Recurrent CNVs in constitutional CMA: a ClinGen-anchored single-center postnatal and prenatal experience

Recurrent copy-number variants (CNVs) mediated by low-copy repeats or segmental duplications constitute an important subset of constitutional chromosomal microarray analysis (CMA) findings. Although ACMG/ClinGen standards have improved CNV classification, the burden of recurrent CNVs is not routinely quantified as a canonical, interval-matched, direction-specific category in clinical CMA output. We applied a strict ClinGen-anchored framework to quantify recurrent genomic disorder intervals in a consecutive single-center CMA series.

Constitutional array-CGH and SNP-array records from a single center (2014-2026) were retrospectively reviewed, including 1,661 postnatal probands and 401 prenatal samples, with aneuploidies excluded. Recurrent regions were defined using the ClinGen Dosage Sensitivity Map and restricted to canonical intervals presumed to be mediated by non-allelic homologous recombination. Assignment required breakpoint concordance with the flanking low-copy repeat blocks and canonical interval size; terminal, unbalanced, and atypical imbalances were excluded. All records were re-curated against a single ClinGen release, with earlier GRCh37 coordinates converted to GRCh38.

Pathogenic or likely pathogenic findings were identified in 278/1,661 postnatal probands (16.74%) and 28/401 prenatal samples (6.98%). Recurrent CNVs were detected in 92/1,661 postnatal (5.54%, 95% CI 4.54-6.74) and 17/401 prenatal cases (4.24%, 95% CI 2.66-6.68). Pathogenic/likely pathogenic CNVs were found in 66/92 postnatal (71.74%) and 7/17 prenatal carriers (41.18%); the remaining findings were of uncertain significance or benign. Pathogenic recurrent CNVs thus accounted for 66/278 (23.74%) of postnatal and 7/28 (25.00%) of prenatal non-aneuploid pathogenic findings. Proximal 22q11.21 and 16p11.2 were most frequent. Deletions predominated postnatally (64/93, 68.82%); duplications comprised 11/17 (64.71%) of prenatal recurrent CNVs. Postnatal pathogenic recurrent-CNV yield was 58/1,589 (3.65%) in phenotype-driven testing (no suspected recurrent-CNV syndrome) and 6/19 (31.58%) with syndrome-specific suspicion; neurodevelopmental referrals contributed most cases.

ClinGen-defined recurrent CNVs were a substantial component of constitutional CMA findings. However, only about three-quarters of postnatal and fewer than half of prenatal carriers harbored pathogenic/likely pathogenic CNVs, indicating that a recurrent-region label alone is insufficient for interpretation. These findings support region- and dosage-specific reporting, with evidence-based classification distinguished from clinical interpretation, particularly for prenatal findings and inherited duplications, where the same label may have different patient-specific implications.