Myelodysplastic syndromes (MDS) are diagnosed and classified based on cytomorphology (CM) and cytogenetics. The accurate identification of dysmyelopoiesis and quantification of bone marrow blasts requires a high degree experience in CM. Multiparameter flow cytometry (MFC) has been shown capable of identifying dysplastic features in all lineages and to reveal prognostic information in MDS. We analyzed 23 bone marrow samples from patients with untreated MDS by MFC, CM, and cytogenetics in parallel. In addition, four normal bone marrow samples and one bone marrow sample with hereditary sideroachrestic anemia (HSA) were analyzed by MFC for comparison. Diagnoses included RA (n=4), RARS (n=1), RAEB-1 (n=7), RAEB-2 (n=6), 5q- syndrome (n=2), and CMML (n=2). CM revealed dysgranulopoiesis and dyserythropoiesis in 16 and 14 cases, respectively, and blasts counts ranging from 3% to 20% (median, 9%). Karyotype aberrations included 5q- (n=3); trisomy 8 (n=2); del(11q) (n=2); t(3;11) (n=1); loss of chromosome X (n=1); and trisomy X, trisomy 8, and 5q- (n=1); while n=13 cases had normal karyotypes. Blast counts as determined by MFC ranged from 1.9% to 14.8% (median, 4.7%; correlation with CM: r=0.466, p=0.029). The median number of aberrant features detected by MFC were 4 for blasts (range, 2 to 9), 3 for granulocytes (1 to 8), 3 for monocytes (0 to 7), and 0 for erythrocytes (0 to 2). The presence of dysgranulopoiesis (CM) correlated with hypogranulation (MFC; 50% vs. 17%, p>0.05) and expression of HLA-DR (38% vs. 17%, p>0.05) and CD7 (13% vs. 0%, p>0.05) in granulocytes as well as with the lack of CD16 expression (63% vs. 0%, p=0.035), the expression of CD2 (25% vs. 0%, p>0.05), and the dim expression of CD66 (63% vs. 40%, p>0.05) on monocytes. Combining these parameters led to strong correlations between dysgranulopoiesis (CM) and at least one dysplastic feature by MFC for granulocytes (75% vs. 17%, p=0.023) and monocytes (81% vs. 33%, p=0.054). Combining the granulocytic and the monocytic MFC scores revealed dysplastic features in 100% cases with dysgranulopoiesis (CM) and in 50% without (p=0.013). No clear-cut correlations were found between dyserythropoiesis (CM) and dysplastic features in erythroid cells as detected by MFC. The most striking MFC features of cases with 5q- syndrome included CD2-positivity (100% vs. 6%, p=0.003) and CD34-positivity (100% vs 28%, p=0.042) in monocytes as well as hypogranulation (100% vs. 32%, p=0.055) and HLA-DR-positivity (100% vs 21%, p=0.023) in granulocytes. Dysplastic features that were detected by MFC were related to each other. Thus, in blasts the expression of CD56 was strongly related to the expression of CD64 (p=0.0003) and the dim expression of CD45 (p=0.0003). In addition, the expression of CD11b was associated with the expression of CD15 (p=0.001). In granulocytes, the lack of CD33 was associated with a lack of CD64 expression (p=0.0004). In monocytes, the dim expression of CD45 was associated with a lack of CD14 expression (p=0.0005). The case with HSA displayed an aberrant pattern of CD71 expression. These data suggest that MFC is capable of reproducibly detecting dysplastic features in different compartments of the bone marrow. Strong correlations between CM and cytogenetic finding and MFC are detectable. Future studies will define the role of MFC in optimizing the diagnosis of MDS and the prognostication in patients with MDS.

Author notes

Corresponding author

Sign in via your Institution