Significance of Flow Cytometric and Mutational Findings in Patients with Cytopenias and Limited or No Signs of Myelodysplasia By Cytomorphology

Introduction:The diagnosis of myelodysplastic syndromes (MDS) has been clearly defined by the WHO classification but remains a challenge in a significant number of cases with cytomorphologically borderline findings and normal karyotype. Furthermore, flow cytometry is capable of identifying MDS-specific aberrant antigen expression yet its value in these borderline cases as well as in those even without cytomorphologic findings of myelodysplasia remains to be clarified. Follow-up analyses as well as extension of diagnostic work-up to screening for molecular mutations may give further insight.

Aims:Assess the significance of cytomorphologically borderline dysplastic changes and of flow cytometric MDS-related findings in the absence of a clear-cut diagnosis of MDS by screening for molecular mutations and by diagnostic reassessment during follow-up.

Patients and methods: Bone marrow samples of 322 patients were assessed for suspected MDS by cytomorphology, flow cytometry and cytogenetics in parallel from 08/2005 to 11/2014 which 1) did not reveal a definite diagnosis of MDS by cytomorphology, 2) had a normal karyotype and 3) had at least one follow-up bone marrow assessment. By cytomorphology, 159 (49%) cases had borderline dysplastic findings while 163 (51%) had no sign of MDS. By flow cytometry, 138 (43%) cases had findings in agreement with MDS according to ELN criteria (Westers et al., Leukemia 2012; at least three aberrantly expressed antigens), 141 (44%) had borderline findings (one or two aberrantly expressed antigens) and 43 (13%) had no signs of MDS. A total of 699 follow-up samples were analyzed (median 2/patient). The median follow-up amounted to 3.0 years. In 147/322 patients (46%) screening for molecular mutations was performed on the initial samples, respectively, targeting a total of 20 genes (median 4 genes/patient, range 1-20). Analyzed genes were ASXL1, TET2, RUNX1, SRSF2, BCOR, DNMT3A, IDH2, NPM1, SF3B1, TP53, ZRSR2, CBL, CSF3R, ETV6, KDM6A, KRAS, MLL, SETBP1, SMC3 and U2AF1.

Results: A total of 145 patients (45%) were diagnosed with MDS by cytomorphology during follow-up. The median duration until diagnosis amounted to 3.4 years. Regarding initial cytomorphology, more cases with borderline dysplastic findings were diagnosed MDS at follow-up than those without any dysplastic findings (82/159 (52%) vs 63/163 (39%), p=0.025). However, the duration until diagnosis of MDS did not differ significantly between the two groups (median 2.6 vs 3.4 years). Regarding initial flow cytometry, more cases with findings in agreement with MDS were diagnosed MDS by cytomorphology at follow-up than those without (80/138 (58%) vs 65/184 (35%), p<0.001) while there was no difference between cases with one or two aberrantly expressed antigens at initial assessment vs those with none (51/141 (36%) vs 14/43 (33%), n.s.). The duration until diagnosis of MDS significantly differed between the groups as defined by flow cytometry and was shortest in cases in agreement with MDS at initial assessment and longest in those without any aberrantly expressed antigen (median 1.9 vs 4.1 vs 5.6 years, p<0.001). Overall survival (OS) for all cases was 80% at 5 years. While initial cytomorphologic results revealed no impact on OS, patients with an initial flow cytometric result in agreement with MDS tended to have a shorter OS (5 year OS 70% vs 88%, p=0.12). Molecular screening revealed mutations in 21/147 patients (14%) at initial assessment. Mutated genes included ASXL1 (mutated in 6 patients), TET2 (6), RUNX1 (3), SRSF2 (3), as well as 2 cases each for BCOR, DNMT3A, IDH2, NPM1, SF3B1, TP53 and ZRSR2 and 1 case each for CBL, CSF3R, ETV6, KDM6A, KRAS, MLL, SETBP1, SMC3 and U2AF1. The percentage of patients with at least one mutation did not differ between cases with borderline dysplastic findings by cytomorphology as compared to those without any dysplastic findings. In contrast, significantly more cases with findings in agreement with MDS by flow cytometry had at least one mutation as compared to those with one or two aberrantly expressed antigens as well as to those with none (15/71 (21%) vs 6/58 (10%) vs 0/18, p=0.012).

Conclusions: This data strongly supports the need to define the role of flow cytometry in the diagnostic work-up in suspected MDS and argues for an integrated approach with cytomorphology and cytogenetics. Implementation also of molecular data on mutations may further improve the validity of MDS diagnostics.