By Monica Heger
This story was originally published on May 2.
Verinata Health, formerly Artemis Health, last week published an initial validation study of a sequencing-based test to detect fetal aneuploidy in maternal blood and said it plans to launch a test by early next year for the detection of trisomy 21 and potentially other aneuploidies.
"The clinical validation study is well underway," Verinata's CEO Caren Mason told Clinical Sequencing News, and "we are currently preparing for launch at the end of 2011 or the first quarter of 2012."
Verinata published the results of its initial validation study last week in Clinical Chemistry. The test is based on technology licensed from Stanford University and developed in Stephen Quake's lab (CSN 4/27/2011).
The team used Illumina's Genome Analyzer for the sequencing in the validation study, but will use both the GA and the HiSeq 2000 for the clinical validation study, in which they will also test sample multiplexing. Verinata plans to commercialize the test on the HiSeq.
The timing of Verinata's commercial test coincides with the planned launches of trisomy 21 tests by Sequenom and LifeCodexx, both of which plan to commercialize their sequencing-based tests at the end of the year or early next year (CSN 4/19/2011). The three will likely compete with each other, although LifeCodexx does not plan to market its test in the US.
Verinata, like Sequenom and LifeCodexx, plans to market its offering as a lab-developed test. It has not yet determined a price for the test, and Mason said that one factor guiding pricing will be "what we are able to offer as a result of the clinical study — the number of aneuploidies we can call affected or nonaffected."
In the initial validation study, the Verinata team correctly identified 13 of 13 trisomy 21 samples and eight of eight trisomy 18 samples. For the clinical validation study, it is testing its methodology on thousands of samples from over 50 sites in the US. It will test its diagnostic capabilities not just for trisomy 21, but for others trisomies, such as triosmy 13 and trisomy 9, which it may include in the commercial test, depending on the results of the clinical validation study.
Additionally, Verinata, which changed its name from Artemis Health in April to emphasize its focus on natal health, plans to develop other sequencing based diagnostic tests in the field of prenatal and maternal health, Mason said, although she declined to elaborate further.
Counting Chromosomes
To perform its study, the Verinata team first collected blood samples from 1,014 patients at 13 US clinics from women scheduled to undergo an invasive prenatal procedure. Of those samples, 119 were chosen to undergo sequencing, 53 of which came from women with an abnormal fetal karyotype.
The researchers then chose a "training set" of 71 samples, 26 of which had abnormal fetal karyotypes, on which to optimize the algorithm. After validating the algorithm, they chose an independent test set of 48 samples, 27 of which had abnormal fetal karyotypes, and achieved 100 percent correct classification of the 13 trisomy 21 samples and the eight trisomy 18 samples.
Gestational age of the fetus ranged from 10 weeks to 28 weeks, with a mean age of 15 weeks, 3 days for the test set. However, the company will include fetal ages as low as 8 weeks in the clinical validation study with the aim of marketing its commercial test for use in the first trimester of pregnancy — a goal that Sequenom also hopes to attain.
Sequenom collaborator Dennis Lo of the Chinese University of Hong Kong said via e-mail, that "it is encouraging to see an additional confirmatory report from an independent group that massively parallel sequencing is indeed a promising approach for the noninvasive prenatal detecting of fetal chromosomal aneuploidies."
However, he noted that the sample size in the Verinata study was not very large. While the team collected more than 1,000 blood samples, only 48 were included in the team's "test set." By comparison, Lo and colleagues sequenced more than 700 pregnant women in a validation study for their fetal trisomy 21 test that they published in the British Medical Journal in January (IS 1/18/2011).
Verinata's method "basically builds on the foundation laid by the papers independently published by our group and that of Stephen Quake's," Lo added.
The Verinata, Sequenom, and LifeCodexx tests are very similar, but use slightly different statistical methods to calculate chromosome numbers. The Sequenom method, based on technology licensed from Lo, employs what it has termed a "z-score" to calculate the number of chromosomes. The z-score compares the proportion of reads coming from chromosome 21 to the total number of reads from all chromosomes. LifeCodexx has not yet published the results of its study, but has said that its method is based on Sequenom's method.
Verinata, meantime, uses a "normalized chromosome value," which calculates the number of reads from chromosome 21 and compares that to a subset of reads from another chromosome or set of chromosomes.
Both methods then establish a threshold value, and if the observed read count for a given sample is above that value for chromosome 21, then it is given a trisomy diagnosis.
According to Richard Rava, Verinata's president and co-scientific founder, normalizing for just one chromosome or set of chromosomes, versus all the chromosomes, might give Verinata's test better power to detect other trisomies, such as trisomy 18.
Rava said that his team also tested a similar statistical method described in Lo's paper, and that there was often "so much noise that you can't [detect] the aneuploidies."
Reanalyzing their data with the Sequenom "z-score" method, the Verinata team detected 10 out of 13 trisomy 21 cases, and five out of eight trisomy 18 cases, Rava said.
However, Lo said that the Verinata team used a different threshold by which to declare an aneuploidy. In his analysis, he used a z-score of 3, while the Verinata team used a z-score of 4. Had they used the z-score of 3, Lo said they would have been able to detect 12 out of 13 trisomy 21 cases.
In Lo's BMJ paper, the researchers did detect other trisomies with their bioinformatics method, but focused only on the diagnostic power of detecting trisomy 21. Sequenom has indicated that in its clinical validation study, some principal investigators are testing the method's feasibility for detecting trisomy 18 and trisomy 13 (CSN 3/16/2011).
Have topics you'd like to see covered by Clinical Sequencing News? Contact the editor at mheger [at] genomeweb [.] com.