Combined Impact of Aberrant Immunophenotypes, Somatic Mutations and Cytogenetic Aberrations on the Probability of Developing Myelodysplasia in Patients with Cytopenias of Undetermined Significance

Background: Myelodysplastic syndromes (MDS) are diagnosed according to WHO based on cytomorphology and cytogenetics. Various conditions without dysplasia have been described which are associated with an increased although low risk of developing myeloid malignancy (Steensma et al. Blood 2015): non-clonal idiopathic cytopenias of undetermined significance (ICUS, cytopenias and no mutations), clonal hematopoiesis of indeterminate potential (CHIP, mutations and no cytopenias) and clonal cytopenias of undetermined significance (CCUS, both cytopenias and mutations). The spectrum of encountered mutations resembles the one of MDS. MDS-typical immunophenotypes have been identified and are recommended by ELN in the diagnostic work-up of patients (pts) with suspected MDS (Malcovati et al., Blood 2013). Similar to mutations described in both MDS and CHIP/CCUS, MDS-typical immunophenotypes are found in cytopenic non-MDS pts. Our study aims at the determination of the diagnostic potential of both of these findings.

Aims: To determine the impact of aberrant immunophenotypes, somatic mutations and cytogenetic aberrations in the absence of dysplasia on the later diagnosis of MDS.

Patients and methods: Between 2005 and 2018 a total of 2279 bone marrow samples of pts with suspected MDS were analyzed by a combination of cytomorphology and cytogenetics as well as NGS and flow cytometry. NGS targeted the following genes in all pts: ASXL1, CBL, DNMT3A, EZH2, JAK2, RUNX1, SF3B1, SRSF2, TET2, TP53 and U2AF1. Further genes analyzed in subsets of pts only were BCOR, CSF3R, CSNK1A1, ETV6, GATA1, GATA2, IDH1, IDH2, KIT, KRAS, NPM1, NRAS, PTPN11, SETBP1, WT1 and ZRSR2. Flow cytometry was performed according to ELN standards (Westers et al., Leukemia 2012). For 368 pts at least one follow-up bone marrow analysis was available, these pts are the subject of the present study. Patients ages ranged from 20 to 90 years (median 70 years), male:female ratio was 1.9:1.

Results: 277/368 (75%) pts were diagnosed with MDS by cytomorphology. In 263/368 (72%) of pts MDS-typical immunophenotypes were present. Findings concordant between cytomorphology and flow cytometry were present in 304/368 (83%) pts reflecting published large series. Focus of further analyses were the 91 pts without diagnosis of MDS by cytomorphology. In 34 of these 91 pts MDS was diagnosed at follow-up, median follow-up time was 16.3 months, median time to MDS was 38.5 months. At initial evaluation, in 25 of these 91 (27%) pts MDS-typical immunophenotypes were present and resulted in a shorter time to MDS as compared to those without MDS-typical immunophenotypes (median 15.9 vs. 39.9 months, p=0.012). Frequencies of mutations with impact on time to MDS were as follows: ASXL1, present in 9/91 (10%), median time to MDS ASXL1 mutated vs non-mutated 14.2 vs. 39.9 months, p=0.010; EZH2, 3/91 (3%), 7.4 vs. 39.9 months, p=0.020; JAK2, 3/91 (3%), 4.9 vs. 39.9 months, p=0.003; RUNX1, 5/91 (5%), 7.4 vs. 39.9 months, n.s.; SRSF2, 7/91 (8%), 11.9 vs. 39.9 months, p=0.008; TET2, 14/91 (15%), 14.2 vs. 57.1 months, p=0.029; TP53, 4/91 (4%), 6.5 vs. 44.1 months, p=0.007. Other genes were less frequently found mutated without an impact on the time to MDS. An aberrant karyotype was present in 9 of 89 evaluable pts and was associated with a shorter time to MDS (median 4.9 vs. 57.1 months, p<0.001). Also age was associated with a shorter time to MDS (HR 1.47 per decade, p=0.029), peripheral blood cell counts were not. Patients with two out of the three parameters 1) aberrant immunophenotype, 2) at least one mutation in the genes ASXL1, EZH2, JAK2, RUNX1, SRSF2, TET2, TP53, 3) aberrant karyotype had a significantly shorter duration to MDS as compared to those without: 1) and 2), 7.4 vs. 44.1 months, p<0.001; 1) and 3), 4.9 vs. 39.9 months, p=0.004; 2) and 3), 16.0 vs. 44.1 months, p=0.041. Patients presenting with any two out of the parameters 1), 2) and 3) had a highly significantly shorter time to MDS as compared to those presenting with only one or none: 6.5 vs. 57.1 months, p<0.0001.

Conclusions: In the absence of cytomorphologic signs of dysplasia, findings commonly present in MDS patients, i.e. aberrant immunophenotypes, somatic mutations and cytogenetic aberrations, occur in patients with cytopenias of undetermined significance. The presence of two of these three findings may be considered as a pre-diagnostic state of MDS. Further studies should clarify the diagnostic potential of this approach.