Meeting the challenge of interpreting high-resolution single nucleotide polymorphism array data: does increased diagnostic power outweigh the dilemma of rare variants

Authors

  • SC Hillman,

    1. School of Clinical and Experimental Medicine, College of Medicine and Dentistry, University of Birmingham, Birmingham, UK
    Search for more papers by this author
  • DJ McMullan,

    1. West Midlands Regional Genetics Laboratories and the Department of Clinical Genetics, Birmingham Women's Foundation Trust, Birmingham, UK
    Search for more papers by this author
  • ER Maher,

    1. School of Clinical and Experimental Medicine, College of Medicine and Dentistry, University of Birmingham, Birmingham, UK
    2. West Midlands Regional Genetics Laboratories and the Department of Clinical Genetics, Birmingham Women's Foundation Trust, Birmingham, UK
    Search for more papers by this author
  • MD Kilby

    1. School of Clinical and Experimental Medicine, College of Medicine and Dentistry, University of Birmingham, Birmingham, UK
    2. Fetal Medicine Centre, Birmingham Women's Foundation Trust, Birmingham, UK
    Search for more papers by this author

Sir,

The Ganesamoorthy et al.[1] publication using two high-resolution chromosomal microarray (CMA) platforms (Affymetrix 2.7M/ Illumina Human cytoSNP-12) to analyse fetal chromosomes in 104 cases (for structural anomaly on ultrasound scan) was of interest to all involved in prenatal diagnosis. The authors examined the benefits of using high-resolution CMA in the detection of pathogenic chromosomal variants known to have a definite link to a phenotype, compared with detection of chromosomal variants of unknown significance with an uncertain link to a phenotype (VOUS). The first can inform difficult decision making, whereas the second may cause unnecessary anxiety that may extend until after the child is delivered.[2]

We would like to draw comparisons to our recently published cohort using a lower resolution targeted CMA (Bluegnome, Focus Constitutional BAC array) in a cohort of 243 cases where a congenital anomaly was identified using fetal ultrasound.[3]

Ganesamoorthy et al.[1] report that in 6.7% of cases (7/104) there were pathogenic/likely pathogenic copy number variations (CNVs) detected by CMA, undetected by conventional karyotyping.[1] This is in comparison to our 3.7% (9/243) detected by the BAC CMA.[3] We would predict that five of seven cases reported by Ganesamoorthy et al. would have been detected by the BAC CMA in our study (i.e. a pathogenic/likely pathogenic CNV detection rate of 4.8%).

The two imbalances not detectable using the BAC array included duplications at 18p and 18q (<250 kilobases) and paternal UPD 14. This latter finding demonstrates the utility of single nucleotide polymorphism arrays in the prenatal setting.

The most pronounced comparison between the studies is the proportion of VOUS detected. Ganesamoorthy et al.[1] detected VOUS in 6.7% of cases (7/104) and reported these back to mothers. However, it was noted that these imbalances did not contain genes previously associated with human phenotype and were presumably followed up because of gene function. In most of these cases the imbalance was inherited. Although demonstrating inheritance cannot exclude pathogenicity, this suggests that they are probably rare benign variants. Our study noted a 0.4% incidence of VOUS (1/243) and in other cases where a VOUS was found, follow-up studies (including parental studies) allowed us to redefine the variant as benign, so reassuring parents.

Using high-resolution CMA and detecting frequent VOUS has significant implications in terms of cost, time and, importantly, difficult counselling scenarios that could theoretically end with termination of a pregnancy for a result which turns out to be benign.

The authors concluded that emphasis should be placed on establishing an evidence-based policy for the interpretation and reporting of CNVs. We agree but would go further to suggest caution in the implementation of very-high-resolution CMA in the prenatal setting. It may be that the evidence base can be built to some degree in the postnatal setting and then transferred to the prenatal.

This debate is only just beginning and cannot fully be informed without the input of patients/patient support groups, but it is a debate that will continue to become increasingly problematic as the resolution of CMA increases and eventually next-generation sequencing is employed in the prenatal arena.

Ancillary