Validation of Whole Genome SNP Array As the Primary Genetic Test in Myelodysplastic Syndrome, to Replace Chromosome and FISH Copy Number Analysis, in a UK Diagnostic Genetic Laboratory

Metaphase cytogenetic (MC) analysis in patients with myelodysplastic syndrome (MDS) remains the only essential recommended genetic test in Europe(Malcovati et al, Blood 2013). The chromosome abnormalities identified provide evidence of clonality and information regarding disease characterisation, risk stratification, treatment strategies and disease monitoring. With the exception of the 3q/MECOM gene rearrangement, observed in 1% MDS, the common, recurrent chromosome abnormalities and all specified abnormalities in the revised International Prognostic Scoring System (IPSS-R) for MDS are copy number changes (deletion, duplication, amplification) (Greenberg et al, Blood 2012). However, the identification of gene mutations is increasingly becoming important to confirm a diagnosis of MDS in problematic cases, refine risk (particularly in those with lower risk IPSS-R scores) and improve patient management (Bejar et al, JCO 2011;Haferlach et al, Leukemia 2014). Currently there is insufficient prognostic evidence to use gene mutation analysis as the sole genetic test, but SNP array technology may be seen as a stepping stone towards providing some gene mutation information as it detects copy number variation (CNV) at single gene, and sometimes intragenic, resolution and copy neutral loss of heterozygosity (CN-LOH), regions which are known to harbor bi-allelic gene mutations. CN-LOH is common in hematological neoplasia, including MDS.

We undertook a prospective validation study using the Affymetrix Cytoscan® HD SNP array, with more than 2.6 million copy number markers, to compare SNP array with MC in patient samples at presentation with confirmed or highly suspected MDS. Subsequent monitoring by SNP array was performed in those patients where samples were received routinely for MC. Expectations were that SNP array analysis would be at least comparable to MC, by detecting all copy number changes detected by chromosome analysis; increase the diagnostic yield of MDS, due to the detection of both sub microscopic CNV and chromosome cryptic CN-LOH; and increase the number of prognostically relevant genetic abnormalities identified.

Samples from 358 patients at disease presentation were received for MC at the West Midlands Regional Genetics Laboratory, UK and were used in this study. Of the 358 patients, 207 had a diagnosis of MDS confirmed and 29 had a diagnosis of MDS/MPN confirmed. The remaining 122 cases were classified as follows: 26 other hematological neoplasia (including MPN, AML, CLL and MM), 51 reactive, 32 non-neoplastic disorders and 13 non-diagnostic.

In the 236 confirmed MDS or MDS/MPN cases, marrow was available in 219 and peripheral blood (PB) only in 17. PB samples were abnormal by SNP array in 11/17 (65%). Of the 219 marrows, 146/219 (67%) had comparable SNP array and MC results. A normal karyotype or failed MC result was observed in 130 samples, of these 35 (27%) had an abnormal SNP array. Cases were abnormal by both SNP array and MC in 84 (38%) samples, of these additional abnormalities were detected by SNP array in 30 (36%), including gene mutations with prognostic significance, such as KMT2A-itd, TP53 deletion and RUNX1 deletion. A normal SNP array was observed in 5 (2%) samples with abnormal MC due to balanced rearrangements (SNP array will not detect balanced rearrangements) and low level (1-3%) abnormal clones. In marrow samples, the diagnostic yield by MC alone was 91/219 (42%) and by SNP array analysis alone 119/219 (54%). Therefore, SNP array increased the diagnostic yield of MDS and MDS/MPN by 12% and identified further abnormalities in an additional 30/219 (14%) samples.

In conclusion, this validation study demonstrates that SNP array may be offered as an alternative test to MC in patients with a confirmed or highly suspected diagnosis of MDS. Patients with features suggesting 3q/MECOM involvement should have FISH for MECOM rearrangements.