Use of array genomic hybridization technology for constitutional genetic diagnosis in Canada

The aim of the present article is to inform the paediatric community of the recommendations set forth by the Canadian College of Medical Geneticists (CCMG) regarding the use of a new technology – array genomic hybridization – for constitutional genetic diagnosis. These recommendations were approved by the CCMG and its Board of Directors in 2009, and are published in their entirety on the CCMG website (www.ccmg-ccgm.org/policy.html). The present commentary represents an abridged version, as it relates to the practice of paediatrics.

BACKGROUND

The search for the etiology of unexplained developmental delay, congenital anomalies, dysmorphism and other features suggestive of a genetic cause has traditionally involved chromosomal investigations using microscopy. The ‘gold standard’ of banded chromosome analysis is limited to a resolution of 5 Mb to 10 Mb (1) and detects abnormalities in 5% to 10% of referred cases, depending on the ascertainment criteria. Locus-specific fluorescence in situ hybridization (FISH), which tests for the number of copies of selected specific chromosomal regions, detects subtelomeric and interstitial submicroscopic chromosomal rearrangements (usually 3 Mb to 5 Mb in length) associated with particular phenotypes. Chromosome abnormalities detected by FISH are found in an additional 3% to 6% of patients (2).

Array genomic hybridization reveals chromosomal imbalances (extra or missing DNA sequences) across the entire genome by comparing the patient’s DNA with reference DNA in a quantitative fashion. Results using this technology have demonstrated its greater sensitivity over traditional microscopic chromosome and locus-specific FISH tests for the identification of unbalanced chromosome anomalies of 1 Mb or smaller (3). Array genomic hybridization analysis of phenotypically normal individuals demonstrated that normal variations in the number of copies of a particular sequence (copy number variants [CNVs]) occur at many loci throughout the human genome (4–7). Therefore, when a CNV is found in a patient’s sample, it may either be pathogenically associated with the patient’s condition or simply a benign copy number alteration. It is imperative to distinguish between pathogenic and benign CNVs by incorporating family studies, information about the gene content of the region in question, and data from CNV databases and current literature.

Multiple reviews (8–15) on the topic of array genomic hybridization have been published. There are advantages beyond increased detection of chromosome anomalies with greater resolution using array genomic hybridization: eg, the equivalent of hundreds or thousands of FISH tests can be performed simultaneously and, therefore, there is less reliance on the clinician’s suspicion of a specific diagnosis; and array genomic hybridization does not require dividing cells and, therefore, studies can be performed on postmortem samples (3). Similarly, there are disadvantages of array genomic hybridization: eg, balanced chromosome rearrangements, such as translocations or inversions (in which genetic material is only rearranged but not lost or gained), cannot be identified; triploidy (69 chromosomes) can only be detected using a particular type of genomic hybridization (single-nucleotide polymorphism-based platform); and the location or orientation of a chromosome segment duplication cannot be established. Conventional cytogenetics or FISH are often required as follow-up studies to abnormal array genomic hybridization results. This can include parental testing to compare the affected person’s results with the parents to assist interpretation of findings. When the biological parent is not available, the clinical significance of the genomic hybridization findings may not always be clear. Furthermore, the interpretation and follow-up of CNVs is labour intensive and requires well-developed data management strategies and resources.

RECOMMENDATIONS FOR TESTING NEONATAL/PAEDIATRIC SAMPLES

• Array genomic hybridization testing involves complex counselling issues including difficulty with interpretation due to CNVs (often requiring parental studies), and the potential of discovering nonpaternity with parental testing. It may also not provide information in cases with single-gene disorders or other syndromes such as cystic fibrosis. There is also the potential to identify, in children, deletions or duplications that will exert their effects only later in life as adult-onset disorders (eg, familial adenomatous polyposis). All clinicians ordering array-based tests should fully understand the limitations and potential pitfalls of this type of testing, and ensure that their test subject or parent/guardian is also made aware so that they can make an informed decision about whether to undergo this type of testing.

• Array genomic hybridization should be the first-line laboratory investigation for the patient whose developmental delay/mental retardation, autism, multiple congenital anomalies or dysmorphic features are unexplained after a thorough history and physical examination. Array genomic hybridization could also be used for the study of postmortem samples, including stillbirths with abnormalities as described above, if future literature supports its usefulness in this regard.

• Chromosome studies and FISH tests are not routinely required for the investigation of the above-mentioned patients who have normal array genomic hybridization studies using a platform that includes whole genome coverage. Array genomic hybridization replaces the previous testing modalities of chromosome analysis by G-banding and FISH for targeted microdeletions or subtelomeric rearrangements.

• For patients with developmental delay and/or dysmorphisms of suspected chromosomal etiology, array genomic hybridization should be the first line of investigation, with the possible exception of those suspected of having a common standard numeric chromosome anomaly. Specifically, array genomic hybridization is not recommended for the investigation of the child or adult suspected of having either Down syndrome, trisomy 13, trisomy 18, Turner syndrome, Klinefelter syndrome, XXX or XYY because, in many cases, confirmatory chromosome studies would be required if the diagnoses were made with array genomic hybridization. This is especially true for Down syndrome and trisomy 13, which can be associated with Robertsonian translocations not detectable by array genomic hybridization. Carriers of Robertsonian translocations are at significant risk of recurrence. Therefore, if the clinician has a strong suspicion of one of these conditions, it is appropriate to request traditional chromosome studies first.

• Array genomic hybridization should not be used for the investigation of suspected triploidy unless a single-nucleotide polymorphism-based array is used.

• When there is a strong clinical suspicion of a disorder known to be caused by CNVs in single genes, such as Duchenne muscular dystrophy, for which less-expensive targeted molecular genetic testing is available, array genomic hybridization should not be used as the first line of investigation.