Chromosomal Microarray

Chromosomal microarray (CMA) detects copy number variants (deletions and duplications) across the whole genome at a resolution of 10 to 100 kilobases, far finer than karyotype (5 to 10 megabases).

Two platforms are in use: array comparative genomic hybridisation (patient DNA labelled in one colour and a reference in another, hybridised to oligonucleotide probes) and single nucleotide polymorphism (SNP) arrays, which also detect long contiguous stretches of homozygosity (regions of absence of heterozygosity), uniparental disomy, triploidy and consanguinity, and low level mosaicism (down to 10 to 20%).

Neither detects balanced rearrangements (balanced translocations, inversions), low level mosaicism under the detection threshold, point mutations or repeat expansions, which need karyotype, sequencing or targeted tests.

CMA is the first line test for developmental delay, intellectual disability, autism spectrum disorder and multiple congenital anomalies, with a diagnostic yield of 15 to 20% (against 3 to 5% for karyotype), and for prenatal diagnosis after an abnormal ultrasound (additional yield of 6% over karyotype, 1.7% when the foetus is structurally normal).

In miscarriage tissue it gives results in over 90% of cases (culture failure does not apply, though maternal contamination does). Common findings are 22q11.2 deletion, 16p11.2 deletion and duplication, 15q11-q13 deletions (Prader-Willi and Angelman), 1q21.1 and 15q13.3 copy number variants, 17q12 deletion, Williams syndrome region 7q11.23, Smith-Magenis syndrome, and Xp deletions.

Indications. Unexplained global developmental delay or intellectual disability. Autism spectrum disorder. Multiple congenital anomalies, congenital heart disease, or dysmorphism. Unexplained epilepsy with developmental problems. Short stature or growth failure with additional features.

Prenatal: structural anomalies on ultrasound, increased nuchal translucency with a normal karyotype, and stillbirth or miscarriage tissue. Normal karyotype with persisting clinical suspicion. Characterising a marker chromosome or an abnormal karyotype. Suspected consanguinity related recessive disease (SNP array).

Not for balanced rearrangement carriers, family history of a single gene disorder or repeat expansion disorders.

Specimen. Peripheral blood 2 to 5 mL in EDTA (extracted DNA 1 to 3 micrograms), or saliva or buccal swab. Prenatal: amniotic fluid (10 to 20 mL, uncultured or cultured) or chorionic villi (10 to 20 mg), with maternal blood for contamination exclusion.

Products of conception and tissue in sterile saline (not formalin; formalin fixed paraffin embedded tissue works only on some platforms). Parental samples are collected at the same time for interpretation of inherited variants. A request form states phenotype, family history and consent for incidental findings.

Reference values. No pathogenic or likely pathogenic copy number variant, and no clinically significant region of homozygosity. Normal copy number is two for autosomes (and one in males for X and Y). Common benign variants occur in every individual, and a normal result is reported as no copy number change in cytogenomic nomenclature.

Interpretation.

• Pathogenic copy number variant: a deletion or duplication of a known disease causing region or gene (for example 22q11.2 deletion, 15q11.2-q13 deletion), explaining the phenotype and allowing recurrence risk counselling.
• Likely pathogenic: overlaps a critical region, large, de novo, gene rich, or contains a dosage sensitive gene, with partial evidence.
• Variant of uncertain significance (VUS): not previously described, unknown dosage effect, or inherited from a parent with no apparent phenotype. Found in 5 to 10% of postnatal and 1 to 3% of prenatal cases, and not used for clinical decisions. Parental testing and database review (reclassification over time) help.
• Likely benign or benign: common population variants.
• Inheritance: de novo variants favour pathogenicity; inherited variants from a normal parent are usually benign, though incomplete penetrance and variable expressivity apply to some (16p11.2, 15q11.2 BP1-BP2, 22q11.2 in mildly affected parents).
• Regions of homozygosity (SNP array): total over 2 to 3% of the genome in a single child suggests consanguinity or uniparental isodisomy, and may identify recessive disease genes in the homozygous regions.
• Mosaicism: detected at about 10 to 20% on SNP arrays.
• Triploidy and uniparental disomy are identified by SNP arrays but not by array comparative genomic hybridisation.
• Large aneuploidies and unbalanced translocations appear as gain or loss of whole chromosome arms.
• Incidental findings (cancer predisposition genes, adult onset disorders) are reported as a pre-test policy sets.

Critical values and findings needing immediate action.

• Prenatal pathogenic copy number variant with severe phenotype: urgent genetic counselling, parental testing, and discussion of options within legal and gestational limits.
• 22q11.2 deletion: check calcium, immune function and echocardiography promptly, and plan for hypocalcaemia and cardiac surgery.
• Deletion of 15q11-q13 in a hypotonic infant (Prader-Willi) or with seizures (Angelman): confirm with methylation analysis and begin supportive management and feeding support.
• Pathogenic variants in cancer predisposition genes (for example 17q11.2 NF1 deletion, 11p13 WAGR region with WT1 deletion): surveillance for tumours (renal ultrasound for WT1).
• Unbalanced result from a parental balanced translocation: recurrence risk counselling, parental karyotype and prenatal options in future pregnancies.
• Maternal cell contamination in prenatal sample: repeat analysis before reporting.

Interferences and pitfalls.

• CMA does not detect balanced translocations, inversions, small insertions, point mutations or small indels, trinucleotide repeat expansions (fragile X, Huntington), methylation defects (Beckwith-Wiedemann, Silver-Russell), or mitochondrial DNA variants. Sequencing, methylation testing and targeted tests are required for those.
• Low level mosaicism under 10 to 20% is missed; a normal array does not exclude a mosaic abnormality in tissue not sampled.
• Variants of uncertain significance generate parental anxiety and clinical uncertainty, and need careful pre-test counselling.
• Resolution depends on probe design and the reporting threshold (commonly 50 to 200 kilobases for pathogenic regions and 400 kilobases to 1 megabase in other regions); a loss of one probe is not reliable.
• DNA from formalin fixed material is often degraded; quality control and flagging is required.
• Interpretation changes over time as databases grow; re-analysis of old results can reclassify variants (benefit of periodic review).
• Findings unrelated to the indication (carrier status, adult onset conditions) require consent policy.
• Parental samples and phenotype data improve interpretation, and a result without clinical context is incomplete.
• Cost and access limit the use of microarray in low resource settings, where karyotype and targeted tests continue to be used.
• Microarray is gradually being supplemented by exome and genome sequencing, which detect copy number variants and single nucleotide variants together.

Special groups.

• Children with developmental delay, autism or multiple anomalies: first line investigation, with higher yield for syndromic and severe cases.
• Prenatal: used when ultrasound shows anomalies and also offered for normal foetuses with counselling about uncertain findings.
• Couples after miscarriage or stillbirth: analysis of tissue clarifies cause and recurrence risk.
• Consanguineous families: SNP arrays identify homozygous regions and direct sequencing for recessive diseases.
• Adults with unexplained intellectual disability, epilepsy or congenital anomalies: diagnostic yield is similar to children.
• Regions: access is concentrated in tertiary centres, and samples are sent to reference laboratories.

Follow up tests. Parental testing (blood for microarray or targeted FISH), karyotype for balanced rearrangements, FISH to confirm and characterise a copy number change, MLPA, methylation analysis (Prader-Willi, Angelman, Beckwith-Wiedemann), exome or genome sequencing for normal arrays, fragile X testing in males with intellectual disability, calcium,

parathyroid hormone and echocardiography in 22q11.2 deletion, renal ultrasound for WT1 deletion, audiology, ophthalmology and developmental review, and genetic counselling.