Interpretation
Data regarding the yield of CMA in pregnancies complicated with short long bones, are sparse. Liu et al,11 recently reported a high yield of 15.6% for pathogenic/likely pathogenic CNV in cases with short femur length.11 This might be explained by the cohort features, including a higher rate of cases with additional sonographic findings: 38/64 (59%) compared to 25/66 (38%) in our cohort. Shaffer et al, reported an incremental yield of 7.3% for CMA over the karyotype19 compared to our incremental yield of only 1.5%.
The rate of abnormal testing in isolated cases, with no additional findings, in our cohort was 8%, which is comparable to the 9.5% reported by Liu et al.11 Of the two isolated cases in which pathogenic CNVs were detected, one had a deletion at 18p11.32-p11.21, which does not include any gene related to reduced growth. However, Chen et al.20 described a 13-year-old girl, who presented with Turner-like syndrome including short stature, with a 18p11.32-p11.21 deletion, identical to the deletion in our case. The other deletion in chromosome 16 encompasses the MAF gene (OMIM #601088). Heterozygous mutations in the MAF gene result in Ayme-Gripp syndrome (OMIM# 601088), in which reduced growth is part of the phenotype, along with other clinical features such as congenital cataracts, which were not reported in this case.21
The additional two cases of pathogenic/likely pathogenic CNV were found in non-isolated cases. The deletion in at 1p31.1 was detected in a fetus with suspected muscular ventricle septum defect. The deletion contains the CDC42 gene (OMIM#616737), and the RPL11 gene (OMIM #612562). Heterozygous mutations in the CDC42 gene cause Takenouchi-Kosaki syndrome (OMIM #616737), an autosomal dominant multisystem disorder with cardiac and skeletal involvement.22 Heterozygous mutations in theRPL11 gene result in Diamond-Blackfan anemia (OMIM # 612562), a variable phenotype syndrome characterized by red blood cell aplasia, growth retardation, craniofacial, upper limb, heart, and urinary system congenital malformations.23
The second case of pathogenic/likely pathogenic CNV with non-isolated short long bones, was diagnosed with persistent right umbilical vein, which is considered a soft marker for increased risk for malformations.24 Duplication of the NSD1 gene was reported to be associated with clinical characteristics of short stature, specific facial features, and intellectual disability in a small number of patients.25, 26 Interestingly, deletions and coding variants of the NSD1 gene cause Sotos syndrome, which is characterized overgrowth, macrocephaly, typical facies and learning disability.27, 28
Chen et al.29 were the first to suggest a gene dose effect of the NSD1 gene, followed by a number of reported cases of which duplications of the 5q35 region encompassing the NSD1 gene resulted in a phenotype of short stature and microcephaly, along with other phenotypic features, including learning disability, mild to moderate intellectual disability, distinctive facies, delayed bone age, microcephaly, seizures, and failure to thrive.25, 26, 30, 31
In our cohort, seven cases presented with sonographic features suggestive of skeletal dysplasia. No clinically significant aberrations were identified by CMA among these cases. Subsequently, two of these cases underwent further genetic testing with whole exome sequencing with no pathogenic aberrations were found in these cases. Liu et al. reported the results of the genetic analysis performed for the 15 cases suggestive of skeletal dysplasia in their cohort.11Genetic aberrations were identified by CMA in two cases (2/15), 1 pathogenic/likely pathogenic CNV and one variant of unknown significance. Genetic sequencing, identified pathogenic aberrations in 53% (8/15) of these cases.11
These results emphasize the challenges of prenatal diagnosis in cases of short long bones. Major, significant disorders will not be diagnosed by CMA alone. Hence, in the presence of sonographic findings suggestive of skeletal dysplasia, with negative CMA, further investigation is recommended using whole exome sequencing or targeted gene panels. Figure 1 depicts a suggested protocol for pregnancies diagnosed with fetal short long bones.
A main issue addressed in our study was whether the risk for abnormal genetic analysis by CMA is significantly higher in cases of short long bones suspected by prenatal scanning, as compared to the background risk. We found that the rate of pathogenic/likely pathogenic CNVs in cases was significantly higher compared to the background population, in a local cohort13 and in a cohort derived from a meta-analysis.14 The fact that the specific CNVs detected are associated with skeletal anomalies emphasize the relevance of our findings.
In an attempt to better characterize groups that would benefit from genetic analysis, we grouped the cases by the presence of additional findings. The yield of CMA was significantly higher than the background risk, in both comparison groups. Additional analysis revealed that cases diagnosed after 22 weeks of gestation but not after 24 weeks had a significantly higher yield of CMA compared to the background risk, suggesting that cases with a diagnosis of short long bones detected after 24 weeks are more likely to be constitutional and genetic evaluation in these cases is less beneficial. However, larger cohorts are needed to determine the accuracy of this suggestion. The yield of CMA was also higher than the background risk in cases where the measured percentile was larger than the 3rd percentile, suggesting that testing cases in the lower portion of the normal range should also be considered.