Improving the differential diagnosis between myelodysplastic syndromes and reactive peripheral cytopenias by multiparametric flow cytometry: the role of B-cell precursors

Since December of 2009, immunophenotyping was included in the diagnostic work-up of peripheral cytopenias in our Institution together with PB counts, BM cytology and karyotyping. The WHO criteria were used for the diagnosis of MDS and exclusion of deficiency anemias, viral infections, auto-imune diseases and renal or hepatic insufficiency was made [1,2]. During the period of the study (December 2009 – February 2013), we could confirm the diagnosis of MDS in 56 cases while in 35 cases the final diagnosis of reactive cytopenias was made (Table 1). Twenty five patients were excluded because of lack of complete clinical data or uncertain diagnosis.

The classification of the MDS cases was made according to the WHO criteria and risk category according to IPSS, IPSS-R [32] and WPSS (using hemoglobin values instead of “transfusion dependency”) [33] was assessed.

Cytogenetic analysis of BM was performed after 24 hours of culture according to standard methods. In each case, at least 20 mitoses were analyzed and the karyotypes were reported according to the International System for Human Cytogenetic Nomenclature [34].

Normal BM samples were obtained from 41 healthy donors for allogeneic bone marrow transplantation (age: 15–69 years) in order to standardize a normal immunophenotypic profile for our laboratory. All samples were collected during the period between July 2009 and January 2013.

All BM samples were obtained after informed consent was given by each person, according to the recommendations of the local Ethics Committee (proc. Nr. 0652.0.146.000-08).

Immunophenotyping was performed as previously described [10,19]. Briefly, the EDTA-anticoagulated BM sample (5-7 × 10⁶ cells in 100 μl per test) was processed using a standardized direct lyse-and-wash technique within 24 hours after bone marrow aspiration [3]. Quality control, calibration and compensation with FACS Comp were performed daily in our equipment.

Antigenic expression was studied using four-color combinations of monoclonal antibodies (MoAbs) conjugated with fluorescein isothiocyanate (FITC), phycoerythrin (PE), peridin clorophyll protein (PerCP) and allophyicocyanin (APC) fluorocromes. The following combinations were used to study the myelomonocytic maturation and progenitor cell populations: HLA-DR/CD14/CD45/CD33; CD16/CD11b/CD45/CD13; CD13/CD34/CD45/CD117; CD10/CD19/CD45/CD34; CD7/CD56/CD45/CD34. The specificity and source of each reagent have already been described in detail [23]. Immediately after staining, samples were acquired in a FACS Calibur flow cytometer (Becton Dickinson – BD Biosciences) using the CellQuest software (BD Biosciences). Instrument quality control, calibration using FacsComp™ (BD) software and spectral compensation were performed daily. Information about at least 100,000 nucleated BM cells was acquired for each sample. Data analysis was made using Infinicyt software (Cytognos). The strategies of analysis were standardized as previously described [10,19,23]. These variables also were assessed in normal BM.

The analysis of the myelomonocytic series was made as previously described [23] according to ELN standardization [3,27,31]. Briefly, maturation pattern of neutrophils was analyzed by their expression of CD13, CD16, CD11b, CD33 and HLA-DR. The side-scatter (SSC) of the granulocytic population and the antigen expressions were considered abnormal if the values of the mean fluorescence intensity (MFI) were above or below the benchmark values obtained for the normal cases (Table 2). Monocytes were analyzed by their expression of HLA-DR, CD64 and CD14. The combination CD16/CD11b/CD45/CD13 was used to quantify the CD16⁺ monocytes. The aberrant expression of CD34, CD7 and CD19 was investigated in each population considering abnormal when at least 10% of the cells expressed these antigens. For expression of CD56 in the myelomonocytic cell line, only values of >20% for granulocytes and >50% for monocytes were considered abnormal.

CD34⁺ cells were separated in the SSC/CD34 dot plot [10,19] (Figure 1) and their co-expression of CD19, CD10, CD13, CD117, CD7 and CD56 was examined. The B-cell precursors (CD34⁺/CD19⁺/CD10⁺) were measured as percentage of the total nucleated cells.

All immunophenotypic features were compared with the values found in normal BM. We computed the total number of granulocytic and monocytic abnormalities, as well as those of CD34⁺ cells. We also computed the sum of all abnormalities found.

The mean values and standard deviation were obtained for all variables analyzed. The difference between groups was assessed by analysis of variance. The normal distribution of the features of normal hemopoiesis was examined by the Kolmogorov-Smirnov test, and those with a non-normal distribution were submitted to a log transformation. The relation of B-cell precursors with age and among the groups studied was examined using the Spearman rank order correlation and multiple regressions. Values were considered significant when p < 0.05. The Winstat and SPSS.15 softwares were used for the calculations.

Concerning patients with MDS, the majority were RCMD (Table 1). In 5 cases no mitoses were available for karyotyping. As some of the patients’ groups in IPSS-R and WPSS classification were rather small, we grouped the cases with <5% BM blast (low risk) and MDS with > 5% BM blast (high risk) for analysis. A major part of the patients were low and intermediate risk in all clinical scores analyzed and had BM blasts <5%.

The cases with non-clonal cytopenias included deficiency anemias (n = 11), drug-induced cytopenias (n = 6), aplastic anemia (n = 3), idiopathic thrombocytopenic purpura (n = 4), auto-immune diseases (n = 4), thyroid dysfunction (n = 3), and infection-associated leucopenia (n = 4). There were 15 men and 20 women with a median age of 60 years (14–86).

The normal values were established by the analysis of 41 normal donors of allogeneic bone marrow transplantation with a median age 32 years (15–69); 25 males and 16 females. All features examined except for MFI for SSC in the granulocytic precursors and total B-cell precursors presented a normal distribution. So we used the mean values obtained ± 2 standard deviations except for MFI SSC of myelomonocytic cell lines and percentage of total B-cell precursors. For these two features we used the 5% and 95% percentiles, due to the large variation observed in the normal controls.

The abnormalities found in MDS are also shown in Tables 2 and 3. Most of the aberrancies studied were more common in MDS than in reactive cases. Concerning CD34⁺ cells, they were increased in 26% of the cases of MDS with BM blast count <5% in cytology and in 83% of the cases with RAEB. The same was true for the cells with the phenoptype CD34⁺/CD13⁺/CD117⁺ (Table 3). Abnormal co-expressions were rare in reactive cytopenias and common in MDS (Table 3).

Considering all the subjects studied, the median age was 55 years. Within the groups, normal subjects had a median age of 32 years (range: 15–69); reactive cytopenias: 60 years (range: 14–86) and MDS: 69 years (range: 15–84).

In the Spearman’s test, there was a negative correlation between age and BCP in all three groups: r = − 0.328; p 0.023 for healthy donors; r = − 0.321; p = 0.032 for reactive cytopenias (Figure 2A) and r = − 0.412; p = 0.002 for MDS cases (Figure 2B). These cells were absent in none of the normal donors, in 8/35 cases of reactive cytopenias and in 33/54 cases of MDS.

Among the subjects with age <55 years, mean B-cell precursors were 0.18%, 0.14% and 0.17% in normal controls, reactive cytopenias and MDS cases respectively, which was not significantly different. Among subjects >55 years old, mean values were 0.28%, 0.075% and 0.015% respectively. These values were significantly different (p < 0.005).

In the multiple regressions considering age, total % CD34⁺ cells, BCP and number of alterations in CD34⁺ cells, the relation could be described in non-clonal subjects (normals and reactive cytopenias) by the equation:

For the MDS cases the equation was: