Evaluation of comprehensive chromosome screening platforms for the detection of mosaic segmental aneuploidy

Goodrich, D, Xing, T, Tao, X, Lonczak, A, Zhan, Y, Landis, J, Zimmerman, R, Scott, R T, Treff, N R,
J Assist Reprod Genet. Aug. 2017 doi: 10.1007/s10815-017-0924-4


PURPOSE: A subset of preimplantation embryos identified as euploid may in fact possess both whole and sub-chromosomal mosaicism, raising concerns regarding the predictive value of current comprehensive chromosome screening (CCS) methods utilizing a single biopsy. Current CCS methods may be capable of detecting sub-chromosomal mosaicism in a trophectoderm biopsy by examining intermediate levels of segmental aneuploidy within a biopsy. This study evaluates the sensitivity and specificity of segmental aneuploidy detection by three commercially available CCS platforms utilizing a cell line mixture model of segmental mosaicism in a six-cell trophectoderm biopsy. METHODS: Two cell lines with known karyotypes were obtained and mixed together at specific ratios of six total cells (0:6, 1:5, 2:4, 3:3, 4:2, 5:1, and 6:0). A female cell line containing a 16.2 Mb deletion on chromosome 5 and a male cell line containing a 25.5 Mb deletion on chromosome 4 were used to create mixtures at each level. Six replicates of each mixture were prepared, randomized, and blinded for analysis by one of the three CCS platforms (SNP-array, VeriSeq NGS, or NexCCS). Sensitivity and specificity of segmental aneuploidy at each level of mosaicism was determined and compared between each platform. Additionally, an alternative VeriSeq NGS analysis method utilizing previously published criteria was evaluated. RESULTS: Examination of the default settings of each platform revealed that the sensitivity was significantly different between NexCCS and SNP up to 50% mosaicism, custom VeriSeq, and SNP-array up to 66% mosaicism, and between NexCCS and custom VeriSeq up to 50% mosaicism. However, no statistical difference was observed in mixtures with >50% mosaicism with any platform. No comparison was made between default VeriSeq, as it does not report segmental imbalances. Furthermore, while the use of previously published criteria for VeriSeq NGS significantly increased sensitivity at low levels of mosaicism, a significant decrease in specificity was observed (66% false positive prediction of segmental aneuploidy). CONCLUSION: These results demonstrate the potential of NGS-based detection methods to detect segmental mosaicism within a biopsy. However, these data also demonstrate that a balance between sensitivity and specificity should be more carefully considered. These results emphasize the importance of vigorous preclinical evaluation of new testing criteria prior to clinical implementation providing a point of departure for further algorithm development and improved detection of mosaicism within preimplantation embryos.