Methods
This retrospective study was based on national records from the Israeli Ministry of Health (MOH). We searched the electronic database of the MOH for CMA tests performed from January 2016 to March 2018 for the indication of short fetal long bones, below the 5thpercentile for gestational age. All testing was financed by the MOH after approval by a clinical geneticist.
The study was approved by the Institutional Review Board for Human Subjects (September 6, 2016, registration number – MOH2016).
Clinical data were obtained from the Israeli national database and from patients’ medical records. Detailed clinical information including maternal age, maternal chronic illness, familial background of genetic conditions, obstetrical history of recurrent spontaneous abortions, elevated nuchal translucency, biochemical screening results, gestational age at diagnosis, lowest bone length percentile recorded during the pregnancy, and the presence of additional sonographic findings were retrieved.
CMA findings were reviewed independently by two authors (R.S.H. and I.M.) and grouped independently into four categories:
  1. Normal (including benign and variants of unknown significance - likely benign categories),
  2. Pathogenic (P)/likely pathogenic (LP) variants,
  3. Microdeletion/duplication with low penetrance, and
  4. Variants of unknown clinical significance (VUS).
The categorization was based on laboratory reports, as well as on new information gained from the medical literature and from the authors’ experience.
For VUS, only cases with deletions ≤1 Mb and duplications ≥2 Mb were included. These variants are reported by the lab, according to the guidelines determined by Israeli Society of Medical Geneticists.
Microarray results were also categorized into ”karyotype-detectable” (i.e., copy number variants of at least 10 MB) or not “karyotype-detectable” in order to assess the incremental yield of CMA over karyotype.
Two cohorts were used to assess the background risk.
  1. A large local cohort of 5,541 cases with normal prenatal ultrasounds in a large, hospital-based clinical laboratory.13 The detection rate for this cohort was 1.4% (78 cases).
  2. A second cohort of 10,614 cases was extrapolated from a meta-analysis by Srebniak et al.14 We calculated the background risk by adding the risk for submicroscopic chromosomal abnormalities to the risk for karyotype detectable chromosomal abnormalities.15 This risk was 1:384, based on the average maternal age of our cohort. The detection rate for this cohort was 1.1%.
Various platforms were used by 12 laboratories to perform CMA. Most medical center laboratories in Israel use the CytoScan 750K array, which is composed of 550,000 nonpolymorphic copy number variant probes and more than 200,000 single-nucleotide polymorphism probes, with an average resolution of 100 Kb.16 One laboratory uses Infinium OmniExpress-24 v1.2 BeadChip, which includes 713,599 genome-wide markers at an average spacing of 4,080 bases and has a targeting minor allele frequency of 5%,17 and one laboratory performs microarray analysis using a Cytochip ISCA 8360K format, BlueGnome.18 Two additional centers (one previously working with the BlueGnome platform and one using BlueGnome and then Illumina) switched to Affymetrix in 2017. One laboratory uses GenetiSure Unrestricted CGH + SNP (43180K) P/N G5976A Agilent. Genomic coordinates were evaluated in accordance with genome build GRCh37/hg19 in all laboratories. All analyses performed in the laboratories met the standards and guidelines of the American College of Genetics and Genomics for constitutional cytogenomic microarray analysis, including postnatal and prenatal applications,13, 15 adopted by the recommendations of the Israeli Medical Genetics Association.16