match the maternal DNA in the samples.
A total of 44 labs registered for the EQA and 40 submitted test reports for evaluation. Participating labs came from a variety of European countries including the UK, Spain, the Netherlands, Belgium, and Germany, as well as from the US, China, and Australia.
Test reports were marked according to a variety of criteria, Deans said, including whether they identified the correct genotype and whether their reported fetal fraction corresponded with the known one. Test interpretation was not scored, but the organizers looked at whether all essential interpretive elements were included, whether the testing methodology was stated, and whether the limitations of the test were provided.
About half the labs reported the fetal DNA fraction for all cases and about 40 percent of labs did not report fetal fraction. For those that did report it, numbers ranged widely and did not always match the known fetal DNA fraction.
Four different laboratories reported erroneous results. Two labs found a high risk for trisomy 21 in the low-risk sample, one lab did not detect a high risk in the trisomy 21 sample, and one lab detected a high risk in the trisomy 21 sample but also reported a borderline high risk for trisomy 13 in that sample.
Deans said among the issues they found were that essential information was sometimes buried deep in the report, that reports did not explain the test limitations, and that the test methodology was not explained in detail.
The EQA providers are currently working on the next pilot NIPT quality assessment scheme, for which they are obtaining real plasma samples from consented patients. That study is expected to be open to all test methodologies, and the plan is to distribute those samples to participating labs in the fall.