The Future is Now for the Laboratory Evaluation of Myelodysplastic Syndromes

Xu, Xinjie; Kelley, Todd

doi:10.1182/hem.V11.5.3179

Technological advances in DNA analysis have revolutionized our understanding of the genetic basis of myelodysplastic syndrome (MDS). As this technology is incorporated into the routine evaluation of patients, the clinical benefits of genetic analysis will continue to expand. Currently available are two widely used and complementary strategies for querying genetic abnormalities in patients with MDS: single-nucleotide polymorphism–based genetic arrays (SNPa) for detection of large-scale genomic gains and losses; and next-generation sequencing (NGS) studies for detection of single-gene abnormalities, including small insertions/deletions and substitution mutations (Table). This technology is revolutionizing the practice of clinical hematology in a variety of ways, including allowing for more accurate and timely diagnosis by the laboratory, particularly in those cases where the morphologic findings and the results of classical karyotyping are inconclusive. And the results of these tests are being used to personalize the management of patients with MDS. For example, the spectrum of genetic abnormalities often varies among patients with the same WHO classification of MDS, and, in such cases, molecular analyses can be used to subclassify patients into categories that have prognostic and therapeutic implications.

Comparison of Common Clinical Strategies for Assessing Genetic Abnormalities in Patients with Myelodysplastic Syndromes

Testing methodology	Whole-genome or targeted analysis?	Resolution	Detection of single base substitutions?	Detection of copy number changes?	Detection of balanced rearrangements?	Detection of aCN-LOH?	Growth in culture required?	Common findings in MDS patients
Metaphase cytogenetics	Whole genome	Low	No	Yes	Yes	No	Yes	Deletions of 5q, 7q, 20q; -Y, -5, -7; +8, i(17q), inv(3)/t(3;3), etc.
FISH	Targeted	Low - Moderate	No	Yes	Yes	No	No	Deletions of 5q, 7q, 20q and trisomy 8
SNP arrays	Whole genome	Moderate	No	Yes	No	Yes	No	Loss: 4q, 5q, 7q, 17p, 20q, 21q; Gain: 8; aCN-LOH: 4q, 7q, 11q, 17p
NGS-based mutation panels	Targeted	High	Yes	No*	No*	No	No	Mutations in: SF3B1, SRSF2, TET2, RUNX1, ASXL1, DNMT3A

Testing methodology	Whole-genome or targeted analysis?	Resolution	Detection of single base substitutions?	Detection of copy number changes?	Detection of balanced rearrangements?	Detection of aCN-LOH?	Growth in culture required?	Common findings in MDS patients
Metaphase cytogenetics	Whole genome	Low	No	Yes	Yes	No	Yes	Deletions of 5q, 7q, 20q; -Y, -5, -7; +8, i(17q), inv(3)/t(3;3), etc.
FISH	Targeted	Low - Moderate	No	Yes	Yes	No	No	Deletions of 5q, 7q, 20q and trisomy 8
SNP arrays	Whole genome	Moderate	No	Yes	No	Yes	No	Loss: 4q, 5q, 7q, 17p, 20q, 21q; Gain: 8; aCN-LOH: 4q, 7q, 11q, 17p
NGS-based mutation panels	Targeted	High	Yes	No*	No*	No	No	Mutations in: SF3B1, SRSF2, TET2, RUNX1, ASXL1, DNMT3A

This table was adapted from a table published in the Journal of Molecular Diagnostics, 16, Nybakken GE, Bagg A. The genetic basis and expanding role of molecular analysis in the diagnosis, prognosis, and therapeutic design for myelodysplastic syndromes, 147, Copyright Elsevier 2014.

*Depending on the informatics pipeline in place, it is possible to detect copy number changes and balanced rearrangements by NGS, although it is still uncommon in clinical testing.

SNPa

Conventional cytogenetic testing, including metaphase cytogenetics (MC) and fluorescence in situ hybridization (FISH) studies, is traditionally considered the gold standard for the genetic evaluation of patients with MDS. (Table) However, these strategies come with limitations. MC requires actively dividing cells and is a method with inherently low resolution. FISH can overcome some of the limitations of MC; but because FISH is a targeted assay, only those abnormalities under analysis (e.g., 5q-) will be detected. In addition, neither MC nor FISH can detect copy neutral loss of heterozygosity (CN-LOH), a frequent finding in MDS that can serve as a clonal disease marker. (Table) These technical limitations impede the discovery of cryptic or low-frequency chromosomal lesions that may have diagnostic and/or prognostic significance. Conventional cytogenetic testing identifies chromosomal abnormalities in approximately 50 percent of patients with MDS. For the remainder with normal-karyotype MDS, genetic abnormalities will go undetected unless samples are analyzed further. The natural history of normal-karyotype MDS is variable, and in this setting, higher-resolution genetic analysis may provide clinically relevant information.

SNPa has emerged as a valuable tool for identifying both copy number changes undetected by MC or FISH and CN-LOH. (Table) Because SNPa does not depend on the availability of actively dividing cells, it is especially useful when MC has failed due to lack of growth of bone marrow cells in culture. (Table) SNPa can detect chromosomal lesions in patients with MDS at a significantly higher rate compared with conventional cytogenetic testing (75% vs. 50%) due to its capacity to identify cryptic chromosomal deletions and gains and acquired CN-LOH (aCN-LOH).^1-4For example, recurrent small deletions beyond the resolution of MC and FISH involving genes such as TET2 on 4q24, DNMT3A on 2p23.3, and RUNX1 on 21q22 discovered by SNPa have been reported in patients with MDS.^2,4 In addition, SNPa detects aCN-LOH in 20 percent of MDS patients.²aCN-LOH represents a major mechanism by which tumor suppressor genes are inactivated or by which oncogenic mutations become homozygous.²For example, aCN-LOH involving chromosomes 7, 11, and 17 underlies one of the processes by which both copies of the tumor suppressor genes EZH2, CBL, and TP53,respectively, are inactivated, and those genotypes are associated with an unfavorable overall survival, similar to that observed when deletions account for the inactivation process.⁴The presence of additional chromosomal lesions detected by SNPa is an independent predictor of adverse clinical outcomes, including worse overall survival and greater risk of progression to acute myeloid leukemia (AML).⁴The capacity of SNPa to identify cryptic chromosomal lesion in patients with MDS not only improves the diagnostic yield of the laboratory testing but also refines risk stratification by identifying patients with more aggressive disease who might benefit from alternative therapies through participation in clinical trials.

NGS

NGS, also called massively parallel sequencing, rapidly generates enormous amounts of DNA sequence data that, when coupled with sophisticated informatics algorithms, can be used to assemble the nucleotide sequence of the entire genome of an individual (composed of ~3 billion base pairs). This procedure is called whole genome sequencing. Alternatively, the procedure can be modified such that only the protein coding components of each gene are sequenced. This process is called whole exome sequencing and reduces the amount of sequence data generated by 99% compared to whole genome sequences as the 180,000 exomes of the human genome consist of approximately 30 million base pairs. NGS can also be used to sequence a specific set of genes (called mutation panels) that are usually related in some way. For example, one sequencing panel might consist of genes known to be frequently mutated in MDS, or another panel might consist of genes involved in a signaling pathway involved in leukemia pathogenesis. (Table) Initially, NGS was used primarily as a research tool, but the combination of lower cost and improvements in both sequencing technology and bioinformatics has made feasible the implementation of NGS protocols in clinical laboratories. In fact, in late 2013, the FDA granted marketing authorization for a commercial high-throughput sequencer. This marketing authorization of a non-disease-specific platform allows any lab to test any sequence for any purpose. At a number of larger medical centers throughout the US, NGS is available for analysis of DNA samples from patients with MDS and other myeloid malignancies, with some results (depending upon the complexity of the analysis) being reported within a week. (Table) A recent study involving DNA samples from 944 patients that used a panel consisting of 104 genes demonstrated that 90 percent of MDS patients have at least one these genes mutated with three being the median number of mutated genes observed in this study.⁵Like the results of SNPa testing, the information derived from NGS can be valuable in establishing a diagnosis in patients without sufficient morphologic evidence of dysplasia or when there is no evidence of cytogenetic abnormalities using traditional methods. Data from NGS can also be used to determine if a patient fits into a prognostically significant disease subclassification.

MDS-associated mutations recurrently affect the same discrete cellular pathways. These include regulators of transcription (RUNX1, ETV6), cell cycle regulators (TP53), components of the RNA splicing complex known as the spliceasome (SF3B1,SRSF2, U2AF1), regulators of epigenetic functions such as methylation and chromatin remodeling (TET2, EZH2, DNMT3A, ASXL1, IDH1, IDH2), and components of the cohesin complex (STAG2). In terms of frequency, mutations in one of the components of the spliceasome complex are present in almost half of cases, with SF3B1 and SRSF2 being most commonly affected.⁶Mutations in TET2, DNMT3A, RUNX1, and ASXL1 also appear commonly in MDS patients with frequencies of at least 10 percent.^5,7 Many other genes are recurrently mutated in MDS, albeit at a lower frequency. These same abnormalities may be found across the spectrum of myeloid malignancies, including in acute myeloid luekemia, in myeloproliferative neoplasms (MPN; particularly TET2 and ASXL1), and in MPN/MDS neoplasms such as chronic myelomonocytic leukemia.

In general, a higher overall number of driver mutations correlates with worse outcomes in MDS.⁸Specific mutations also have independent prognostic significance. For example, SRSF2 mutations are associated with a higher frequency of progression to AML,⁹while mutations in SF3B1 are associated with longer event-free survival.¹⁰Specific pairwise associations between mutations have also been observed, suggesting co-dependence or co-evolution during disease progression. Mutations may also demonstrate mutual exclusivity. Although the revised international prognosis scoring system (IPSS-R) does not incorporate mutation status of individual genes, in many cases IPSS-R-independent prognostic information is gained through this type of analysis. With the goal of developing more informative scoring systems, comprehensive prognostic models have been proposed that incorporate the entire spectrum of clinical, laboratory, and genetic information.⁵

SNPa testing and NGS-based mutation studies are revolutionizing the clinical and laboratory evaluation of patients with hematologic malignancies including MDS. The analyses can usually be performed on either peripheral blood or bone marrow, thus potentially eliminating the need for repeated bone marrow aspirations to obtain samples.¹¹With the use of these methods, it is now possible to detect clonal genetic abnormalities in almost all patients with MDS. Although challenges remain, including cost and reimbursement issues as well as technical and data analysis hurdles, comprehensive genetic testing is a notable improvement over classical karyotyping alone, facilitating more accurate and timely diagnoses and adding prognostic information. As use of the technology becomes more widespread, pathologists and hematologists will be relied upon to ensure the appropriate integration of newly available genetic information with traditional clinical, laboratory, morphologic, and histopathologic findings.