Clinical utility of array comparative genomic hybridisation for prenatal diagnosis: a cohort study of 3171 pregnancies


Authors’ Reply


We thank Dr Hillman et al.1 for their comments on our article.2 Firstly, we would like to apologise for the misleading sentences in the abstract. Array comparative genomic hybridisation (CGH) performed in 194 fetuses with abnormal ultrasound could identify 33 (17%) genomic imbalances, including 16 submicroscopic imbalances, 12 numerical abnormalities and five large deletion/duplications. Secondly, the columns in the row of pathologic karyotyping/array CGH reports and live births were shifted one space to the right in table 2.

All of the enrolled cases received both traditional karyotyping and array CGH. The pathogenic karyotype reports were made up of 23 fetuses with derived chromosomes and six with small supernumerary marker chromosomes (sSMCs), whereas pathogenic array reports consisted of eight fetuses with large del/dup, three with micro-del/dup and two with variants of unknown significance. There were 18 fetuses with abnormal chromosome reports and normal array CGH reports, including 15 with derived chromosomes and three with sSMCs. All of the 18 babies were born alive and healthy.

The nomenclature of the array CGH report in table 3 corresponds to the specific array CGH platforms used. If the result of array CGH was named by bacterial artificial chromosome (BAC) clones, then BAC array CGH was used. If the result of array CGH was named according to nucleotide positions, then oligonucleotide array CGH was used. In our cohort, the smallest change was 0.21-Mb microduplication at 2q37.3 in case 8, illustrated by BAC array CGH. We predict that all of the abnormal results noted by 60-Kb oligonucleotide array CGH would be detected by BAC array CGH. As to the five fetuses with variations of unknown significance (VOUS), four were identified by BAC array CGH (cases 10, 14, 30 and 31) and one (case 11, 4.8-Mb duplication at chromosome Yp) was detected by oligonucleotide array CGH. Because both of the array CGH platforms were target arrays, focusing on over 260 known constitutional congenital disease regions, they may provide a lower rate of abnormalities of uncertain significance than that expected from more complex tiling arrays. We did not demonstrate the benefits or drawbacks of BAC array CGH over oligonucleotide array CGH when applied to prenatal diagnosis. But the importance of choosing an appropriate array platform that detects pathogenic genomic imbalance without reporting too many copy number variations cannot be overemphasised.

The interpretation and genetic counselling of VOUS in the prenatal setting is challenging and important. In our cohort, the risk of VOUS is 2/1000. The three fetuses with VOUS had marker chromosomes (cases 10 and 11) and abnormal sonographic findings (case 14). Another large-scale prenatal study reported by Fiorentino et al. showed that the risk of VOUS is 1/1037 (0.1%).3 On the other hand, the reported detection rates of pathologic microdeletions or microduplications in all pregnant women were 34/3171 (1.1%) and 9/1037 (0.9%). These differences between the detection rates of VOUS and pathogenic copy number changes further justify the use of chromosomal array technology in the prenatal setting without a specific clinical indication.