Prognostic impact of SNP array karyotyping in myelodysplastic syndromes and related myeloid malignancies

Abstract

Single nucleotide polymorphism arrays (SNP-As) have emerged as an important tool in the identification of chromosomal defects undetected by metaphase cytogenetics (MC) in hematologic cancers, offering superior resolution of unbalanced chromosomal defects and acquired copy-neutral loss of heterozygosity. Myelodysplastic syndromes (MDSs) and related cancers share recurrent chromosomal defects and molecular lesions that predict outcomes. We hypothesized that combining SNP-A and MC could improve diagnosis/prognosis and further the molecular characterization of myeloid malignancies. We analyzed MC/SNP-A results from 430 patients (MDS = 250, MDS/myeloproliferative overlap neoplasm = 95, acute myeloid leukemia from MDS = 85). The frequency and clinical significance of genomic aberrations was compared between MC and MC plus SNP-A. Combined MC/SNP-A karyotyping lead to higher diagnostic yield of chromosomal defects (74% vs 44%, P < .0001), compared with MC alone, often through detection of novel lesions in patients with normal/noninformative (54%) and abnormal (62%) MC results. Newly detected SNP-A defects contributed to poorer prognosis for patients stratified by current morphologic and clinical risk schemes. The presence and number of new SNP-A detected lesions are independent predictors of overall and event-free survival. The significant diagnostic and prognostic contributions of SNP-A-detected defects in MDS and related diseases underscore the utility of SNP-A when combined with MC in hematologic malignancies.

Figure 1

Comparison of exemplary lesions detected by genomewide SNP-A 6.0. (A) Whole-genome view shows the complete diploid set of human chromosomes. A close-up view of chromosome 21 shows the presence of a deletion on its long arm, illustrated by a decrease in the copy number line (blue line and red dots) and LOH (paucity of green bars, indicating the frequency of heterozygous loci [red box]). A closer look at the region shows loci potentially involved in disease pathogenesis. (B) Illustration of the 3 exemplary characteristic lesions seen by 6.0 SNP-A: deletion, duplication, and aCN-LOH. For deletions, LOH as indicated by contraction of the AA and BB alleles (blue lines) and reduction of the copy number (top) marks the affected region. For duplications, the doubling of the heterozygous AB allele and an increase in copy number delineate the abnormality. For aCN-LOH, the loss of the AB allele without contraction of the AA and BB alleles and a normal copy number identify the lesion. In each example, the normal copy number and allele calls for CD3⁺ cells (green lines) shows the absence of a germline abnormality, thus confirming the somatic nature of the lesions. (C) An exemplary microdeletion in chromosome 7, seen as a drop in the diploid copy number line (blue line in red box) and the absence of this abnormality in the CD3⁺ lymphocytes, is shown.

Figure 2

Types and genomic distribution of chromosomal lesions detected by SNA-A karyotyping in MDSs. Gains (green), deletions (red), and acquired somatic uniparental disomy (blue) are shown.

Figure 3

Clinical effect of lesions detected by SNP-A karyotyping analysis in combination with MC and new lesions detected by SNP-A regardless of MC results. Patients with new defects (red line) detected by the combination MC or SNP-A had worse OS (A), EFS (B), and PFS (C) than patients with no defects (black line) detected by both techniques. Patients with new lesions detected by SNP-A regardless of MC results showed worse OS (D), EFS (E), and PFS (F) than patients with an unchanged karyotype. In addition, SNP-A–identified lesions had an effect on OS in patients in the low-risk IPSS group (G) but not on EFS (H) and PFS (I). New SNP-A–detected defects did not affect outcomes in the high-risk IPSS group (J, K, and L). Tables represent the number at risk over time. Characteristics of censored patients have been presented in supplemental Table 4.

Figure 4

New prognostic risk groups based on combined MC and SNP-A analyses. SNP-A was able to further delineate cytogenetic risk groups determined by MC as defined by IPSS (A-B). Four prognostic risk groups, namely favorable, intermediate-1, intermediate-2, and unfavorable were defined with the use of the combined karyotypic analysis by MC and SNP-A (C-D). †Favorable indicates good cytogenetics and no additional lesions detected by SNP-A. Intermediate-1 indicates good cytogenetics and additional SNP-A lesions detected, or intermediate cytogenetics and no additional lesions by SNP-A. Intermediate-2 indicates intermediate cytogenetics and additional lesions detected, or poor cytogenetics and no additional lesions detected by SNP-A. Unfavorable indicates poor cytogenetics by MC and additional lesions detected by SNP-A.