The prognostic significance of antigen expression in leukaemia

Best Practice & Research Clinical Haematology Vol. 16, No. 4, pp. 613 –628, 2003 doi:10.1053/ybeha.2003.286, www.elsevier.com/locate/jnlabr/ybeha

5 The prognostic significance of antigen expression in leukaemia Richard Schabath MD Richard Ratei MD Wolf-Dieter Ludwig*

MD, PhD

Professer Robert-Ro¨ssle-Clinic, Department of Haematology, Oncology and Tumour Immunology, HELIOS Clinic Berlin, Charite´, Campus Berlin-Buch, Lindenberger Weg 80, Berlin D-13122, Germany

Numerous immunophenotypic features have been examined for their potential prognostic significance in predicting treatment outcome in leukaemias. These include immunophenotypic subgroups of acute lymphoblastic leukaemia (ALL) and immature acute myeloid leukaemia, expression of individual surface antigens or combined immunophenotypic features, and more recently, molecules mediating the multidrug resistance phenotype or being involved in the regulation of drug-induced apoptosis. Most previous studies investigating the prognostic significance of antigen expression in leukaemia have not used the information provided by multiparameter flow cytometry and have chosen rather arbitrary cut-off points for marker positivity. Moreover, given significant associations between immunophenotypic features and genetic abnormalities in leukaemic cells, immunophenotyping as an independent predictor of treatment outcome has been questioned. Thus, except for lineage assignment of leukaemic blasts and definition of maturational status in ALL, information provided by immunophenotyping is currently not applied to risk-classification systems or used for planning patient treatment in leukaemia. We review some of the recent findings regarding the prognostic impact of distinct immunophenotypic features in acute leukaemias and myelodysplastic syndrome. Key words: immunophenotyping; leukaemia; prognosis; flow cytometry; genotype –phenotype associations; drug-resistance; apoptosis-regulating proteins.

In the past two decades, the impact of immunophenotyping by flow cytometry in the diagnosis and management of acute leukaemia has expanded rapidly. This advance can be attributed mainly to significant progress in laser and computer technologies, the production of several hundred monoclonal antibodies (mAbs) to a variety of antigens expressed by haematopoietic cells, and the availability of distinct fluorochromes conjugated to mAbs, allowing the simultaneous measurement of at least three or four * Corresponding author. Tel.: þ49-3094171314; Fax: þ 49-3094171307. E-mail address: [email protected] 1521-6926/03/$ - see front matter Q 2003 Elsevier Ltd. All rights reserved.

614 R. Schabath, R. Ratei and W.-D. Ludwig

cellular antigens in combination with two intrinsic parameters (cell size and cytoplasmic complexity) as determined on the basis of the light-scattering properties of leukaemic blasts (i.e. forward- and side-scatter characteristics, FSC and SSC) (reviewed in Refs. 1– 6). There is now general consensus that multiparameter flow cytometry is a powerful diagnostic tool for the immunophenotypic characterization of acute leukaemias and chronic lymphoproliferative disorders that can be applied to define immunophenotypic subsets, to further characterize leukaemic cells in ways that correlate to prognosis, to detect minimal residual disease (MRD), and, more recently, to develop and monitor antibody-based treatment strategies.1 – 5 Numerous immunophenotypic features have been examined for their potential prognostic significance in predicting treatment response, remission duration, and survival. These include immunophenotypic subgroups of precursor B-, T-cell acute lymphoblastic leukaemia (ALL), and immature acute myeloid leukaemia (AML), expression of individual antigens or composite immunophenotypic patterns, as well as features of aberrant antigen expression which do not obey the ontogenic rules of normal haematopoietic differentiation (reviewed in Refs. 1,3 – 5). More recently, functional characteristics such as molecules reflecting cellular resistance mechanisms (e.g. multidrug resistance phenotype) or apoptosis-regulating proteins have also been analysed as to their prognostic impact on treatment outcome in acute leukaemias, myelodysplastic syndrome (MDS) and chronic lymphoproliferative disorders (reviewed in Ref. 7). Most previous studies investigating the diagnostic and prognostic impact of immunophenotyping in acute leukaemias have used 20% of cells stained with mAbs for surface markers and 10% for more specific markers usually expressed in the cytoplasm (e.g. myeloperoxidase, CD79a, cytoplasmic CD3) as the general cut-off points for marker positivity.8 Obviously, these threshold values were chosen arbitrarily and have been criticized9, because they may obscure the biological importance of weakly expressed immunophenotypic features and they are not based on physiological knowledge but rather serve as a convenient means of data collection. Moreover, many clinical studies (see below) describing immunophenotypic features of acute leukaemias and correlating prognosis with immunophenotyping in ALL and AML have been performed with single-colour analyses. It is obvious that these studies have not always been adequate to distinguish malignant from normal haematopoietic cells, and, more importantly, have not made use of the information provided by multiparameter flow cytometry.2 These problems are further complicated by many technical considerations, including sample preparation methodology, selection of fluorochrome-antibody conjugates, gating strategies, optical configurations of flow cytometers, and other instrument parameters. Despite several published guidelines addressing distinct technical issues in flow-cytometric immunophenotyping of haematopoietic neoplasms8,10,11 methods have not yet been adequately standardized. More recently, novel outcome-driven statistical methods, such as classification and regression trees (CART), have been proposed to retrospectively identify prognostically significant antigen expression levels or cut-off points, especially in AML.9,12 Moreover, the intensity of antigen expression as measured by the mean (or median) fluorescence intensity and rather complex scoring strategies based on percentages of positive cells together with intensity and/or homogeneity of antigen expression have been applied to identify prognostically relevant immunophenotypic subsets and to predict molecular genetic subtypes.13 – 16 Future studies in leukaemias will show whether these novel parameters and scoring strategies, using standardized flow-cytometric procedures, provide clinical relevant information.

Prognostic significance of antigen expression 615

In the following paragraph we focus on the prognostic impact of distinct immunophenotypic subsets and specific antigens in ALL, AML, acute leukaemias of ambiguous lineage and MDS.

PROGNOSTIC IMPACT OF IMMUNOPHENOTYPING IN ALL Immunophenotyping has become essential in the diagnosis of ALL, and has substantially contributed to a more precise and biologically oriented classification of the disease (reviewed in Refs. 1,3 – 5,17). During the last two decades, immunophenotyping in ALL has been applied mainly to distinguishing ALL from AML, lineage assignment of leukaemic blasts, phenotypic characterization of pathological cell subsets, and for examining the role of membrane antigen expression in predicting treatment response (reviewed in Refs. 4,5,17). Additionally, based on observations that leukaemic blasts frequently show aberrant or asynchronous antigen expression compared with normal haematopoietic cell differentiation, leukaemia-associated phenotypic features have been routinely used to detect MRD in ALL (reviewed in Ref. 6). More recently, immunophenotyping in conjunction with cytogenetic and molecular genetic studies has identified significant associations between immunophenotypic features and numerical and/or structural chromosomal abnormalities16 that have contributed to a refined ALL classification, especially in precursor B-cell ALL (reviewed in Refs. 4,5,17,18). In the past, the lack of standardized criteria for classification of immunophenotypic subgroups, the paucity of controlled prospective studies on the treatment outcome of precursor B- and T-cell ALL subsets, and the different treatment strategies complicated the assessment of the prognostic impact of immunophenotyping studies in ALL (reviewed in Refs. 4,17). In addition, the strong correlation between certain immunophenotypic subgroups and cytogenetic or clinical features has called into question the value of immunophenotyping as an independent predictor of treatment outcome. Moreover, several studies have shown that the prognostic impact of immunophenotypic subgroups as well as chromosomal abnormalities is diminished by the improved efficacy of chemotherapy; hence, prognostic factors must be evaluated in the context of therapy delivered.19 – 22 In precursor B-cell ALL, no substantial differences in remission rates were recorded for immunophenotypic subgroups, but several studies revealed an association between the maturational stage of B-lymphoblasts and the duration of remission. Most studies in both childhood and adult ALL have reported a worse prognosis for patients whose leukaemic blasts express an immature CD10-negative pro-B phenotype21,23 – 26 which, however, was frequently associated with adverse biological (e.g. 11q23 re-arrangements) and clinical features (e.g. high tumour burden, age , 1 year). Cytogenetic and molecular genetic studies have provided conclusive evidence that children and adults with common and pre-B ALL differ significantly with respect to the incidence of the known favourable or unfavourable chromosomal translocations. For instance, t(9;22), accounts for up to 55% of adult and , 5% of children with CD10þ precursor B-cell ALL, whereas the reported frequency for t(12;21), associated with a good prognosis in most recent studies, ranges between 12 and 36% in childhood common or pre-B ALL, and t(12;21) rarely occurs in adult patients (reviewed in Refs. 4,27). These findings may partially explain the striking differences observed in treatment outcome between children and adults with common or pre-B ALL.

616 R. Schabath, R. Ratei and W.-D. Ludwig

Early studies of childhood precursor B-cell ALL indicated a poorer treatment outcome for pre-B ALL compared with pro-B and common ALL.28 Subsequent combined immunophenotypic and cytogenetic findings showed that this difference in outcome was due to a chromosomal t(1;19)(q23;p13) that is almost exclusively associated with pre-B ALL.19 Confirmation of the prognostic importance of the pre-B ALL immunophenotype has been limited to sequential studies of the Pediatric Oncology Group (POG) because, until recently, this was the only group performing cytoplasmic m testing in the context of large prospective clinical trials. However, recent data show that more intensive therapy of pre-B ALL with the t(1;19) translocation now results in treatment outcomes approaching that of common ALL (reviewed in Ref. 5). Similarly, the German ALL-BFM trials, and the analysis of the Medical Research Council (MRC) UKALL trial XI, did not reveal any significant differences in the duration of remission between childhood common and pre-B ALL.29,30 In children with precursor B-cell ALL, the prognosis has been linked with other immunophenotypic features, such as expression of CD20, CD34 and CD45, and it has been suggested that the lack of CD20 and CD45 antigens or the presence of CD34 on leukaemic blasts may be associated with a longer event-free survival (EFS) (reviewed in Refs. 4,5). However, in view of the relationship of these immunophenotypic features (e.g. absence of CD45) to other biologically favourable characteristics14,31, their prognostic significance has to be evaluated in further studies by adjusting results for the presence of other risk factors. Several studies in childhood and adult ALL have shown that a remarkable prognostic improvement in B-ALL is achieved by the development of intensive treatment strategies, especially adapted to the biological and clinical features of this disease. 20,22 These data impressively illustrate that more effective treatment can offset the negative prognostic impact of biological characteristics, such as the immunophenotype or chromosomal translocations. In precursor T-cell ALL, various immunophenotypic features seem to be associated with an increased risk of treatment failure, including an immature pro-/pre-T-ALL phenotype, membrane expression of CD3 or MHC class II antigen, and negativity for CD2, CD5, THY antigens (similar to CD1), or CD10 (reviewed in Refs. 4,32). The prognostic impact of these factors, however, has differed according to the treatment strategies used, and immunophenotyping still represents a controversial prognostic factor that has not been routinely used for risk-classification or assignment to novel treatment strategies in high-risk precursor T-cell ALL patients. Further attempts to identify additional prognostically relevant subgroups of precursor T-cell ALL have been largely unsuccessful in both childhood and adult ALL. 33,34 However, at least three multicentre trials in childhood ALL, using similar maturational staging systems, have recently lent strong support to evidence that children with cortical (CD1aþ) precursor T-cell ALL have a better early response to treatment, as illustrated, for instance, by the in vivo response to corticosteroids, and a significantly longer duration of EFS than in those with an immature or mature precursor T-cell phenotype.29,35,36 Similar data showing a significant improvement in survival of adult patients who express CD1, CD2, CD4 and CD5 antigens compared with patients not expressing these antigens, have been recently published by the Cancer and Leukaemia Group.21 Although, at present it is unclear why patients with a cortical immunophenotype respond better to treatment, recent investigations of apoptosisrelated parameters, including spontaneous apoptosis in vitro and modulation of apoptosis by IL-7, suggested that maturational stages of precursor T-cell ALL may differ as to their accessibility to apoptotic programmes, with lymphoblasts expressing CD1a

Prognostic significance of antigen expression 617

or exhibiting a selection-related phenotype being more susceptible to apoptosis than leukaemic lymphoblasts with an immature phenotype.35,37,38 Recently, by using oligonuleotide micro-arrays, distinct gene expression signatures have been identified in precursor T-cell ALL that correlated with aberrant expression of T-cell oncogenes (e.g. LYL1, HOX11, TAL1) specific stages of normal thymocyte development and prognosis.39 HOX11 þ cases exhibited an early thymocyte phenotype with increased expression of CD1a and were associated with a favourable prognosis, whereas oncogenes aberrantly expressed in cases with an immature or more mature thymocyte phenotype (i.e. LYL1, TAL1) were associated with a much worse response to treatment. Several studies have suggested that a subclassification of membrane CD3þ precursor T-cell ALL according to T-cell antigen receptor (TCR) ab or gd expression provides valuable clinical information, because TCR-gdþ cases represent an important, albeit rare, subgroup of precursor T-cell ALL with distinctive clinicopathological features and prognosis.40 – 42 Further prospective studies are needed to characterize more thoroughly the cell-biological features of TCR-gdþ lymphoblasts and to confirm the better prognosis of this subgroup as compared with TCR-abþ precursor T-cell ALL.42 Recent basic and translational research on mechanisms of cellular chemoresistance and regulation of programmed cell death (apoptosis) identified a number of potentially relevant proteins and protein families, and numerous studies addressed their prognostic impact in ALL (reviewed in Ref. 7). Although differential expression patterns of these factors in ALL patients have been suggested, their prognostic significance could not yet be validated. Constitutive expression and functional activity of the multidrug resistance P-glycoprotein (P-gp) as well as other drug transporter proteins (e.g. MRP, LRP, BCRP) did not predict treatment response and overall survival in ALL (reviewed in Ref. 43). Additionally, various apoptosis-related markers, including mitochondrial pro- and anti-apoptotic proteins, such as Bcl-2 and Bax, and the family of cell death receptors (e.g. CD95 and TRAIL) have been investigated (reviewed in Ref. 7). Except for one report on Bcl-2 in a small series of patients44, studies failed to demonstrate any clinical significance of these parameters in ALL. More recently, constitutive overexpression of the inactive forms of caspases-3 and -2, members of the pro-apoptotic family of cystein proteases involved in the execution phase of apoptosis, has been reported in ALL, but this observation remains to be validated in further studies involving larger series of patients.45

PROGNOSTIC IMPLICATIONS OF IMMUNOPHENOTYPING IN AML Immunophenotyping by flow cytometry has been instrumental in recognizing minimally differentiated AML (AML-M0), acute megakaryoblastic leukaemia (AML-M7), and AML co-expressing lymphoid-associated antigens.8,46,47 The diagnostic sensitivity of a comprehensive panel of mAbs to myeloid and lymphoid lineage as well as progenitorcell-associated antigens has been demonstrated in both childhood and adult AML. Attempts to correlate immunophenotypic features with the various AML subtypes (AML-M1 through AML-M6) according to the FAB classification have been largely unsuccessful (reviewed in Refs. 1,4,5,48). In AML, interpretation of immunophenotyping studies may be confusing because leukaemic blasts in bone marrow and peripheral blood specimens are frequently admixed with normal haematopoietic cells, and the blast cell population can be also heterogeneous. Therefore, various multiparameter flow cytometry techniques have

618 R. Schabath, R. Ratei and W.-D. Ludwig

been proposed that facilitate the identification of leukaemic cells. Among these techniques, the leukocyte common antigen (CD45)/side-scatter (SSC) gating procedure allows an efficient discrimination between the blast cell population and normal cells and facilitates the analysis of leukaemic blasts present in low proportions. The use of CD45/SSC gating, primarily gated on blast cells identified by virtue of the low/intermediate CD45 density, correlates with bone marrow differential and provides characteristic flow-cytometric profiles for most subtypes of AML.49 The prognostic significance of surface antigen expression in AML is still a matter of controversy. Although some investigators, especially in childhood AML, could not show any correlation between the expression of individual progenitor-, myeloid- or lymphoid-associated antigens and treatment outcome50 – 52, others suggested a significant influence of specific antigens or combined phenotypic features on the complete remission (CR) rate and/or CR duration and survival. Among the antigens implicated in having an adverse prognostic effect are CD7, CD9, CD11b, CD13, CD14, HLA-DR, CD34 and TdT.12,53 – 55 On the other hand, the presence of CD15, CD65 and CD2 has been associated with a better treatment outcome.56,57 Other authors could not confirm these findings (e.g. CD2, CD7, CD34, TdT).50 – 52,57 – 59 The comparability of most of these results, however, is hampered by methodological differences such as the choice of mAbs and techniques applied to the detection of antigen expression, inconsistencies in criteria for defining antigen positivity, and variation in the patient populations studied (i.e. children and/or adults) or the treatment administered. Moreover, the prognostic value of correlating clinical outcome with specific antigens rather than evaluating the composite immunophenotype must be questioned in view of recent findings demonstrating that expression of particular antigens can be associated with favourable as well as poor prognostic genetic aberrations. For example, t(8;21), inv(16), chromosome 5 and 7 aberrations, and complex karyotypes, were more frequently observed in CD34þ AML58,60, and CD19 co-expression may occur in AML with either t(8;21) or t(9;22).52,61,62 These results suggest that CD34þ and/or CD19þ AML comprise a heterogeneous group of patients with good as well as poor risk factors. Recent data, suggesting a prognostic role of CD56 expression in AML with t(8;21)63 and acute promyelocytic leukaemia (APL)64 but not in AML cases with 11q23 translocations65 are in line with this statement. In adult AML, immunophenotyping studies using multiparameter flow cytometry suggested that CD56 expression on leukaemia cells in AML with t(8;21) is associated with a significantly shorter CR duration and survival, and thus may be useful in stratifying therapy for this subtype of AML.65 Similarly, a MEDLINE-based analysis of all reported cases of CD56-positive APL suggested that CD56 expression in APL may be used as an indicator of poor treatment outcome.64 These data, however, were obtained in a retrospective manner and should be validated in prospective studies. Our own results in a large series of untreated children and adults with de novo AML enrolled in the German AML-BFM and AMLCG studies do not show any influence of the expression of individual myeloid-, lymphoid-, and progenitor-cellassociated antigens on prognosis52,58,66,67 and thus do not indicate that immunophenotyping alone can be applied to risk stratification in AML at diagnosis. These findings are in line with other recent studies in children50,51 and adults48,59 with AML. Other studies suggested that combined immunophenotypic features, such as CD14þ/HLA-DR2 or CD11bþ, CD117þ, CD142, CD1222, may be more informative for the early identification of adult patients with AML with poor response to induction chemotherapy and shorter survival.12,53 The poor prognosis

Prognostic significance of antigen expression 619

of these immunophenotypic features could not be attributed to other well-known clinical or biological risk factors, including age, cytogenetic abnormalities, or P-gp expression.12 Given the conflicting results regarding the prognostic implications of the expression of single antigens in AML, a prognostic score was recently proposed based on the expression of five myeloid antigens (i.e. MPO, CD13, CD33, CD65s, and CD117).59 Adult AML patients disclosing a panmyeloid phenotype (defined by the expression of all five myeloid-lineage associated antigens) had a significantly higher CR rate as well as a longer disease-free and overall survival. Interestingly, in multivariate analysis, only age, the panmyeloid phenotype, and P-gp activity influenced treatment outcome. However, the results of this single-centre study were obtained in a rather small cohort of patients and have not yet been confirmed in larger multicentre studies. Leukaemias of the M0 subtype, which cannot be recognized on morphological grounds alone, comprise 3– 6% of paediatric AML and up to 10% of adult AML. In 1991, the FAB Cooperative Group listed morphological, cytochemical and immunophenotypic diagnostic criteria, and proposed the designation ‘M0’ for these leukaemias.47 More recently, stricter guidelines for excluding lymphoblastic and megakaryoblastic leukaemias have been proposed. They are based on the availability of more specific lineage-restricted mAbs, the use of multicolour flow cytometry, and the cytoplasmic detection of myeloid antigens in fixed cells (e.g. CD13, MPO) (reviewed in Refs. 4,5,8). According to these criteria, acute leukaemias devoid of detectable MPO can be classified as AML-M0 only in the absence of lineagerestricted lymphoid (e.g. CD3, CD22, CD79a, TCRb) and megakaryocytic antigens (e.g. CD41, CD61). The poor treatment outcome which has been observed in most studies of adult de novo AML-M0 should consider the cytogenetic findings in this immature subset of AML revealing a variety of clonal abnormalities (such as complex karyotypes, anomalies of chromosome 5 and/or 7, trisomy 8, and trisomy 13 reflecting the heterogeneity of minimally differentiated AML).68 – 70 These findings indicate that AML-M0, very similar to AML-M1, is not a unique subtype of leukaemia but probably includes distinct malignant myeloid processes with different underlying cyto- and/or molecular genetic defects. The cytogenetic abnormalities and the higher level of P-gp expression described in most but not all studies may contribute to the poor treatment outcome which has been observed in adults with AML-M0.69,71,72 Expression and functional activity of chemoresistance factors, including drug transporter proteins and members of the apoptosis-regulating protein families, have been extensively studied and appear to be of higher prognostic relevance in AML than in ALL (reviewed in Ref. 7). Functional activity of P-gp and MRP, as well as P-gp expression in combination with expression of Bcl-2, disclosed a prognostic impact in response to induction therapy and overall survival.60,73,74 More recently, striking inverse correlations between CD34 or CD117 mean fluorescence intensity and Bax/ Bcl-2 ratio has been described, suggesting that a lower Bax/Bcl-2 ratio is consistent with immaturity.75 Accordingly, a lower Bax/Bcl-2 levels were correlated with poor-risk cytogenetics, a lower CR rate, and shorter overall survival. Our own studies on constitutive expression and function of the CD95 cell death receptor showed a correlation of the latter parameter with response to induction chemotherapy.60 Investigation of expression levels of caspases-3 and -2, as well as of XIAP, a member of a novel family of caspase inhibitors, pointed to a potential prognostic relevance of these factors.76 Despite these promising results, integration of apoptosis- and chemoresistance-related markers into current treatment protocols requires further prospective clinical studies.

620 R. Schabath, R. Ratei and W.-D. Ludwig

PROGNOSTIC IMPLICATIONS OF IMMUNOPHENOTYPING IN ACUTE LEUKAEMIAS OF AMBIGUOUS LINEAGE The widespread application of flow-cytometric immunophenotyping with a large panel of mAbs to myeloid- and lymphoid-associated differentiation antigens has led to the recognition of acute leukaemias with blasts co-expressing antigens associated with different lineages. Unfortunately, much controversy has surrounded the criteria for identifying such leukaemias, and a variety of terms have been used to refer to these acute leukaemias, such as hybrid, biphenotypic, and mixed-lineage acute leukaemias, thus causing considerable confusion and complicating the assessment of the clinical importance of these observations. More recently, the term acute leukaemias of ambiguous lineage18 has been proposed to describe forms of acute leukaemia in which the morphological, cytochemical and immunophenotypic features of the leukaemic blasts lack sufficient evidence to be classified as of myeloid or lymphoid origin (acute undifferentiated leukaemia) or which have morphological and/or immunophenotypic features of both myeloid and lymphoid cells or, rarely, both B and T lineages (biphenotypic acute leukaemia, BAL). These are rare leukaemias and account for less than 4% of all cases of acute leukaemia. Investigations of the past decade support the concept of two broad categories of acute leukaemias with aberrant or disparate expression of lineage-associated features (reviewed in Refs. 5,18). Acute leukaemias in the most common category, also referred to as Myþ ALL and Lyþ AML, have distinct immunophenotypic, genotypic and clinical features characteristic of a strong commitment to a single lineage but with one or several aberrant features of another lineage. The second category of acute leukaemias display a mixture of antigenic and often also genotypic features that make it unclear whether the leukaemic blasts are committed to a single lineage of differentiation (i.e. true BAL). More recently, strict and well-defined criteria have been proposed that were aimed at distinguishing BAL from those cases with aberrant expression of one or more markers from another lineage (e.g. Myþ ALL and Lyþ AML).5,8,77 Criteria and scoring systems applied to the diagnosis of BAL are based on the number and degree of specificity of the markers (lymphoid and myeloid) expressed by the leukaemic blasts and have been described in detail elsewhere.8,77 The diagnosis of both BAL and Myþ ALL or Lyþ AML requires multiparametric flow cytometry with at least two fluorochromes conjugated with different mAbs to demonstrate co-expression of lineage-specific (e.g. MPO, CD22, CD79a, CD3) and/or lineage-associated antigens. Although extensive data on the cell-biological features and response to treatment of BAL are not yet available, preliminary results suggest that they represent an uncommon subtype with distinct genetic (e.g. Ph translocation, 11q23 re-arrangements, complex cytogenetic abnormalities) and clinical features, as well as a poor prognosis.78 – 81 Based on immunophenotyping, cytogenetic and molecular genetic findings, and the documented phenomenon of in vivo as well as in vitro phenotypic switches in some cases of BAL, it has been suggested that these leukaemias arise in a multipotent progenitor-cell with the capability of differentiating along both myeloid and lymphoid lineages.80 In contrast to BAL, Myþ ALL and Lyþ AML occur frequently. Their incidence has varied considerably between independent studies, both overall and with regard to individual antigens (reviewed in Refs. 82,83), ranging from 5% to more than 50% for Myþ ALL and from 10 to 30% for Lyþ AML. This wide variability has been attributed to a number of reasons, including the lack of consistent criteria for the diagnosis of Myþ ALL or Lyþ AML and defining positive results, the utilization of various panels of

Prognostic significance of antigen expression 621

mAbs, the lack of lineage specificity of most of the mAbs used, and several technical factors.78,82 – 84 In view of previous studies pointing to myeloid-antigen expression as a predictor of poor prognosis in both childhood and adult ALL (reviewed in Refs. 4,17), considerable interest has focused on the cell-biological features and clinical significance of this subgroup of acute leukaemias. Several recent studies, including more than 4000 paediatric patients with ALL85 – 87 and our own data in almost 5000 children treated within the ALL-BFM 86, 90 and 95 trials29,88, have failed to demonstrate an association between Myþ ALL and poor outcome. In some of these studies, myeloid-associated antigen expression was clearly associated with certain genetic features of leukaemic cells, particularly MLL and ETV6-AML1 rearrangements.87,88 (reviewed in Refs. 4,5). In contrast to childhood Myþ ALL, the clinical importance of myeloid-associated antigen expression in adult ALL is still unknown. The presence of myeloid-associated antigens has been associated with a poor outcome in some, but not all studies (reviewed in Ref. 4). Most of these studies, however, included only a relatively small number of patients, have not always carefully excluded minimally differentiated AML (AML-M0), and, most important, have not adequately taken into account the prognostic importance of specific genetic abnormalities frequently found in adult patients with Myþ ALL, such as Ph positivity or 11q23 re-arrangements.21,23,84 Further prospective studies, consistently based on well-defined diagnostic criteria, are urgently needed to elucidate more accurately the biological heterogeneity of Myþ ALL and to establish its clinical relevance in adult patients. A critical review of data published in the literature revealed that most retrospective and prospective studies failed to demonstrate any prognostic significance for Lyþ AML, except for CD7þ AML.89 The latter subgroup has been associated with more frequent expression of progenitor-associated markers (e.g. CD34, CD117, HLA-DR, TdT), concomitant re-arrangements of Ig and/or TCR gene re-arrangements, and poor prognosis in most (but not all) studies in both childhood and adult AML (reviewed in Refs. 4,5,83). It should be noted that immature CD7þ AML and pro-/pre-T ALL occasionally show biological similarities such as reactivity with mAbs recognizing antigens expressed on both immature T-cell ALL and AML, responsiveness to several growth factors, expression of c-kit at the mRNA and protein levels, expression of the multidrug resistance phenotype, and similar TCRd gene re-arrangements, suggesting that, in at least some CD7þ acute leukaemias, malignant transformation has arisen in a pluripotent progenitor-cell with variable differentiation potential along both myeloid and T-lymphoid lineages (reviewed in Ref. 4). Given the significant associations between expression of several lymphoidassociated antigens by AML and specific genetic abnormalities, such as CD19 in AML with t(8;21) and CD2 in AML with t(15;17) as well as AML with inv(16) or t(16;16), cytogenetic and molecular data have to be incorporated into the classification of Lyþ AML, and future studies evaluating the prognostic significance of Lyþ AML have to take into consideration its genetic backgound.

PROGNOSTIC IMPLICATIONS OF IMMUNOPHENOTYPING IN MYELODYSPLASTIC SYNDROME (MDS) Myelodysplastic syndrome (MDS) is a heterogeneous group of malignant disorders of haematopoietic progenitors in which the bone marrow is composed of clonal haematopoietic cells showing various degree of differentiation. Classification of MDS

622 R. Schabath, R. Ratei and W.-D. Ludwig

according to the recommendations by the FAB co-operative group or, more recently, by the WHO, is based primarily upon morphological features and cytogenetics. The prognosis of MDS, significantly differing among FAB and WHO subtypes18, is more accurately predicted by the International Prognostic Scoring System (IPSS) which is composed of three parameters, i.e. the percentage of bone marrow (BM) blasts, the degree of cytopenia, and the karyotype.90 In contrast to de novo AML, relatively few data have been compiled regarding the diagnostic and prognostic impact of immunophenotyping in MDS. Moreover, conflicting results have been reported as to the clinical relevance of surface marker abnormalities in MDS.91 More recently, the diagnostic utility of flow-cytometric immunophenotyping in MDS has been demonstrated by showing myeloid, erythroid, and megakaryocytic abnormalities in 45 patients with MDS.92 Immunophenotyping by multiparameter flow cytometry was especially informative in cases in which the aspirates collected for morphology were inadequate and in differentiating hypocellular MDS from aplastic anaemia. The prognostic significance of immunophenotypic features in patients with MDS was recently shown by two independent studies.93,94 The Groupe d’Etude Immunologique des Leuce´mies (GEIL) showed that specific immunophenotypic clustering of patients with MDS could be achieved on the basis of a primary gating strategy of myeloid cells according to CD45 antigen expression and SSC as well as the quantification of the fluorescence ratio between the measured mean fluorescence intensity of the marker tested and the measured mean autofluorescence of the cells.93 Clusters related to high levels of CD36 expression on CD45low blast cells and CD45high/SSChigh granulocytes were associated with a poor International Prognostic Scoring System (IPSS). Increased CD34 expression was associated with a poor IPSS, an unfavourable cytogenetic risk factor, and high levels of blast cells on BM smear. Valuable information regarding the clinical significance of immunophenotypic data in MDS was also obtained by a study using a new density gradient centrifugation reagent and a CD45 gating strategy for preparing blast-rich specimens.94 Immunophenotyping by multiparameter flow cytometry revealed that a high proportion of the enriched blast cells (EBC) from almost all patients showed an immunophenotype of committed myeloid precursors (i.e. CD34þCD38þHLA-DRþCD13þCD33þ). Markers for myeloid cell immaturity (e.g. CD7, CD117) were more prevalent on EBCs from high-risk MDS and acute leukaemia transformed from MDS. CD7 positivity of EBCs was an independent variable for poor prognosis in MDS. These results, if confirmed in further prospective studies, open new perspectives for both the diagnostic and prognostic application of immunophenotyping by multiparameter flow cytometry to blast cells in MDS.

SUMMARY Several studies, in both children and adult patients, have shown the prognostic impact of immunophenotyping in ALL, especially by identifying certain immunophenotypic subgroups (e.g. pro-B ALL, cortical CD1aþ precursor T-cell ALL) with differing prognosis and by defining maturational subtypes (i.e. B-cell ALL) whose prognosis could be significantly improved by intensive treatment strategies. Additionally, significant associations between immunophenotypic features and genetic abnormalities and/or oncogene expression patterns have recently contributed to a refined classification of precursor B- and T-cell ALL that may contribute to risk-adapted treatment protocols in the future. By contrast, the prognostic impact of immunophenotypic features in AML is still a matter of controversy. Recent findings demonstrating that expression of

Prognostic significance of antigen expression 623

particular antigens can be associated with favourable and poor prognostic genetic aberrations have questioned the value of immunophenotyping as an independent predictor of treatment outcome in AML. The prognostic impact of BAL, an uncommmon subtype displaying a mixture of antigenic and genotypic features that often do not allow a definite lineage assignment, is still unclear. Most recent studies have failed to demonstrate an association of other subtypes of acute leukaemias with aberrant immunophenotypic features (i.e. Myþ ALL and Lyþ AML) with treatment outcome. Recent studies in MDS suggest that immunophenotypic clustering of patients with prognostic relevance can be achieved by multiparameter flow cytometry using specific gating strategies and/or quantification of antigen expression. These results have to be confirmed in further clinical trials and may open new perspectives for the application of immunophenotyping in MDS.

ACKNOWLEDGEMENTS This review was supported by the Bundesministerium fu¨r Forschung und Technologie (Competence Network: Acute and Chronic Leukaemias).

Practice points † multiparameter flow cytometry has been shown to be crucial for the characterization of immunophenotypic subgroups of acute leukaemias and for the detection of MRD † immunophenotyping has become essential in the diagnosis and classification of ALL. In conjunction with cytogenetic and molecular genetic studies, significant associations between immunophenotypic features and numerical and/or structural chromosomal abnormalities have been demonstrated in precursor B-cell ALL † in precursor B-cell ALL, the worse prognosis of specific immunophenotypic subgroups or antigens is frequently due to their association with adverse cellbiological and/or clinical features † recent studies in children and adults have shown that patients with a cortical (CD1aþ) precursor T-cell immunophenotype have a better treatment outcome than those with an immature or mature immunophenotype † most studies in childhood and adult AML have questioned the prognostic value of correlating clinical outcome with expression of specific antigens † AML of the M0 subtype includes distinct malignant myeloid processes with different underlying cytogenetic and/or molecular genetic subtypes that may contribute to the worst prognosis of this subtype † the diagnosis of the two broad categories of acute leukaemias with aberrant or disparate antigen expression patterns (i.e. Myþ ALL/Lyþ AML and BAL) requires immunophenotypic characterization by multiparametric flow cytometry † most recent studies have failed to demonstrate any independent prognostic significance of Myþ ALL and Lyþ AML

624 R. Schabath, R. Ratei and W.-D. Ludwig

Research agenda † the diagnostic and prognostic role of novel immunophenotypic parameters (e.g. antigen intensity, scoring systems) should be evaluated by multiparameter flow cytometry within prospective clinical studies † significant associations between specific immunophenotypic features and other cell-biological characteristics, especially genetic aberrations, require that the prognostic significance of antigen expression in ALL and AML be prospectively evaluated in multicentre trials by adjusting immunophenotyping results for the presence of other cell-biological risk factors † investigations of apoptosis-related parameters and gene expression signatures suggest that maturational stages of precursor T-cell ALL differ as to their expression of T-cell oncogenes and accessibility to apoptotic programmes † the use of apoptosis- and chemoresistance-related parameters as risk features within current treatment protocols for acute leukaemias and MDS require further prospective clinical studies † future investigations evaluating the cell-biological heterogeneity and prognostic significance of BAL require standardized procedures of multiparametric flow cytometry and have to take into consideration its genetic background † the diagnostic and prognostic impact of immunophenotyping in MDS has been recently demonstrated. However, the clinical significance of immunophenotypic clustering of MDS patients by specific gating strategies, and the inclusion of qualitative as well as quantitative parameters of antigen expression, has to be corroborated by ongoing studies

REFERENCES * 1. Jennings CD & Foon KA. Recent advances in flow cytometry: application to the diagnosis of hematologic malignancy. Blood 1997; 90: 2863–2892. 2. Orfao A, Schmitz G, Brando B et al. Clinically useful information provided by the flow cytometric immunophenotyping of hematological malignancies: current status and future directions. Clinical Chemistry 1999; 45: 1708– 1717. 3. Be´ne´ MC, Bernier M, Castoldi G et al. Impact of immunophenotyping on management of acute leukemias. Haematologica 1999; 84: 1024–1034. * 4. Ludwig WD, Haferlach T & Schoch C. Classification of acute leukemias: perspective 1. In Pui CH (ed.) Treatment of Acute Leukemias: New Directions for Clinical Research. Totowa, NJ: Humana Press Inc, 2003, pp 3– 41. * 5. Behm F. Classification of acute leukemias: perspective 2. In Pui CH (ed.) Treatment of Acute Leukemias: New Directions for Clinical Research. Totowa, NJ: Humana Press Inc, 2003, pp 43–58. 6. Campana D & Coustan-Smith E. Advances in the immunological monitoring of childhood acute lymphoblastic leukemia. Clinical Haematology 2002; 15: 1– 19. 7. Schimmer AD, Hedley DW, Penn LZ & Minden MD. Receptor- and mitochondrial-mediated apoptosis in acute leukemia: a translational view. Blood 2001; 98: 3541–3553. * 8. Be´ne´ MC, Castoldi G, Knapp W et al. Proposals for the immunological classification of acute leukemias. European Group for the Immunological Characterization of Leukemias (EGIL). Leukemia 1995; 9: 1783–1786. 9. Paietta E, Andersen J & Wiernik PH. A new approach to analyzing the utility of immunophenotyping for predicting clinical outcome in acute leukemia. Eastern Cooperative Oncology Group. Leukemia 1996; 10: 1–4. 10. Rothe G & Schmitz G. Consensus protocol for the flow cytometric immunophenotyping of hematopoietic malignancies. Working Group on Flow Cytometry and Image Analysis. Leukemia 1996; 10: 877 –895.

Prognostic significance of antigen expression 625 11. Davis BH, Foucar K, Szczarkowski W et al. U.S. –Canadian Consensus recommendations on the immunophenotypic analysis of hematologic neoplasia by flow cytometry: medical indications. Cytometry 1997; 30: 249– 263. 12. Paietta E, Andersen J, Yunis J et al. Acute myeloid leukemia expressing the leucocyte integrin CD11b—a new leukaemic syndrome with poor prognosis: result of an ECOG database analysis. Eastern Cooperative Oncology Group. British Journal of Haematology 1998; 100: 265–272. 13. Porwit MacDonald A, Janossy G, Ivory K et al. Leukemia-associated changes identified by quantitative flow cytometry. IV. CD34 overexpression in acute myelogenous leukemia M2 with t(8;21). Blood 1996; 87: 1162– 1169. 14. De Zen L, Orfao A, Cazzaniga G et al. Quantitative multiparametric immunophenotyping in acute lymphoblastic leukemia: correlation with specific genotype. I. ETV6/AML1 ALLs identification. Leukemia 2000; 14: 1225–1231. 15. Tabernero MD, Bortoluci AM, Alaejos I et al. Adult precursor B-ALL with BCR/ABL gene rearrangements displays a unique immunophenotype based on the pattern of CD10, CD34, CD13 and CD38 expresssion. Leukemia 2001; 15: 406– 414. * 16. Hrusak O & Porwit MacDonald A. Antigen expression patterns reflecting genotype of acute leukemias. Leukemia 2002; 16: 1233–1258. 17. Ludwig WD, Raghavachar A & Thiel E. Immunophenotypic classification of acute lymphoblastic leukemia. Baillie`re’s Clinical Haematology 1994; 7: 235–262. * 18. WHO, Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France: IARC Press, 2001. 19. Raimondi SC, Behm FG, Roberson PK et al. Cytogenetics of pre-B-cell acute lymphoblastic leukemia with emphasis on prognostic implications of the t(1;19). Journal of Clinical Oncology 1990; 8: 1380–1388. 20. Hoelzer D, Ludwig WD, Thiel E et al. Improved outcome in adult B-cell acute lymphoblastic leukemia. Blood 1996; 87: 495 –508. 21. Czuczman MS, Dodge RK, Stewart CC et al. Value of immunophenotype in intensively treated adult acute lymphoblastic leukemia: cancer and leukemia Group B study 8364. Blood 1999; 93: 3931–3939. 22. Reiter A, Schrappe M, Tiemann M et al. Improved treatment results in childhood B-cell neoplasms with tailored intensification of therapy: a report of the Berlin-Frankfurt-Munster Group Trial NHL-BFM 90. Blood 1999; 94: 3294–3306. 23. Ludwig WD, Rieder H, Bartram CR et al. Immunophenotypic and genotypic features, clinical characteristics, and treatment outcome of adult pro-B acute lymphoblastic leukemia: results of the German multicenter trials GMALL 03/87 and 04/89. Blood 1998; 92: 1898– 1909. 24. Uckun FM, Sather H, Gaynon P et al. Prognostic significance of the CD10 þ CD19 þ CD34 þ Bprogenitor immunophenotype in children with acute lymphoblastic leukemia: a report from the Children’s Cancer Group. Leukemia & Lymphoma 1997; 27: 445 –457. 25. Lenormand B, Be´ne´ MC, Lesesve JF et al. PreB1 (CD102) acute lymphoblastic leukemia: immunophenotypic and genomic characteristics, clinical features and outcome in 38 adults and 26 children. The Groupe dEtude Immunologique des Leucemies. Leukemia & Lymphoma 1998; 28: 329– 342. 26. Consolini R, Legitimo A, Rondelli R et al. Clinical relevance of CD10 expression in childhood ALL. The Italian Association for Pediatric Hematology and Oncology (AIEOP). Haematologica 1998; 83: 967– 973. 27. Pui CH & Evans WE. Acute lymphoblastic leukemia. New England Journal of Medicine 1998; 339: 605– 615. 28. Crist W, Boyett J, Jackson J et al. Prognostic importance of the pre-B-cell immunophenotype and other presenting features in B-lineage childhood acute lymphoblastic leukemia: a Pediatric Oncology Group study. Blood 1989; 74: 1252–1259. 29. Ludwig WD, Harbott J, Bartram CR et al. Incidence and prognostic significance of immunophenotypic subgroups in childhood acute lymphoblastic leukemia: experience of the BFM study 86. Recent Results in Cancer Research 1993; 131: 269 –282. 30. Hann IM, Richards SM, Eden OB & Hill FG. Analysis of the immunophenotype of children treated on the Medical Research Council United Kingdom Acute Lymphoblastic Leukemia Trial XI (MRC UKALLXI). Medical Research Council Childhood Leukemia Working Party. Leukemia 1998; 12: 1249– 1255. 31. Ratei R, Sperling C, Karawajew L et al. Immunophenotype and clinical characteristics of CD45-negative and CD45-positive childhood acute lymphoblastic leukemia. Annals of Hematology 1998; 77: 107 –114. 32. Uckun FM, Sensel MG, Sun L et al. Biology and treatment of childhood T-lineage acute lymphoblastic leukemia. Blood 1998; 91: 735 –746. 33. Crist WM, Shuster JJ, Falletta J et al. Clinical features and outcome in childhood T-cell leukemia-lymphoma according to stage of thymocyte differentiation: a Pediatric Oncology Group Study. Blood 1988; 72: 1891– 1897. 34. Thiel E, Kranz BR, Raghavachar A et al. Prethymic phenotype and genotype of pre-T (CD7þ/ER2)-cell leukemia and its clinical significance within adult acute lymphoblastic leukemia. Blood 1989; 73: 1247– 1258.

626 R. Schabath, R. Ratei and W.-D. Ludwig 35. Niehues T, Kapaun P, Harms DO et al. A classification based on T cell selection-related phenotypes identifies a subgroup of childhood T-ALL with favourable outcome in the COALL studies. Leukemia 1999; 13: 614–617. 36. Pullen J, Shuster JJ, Link M et al. Significance of commonly used prognostic factors differs for children with T cell acute lymphocytic leukemia (ALL), as compared to those with B-precursor ALL. A Pediatric Oncology Group (POG) study. Leukemia 1999; 13: 1696–1707. 37. Karawajew L, Ruppert V, Wuchter C et al. Inhibition of in vitro spontaneous apoptosis by IL-7 correlates with upregulation of Bcl-2, cortical/mature immunophenotype, and better early cytoreduction in childhood T-ALL. Blood 2000; 96: 297 –306. 38. Wuchter C, Ruppert V, Schrappe M et al. In vitro susceptibility to dexamethasone- and doxorubicininduced apoptotic cell death in context of maturation stage, responsiveness to interleukin 7, and early cytoreduction in vivo in childhood T-cell acute lymphoblastic leukemia. Blood 2002; 99: 4109–4115. * 39. Ferrando AA, Neuberg DS, Staunton J et al. Gene expression signatures define novel oncogenic pathways in T cell acute lymphoblastic leukemia. Cancer Cell 2002; 1: 75–87. 40. Campana D, van Dongen JJ, Mehta A et al. Stages of T-cell receptor protein expression in T-cell acute lymphoblastic leukemia. Blood 1991; 77: 1546–1554. 41. Alfsen GC, Beiske K, Holte HT et al. T-cell receptor gammadelta þ /CD3 þ 4 2 8 2 T-cell acute lymphoblastic leukemias: a distinct subgroup of leukemias in children. A report of five cases. Blood 1991; 77: 2023–2030. 42. Schott G, Sperling C, Schrappe M et al. Immunophenotypic and clinical features of T-cell receptor gammadelta þ T-lineage acute lymphoblastic leukemia. British Journal of Haematology 1998; 101: 753–755. 43. Ross DD. Novel mechanisms of drug resistance in leukemia. Leukemia 2000; 14: 467 –473. 44. Maung ZT, MacLean FR, Reid MM et al. The relationship between bcl-2 expression and response to chemotherapy in acute leukemia. British Journal of Haematology 1994; 88: 105–109. 45. Faderl S & Estrov Z. The clinical significance of caspase regulation in acute leukemia. Leukemia & Lymphoma 2001; 40: 471 –481. 46. Bennett JM, Catovsky D, Daniel MT et al. Criteria for the diagnosis of acute leukemia of megakaryocyte lineage (M7). A report of the French-American-British Cooperative Group. Annals of Internal Medicine 1985; 103: 460 –462. 47. Bennett JM, Catovsky D, Daniel MT et al. Proposal for the recognition of minimally differentiated acute myeloid leukemia (AML-M0). British Journal of Haematology 1991; 78: 325 –329. 48. Orfao A, Vidriales B, Gonzalez M et al. Diagnostic and prognostic importance of immunophenotyping in adults with acute myeloid leukemia. In: Recent Results in Cancer Research 1993; 131: 369–379. 49. Borowitz MJ, Guenther KL, Shults KE & Stelzer GT. Immunophenotyping of acute leukemia by flow cytometric analysis. Use of CD45 and right-angle light scatter to gate on leukaemic blasts in three-color analysis. American Journal of Clinical Pathology 1993; 100: 534–540. 50. Smith FO, Lampkin BC, Versteeg C et al. Expression of lymphoid-associated cell surface antigens by childhood acute myeloid leukemia cells lacks prognostic significance. Blood 1992; 79: 2415–2422. 51. Kuerbitz SJ, Civin CI, Krischer JP et al. Expression of myeloid-associated and lymphoid-associated cellsurface antigens in acute myeloid leukemia of childhood: a Pediatric Oncology Group study. Journal of Clinical Oncology 1992; 10: 1419–1429. 52. Creutzig U, Harbott J, Sperling C et al. Clinical significance of surface antigen expression in children with acute myeloid leukemia: results of study AML-BFM-87. Blood 1995; 86: 3097–3108. 53. Solary E, Casasnovas RO, Campos L et al. Surface markers in adult acute myeloblastic leukemia: correlation of CD19þ, CD34þ and CD14þ/DR2 phenotypes with shorter survival. Groupe d’Etude Immunologique des Leucemies (GEIL). Leukemia 1992; 6: 393–399. 54. Del Poeta G, Stasi R, Venditti A et al. Prognostic value of cell marker analysis in de novo acute myeloid leukemia. Leukemia 1994; 8: 388 –394. 55. Bradstock K, Matthews J, Benson E et al. Prognostic value of immunophenotyping in acute myeloid leukemia. Australian Leukemia Study Group. Blood 1994; 84: 1220–1225. 56. Schwarzinger I, Valent P, Ko¨ller U et al. Prognostic significance of surface marker expression on blasts of patients with de novo acute myeloblastic leukemia. Journal of Clinical Oncology 1990; 8: 423–430. 57. Ball ED, Davis RB, Griffin JD et al. Prognostic value of lymphocyte surface markers in acute myeloid leukemia. Blood 1991; 77: 2242– 2250. 58. Sperling C, Bu¨chner T, Creutzig U et al. Clinical, morphologic, cytogenetic and prognostic implications of CD34 expression in childhood and adult de novo AML. Leukemia & Lymphoma 1995; 17: 417–426. 59. Legrand O, Perrot JY, Baudard M et al. The immunophenotype of 177 adults with acute myeloid leukemia: proposal of a prognostic score. Blood 2000; 96: 870–877.

Prognostic significance of antigen expression 627 60. Wuchter C, Karawajew L, Ruppert V et al. Clinical significance of CD95, Bcl-2 and Bax expression and CD95 function in adult de novo acute myeloid leukemia in context of P-glycoprotein function, maturation stage, and cytogenetics. Leukemia 1999; 13: 1943–1953. 61. Cuneo A, Michaux JL, Ferrant A et al. Correlation of cytogenetic patterns and clinicobiological features in adult acute myeloid leukemia expressing lymphoid markers. Blood 1992; 79: 720–727. 62. Reading CL, Estey EH, Huh YO et al. Expression of unusual immunophenotype combinations in acute myelogenous leukemia. Blood 1993; 81: 3083–3090. 63. Baer MR, Stewart CC, Lawrence D et al. Expression of the neural cell adhesion molecule CD56 is associated with short remission duration and survival in acute myeloid leukemia with t(8;21)(q22;q22). Blood 1997; 90: 1643–1648. 64. Murray CK, Estey E, Paietta E et al. CD56 expression in acute promyelocytic leukemia: a possible indicator of poor treatment outcome? Journal of Clinical Oncology 1999; 17: 293–297. 65. Baer MR, Stewart CC, Lawrence D et al. Acute myeloid leukemia with 11q23 translocations: myelomonocytic immunophenotype by multiparameter flow cytometry. Leukemia 1998; 12: 317–325. 66. Sperling C, Schwartz S, Bu¨chner T et al. Expression of the stem cell factor receptor C-KIT (CD117) in acute leukemias. Haematologica 1997; 82: 617–621. 67. Schwartz S, Heinecke A, Zimmermann M et al. Expression of the C-kit receptor (CD117) is a feature of almost all subtypes of de novo acute myeloblastic leukemia (AML), including cytogenetically good-risk AML, and lacks prognostic significance. Leukemia & Lymphoma 1999; 34: 85– 94. 68. Cuneo A, Ferrant A, Michaux JL et al. Cytogenetic profile of minimally differentiated (FAB M0) acute myeloid leukemia: correlation with clinicobiologic findings. Blood 1995; 85: 3688–3694. 69. Venditti A, Del Poeta G, Buccisano F et al. Minimally differentiated acute myeloid leukemia (AML-M0): comparison of 25 cases with other French-American-British subtypes. Blood 1997; 89: 621 –629. 70. Be´ne´ MC, Bernier M, Casasnovas RO et al. Acute myeloid leukemia M0: haematological, immunophenotypic and cytogenetic characteristics and their prognostic significance: an analysis in 241 patients. British Journal of Haematology 2001; 113: 737–745. 71. Stasi R, Del Poeta G, Venditti A et al. Analysis of treatment failure in patients with minimally differentiated acute myeloid leukemia (AML-M0). Blood 1994; 83: 1619–1625. 72. Cohen PL, Hoyer JD, Kurtin PJ et al. Acute myeloid leukemia with minimal differentiation. A multiple parameter study. American Journal of Clinical Pathology 1998; 109: 32–38. 73. Legrand O, Simonin G, Perrot JY et al. Pgp and MRP activities using calcein-AM are prognostic factors in adult acute myeloid leukemia patients. Blood 1998; 91: 4480–4488. 74. Kasimir Bauer S, Beelen D, Flasshove M et al. Impact of the expression of P glycoprotein, the multidrug resistance-related protein, bcl-2, mutant p53, and heat shock protein 27 on response to induction therapy and long-term survival in patients with de novo acute myeloid leukemia. Experimental Hematology 2002; 30: 1302–1308. * 75. Del Poeta G, Venditti A, Del Principle MI et al. Amount of spontaneous apoptosis detected by Bax/Bcl-2 ratio predicts outcome in acute myeloid leukemia (AML). Blood 2003; 101: 2125–2131. 76. Tamm I, Kornblau SM, Segall H et al. Expression and prognostic significance of IAP-family genes in human cancers and myeloid leukemias. Clinical Cancer Research 2000; 6: 1796–1803. 77. Be´ne´ MC, Bernier M, Casasnovas RO et al. The reliability and specificity of c-kit for the diagnosis of acute myeloid leukemias and undifferentiated leukemias. The European Group for the Immunological Classification of Leukemias (EGIL). Blood 1998; 92: 596–599. 78. Hanson CA, Abaza M, Sheldon S et al. Acute biphenotypic leukemia: immunophenotypic and cytogenetic analysis. British Journal of Haematology 1993; 84: 49–60. 79. Carbonell F, Swansbury J, Min Tet al. Cytogenetic findings in acute biphenotypic leukemia. Leukemia 1996; 10: 1283–1287. 80. Matutes E, Morilla R, Farahat N et al. Definition of acute biphenotypic leukemia. Haematologica 1997; 82: 64–66. 81. Legrand O, Perrot JY, Simonin G et al. Adult biphenotypic acute leukemia: an entity with poor prognosis which is related to unfavourable cytogenetics and P-glycoprotein over-expression. British Journal of Haematology 1998; 100: 147–155. 82. Drexler HG & Ludwig WD. Incidence and clinical relevance of myeloid antigen-positive acute lymphoblastic leukemia. Recent Results in Cancer Reseach 1993; 131: 53–66. 83. Drexler HG, Thiel E & Ludwig WD. Acute myeloid leukemias expressing lymphoid-associated antigens: diagnostic incidence and prognostic significance. Leukemia 1993; 7: 489–498. 84. Khalidi HS, Chang KL, Medeiros LJ et al. Acute lymphoblastic leukemia. Survey of immunophenotype, French-American-British classification, frequency of myeloid antigen expression, and karyotypic abnormalities in 210 pediatric and adult cases. American Journal of Clinical Pathology 1999; 111: 467 –476.

628 R. Schabath, R. Ratei and W.-D. Ludwig 85. Uckun FM, Sather HN, Gaynon PS et al. Clinical features and treatment outcome of children with myeloid antigen positive acute lymphoblastic leukemia: a report from the Children’s Cancer Group. Blood 1997; 90: 28–35. 86. Putti MC, Rondelli R, Cocito MG et al. Expression of myeloid markers lacks prognostic impact in children treated for acute lymphoblastic leukemia: Italian experience in AIEOP-ALL 88-91 studies. Blood 1998; 92: 795–801. * 87. Pui CH, Rubnitz R, Hancock ML et al. Reappraisal of the clinical and biologic significance of myeloidassociated antigen expression in childhood acute lymphoblastic leukemia. Journal of Clinical Oncology 1998; 16: 3768–3773. 88. Ludwig WD, Harbott J, Rieder H, et al. Incidence, biologic features and treatment outcome of myeloidantigen-positive acute lymphoblastic leukemia (My þ ALL). In Bu¨chner T, et al. (eds) Acute Leukemias IV. Prognostic Factors. Berlin: Springer, 1994, pp 24. 89. Ogata K, Yokose N, Shioi Y et al. Reappraisal of the clinical significance of CD7 expression in association with cytogenetics in de novo acute myeloid leukemia. British Journal of Haematology 2001; 115: 612–615. 90. Greenberg P, Cox C, LeBeau MM et al. International scoring system for evaluating prognosis in myelodysplastic syndromes. Blood 1997; 89: 2079–2088. 91. Elghetany MT. Surface marker abnormalities in myelodysplastic syndromes. Haematologica 1998; 83: 1104–1115. 92. Stetler Stevenson M, Arthur DC, Jabbour N et al. Diagnostic utility of flow cytometric immunophenotyping in myelodysplastic syndrome. Blood 2001; 98: 979 –987. 93. Maynadie M, Picard F, Husson B et al. Immunophenotypic clustering of myelodysplastic syndromes. Blood 2002; 100: 2349–2356. * 94. Ogata K, Nakamura K, Yokose N et al. Clinical significance of phenotypic features of blasts in patients with myelodysplastic syndrome. Blood 2002; 100: 3887–3896.