Four-Color Flow Cytometry Shows Strong Concordance With Bone Marrow Morphology and Cytogenetics in the Evaluation for Myelodysplasia

Transcription

1 Hematopathology / FLOW CYTOMETRY TO ASSESS MYELODYSPLASIA Four-Color Flow Cytometry Shows Strong Concordance With Bone Marrow Morphology and Cytogenetics in the Evaluation for Myelodysplasia Steven J. Kussick, MD, PhD, 1 Jonathan R. Fromm, MD, PhD, 1 Anthony Rossini, PhD, 2 Ying Li, MD, PhD, 1 Anthony Chang, MD, 1 Thomas H. Norwood, MD, 3 and Brent L.Wood, MD, PhD 1 Key Words: Myelodysplasia; Myelodysplastic syndromes; Flow cytometry; Antigen expression; Cytogenetics; Bone marrow DOI: /6PBP78G4FBA1FDG6 Abstract The ability of 4-color flow cytometry (FC) to help identify myelodysplastic syndromes (MDSs) was evaluated in 124 bone marrow aspirates from unselected patients with unexplained cytopenias and/or monocytosis. The morphologic features of bone marrow aspirate smears were correlated with FC and cytogenetic findings blindly, and patterns of antigen expression were compared with patterns seen in nonneoplastic and normal marrow specimens. Of 124 cases, 58 (46.7%) had definitive FC abnormalities ( flow-abnormal ), 19 cases (15.3%) had mild FC abnormalities of indeterminate significance, and 47 cases (37.9%) had essentially normal FC. Highly significant differences were identified between the flowabnormal group and other groups in mean myeloid blast percentages and numbers of abnormal antigens expressed, even when the analysis was limited to cases with fewer than 5% myeloid blasts. Strikingly, flowabnormal cases constituted 50 (89%) of the 56 morphologically abnormal cases and 31 (94%) of the 33 cytogenetically abnormal cases, demonstrating the strong concordance of FC-identified antigenic abnormalities with morphologic features and cytogenetics in the evaluation of patients with unexplained cytopenias. The diagnosis of the myelodysplastic syndromes (MDSs) historically has relied on combining the clinical history, morphologic features of the peripheral blood and/or bone marrow sample, and cytogenetic information. However, because morphologic evaluation is inherently subjective, and cytogenetics, while objective, identifies abnormalities in only 30% to 40% of MDSs, 1 additional objective correlates of MDS are needed. Multiparametric flow cytometry (FC) represents a highly reproducible and objective way of assessing the expression of multiple antigens on a single cell. By comparing patterns of antigen expression on a given cell population with the patterns identified on normal cells of that type, one potentially can identify abnormalities that, if sufficiently great, might substitute for clonality studies in identifying malignancy. Patterns of expression of a number of antigens during normal myelopoiesis have been described in relatively great detail. 2-4 A number of studies during the past 15 years have applied FC to the study of MDSs. Most of the earlier studies investigated a relatively small number of surface antigens, and these studies are well reviewed and critiqued by Elghetany. 5 Some of the described antigenic abnormalities have included the following: (1) loss of erythrocyte A, B, and H antigens in MDSs 6 ; (2) decreased expression of c-mpl, glycoprotein IIb/IIIa, and glycoprotein Ib on platelets from patients with refractory anemia (RA) 7 ; (3) dyssynchronous expression of CD11b and CD16 in the developing neutrophils of patients with MDS 8 ; (4) aberrant coexpression of CD14 and CD66 on the myeloid cells in a subset of MDSs 9 ; (5) decreased CD10 on neutrophils in MDSs 10 ; (6) changes in a variety of leukocyte activation antigens, including FcR I, FcR II, and FcR III, in MDSs 11 ; (7) greater variability in the expression of CD38, CD71, CD13, and CD33 in RA vs normal marrow or marrow 170 Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6

2 Hematopathology / ORIGINAL ARTICLE involved by aplastic anemia 12 ; and (7) aberrant coexpression of CD56 on myeloid blasts in MDSs. 13,14 Two recent studies identified immunophenotypic abnormalities among relatively mature myeloid cells in the bone marrow 15 and peripheral blood neutrophils 16 of patients with MDSs, including decreased side light scatter, 15,16 decreased CD10, 15 and increased HLA-DR, 15 CD11a, 16 and CD A variety of studies have used FC to document abnormalities in apoptosis and/or proliferation in MDSs. The latter studies generally have found abnormally increased apoptosis on bone marrow precursors in RA and RA with ringed sideroblasts, with associated increases in caspase 3 activation 20,21 and in expression of proapoptotic bcl-2 family members such as bad, bak, and bcl-xs. 22,23 In contrast, in RA with excess blasts (RAEB), RA with excess blasts in transformation (RAEB-T), and chronic myelomonocytic leukemia (CMML), there tends to be a relative increase in proliferation. 17 Several studies have looked more generally at the role of 3-color FC in the diagnosis or classification of MDSs. Stetler- Stevenson and colleagues 24 used 3-color FC to identify immunophenotypic abnormalities in the bone marrow of 45 patients with well-established MDSs and then applied FC to a series of 20 equivocal cases to demonstrate that FC is helpful in making the final diagnosis of MDS. Ogata and colleagues 25 used 3-color FC to evaluate enriched populations of blasts from 95 patients with MDSs and 21 patients with acute myeloid leukemia arising from an MDS. They found that blasts from patients with a low-risk MDS (RA or RA with ringed sideroblasts) often expressed antigens associated with some degree of myeloid maturation, whereas blasts from patients with a high-risk MDS (RAEB, RAEB-T, and CMML) tended to have a less mature immunophenotype. Maynadie and coworkers 26 used 3-color FC along with a hierarchical clustering algorithm to identify immunophenotypic features that could help distinguish different types of MDS as defined by the French-American-British classification. 27 Additional immunophenotypic clusters were found to correlate with the International Prognosis Scoring System score. Del Canizo et al 28 used 3-color FC to evaluate 7 myeloid-associated antigens and noted that 90% of samples from patients with an MDS showed aberrant immunophenotypes, compared with control subjects. Finally, Wells and coworkers 29 used 3-color FC to develop a scoring system for quantifying the overall extent of immunophenotypic abnormalities. This group found that the FC scores correlated inversely with leukocyte and absolute neutrophil counts and directly with International Prognostic Scoring System risk categorization and the likelihood of relapse following stem cell transplantation for MDS. In the University of Washington Hematopathology Laboratory (UWHL), Seattle, we have used 4-color FC since 1998 to evaluate bone marrow aspirates from patients suspected of having myeloid stem cell neoplasms and have accumulated a large data set of more than 800 normal and abnormal bone marrow samples. This data set has allowed us to identify reproducible patterns of antigen expression in normal granulocytic and monocytic maturation, including changes seen in benign and reactive settings such as marrow regeneration. 30 This understanding of benign patterns of myeloid antigen expression, in turn, has allowed us to use 4-color FC to identify antigenic abnormalities that, if sufficiently extensive, can be used to support the possibility of a myeloid stem cell neoplasm, as previously demonstrated in the workup of myeloproliferative disorders of the nonchronic myelogenous leukemia type. 31 We now validate our 4-color method for assisting in the diagnosis of MDS by showing strong concordance of the FC results with the morphologic and cytogenetic features of an unselected series of patients with unexplained cytopenias or monocytosis. Materials and Methods This study was based on the retrospective identification of 180 bone marrow aspirate specimens evaluated in our laboratory by FC to rule out MDS and/or CMML between June 1998 and December 2001, and in which concurrent cytogenetic information was available. For 124 of these cases, adequate aspirate smears existed in the UWHL to enable morphologic correlation with the FC and cytogenetic findings; detailed CBC count data from the time of bone marrow evaluation were available for the large majority of these patients. The University of Washington Human Subjects Review Committee approved this study (application E 01), as did the institutional review board of the Seattle Veterans Affairs Medical Center and the institutional review boards of 7 community hospitals that refer bone marrow specimens to the UWHL for FC evaluation. Flow Cytometry In all cases, erythroid cells were bulk lysed with buffered ammonium chloride, washed once with phosphate-buffered saline (PBS) bovine serum albumin (BSA)-azide, ph 7.4, and resuspended to the desired cell concentration in PBS-BSA or RPMI. We incubated 100 µl of the cell suspension ( to cells) with appropriate amounts of titered antibodies for 15 minutes at room temperature in the dark, washed the suspension once with PBS-BSA-azide, and resuspended it in 1% paraformaldehyde. In most cases, 100,000 to 150,000 cells were analyzed from each tube of cells and antibodies. One of the following 2 panels of antibodies was used in virtually all cases: (1) IgG2a-fluorescein isothiocyanate (FITC)/IgG1-phycoerythrin (PE)/CD45-PE Texas red (ECD)/IgG2a-PE cyanine 5 (Cy5); HLA-DR-FITC/CD33-PE/CD45-ECD/CD15- PE-Cy5; CD2-FITC/CD56-PE/CD45-ECD/CD14-PE-Cy5; Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6 171

3 Kussick et al / FLOW CYTOMETRY TO ASSESS MYELODYSPLASIA CD10-FITC/CD19-PE/CD45-ECD/CD5-PE-Cy5; CD34- FITC/CD11b-PE/CD45-ECD/CD16-PE-Cy5; CD7- FITC/CD13-PE/CD45-ECD/CD117-PE-Cy5; or (2) HLA- DR-FITC/CD33-PE/CD45-ECD/CD14-PE-Cy5; CD15- FITC/CD11b-PE/CD45-ECD/CD34-PE-Cy5; HLA-DR- FITC/CD13-PE/CD45-ECD/CD16-PE-Cy5; CD38-FITC/ CD117-PE/CD45-ECD/CD34-PE-Cy5; CD5-FITC/CD56- PE/CD19-ECD/CD45-PE-Cy5; CD2-FITC/CD7-PE/ CD45-ECD/CD34-PE-Cy5. The first panel was used from approximately June 1998 through October 2000; the second panel was used from November 2000 through end of the study period in December Anti-CD45 antibodies were included in each tube for gating purposes, 32 whereas the combinations of the other 3 antibodies were designed empirically to maximize our ability to identify antigenic abnormalities among the myeloid blasts, maturing myeloid cells, and maturing monocytes. The specific clones used and strategies for assembling our panels of antibodies have been described previously. 30,31 In each case, the amount of antibody used was based on the manufacturer s suggestion or titration experiments to optimize the signal/noise ratio. Once the cases were identified, archival 4-color FC data from these cases were reanalyzed in a standardized manner (by S.J.K.) on Macintosh G3 or G4 computers (Apple, Cupertino, CA) using software developed in our laboratory (by B.L.W.) as described previously. 30,31 The patterns of antigen expression on the myeloid blasts, maturing myeloid cells, and monocytes were compared with the patterns typically seen in normal myeloid populations by 3 independent observers (S.J.K., Y.L., and B.L.W.) who were blinded to the morphologic and cytogenetic findings. A negative bone marrow sample from a patient with iron deficiency anemia was chosen as the gold standard for comparison with the case specimens Image 1, although the patterns of myeloid antigen expression in this control marrow sample were essentially identical to those observed in 11 additional negative bone marrow aspirates obtained at the time of initial staging for non-hodgkin lymphoma 31 and in virtually all presumed-normal marrow samples that we have evaluated during the 6 years we have been doing this assay. Classification of FC Results Cases were classified as flow-normal, flow-indeterminate, and flow-abnormal, depending on a global assessment of the FC findings in each case, as described previously. 31 In this assessment, 5 major types of antigenic abnormalities were identified: (1) deviations in myeloid antigen intensity, defined as an increase or decrease of at least one third of a decade on a log scale compared with normal, in at least 10% of the cells in the population of interest; (2) abnormally homogeneous antigen expression, without the maturational spectrum normally seen in the population of interest; (3) asynchronous expression of 2 myeloid-associated antigens in most of the Blast to myeloid Blast to myeloid Blast to myeloid Blast to myeloid Blast to myeloid CD11b PE CD16 PE-Cy5 CD15 FITC CD38 FITC Blast to mono Blast to mono Blast to mono Blast to mono Blast to mono CD14 PE-Cy5 CD11b PE CD16 PE-Cy5 CD15 FITC CD38 FITC Image 1 Four-color flow cytometric analysis of the progression from blasts to mature neutrophils (upper row of dot-plots) and monocytes (lower row of dot-plots) during normal maturation in the bone marrow. Red, blast population; green, maturing myeloid-neutrophil population; and lavender, the maturing monocytic population. Arrows denote the changes in antigen expression as maturation progresses from the early myeloid blasts at the base of each arrow to the mature neutrophils or monocytes at the arrowheads. Cy5, cyanine 5; FITC, fluorescein isothiocyanate; PE, phycoerythrin. 172 Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6

4 Hematopathology / ORIGINAL ARTICLE cells in a population of interest; (4) aberrant expression of nonmyeloid antigens, also defined as expression of the abnormal antigen by at least 10% of the cells in the population of interest; and (5) disproportionately decreased side light scatter on the granulocytes (a correlate of hypogranularity), such that the difference between the median side scatter among the granulocytes and among the lymphocytes was one-half log or less. In practice, the identification of any 1 of these 5 types of abnormalities in any individual antigen eliminated a case from the normal category. However, the decision to classify a case as flow-indeterminate or flow-abnormal depended on a global view of the overall number and severity of the antigenic abnormalities by the observers. Certain antigenic alterations, such as convincing expression of the nonmyeloid antigens CD5, CD7, or CD56 on 10% or more of the myeloid blasts, were considered sufficient to put a case into the abnormal category. In contrast, relatively low-level expression of CD56 on 10% to 25% of the maturing granulocytes and/or monocytes, in isolation, was not sufficient for classification as abnormal because such expression of CD56 may be seen in the setting of bone marrow regeneration with or without granulocyte colony-stimulating factor therapy. 30 Abnormal patterns of myeloid antigen expression were evaluated on a case-by-case basis. Complete absence of expression of a myeloid antigen, such as CD13 or CD33, was considered abnormal, whereas less severe decreases (or increases) in the level of antigen expression generally were not considered sufficient for classifying a specimen as abnormal. Although we did not have rigid classification criteria based on the number of abnormalities of myeloid or nonmyeloid antigen expression, in practice, virtually all of the flow-indeterminate cases expressed 3 or fewer abnormal myeloid antigens of mild degree and no more than 1 aberrant nonmyeloid antigen, whereas many of the flow-abnormal cases had 5 or more abnormal myeloid antigens and/or 2 aberrant nonmyeloid antigens. Cytogenetics Concurrent conventional cytogenetic results from G- banded preparations were available from the University of Washington Cytogenetics Laboratory. In most cases, the karyotypic analysis was based on 20 metaphases. Morphologic Evaluation We found 124 cases with Wright-Giemsa stained bone marrow aspirate smears adequate for morphologic evaluation to assess the extent of dysplasia. Because not all of these cases had accompanying peripheral blood smears and/or bone marrow biopsy sections, for the sake of comparability, the morphologic review was limited to the bone marrow aspirate smears. The aspirate smears were evaluated blindly by 2 UWHL hematopathologists with extensive experience in the morphologic diagnosis of myelodysplasia (S.J.K. and B.L.W.). A third investigator (J.R.F.) organized and tabulated the morphologic data generated by this blinded review. To standardize the morphologic assessment, the presence of 10% dysplastic forms was required among the erythroid, myeloid, and/or megakaryocytic lineages for the morphologic changes to be considered consistent with MDS ( positive ). Cases with lesser amounts of dysplasia on the aspirate smears were considered to have indeterminate morphologic features, whereas cases with no significant dysplastic features were considered negative for MDS. Erythroid hyperplasia alone, in the absence of significant dyserythropoiesis, was not considered a dysplastic feature. Because the morphologic review was limited to the aspirate smears, no effort was made to subclassify the positive cases among the various types of MDS or myelodysplastic/myeloproliferative overlap syndromes recognized by the World Health Organization classification system. When there were discrepancies between the 2 morphologic evaluations, the 2 observers together reviewed each discrepant case blinded to all other information about the case, and a consensus morphologic interpretation was achieved. Statistics Statistical comparisons were performed as follows: (1) Continuous numeric data, such as mean patient ages among the different FC-defined groups, were compared using the 2-sided Student t test. (2) Proportions, such as the percentage of cytogenetically abnormal cases among the different FC-defined groups, were compared using the χ 2 test based on the 1 degree of freedom (for pairwise comparisons) or the Fisher exact test (for more complex comparisons). For all statistical tests, statistical significance was defined by a P value of.05 or less. Results Normal Patterns of Antigen Expression With Maturation A representative example of an immunophenotypically normal bone marrow aspirate from a patient with iron deficiency anemia, separately showing the maturational progression from blasts to neutrophils (upper row) and blasts to monocytes (lower row), is shown in Image 1. In the granulocytic and monocytic series, the progression demonstrated in Image 1 occurs in parallel with an increase in CD45 expression. Note that very similar patterns of antigen expression among the myeloid blasts, maturing granulocytes, and monocytes were identified in a series of 11 additional, presumed normal marrow samples from negative FC staging studies for lymphoma (data not shown). Our 4-color FC findings in normal myelopoiesis have been described in detail elsewhere 30,31 and Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6 173

5 Kussick et al / FLOW CYTOMETRY TO ASSESS MYELODYSPLASIA are similar to those described by Loken and Wells, 2 Terstappen and Loken, 3 and Terstappen et al 4 using 3-color FC. Characteristics of Flow-Normal, Flow-Indeterminate, and Flow-Abnormal Groups Compared with the normal reference marrow shown in Image 1, 58 of 124 specimens represented abnormal cases demonstrating unequivocally aberrant patterns of antigen expression with maturation, 19 represented indeterminate cases demonstrating relatively mild abnormalities of myeloid antigen expression, and 47 represented normal cases demonstrating no significant antigenic abnormalities. Among the abnormal cases, 32 (55%) had FC-defined myeloid blast percentages of less than 5%. Image 2, Image 3, and Image 4 demonstrate representative abnormal cases that all demonstrated clonal cytogenetic abnormalities and, based on their morphologic features, seemed to represent RA, refractory cytopenia with multilineage dysplasia, and RAEB, respectively, under the World Health Organization classification system. 33 The features of these FC-defined groups are summarized in Table 1. Overall, patients in the abnormal group were significantly older than those in the combined normal and indeterminate groups, although the age difference did not reach statistical significance when the abnormal cases with fewer than 5% FC-defined myeloid blasts were compared with the combined normal and indeterminate groups. When hematologic parameters obtained at the time of bone marrow evaluation were compared, the overall abnormal group demonstrated significantly lower hemoglobin levels and platelet counts and a higher mean corpuscular volume (MCV) than the combined normal and indeterminate groups. The hemoglobin levels were also significantly lower and the MCV significantly higher when the abnormal cases with fewer than 5% myeloid blasts were compared with the combined normal and indeterminate groups, while the platelet count difference did not reach statistical significance (P =.07). The normal mean neutrophil count in the flow-abnormal group is likely to be due, in part, to the presence of neutrophilia among a number of putative CMML cases in this group (data not shown). As would be expected, the overall abnormal group demonstrated a significantly higher FC-defined mean myeloid blast percentage per case and significantly greater mean numbers of abnormal myeloid and nonmyeloid antigens per case than the normal and indeterminate groups combined. This significant excess in the blast percentage and numbers of abnormal antigens persisted when only the abnormal cases with fewer than 5% myeloid blasts were compared with the combined normal and indeterminate groups. Viable cells CD45 ECD SS Log CD7 FITC CD38 FITC Monocytes CD11b PE CD16 PE-Cy5 CD15 FITC Image 2 Four-color flow cytometric analysis of myeloid and monocytic antigenic expression in a case of refractory anemia. The blast, maturing myeloid-neutrophil, and monocytic populations are colored as in Image 1. The normal patterns of antigen expression are gray in the background of a number of the dot plots. Immunophenotypic abnormalities in this case include decreased CD33 on all myeloid cell populations; decreased HLA-DR on the myeloid blasts; and abnormally homogeneous expression of CD34 and CD38 on the blasts. Cy5, cyanine 5; FITC, fluorescein isothiocyanate; ECD, PE Texas red; PE, phycoerythrin; SS, side scatter. 174 Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6

6 Hematopathology / ORIGINAL ARTICLE Viable cells CD45 ECD SS Log CD7 FITC CD38 FITC Monocytes CD56 PE CD16 PE-Cy5 CD56 PE CD56 PE CD5 FITC CD5 FITC CD5 FITC Image 3 Four-color flow cytometric analysis of myeloid and monocytic antigenic expression in a case of refractory cytopenia with multilineage dysplasia. The blast, maturing myeloid-neutrophil, and monocytic populations are colored as in Image 1. The normal patterns of antigen expression are gray in the background of a number of the dot plots. Immunophenotypic abnormalities in this case include decreased side scatter (SS log) on the myeloid blasts and maturing granulocytes; abnormally homogeneous HLA-DR and CD33 and decreased CD38 on the myeloid blasts; aberrant CD56 on the myeloid blasts, granulocytes, and monocytes; and increased CD13 with abnormal acquisition of CD16 on the granulocytes. Cy5, cyanine 5; FITC, fluorescein isothiocyanate; ECD, PE Texas red; PE, phycoerythrin. Viable cells CD45 ECD SS Log CD15 FITC CD16 PE-Cy5 CD38 FITC CD7 PE Image 4 Four-color flow cytometric analysis of myeloid and monocytic antigenic expression in a case of refractory anemia with excess blasts. The blast, maturing myeloid-neutrophil, and monocytic populations are colored as in Image 1. The normal patterns of antigen expression are gray in the background of a number of the dot plots. Immunophenotypic abnormalities in this case include decreased side scatter (SS log) among the granulocytes; abnormally homogeneous expression of CD33 and HLA-DR on the myeloid blasts; increased CD13 on the myeloid blasts and granulocytes; increased and abnormally homogeneous expression of CD34 and CD38 on the blasts; and aberrant expression of low-level CD15 and strong CD7 on the blasts. Cy5, cyanine 5; FITC, fluorescein isothiocyanate; ECD, PE Texas red; PE, phycoerythrin. Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6 175

7 Kussick et al / FLOW CYTOMETRY TO ASSESS MYELODYSPLASIA Table 1 Flow Cytometrically Defined Patient Groups * Flow Cytometric Results Indeterminate Abnormal Abnormal (<5% Normal (n = 47) (n = 19) (All Cases) (n = 58) Blasts) (n = 32) Mean age (y) No. (%) males 23 (49) 11 (58) 31 (53) 19 (59) Hemoglobin, g/dl (g/l) 11.8 (118) 11.2 (112) 9.8 (98) 9.8 (98) MCV, µm 3 (fl) 91.7 (91.7) 92.1 (92.1) 96.6 (96.6) 97.4 (97.4) Neutrophil count, /µl ( 10 9 /L) 3,100 (3.1) 4,000 (4.0) 4,300 (4.3) 4,200 (4.2) Platelet count, 10 3 /µl ( 10 9 /L) 190 (190) 163 (163) 103 (103) 130 (130) Mean myeloid blast % Mean abnormal myeloid antigens Mean nonmyeloid antigens No. (%) of cases with Abnormal morphologic features 3 (6) 3 (16) 50 (86) 26 (81) Abnormal cytogenetics 1 (2) 1 (5) 31 (53) 16 (50) Mean No. of cytogenetic abnormalities # 0.02 (1) 0.05 (1) 1.6 (2.9) 1.3 (2.6) No. (%) of cases meeting gold standard 4 (9) 4 (21) 51 (88) 27 (84) criteria ** MCV, mean corpuscular volume; MDS, myelodysplastic syndrome. * Statistical significance tests compared all flow cytometrically abnormal cases with combined normal and indeterminate cases or flow cytometrically abnormal cases with <5% myeloid blasts (by flow cytometry) with combined normal and indeterminate cases. Flow cytometrically abnormal groups also were compared separately with the flow cytometrically indeterminate group with respect to myeloid blast percentage, mean abnormal myeloid antigens, and mean nonmyeloid antigens. P <.05; Student t test. P <.001; Student t test. P <.01; Student t test. Flow cytometrically defined myeloid blast percentage. P <.001; χ 2 test, 1 df. # Numbers in parentheses are the mean number of chromosomal abnormalities among the cytogenetically abnormal cases. ** Cases satisfying gold standard (traditional) criteria for the diagnosis of MDS had a clonal cytogenetic abnormality and/or demonstrated morphologic features of MDS. Most important, both the overall abnormal group and the abnormal cases with fewer than 5% myeloid blasts were significantly more likely to have abnormal bone marrow morphologic features and abnormal bone marrow cytogenetics than the combined normal and indeterminate groups. Accordingly, 88% (51/58) of the total flow-abnormal cases, and 84% (27/32) of the flow-abnormal cases with fewer than 5% blasts Table 2 Cytogenetically Defined Patient Groups Cytogenetic Results Normal Abnormal (n = 91) (n = 33) Mean age (y) * No. (%) males 48 (53) 17 (52) Hemoglobin, g/dl (g/l) 11.2 (112) 9.7 (97) MCV, µm 3 (fl) 93.5 (93.5) 95.7 (95.7) Neutrophil count, /µl ( 10 9 /L) 3,800 (3.8) 3,700 (3.7) Platelet count, 10 3 /µl ( 10 9 /L) 170 (170) 76 (76) No. (%) with abnormal flow 27 (30) 31 (94) cytometric results Mean myeloid blast % Mean No. of abnormal myeloid antigens Mean No. of nonmyeloid antigens * No. (%) with abnormal morphologic 26 (29) 30 (91) findings MCV, mean corpuscular volume. * P =.05; Student t test (abnormal vs normal). P <.001; Student t test (abnormal vs normal). P <.001; χ 2 test, 1 df (abnormal vs normal). Flow cytometrically defined myeloid blast percentage. satisfied traditional, or gold standard, criteria for the diagnosis of MDS (ie, had a clonal cytogenetic abnormality and/or demonstrated morphologic features of MDS) vs only 21% of the indeterminate cases and 9% of the normal cases. Note that the statistically significant relationships between FC status and age, hemoglobin level, platelet count, mean myeloid blast percentage, mean numbers of abnormal myeloid and nonmyeloid antigens, likelihood of a clonal cytogenetic abnormality, and mean number of cytogenetic abnormalities all persisted when all 180 cases with complete FC and cytogenetic data were analyzed independent of the morphologic data (data not shown). In the 180-case data set, more than 95% of the cytogenetically abnormal bone marrow samples also were flow-abnormal. Characteristics of the Cytogenetically Normal and Cytogenetically Abnormal Groups Table 2 gives the results according to the cytogenetic features of the cases. As a group, the patients with abnormal cytogenetics were significantly older and had significantly lower hemoglobin levels and lower platelet counts than the group with normal cytogenetics. The abnormal group had significantly higher FC-defined myeloid blast percentages and greater numbers of abnormal myeloid and nonmyeloid antigens. Finally, the abnormal cases were much more likely to demonstrate abnormal morphologic features in bone marrow aspirates than the normal cases. 176 Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6

8 Hematopathology / ORIGINAL ARTICLE Characteristics of Morphologically Normal, Indeterminate, and Abnormal Groups Table 3 gives the results according to the morphologic features of the cases. When grouped together, patients in the abnormal group demonstrated a tendency to be older (P =.06) than the combined normal and indeterminate groups and had significantly higher MCV and lower hemoglobin levels and platelet counts. Strikingly, 89% of cases (50/56) in the abnormal group were definitively abnormal by FC vs only 32% of the indeterminate group (8/25) and 0% of the normal group. Accordingly, the abnormal group demonstrated a significantly greater FC-defined myeloid blast percentage, greater mean numbers of abnormal myeloid and nonmyeloid antigens, and a greater percentage of cytogenetically abnormal cases than the combined normal and indeterminate groups. Specific Immunophenotypic Abnormalities Table 4 gives the frequencies of specific antigenic abnormalities among the gated myeloid blasts, granulocytes, and monocytes for the FC-defined (flow-indeterminate and flow-abnormal groups only), morphologically defined, cytogenetics-defined, and MDS gold standard defined groups. The listed antigens include only those found to be abnormal in at least 10% of the cases in at least 1 of the defined patient groups. Antigens found to be expressed abnormally by 50% or more of the cases in at least 1 of the defined patient groups are indicated and include HLA-DR, CD13, CD33, CD38, and CD117. Features of the Flow-Indeterminate Group in Relation to the Likelihood of MDS The 19 flow-indeterminate cases were evaluated in detail in an effort to identify features that correlated with the likelihood of satisfying the gold standard criteria for MDS. Although the small number of cases in this group prevented formal demonstration of statistical significance for any of the comparisons in Table 5, there was a tendency for gold standard positive patients to be older and to have a higher mean MCV and lower mean hemoglobin and platelet counts than the gold standard negative patients. Thus, clinical features suggestive of MDS tended to correlate with the presence of cytogenetic and/or morphologic abnormalities in the flowindeterminate group. Among the FC-defined parameters, there was some excess of the mean myeloid blast percentage and of the mean number of abnormal myeloid and nonmyeloid antigens in the gold standard positive cases compared with the gold standard negative cases. Summary of Relationships Between FC, Morphologic Features, and Cytogenetics Table 6 summarizes the relationships among these 3 key parameters in the diagnosis of MDS. There was a highly significant (P <.001) association between abnormal FC and abnormal morphologic findings, independent of cytogenetics; similarly, there was a highly significant (P <.001) association between abnormal FC and abnormal cytogenetic findings, independent of morphologic findings. Table 3 Morphologically Defined Patient Groups * Morphologic Results Normal (n = 43) Indeterminate (n = 25) Abnormal (n = 56) Mean age (y) No. (%) males 24 (56) 11 (44) 30 (54) Hemoglobin, g/dl (g/l) 12.3 (123) 10.3 (103) 9.7 (97) MCV, µm 3 (fl) 90.6 (90.6) 93.0 (93.0) 97.1 (97.1) Neutrophil count, /µl ( 10 9 /L) 2,800 (2.8) 4,200 (4.2) 4,400 (4.4) Platelet count, 10 3 /µl ( 10 9 /L) 194 (194) 139 (139) 109 (109) Mean myeloid blast % Mean No. of abnormal myeloid antigens Mean No. of nonmyeloid antigens No. (%) with Abnormal flow cytometric results 0 (0) 8 (32) 50 (89) Abnormal cytogenetics 0 (0) 3 (12) 30 (54) Mean No. of cytogenetic abnormalities # (1.3) 1.6 (3.0) MCV, mean corpuscular volume. * Statistical significance tests were performed only for comparison of all morphologically abnormal cases vs combined normal and indeterminate cases. P =.06; Student t test. P <.001; Student t test. P <.01; Student t test. Flow cytometrically defined myeloid blast percentage. P <.001; χ 2 test, 1 df. # Numbers in parentheses are the mean number of chromosomal abnormalities among the cytogenetically abnormal cases. Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6 177

9 Kussick et al / FLOW CYTOMETRY TO ASSESS MYELODYSPLASIA Based on the summary data, an abnormal 4-color FC result was 89% sensitive and 88% specific for identifying a bone marrow aspirate satisfying the gold standard criteria for a diagnosis of MDS. The overall sensitivity of 4-color FC for identifying any abnormality (ie, yielding an indeterminate or definitively abnormal result) in a gold standard case of MDS was 95%, whereas the overall specificity of identifying any such abnormality was 67%. Table 4 Specific Antigenic Abnormalities on Blasts, Maturing Granulocytes, and Maturing Monocytes in Flow Cytometrically Defined, Morphologically Defined, and Cytogenetically Defined Patient Groups * Flow Cytometry Morphologic Studies Cytogenetic Studies Gold Standard Indeterminate Abnormal Normal Indeterminate Abnormal Normal Abnormal Negative Positive (n = 19) (n = 58) (n = 43) (n = 25) (n = 56) (n = 91) (n = 33) (n = 66) (n = 58) Blasts HLA-DR 4 (21) 37 (64) 1 (2) 5 (20) 35 (63) 21 (23) 20 (61) 6 (9) 35 (60) CD11b 0 (0) 19 (33) 0 (0) 2 (8) 17 (30) 8 (8) 12 (36) 2 (3) 17 (29) CD13 4 (21) 35 (60) 3 (7) 4 (16) 32 (57) 21 (23) 18 (55) 7 (11) 32 (55) CD15 0 (0) 6 (10) 0 (0) 0 (0) 6 (11) 2 (2) 4 (12) 0 (0) 6 (10) CD33 9 (47) 44 (76) 6 (14) 8 (32) 39 (70) 29 (32) 24 (73) 13 (20) 40 (69) CD34 2 (11) 15 (26) 1 (2) 2 (8) 14 (25) 11 (12) 6 (18) 3 (5) 14 (24) CD38 1 (7) 8 (36) 0 (0) 0 (0) 9 (45) 4 (7) 5 (50) 0 (0) 24 (41) CD45 0 (0) 10 (17) 0 (0) 1 (4) 9 (16) 7 (8) 3 (9) 1 (2) 9 (16) CD117 2 (11) 37 (64) 0 (0) 7 (28) 32 (57) 21 (23) 18 (55) 7 (11) 32 (55) CD5 0 (0) 4 (7) 0 (0) 0 (0) 4 (7) 3 (3) 1 (3) 0 (0) 4 (7) CD7 1 (5) 12 (21) 1 (2) 2 (8) 10 (18) 7 (8) 6 (18) 3 (5) 10 (17) CD56 1 (5) 10 (17) 0 (0) 1 (4) 10 (18) 8 (9) 3 (9) 1 (2) 10 (17) Granulocytes Scatter 3 (16) 26 (45) 1 (2) 5 (20) 23 (41) 13 (14) 16 (48) 5 (8) 24 (41) CD11b 4 (21) 14 (24) 3 (7) 3 (12) 12 (21) 7 (8) 11 (33) 5 (8) 13 (22) CD13 4 (21) 27 (47) 2 (5) 7 (28) 22 (39) 16 (18) 15 (45) 8 (12) 23 (40) CD14 0 (0) 8 (14) 0 (0) 0 (0) 8 (14) 1 (1) 7 (21) 0 (0) 8 (14) CD15 0 (0) 1 (2) 0 (0) 0 (0) 1 (2) 0 (0) 1 (3) 0 (0) 1 (2) CD16 3 (16) 21 (36) 3 (7) 4 (16) 17 (30) 12 (13) 12 (36) 7 (11) 17 (29) CD33 9 (47) 35 (60) 6 (14) 7 (28) 31 (55) 24 (26) 20 (61) 12 (18) 32 (55) CD45 2 (11) 15 (26) 1 (2) 3 (12) 13 (23) 9 (10) 8 (24) 4 (6) 13 (22) CD56 0 (0) 9 (16) 0 (0) 0 (0) 9 (16) 6 (7) 9 (9) 0 (0) 9 (16) Monocytes HLA-DR 1 (5) 4 (7) 1 (2) 2 (8) 2 (4) 3 (3) 2 (6) 3 (5) 2 (3) CD11b 1 (5) 5 (9) 0 (0) 3 (12) 3 (5) 2 (2) 4 (12) 2 (3) 4 (7) CD13 1 (5) 12 (21) 1 (2) 1 (4) 11 (20) 5 (5) 8 (24) 2 (3) 11 (19) CD14 0 (0) 3 (5) 0 (0) 0 (0) 3 (5) 0 (0) 3 (9) 0 (0) 3 (5) CD15 0 (0) 6 (10) 0 (0) 1 (4) 5 (9) 3 (3) 3 (9) 1 (2) 5 (9) CD33 8 (42) 10 (17) 5 (12) 4 (16) 9 (16) 13 (14) 5 (15) 8 (12) 10 (17) CD56 1 (5) 16 (28) 0 (0) 2 (8) 15 (27) 10 (11) 7 (21) 2 (3) 15 (26) * Data are given as number (percentage) of cases in each group with the specific abnormality. CD38 was evaluated only in cases later in the study period (September 2000 and later), so total numbers for each group are lower. Antigens with abnormalities in 50% or more cases in any category. Table 5 Features of the Flow Cytometrically Indeterminate Group by Gold Standard Status Gold Standard Status Negative Positive (n = 15) (n = 4) Mean age (y) No. (%) males 10 (67) 1 (25) Hemoglobin, g/dl (g/l) 11.7 (117) 9.6 (96) MCV, µm 3 (fl) 90.9 (90.9) 95.9 (95.9) Neutrophil count, /µl ( 10 9 /L) 4,500 (4.5) 2,400 (2.4) Platelet count, 10 3 /µl ( 10 9 /L) 187 (187) 85.3 (85.3) Mean myeloid blast % Mean No. of abnormal myeloid antigens Mean No. of nonmyeloid antigens Table 6 Distribution of Cases by Cytogenetic, Morphologic, and Flow Cytometric Findings Flow Cytometry Morphologic Cytogenetics Studies Normal Indeterminate Abnormal Normal * Normal Indeterminate Abnormal Abnormal * Normal Indeterminate Abnormal * The Fisher exact test showed a highly significant association (P <.001) between the morphologic and flow cytometric results within the cytogenetically normal and abnormal groups. MCV, mean corpuscular volume. 178 Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6

10 Hematopathology / ORIGINAL ARTICLE Discussion We demonstrated the usefulness of 4-color FC in the routine evaluation for MDSs in an unselected series of 124 bone marrow aspirates from patients with unexplained cytopenias or monocytosis. The results of FC were compared directly in a blinded manner with those of the current gold standard for diagnosing MDSs: morphologic and cytogenetic evaluation. For a number of reasons, we believe that 4-color FC represents a valuable addition to the evaluation of MDSs. First, in experienced laboratories, FC represents an objective method to aid in the diagnosis of MDSs. Because normal hematopoiesis is characterized by highly reproducible patterns of antigen expression during myeloid maturation in the marrow, 2-4,30,31 it is relatively easy to compare the patterns in patient specimens with the normal patterns to determine the likelihood of an underlying stem cell neoplasm. At our institution, the normal patterns of maturation have been reproducible enough that evaluation of 10 to 15 normal bone marrow samples is sufficient to establish a reference range to begin using in the evaluation of diagnostic specimens. By adding FC to the workup of MDSs, the historic reliance on potentially subjective morphologic criteria might be lessened, which could be of particular benefit when the blast percentage is low, the morphologic features are equivocal, or both. It is important to note that our 4-color FC results correlated strongly with the results of the traditional objective method for evaluating MDSs cytogenetics and with the results of blinded bone marrow evaluations by experienced hematopathologists. Second, 4-color FC shows high sensitivity and specificity in the evaluation for MDSs. In our assay, an abnormal 4-color FC result was 89% sensitive and 88% specific for identifying a bone marrow aspirate satisfying gold standard morphologic or cytogenetic criteria for the diagnosis of MDS. The overall sensitivity of 4-color FC for identifying any abnormality (indeterminate or definitively abnormal) in a gold standard case of MDS was 95%, whereas the overall specificity was 67%. Third, patterns of myeloid maturation can be followed up easily in multiple bone marrow specimens from a patient over time. Such sequential evaluation can be helpful in supporting or excluding the possibility of an underlying myeloid stem cell neoplasm in equivocal clinical settings. In addition, in patients with well-documented MDS, 4-color FC offers an objective method for monitoring the course of the patient s disease. We have seen cases in which the immunophenotypic abnormalities have remained remarkably stable and others in which the immunophenotypic abnormalities have progressed over time. Fourth, unlike conventional cytogenetics, FC does not require the cells of interest to grow in culture. This feature is particularly useful in evaluating low-grade MDSs, in which the abnormal blasts might be unlikely to grow in culture owing to an increased rate of apoptosis Fifth, in contrast with turnaround times of 1 week or more for conventional cytogenetics, FC evaluation of a specimen may be completed within hours of receipt. Therefore, clinical decision making is not delayed significantly by FC. Sixth, FC can readily identify nonmyeloid malignant neoplasms that can clinically mimic MDSs, such as hairy cell leukemia and large granular lymphocytic leukemia. To enable identification of such lymphoid neoplasms, our current standard FC panel examines a variety of lymphoid antigens, as indicated in the Materials and Methods section. In most cases, we also evaluate surface light chain expression on the B lymphocytes to rule out a clonal population. Although our study focused on determining antigenic abnormalities helpful in discriminating MDSs from cytopenias associated with other (nonneoplastic) causes, it is important to remember that certain antigenic abnormalities are seen commonly in nonneoplastic settings. For example, low-level aberrant CD56 expression commonly is seen on maturing granulocytes and/or monocytes in the setting of marrow regeneration and/or granulocyte colony-stimulating factor therapy, 29 whereas increased HLA-DR can be seen on neutrophils in the setting of activation, such as following growth factor therapy. 34 Given the large number of myeloid and nonmyeloid antigens we have evaluated, our work goes beyond a number of the previously published FC studies of myelopoiesis that have evaluated relatively few antigens. 5-14,28 Our study also differs from the more extensive previous studies that, like ours, have evaluated a broad range of myeloid and nonmyeloid antigens, 24-26,29 in that those studies all used 3-color FC. The 4-color approach permits the evaluation of more complex relationships among antigens during the course of myeloid maturation. For example, by combining antibodies to HLA-DR, CD33, CD45, and CD14 in a single tube, one can evaluate the full range of maturation from myeloid blasts (low-intermediate positivity for HLA-DR and CD33, without CD14) to both mature neutrophils (low-level positivity for CD33 and CD14, negative for HLA-DR) and mature monocytes (strongly positive for HLA-DR, CD33, and CD14). Therefore, one can use the data from this single tube to identify immunophenotypic abnormalities on myeloid blasts, maturing granulocytes, and maturing monocytes. Similarly, the tubes in which we combine antibodies to CD15, CD11b, CD45, and CD34 and to HLA-DR, CD13, CD45, and CD16 also are useful for evaluating the full range of myeloid maturation. Another difference between our study and the relatively large previous studies, with the exception of that by Stetler- Stevenson and colleagues, 24 was our primary interest in assessing the usefulness of FC in making the initial diagnosis of MDSs. In contrast, Ogata et al 25 and Maynadie et al 26 were most interested in identifying immunophenotypic differences among the blasts in the different categories of MDSs and identifying Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6 179

11 Kussick et al / FLOW CYTOMETRY TO ASSESS MYELODYSPLASIA immunophenotypic correlates of prognosis in MDSs. Although Wells and coworkers 29 performed a detailed immunophenotypic evaluation of a large number of known MDS cases and non- MDS control cases, their principal interest was to establish FC data as part of a scoring system to predict prognosis. Although our interest in assessing the role of FC in the initial diagnosis of MDSs seems most similar to that of Stetler-Stevenson and colleagues, 24 our study designs were quite different. Stetler-Stevenson et al 24 applied the data from a test set of 45 well-characterized MDS cases to aid in the diagnosis of 20 challenging cases, whereas we examined a series of 124 essentially unselected cases (selected only to have adequate FC, cytogenetic, morphologic, and clinical [CBC] data available). As in our study, Stetler-Stevenson and colleagues 24 found that abnormal patterns of expression of CD11b vs CD16 or CD13 vs CD16 among the maturing granulocytes were particularly useful in distinguishing benign from dysplastic processes, whereas aberrant expression of nonmyeloid antigens was a less common finding. Their study found that loss of CD64 among the granulocytes was present in about two thirds of the MDS cases, whereas CD64 was not one of the antigens routinely evaluated in our panels. In addition, they attempted to evaluate maturing erythrocytes and megakaryocytes by FC, which we have not attempted to do because of methodological concerns. Our 4-color FC assay demonstrates a high degree of sensitivity and specificity for identifying patients with a high likelihood of having an MDS. Therefore, in laboratories experienced at evaluating myeloid maturation, 4-color FC represents a very useful adjunct in the diagnostic evaluation of unexplained cytopenia and monocytosis and seems to be of particular usefulness when the myeloid blast count is low or morphologic features are equivocal. From the Departments of 1 Laboratory Medicine, 2 Medical Education and Biomedical Informatics, and 3 Pathology, University of Washington, Seattle. Address reprint requests to Dr Kussick: Box , University of Washington Medical Center NE Pacific St, Seattle, WA Acknowledgments: We thank the technical staff of the University of Washington Hematopathology Laboratory for invaluable efforts in the generation of the flow cytometric data and the University of Washington Cytogenetics Laboratory for providing nearly all of the cytogenetic data used in the study. References 1. Foucar K. Bone Marrow Pathology. Chicago, IL; ASCP Press; 2001: Loken MR, Wells DA. Normal antigen expression in hematopoiesis: basis for interpreting leukemia phenotypes. In: Stewart CC, Nicholson JKA, eds. Immunophenotyping. New York, NY: Wiley-Liss; 2000: Terstappen LW, Loken MR. cell differentiation in normal bone marrow and acute myeloid leukemia assessed by multi-dimensional flow cytometry. Anal Cell Pathol. 1990;2: Terstappen LW, Safford M, Loken MR. Flow cytometric analysis of human bone marrow, III: neutrophil maturation. Leukemia. 1990;4: Elghetany MT. Surface marker abnormalities in myelodysplastic syndromes. Haematologica. 1998;83: Bianco T, Farmer BJ, Sage RE, et al. Loss of red cell A, B, and H antigens is frequent in myeloid malignancies. Blood. 2001;97: Izumi M, Takeshita A, Shinjo K, et al. Decreased amount of mpl and reduced expression of glycoprotein IIb/IIIa and glycoprotein Ib on platelets from patients with refractory anemia: analysis by a non-isotopic quantitative ligand binding assay and immunofluorescence. Eur J Haematol. 2001;66: Bowen KL, Davis BH. Abnormal patterns of expression of CD16 (FcR-III) and CD11b (CRIII) in the bone marrow of patients with myelodysplastic syndrome. Lab Hematol. 1997;3: Hansen I, Meyer K, Hokland P. Flow cytometric identification of myeloid disorders by asynchronous expression of the CD14 and CD66 antigens. Eur J Haematol. 1998;61: Chang CC, Cleveland RP. Decreased CD10-positive mature granulocytes in bone marrow from patients with myelodysplastic syndrome. Arch Pathol Lab Med. 2000;124: Ohsaka A, Saionji K, Igari J, et al. Altered surface expression of effector cell molecules on neutrophils in myelodysplastic syndromes. Br J Haematol. 1997;98: Otawa M, Kawanishi Y, Iwase O, et al. Comparative multicolor flow cytometric analysis of cell surface antigens in bone marrow hematopoietic progenitors between refractory anemia and aplastic anemia. Leuk Res. 2000;24: Mann KP, DeCastro CM, Liu J, et al. Neural cell adhesion molecule (CD56)-positive acute myelogenous leukemia and myelodysplastic and myeloproliferative syndromes. Am J Clin Pathol. 1997;107: Schlesinger M, Silverman LR, Jiang JD, et al. Analysis of myeloid and lymphoid markers on the surface and in the cytoplasm of mononuclear bone marrow cells in patients with myelodysplastic syndrome. J Clin Lab Immunol. 1996;48: Xu D, Schultz C, Akker Y, et al. Evidence for expression of early myeloid antigens in mature, non-blast myeloid cells in myelodysplasia. Am J Hematol. 2003;74: Cherian S, Moore J, Bantly A, et al. Peripheral blood MDS score: a new flow cytometric tool for the diagnosis of myelodysplastic syndromes. Cytometry B Clin Cytom. 2005;64: Parker JE, Mufti GJ, Rasool F, et al. The role of apoptosis, proliferation, and the bcl-2-related proteins in the myelodysplastic syndromes and acute myeloid leukemia secondary to MDS. Blood. 2000;96: Matthes TW, Meyer G, Samii K, et al. Increased apoptosis in acquired sideroblastic anaemia. Br J Haematol. 2000;111: Merchant SH, Gonchoroff NJ, Hutchison RE. Apoptotic index by annexin V flow cytometry: adjunct to morphologic and cytogenetic diagnosis of myelodysplastic syndromes. Cytometry. 2001;46: Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6

12 Hematopathology / ORIGINAL ARTICLE 20. Boudard D, Sordet O, Vasselon C, et al. Expression and activity of caspases 1 and 3 in myelodysplastic syndromes. Leukemia. 2000;14: Hellstrom-Lindberg E, Schmidt-Mende J, Forsblom AM, et al. Apoptosis in refractory anaemia with ringed sideroblasts is initiated at the stem cell level and associated with increased activation of caspases. Br J Haematol. 2001;112: Delia D, Aiello A, Soligo D, et al. bcl-2 proto-oncogene expression in normal and neoplastic human myeloid cells. Blood. 1992;79: Boudard D, Vasselon C, Bertheas MF, et al. Expression and prognostic significance of Bcl-2 family proteins in myelodysplastic syndromes. Am J Hematol. 2002;70: Stetler-Stevenson M, Arthur DC, Jabbour N, et al. Diagnostic utility of flow cytometric immunophenotyping in myelodysplastic syndrome. Blood. 2001;98: Ogata K, Nakamura K, Yokose N, et al. Clinical significance of phenotypic features of blasts in patients with myelodysplastic syndrome. Blood. 2002;100: Maynadie M, Picard F, Husson B, et al. Immunophenotypic clustering of myelodysplastic syndromes. Blood. 2002;100: Bennett JM, Catovsky D, Daniel MT, et al. Proposals for the classification of the myelodysplastic syndromes. Br J Haematol. 1982;51: Del Canizo MC, Fernandez ME, Lopez A, et al. Immunophenotypic analysis of myelodysplastic syndromes. Haematologica. 2003;88: Wells DA, Benesch M, Loken MR, et al. and monocytic dyspoiesis as determined by flow cytometric scoring in myelodysplastic syndrome correlates with the IPSS and with outcome after hematopoietic stem cell transplantation. Blood. 2003;102: Kussick SJ, Wood BL. Using four-color flow cytometry to identify abnormal myeloid populations. Arch Pathol Lab Med. 2003;127: Kussick SJ, Wood BL. Four-color flow cytometry identifies virtually all cytogenetically abnormal bone marrow samples in the workup of non-cml myeloproliferative disorders. Am J Clin Pathol. 2003;120: Stelzer GT, Shults KE, Loken MR. CD45 gating for routine flow cytometric analysis of human bone marrow specimens. Ann N Y Acad Sci. 1993;677: Jaffe ES, Harris NL, Stein H, et al, eds. Pathology and Genetics of Tumours of the Hematopoietic and Lymphoid Tissues. Lyon, France: IARC Press; 2001:20. World Health Organization Classification of Tumours. 34. Smith WB, Guida L, Sun Q, et al. Neutrophils activated by granulocyte-macrophage colony-stimulating factor express receptors for interleukin-3 which mediate class II expression. Blood. 1995;86: Am J Clin Pathol 2005;124: DOI: /6PBP78G4FBA1FDG6 181