FAQs - Section 5: The Accuracy of Harmony
Overall, Harmony is a much more accurate screening test for aneuploidy of the tested chromosomes than first trimester screening. The accuracy varies according to the chromosome under consideration.
The accuracy of a test reflects its ability to identify an abnormal situation correctly, i.e. sensitivity, as well as its ability to identify a normal situation correctly, i.e. specificity.
Harmony Published Clinical Performance Data for Aneuploidy and Sex Chromosomes
Compared with traditional prenatal screening that yields a 5% false positive rate. This means that one in 20 women experiences false positive results with conventional trisomy 21 tests.
The Harmony test performed by Sonic Genetics in Australia is based on microarray detection of targeted DNA. This is an updated version of the original Harmony assay, and is accredited by the Australian Therapeutic Goods Administration.
Clinical performance of the microarray-based Harmony TGA-accredited test is consistent with the original Harmony assay. Patient numbers from microarray-only clinical performance studies are shown below.
Harmony Microarray IVD Performance Data for Aneuploidy and Sex Chromosomes
†Based on information from more than 20,000 pregnancies.
‡Based on information from more than 1,000 pregnancies.
*Validated in singleton pregnancies only.
Defining the accuracy of an NIPT for detecting the 22q11.2 chromosome deletion is more challenging than defining the accuracy for detecting trisomy 21. There are number of reasons for this:
- The frequency of the 22q11.2 deletion was not well defined until a few years ago. In the past, most people with a 22q11.2 deletion were never diagnosed.
- The 22q11.2 deletion is the most common chromosome deletion (approximately one in 1,000 pregnancies) but it is still less common than trisomy 21.
- The size of the deletion varies in different patients. This makes it difficult to design a test that will detect all deletions in this region.
- Prenatal tests that can confirm a diagnosis of 22q11.2 deletion have only recently been widely available. This means that some claims regarding the accuracy of detecting the 22q11.2 deletion by NIPT have not necessarily been validated using a confirmatory test.
For these reasons, there have been very few published studies of the accuracy of NIPT in detecting chromosome deletions. A recent publication has documented the accuracy of the Harmony test for detecting this deletion.
- Harmony correctly detected >75% of samples with a 22q11.2 deletion (sensitivity).
- Harmony correctly detected >99.6% of samples without a 22q11.2 deletion (specificity).
Harmony Published Clinical Performance Data for 22q11.2 Deletion
The accuracy is less than that reported for detecting trisomies, but much greater than previous ultrasound methods which failed to identify the majority of 22q11.2 deletions. The Harmony test is exceptional in that the test accuracy for detecting the 22q11.2 deletion has been rigorously determined and the results published.
The predictive value of NIPT refers to the accuracy of the test when the actual status of the pregnancy is unknown. What proportion of women with a ‘high probability result’ really have a pregnancy with a chromosome disorder (positive predictive value)? What proportion of women with a ‘low probability result’ have a chromosomally normal pregnancy (negative predictive value)?
The predictive value of NIPT varies according to the underlying probability of there being a chromosome disorder. For example, the predictive value for Down syndrome in mothers in their 20s will be different to the predictive value in mothers over 40. This variability makes it impossible to define the predictive value of NIPT for a particular woman.
The predictive value of Harmony has been reported for groups of women with different risk characteristics (see below), but Harmony does not provide such a figure for an individual patient.
In the NEXT study reported in the NEJM, the following predictive values using Harmony were observed in a cohort of more than 15,000 women.
Trisomies 18 and 13 are much less common than trisomy 21, and it was not possible to provide predictive values for women with different risk characteristics. The observed predictive values from the NEXT study (below) are less certain.
There are no large published studies of the predictive value of testing for the 22q11.2 deletion. Assuming that the actual frequency of the 22q11.2 deletion is 1:1,000, we can calculate that the positive predictive value would be approximately 15%; the negative predictive value would be >99.9%.
It is important to note that in all these settings, the positive predictive value was less than 100%, i.e. it is always possible that, despite an abnormal Harmony result, the fetus is chromosomally normal. This caution applies to all forms of NIPT and is due to both technical and biological reasons. For this reason, it is essential that a ‘high probability’ NIPT result be confirmed before irreversible decisions are made regarding the further management of the pregnancy.
It is also important to note that the negative predictive value of NIPT, while typically very high, is less than 100% in all settings. It is possible that a fetus may have one of the screened chromosomal abnormalities, despite a low risk Harmony result. This possibility should be particularly considered when other factors strongly suggest that one of the screened chromosomal abnormalities is present.
For more information on predictive values, see the white paper from the manufacturer of the Harmony Prenatal Test.
There is limited information about the performance of the Harmony test in identical (monozygotic) twins.
From first principles, identical twins have the same genetic makeup and both will be releasing DNA into the maternal circulation. It would not be possible to distinguish the two fetuses because their DNA will be the same. If one twin has a chromosome abnormality, the other will have the same chromosome abnormality.
With both fetuses releasing the same DNA into the maternal circulation, the accuracy of the Harmony test in identifying the genetic status of the twins should be as good as in a singleton pregnancy.
The situation is more complicated with non-identical (dizygotic) twins. In this situation, each fetus is releasing its own DNA into the maternal circulation. Non-invasive prenatal testing (whichever assay is used) must distinguish the proportions of chromosomes from three people (mother, twin 1 and twin 2) rather than two (mother and fetus in a singleton pregnancy). The amount of DNA being released from each twin into the maternal circulation can be markedly different.
In addition, if one fetus has a chromosome abnormality, it is unlikely that the other fetus will have the same abnormality. These factors complicate the analysis and interpretation of any form of NIPT in non-identical twins.
The published experience of the Harmony test with non-identical twins is that there is an increased chance of the fetal fraction of at least one twin being insufficient for the test to be reported; this reflects the biology of a twin pregnancy and is not a failing of the assay. The data regarding test performance are also limited. The detection rate rate for trisomies in twin pregnancies may also be lower. In Harmony twin validation studies, the detection rate for trisomy 21, 18 and 13 was 29/31 (94%). This detection rate is higher than the sensitivity of conventional first trimester screening in singleton pregnancies, but not as high as the Harmony test in singleton pregnancies.
There is very limited published information about the performance of non-invasive prenatal testing in detecting other aneuploidies in twin pregnancies.
In pregnancies with non-identical twins (or if zygosity is not known), invasive cytogenetic testing may be warranted if either Harmony or conventional first trimester screen indicate high risk of fetal aneuploidy.
Please note the presence of twins on the request form.
For further information, please refer to the white paper: Cell-Free DNA Screening in Twin Pregnancies using the Harmony Prenatal Test
Harmony has not been not validated for triplet or higher order multiple pregnancies. We do not recommend NIPT in this situation.
Harmony has been validated for use in IVF pregnancies. This includes the use of donor egg or sperm, and in a surrogate pregnancy.
Please provide details of the IVF process on the request form, as these are required for test analysis.