Abstract 310

CMML is characterized by monocytic proliferation, cytomorphologic dysplasia, and frequent progression to AML. Heterogeneity in or subtlety of presentation can make diagnosis challenging. Recent advances in molecular technology set the stage for systematic study of genetic and genomic lesions associated with CMML. Initially, RAS and RUNX1 mutations were identified in CMML; subsequently, mutations in TET2, CBL, ASXL1 and EZH2 have been discovered. Most recently, recurrent mutations in various genes of the spliceosomal machinery have been added, with SRSF2 likely the most common mutation in this condition. We hypothesized that more precise analysis of molecular lesions in CMML may allow for better categorization of this condition according to molecular pathogenesis and may provide clues to target therapy of this rather refractory condition. We identified 136 patients with CMML or secondary AML (sAML) with antecedent CMML: 87 CMML-1, 20 CMML-2 and 29 post-CMML sAML. The original cohort has been expanded by an additional 53 patients since first reported. The mean follow up period was 16 months (range, 0–114). Abnormal cytogenetics were found in 50% of the cohort by both metaphase and SNP-array-based karyotyping. In a representative subset of 27 patients, we have applied whole genome sequencing (WES) for which paired tumor/germ line DNA was used. To minimize false positives and focus on the most prevalent/relevant somatic events, we implemented a rational bioanalytic filtering approach and results were aligned using Burrows-Wheeler Aligner and variants detected using the GATK pipeline (Best Practice Variant Detection from Broad Institute). We focused on somatic defects with a frequency of >5% of the cohort. For the most commonly affected genes, results were validated using an expanded panel of 18 genes in 72 additional patients and, thus, for the most relevant genes a cohort of 95 patients was studied. The most frequently mutated genes were TET2 (48%), SRSF2 (35%), ASXL1 (17%) and RUNX1 (17%), whereas CBL (13%), EZH2 (13%), UTX (8%) and U2AF1 (8%), SETPB1 (10%), and RIT1 (9%) were less frequent. We also found TP53 and RUNX1 mutations in 5% and 16% of patients, respectively. A JAK2 V617F mutation was present in one case of seemingly typical CMML. BCOR and STAG2 mutations were found in 13% and 9% of patients, respectively; KRAS/NRAS mutations were in 10%. Spliceosomal gene mutations seem to be mutually exclusive, but were frequently associated with other non-spliceosomal gene mutations examined. Within the cohort of 28 SRSF2 mutant cases, 15 had coexisting TET2 mutations, 22 had ASXL1 mutations, 7 had RUNX1 and 5 had CBL mutations. Among 10 U2AF1 mutant cases, 3, 5 and 2 had TET2, ASXL1, and RUNX1 mutations, respectively. SETBP1 mutations were present in 34% of CMML-1/2 and frequently associated with RUNX1, SRSF2, CBL (approximately 2% each) and ASXL1 (4%) mutations. Cohesin mutations were less frequent (10%) because RAD21 and SMC mutations were absent. Mutations of PTPN11 and NF1 were less frequent in adult CMML than those reported in JMML. We also identified several less-recurrent gene mutations that likely modify pathogenesis or clinical outcomes of specific cases. Serial studies performed on 6 cases showed insight into the clonal architecture, producing a series of putative ancestral and secondary events, including uniparental disomy and acquisition of KRAS/NRAS or SETPB1 mutations.

Association between mutational status and overall survival (OS) was assessed using Kaplan-Meier statistics. While all permutations were tested, we highlight here only significant positive and relevant negative results. In the whole cohort, presence of CBL mutations conferred worse OS (p=.018; HR 2.44, 95%CI 1.18–4.69). Median OS was 16 months for CMML-1, 6 months for CMML-2, and 14-months for sAML. In subgroup analyses, CBL mutations were also significant worse prognostic factor in CMML-1 cohort (p=.037; HR 3.23, 95%CI 1.07–8.04).

In sum, WES provides intricate information on the molecular pathogenesis of CMML and the wide mutational spectrum correlates with the clinical diversity. Expert-based analysis of the genomic data may be supplanted by unsupervised and unbiased approaches which would cluster patients based on molecular similarities.


Maciejewski:NIH: Research Funding; Aplastic Anemia&MDS International Foundation: Research Funding. Makishima:Scott Hamilton CARES Initiative: Research Funding.

Author notes


Asterisk with author names denotes non-ASH members.