Superior Identification of Prognostic Relevant Copy Number Abnormalities By SNP-Based Genomic Arrays As Compared to Interphase FISH in Multiple Myeloma

Multiple myeloma (MM) is a neoplasm that exhibits a broad heterogeneity in both biological behavior and clinical presentation. Specific copy number abnormalities (CNAs) such as hyperdiploidy, 1p loss, 1q gain, 13q loss and 17p loss (including the TP53 gene), and IGH translocations, such as t(4;14)(p16;q32) and t(14;16)(q32;q23), provide important information regarding prognosis and treatment response.

Interphase fluorescence in situ hybridization (FISH) on enriched plasma cells, currently used in clinical diagnostics of MM, is a targeted test aimed at specific genomic loci. However, it is laborious and provides only genetic information of the probe targets. Microarray-based genomic profiling is a high-resolution tool that enables genome-wide analyses for copy number alterations (CNA), including focal CNA (<5 Mb) and regions of copy neutral loss of heterozogosity (CNLOH) that cannot be identified by FISH. A limitation of SNP-based array is its inability to identify balanced translocations.

The aim of this study was to compare FISH with SNP-based genomic arrays with respect to the detection yield for prognostic relevant genetic copy number abnormalities in enriched plasma cell samples from MM patients. In addition we have set up a diagnostic work flow in which on one sample of enriched plasma cells interphase FISH for (balanced) IGH translocations as well as SNP-based array for identification of CNA can be performed.

SNP-based genomic array profiling and FISH were performed in 37 MM patients. After enrichment of CD138 plasma cells half of each sample was treated with 0.075M KCl and, subsequently, fixed with 3:1 methanol/acetic acid and transferred to a microscopic slide for subsequent FISH. From the remaining part of the CD138-enriched plasma cell fraction DNA was extracted to perform SNP-based genomic array. Interphase FISH was performed according to standard methods using the following probes D5S23/D5S721/CEP9/CEP15, LSI13 (13q14), LSI TP53 (17p13.1) (all from Abbott Molecular, USA), and CDKN2C/CKS1B (from Cytocell, UK). 200 nuclei were analyzed by two different investigators and the detection limit was set at 20% as proposed by the EMN (Ross et al 2012; Haematologica 97:1272-1277). SNP-based array was performed using the CytoScan HD array platform (Affymetrix, USA), using the interpretation criteria as proposed by Schoumans (Schoumans et al 2016; Genes Chromosomes Cancer 55:480-491). Data regarding FISH and SNP-based array were obtained in a fully blinded fashion.

All prognostic relevant CNA as observed by FISH were also identified when only SNP-based genomic arrays would have been performed, including 4 cases with loss of 1p, 19 cases with gain of 1q, 14 cases with loss of 13q, 4 cases with loss of 17p, and 20 cases with a hyperdiploid karyotype. However, SNP-based arrays identified 20 additional prognostic relevant abnormalities which were not observed by interphase FISH for several reasons. Due to a higher detection limit of the applied SNP-based array platform, 2 cases with loss of 17p (abnormality present in 15-20% of the cells) and 1 case with loss of 13q and a hyperdiploid karyotype (present in 15% of the cells) were observed by SNP-array only. Four cases showed a 1p21 or 1p16 loss, which were not observed by FISH since these deleted regions were outside the 1p32 probe target region. In 3 cases tumor-associated regions of CNLOH were observed involving the regions 13q and 17p. Finally, in 4 cases in which FISH was suggestive for a hyperdiploid karyotype, the SNP-based array information regarding whole genome analysis and allele frequencies demonstrated that these 4 cases appeared to have a doubled up DNA content in their plasma cells, and therefore the losses of 1p, 13q and 17p were not observed by interphase FISH.

In conclusion, we demonstrate that SNP-based arrays are superior in the identification of prognostic relevant CNA in MM. SNP-based array do identify all CNA as observed by FISH, and in addition, identifies additional prognostic relevant abnormalities, such as loss of 1p, 13q, and 17p, that escaped the detection by FISH. The prognostic relevance of the CNLOH and the loss of 1p21 and 1p16 regions requires further evaluation in prospective clinical trials.