Utility of SNP Arrays in Chronic Myelomonocytic Leukemia with Low Risk Cytogenetic Features or No Metaphases

Background: Clinical, morphological and genetic characteristics of chronic myelomonocytic leukemia (CMML) are heterogeneous and vary from a myelodysplastic predominant profile to a myeloproliferative one. CMML has a highly variable course, with a median overall survival (OS) of 20 months and 15-30% of progression to acute myeloid leukemia (AML). Cytogenetic abnormalities are present in only 20-40% of cases. CMML-specific cytogenetic risk classification stratifies karyotypes into three groups: low risk (normal karyotype and isolated loss of Y chromosome, -Y), poor risk (trisomy 8, monosomy 7, 7q deletion and complex karyotype) and intermediate risk (all other chromosomal abnormalities). According to this model, 65-85% of patients fall into the low risk cytogenetic category. The aim of this study was to characterize type, frequency and prognostic impact of cytogenetic alterations detected by SNP arrays (SNP-A) in a series of 128 patients with CMML and low risk cytogenetic features or no metaphases.

Methods: A retrospective study was performed on 128 patients with CMML. Cases with normal karyotype (n=120), isolated -Y (n=4) and no metaphases (n=4) were selected. Median age at diagnosis was 73 years (range 39-98), there was a 2.4:1 male predominance and 22.4% (28/125) of cases progressed to AML. Median follow up of patients was 26 months (range 1-115). Morphological WHO subtypes were CMML-1 in 104 (81%) cases and CMML-2 in 24 (19%). According to the FAB criteria 82 (64%) cases were included in the myelodysplastic variant (CMML-MD) and 46 (36%) in the myeloproliferative one (CMML-MP). High density SNP-A (Cytoscan HD, Affymetrix) were performed using DNA extracted from bone marrow (n=124) or peripheral blood (n=4) samples at diagnosis. The statistical analysis was performed with SPSS. Kaplan-Meier method was used for OS and progression-free survival (PFS) analysis and log-rank test was used for comparisons between groups.

Results: SNP-A revealed novel chromosomal alterations (copy number alterations, CNA, and loss of heterozygosity, LOH) in 66% (85/128) of cases. Among the abnormal cases (CNA plus LOH), 1 alteration was detected in 65% (55/85) of cases, 2 in 20% (17/85) and ≥3 in 15% (13/85). The median size of the affected genome for CNA and LOH was 759 Kb (range 0-142Mb).

CNA were detected in 38% (48/128) of cases, most of them being gains and losses smaller than 10Mb. Only 7 CNA were larger than 10Mb, four of them corresponded to patients with isolated -Y and three were novel alterations that had not been detected by conventional G-banding cytogenetics. Most affected regions (detected in 5 cases) were gains in chromosomes 3q, 8p and 21q as well as losses in chromosomes 10q and 12p.

LOH were detected in 39% (50/128) of patients. Interstitial LOH larger than >25Mb were detected in 30% (39/128) of cases. Recurrent interstitial LOH were detected in: 4q24-4q35 region (12 cases), involving TET2 gene, which is mutated in 40-50% of CMML; and in chromosome 11 (9 cases), 7 of which included 11q13.3-11q25 region, involving CBLgene, mutated in 5-20% of CMML.

Although the number of total alterations (1, 2 or ≥3) do not have a survival impact, we found a significant correlation between the size of the affected genome (≥30 Mb, including CNA and LOH) and a poorer OS (OS at 3 years 26% vs. 46%, P=0.039). Regarding type of alterations, cases with interstitial LOH had lower OS than cases without LOH (OS at 3 years 19% vs. 45%, P=0.049). Cases with high risk alterations as defined by CMML-specific cytogenetic risk classification (losses in 7q, gains in 8q and ≥3 CNA) had lower PFS than patients with other aberrations (PFS at 1 year 76% vs. 93%, P=0.036), although statistical significance was not reached for OS (OS at 3 years, 27% vs.49%, P=0.096).

Conclusions: SNP-A in CMML patients with low risk cytogenetic features or no metaphases has led to the detection of chromosomal alterations in 66% of cases. Using this technique the prognosis can be better defined, and we can detect a group of patients with worse outcome who could be considered for more intensive treatment.

Acknowledgments: Instituto de Salud Carlos III, Ministerio de Sanidad y Consumo, Spain (PI 11/02519; PI 11/02010); RTICC, FEDER (RD12/0036/0044; RD12/0036/0014); 2014 SGR225 (GRE) Generalitat de Catalunya; Fundació Internacional Josep Carreras, Obra Social “La Caixa” and Celgene Spain; NHRI-EX103-10003NI, Taiwan.

Footnote: Francesc Solé and Lurdes Zamora contributed equally