immunophenotyping of lymphoproliferative disorders: state...

P R A C T I C A L A S P E C T S O F D I A G N O S T I C H I S T O P A T H O L O G Y

Immunophenotyping of lymphoproliferative disorders: state of the art

EMMA J. GUDGIN AND WENDY N. ERBER

Haematology Department, Addenbrooke’s Hospital, Cambridge, United Kingdom

Summary

Immunophenotyping was introduced into diagnostic pathol-ogy over 30 years ago to assist in the diagnosis andclassification of lymphoproliferative disorders. Today therole of immunophenotyping has been expanded beyondthis to include the detection of markers of prognosis,determination of disease phenotypes associated withspecific chromosomal abnormalities, detection of targetsfor immunotherapy and to monitor residual disease.Immunoperoxidase detection methods remain the mostpopular in histopathology, whilst flow cytometry is mostcommonly applied for haematological samples. The range ofmonoclonal antibodies available, including those which workin routinely performed tissue specimens, continues toincrease. This is in part a result of gene expression studiesidentifying precise genetic signatures for certain lymphopro-liferative disorders and the generation of new proteinmarkers to gene products of upregulated genes. This reviewsummarises the current status and applications of immuno-phenotyping in the assessment of many of the lymphoidmalignancies.

Key words: Phenotype, immunohistochemistry, flow cytometry,

lymphoproliferative, lymphoma.

Received 10 June, revised 8 July, accepted 11 July 2005

INTRODUCTION

Immunophenotyping characterises cellular antigen expres-sion by their ability to bind antibodies. This has an essentialrole in the diagnosis and classification of lymphoprolifera-tive disorders. Although immunophenotyping has beenavailable since the 1970s, it is still an evolving technique.Both new antibodies, which are being added on an almostweekly basis, and new applications are appearing. Therecent development of gene expression profiling, ormicroarray studies, has shown that some disorders havespecific genetic signatures. Monoclonal antibodies are nowbeing developed to the products of those genes that areupregulated in neoplastic disorders, and novel immuno-phenotypical markers are thereby being created. These newantibodies will enable further refinement of the classifica-tion of lymphoproliferative disorders and help us gainimportant prognostic information by identifying biologi-cally distinct subgroups. In addition, new therapeutictargets may be identified.

The applications of immunophenotyping therefore haveexpanded beyond diagnosis and disease classification to

predicting prognosis, detecting therapeutic targets anddisease monitoring. Phenotypical profiles of neoplasticlymphoid cells are increasingly being used as surrogatemarkers of genetic abnormalities related to malignanttransformation. Phenotyping is also being applied to assesswhether a neoplastic cell expresses a specific target antigenfor immunotherapy (i.e., monoclonal antibodies boundto cytotoxic agents). During and after treatment of alymphoid malignancy, minimal residual disease activitymay be assessed by following the expression of a particularabnormal immunophenotypical profile. These novel appli-cations highlight the increasing importance of immuno-phenotyping in the analysis of lymphoproliferativedisorders.

Immunophenotyping of lymphoproliferative disorders,be it by flow cytometry or immunohistochemistry, there-fore has a number of roles including:

1. Determining the B- or T-cell lineage of an abnormallymphoid population.

2. Determining the clonality of lymphoid proliferations,especially those of B-cell origin.

3. Determining whether lymphoid cells have a phenotypeassociated with a specific disorder or a particularchromosomal abnormality.

4. Establishing the phenotype of neoplastic lymphoidcells which may be useful for monitoring residualdisease.

5. Assessing the expression of specific molecules whichmay be targets for immunotherapy (e.g., CD20).

6. Assessing cell proliferation and turnover rate.

7. Detection of markers which may assist in determiningprognosis.

In this review, some general comments will be made onimmunophenotyping methods, followed by a more exten-sive discussion of those lymphoproliferative disorderswith distinctive phenotypical features. We have not setout to give an exhaustive list of data on the immunophe-notype of all lymphoproliferative disorders or all methodsavailable for detecting antigen expression. Data arepresented on some of the newer antibodies that have beenshown, at least in initial studies, to have prognosticsignificance or to be useful targets for immunotherapy.We accept that, in this rapidly evolving field, someconflicting data are emerging, especially in regard to theprognostic significance of some novel immunophenotypicalprofiles.

Pathology (December 2005) 37(6), pp. 457–478

ISSN 0031-3025 printed/ISSN 1465-3931 # 2005 Royal College of Pathologists of Australasia

DOI: 10.1080/00313020500368287

IMMUNOPHENOTYPING METHODS

Immunophenotyping was first described in 1941 by AlbertCoons, who demonstrated the use of a fluorescence labelledantibody to localise cellular antigens in tissue sections. Thisimmunofluorescence technique began to be used inleukaemia diagnosis in the 1970s, but was not universallyapplicable to routine diagnostic laboratories as theyrequired fluorescence microscopy or large flow cytometers.1

These drawbacks were largely overcome by the develop-ment of immunoenzymatic techniques using a numberof different enzyme labels, such as acid phosphatase,horseradish peroxidase and alkaline phosphatase.2,3

Immunoperoxidase techniques were most applicable totissue samples, but due to the high levels of endogenousperoxidase in haemopoietic cells (especially erythroid andmyeloid cells) they were not as suitable for blood and bonemarrow samples. Immunoalkaline phosphatase techniqueswere developed and found to be preferable for use withthese peroxidase-rich samples.4 Immunoperoxidase proce-dures became widely used on tissue samples and remain themethod of choice for immunohistology. With furtherdevelopments in flow cytometry and the availability ofbench-top flow cytometers applicable to routine labora-tories, fluorescence labelling has largely replaced immuno-alkaline phosphatase for the analysis of haematologicalsamples.

Flow cytometry

Flow cytometry is widely used to assist in the diagnosis,classification, detection and monitoring of minimal residualdisease in the majority of lymphoproliferative disorders.This method allows the rapid and simultaneous analysis ofmultiple cell parameters, including cell size, complexity,and both surface membrane and intracellular antigens, onlarge numbers of fresh viable cells (Table 1). The majoradvantage of flow cytometry is that a number of cellularantigens can be analysed simultaneously on a cell popula-tion using multiple monoclonal antibodies labelled withfluorochromes, each of which emits light at a differentwavelength. The large range of antibodies available makesflow cytometry the method of choice for establishing the

phenotype of lymphoid malignancies and detecting clon-ality of B-cell disorders.

Flow cytometry requires cells to be in suspension andcan be performed on a wide range of specimens includingperipheral blood, bone marrow aspirate, fine needleaspirate (FNA) and fluid samples. Cell suspensions canalso be prepared from cells extracted from solid tissuebiopsies (e.g., lymph nodes) and analysed by flowcytometry. The mechanical dissociation of cells is done byteasing out cells from fresh unfixed tissue using a scalpeland forceps, needle and syringe and wire mesh, or using anautomated device (e.g., Medimachine, BD Biosciences,USA; DakoCytomation, Denmark). Enzymatic digestionusing a proteolytic enzyme (e.g., pepsin or trypsin) can beused to assist with separation of cells from fibrotic samples.Red cell lysis can be used to remove contaminating red cellsthat may interfere with the flow cytometric analysis; thisdoes not denature or destroy cellular antigens. Densitygradient centrifugation (e.g., Ficoll-Hypaque) can also beused to remove red cells and cell debris from specimens.This method also concentrates the cells of interest. As flowcytometers are now readily available, are simple to use andcan be used to analyse all sample types, flow cytometry isbecoming the method of choice for the phenotypicalanalysis of lymphoproliferative disorders.

However, there remains a major drawback of flowcytometry: the inability to directly assess cellular morphol-ogy and to correlate this with antigen detection. Therefore,it is critical that the cell sample to be analysed by flowcytometry is assessed morphologically (i.e., cell smear orcytocentrifuge preparation of the extracted cells). This notonly ensures that it morphologically resembles the initialspecimen and that there are sufficient intact cells, but alsoguides antibody selection. Typically, many nodal large celllymphomas, fibrotic lesions (e.g., mediastinal large B-celllymphoma) and Hodgkin lymphomas do not yield diag-nostic samples suitable for flow cytometry due to therelatively low numbers of viable malignant cells extractedcompared with the reactive surrounding cells.

As morphology cannot be used to ‘isolate’ the cell ofinterest, other parameters must be used. Identification ofthe cell population of interest by flow cytometry is

TABLE 1 Comparison of immunohistochemistry and flow cytometry

Immunohistochemistry Flow cytometry

Morphological correlation Direct assessment of morphology with antigenexpression

No visual correlation, depends on other parameters forrecognition of cell types

Automation No equipment is needed; automation isavailable

Flow cytometer

Specimen types Paraffin and frozen sections; smears of blood,bone marrow; body fluids

Fresh unfixed cell samples: blood; bone marrow; body fluids;fine needle aspirates; fresh solid tissues (requires tissuedisaggregation)

Antibodies Limited (for paraffin sections) but expanding UnlimitedTurn-around time Slow; only performed after tissue processing

and histology reviewedHours, i.e., same day

Permanent record Yes Listmode analysis for re-analysing dataSimultaneous detection of multiple

antigensDifficult Routine – enables aberrant phenotypes to be detected

T/B cell clonality Difficult Routine (for B cells)Interpretation Subjective Objective; quantitative. Depends on gating correct cell

populationMinimal residual disease assessment

and rare cell analysisDifficult Routine, especially if the neoplastic cells have a specific

phenotype

458 GUDGIN AND ERBER Pathology (2005), 37(6), December

performed by ‘gating’. This electronically selects thepopulation of interest based on cell parameters, such asforward light scatter which separates cells depending ontheir size, and side scatter which correlates with internal cellcomplexity (including nuclear configuration and cytoplas-mic granularity). Cell populations can also be ‘gated’ bytheir ability to bind a particular antibody. CD45 (leucocytecommon antigen) is most commonly used in this regard togate lymphoid cell populations (Fig. 1). Differential inten-sity of expression of this leucocyte common antigentogether with forward and/or side scatter can isolatedistinct lymphoid cell subpopulations (e.g., lymphoblastswith weak CD45 expression; hairy cell leukaemia withstrong CD45). Lineage-specific antibodies, such as CD19

for B cells or CD3 for T cells, may also be used in gatingstrategies. Isotype controls should be included for allanalyses. These are negative controls which ensure thatthere is no non-specific binding of the primary antibodyto the cell population of interest. Positive and negativequadrants are set from the negative isotype control. Mostsamples analysed will contain some negative cells (i.e.,normal cells that do not express the antigen of interest)which also act as internal negative controls.

Interpretation of the data obtained by flow cytometry isquantitative. The percentage of the cells of interest (i.e.,gated events) positive for a particular antigen is providedtogether with the fluorescence intensity (mean channel offluorescence). The latter is proportional to the density of

Fig. 1 Flow cytometry of lymphoblasts in ALL. (A) (B) Variable intensity of CD45 expression on CD45/side scatter gating (boxes). (C) Dual cytoplasmicCD79a and nuclear TdT in a precursor B-lymphoblastic leukaemia. (D) Cytoplasmic CD3 (and lack of myeloperoxidase) in a case of precursor T-lymphoblastic leukaemia.

IMMUNOPHENOTYPING OF LYMPHOPROLIFERATIVE DISORDERS 459

antigen expression by the cell. In most analyses there isclear separation of positive from negative cells based onfluorescence intensity (Fig. 1). However, with some neo-plastic lymphoid cells, where antigens are only weaklyexpressed, there is no clear cut-off between cells in thepositive and negative quadrants (Fig. 2). In this situation,interpretation is based on the percentage of cells in thepositive quadrant. An arbitrary cut-off level is used todefine positive expression; this is usually 20% of cells.

Flow cytometry is being increasingly used for lymphnode analysis but has still not gained routine acceptance.This is despite the ability to rapidly assess the cell sample(i.e., within hours), determine clonality by light chainrestriction, and simultaneously detect multiple antigens ona cell (e.g., CD5 expression by B cells). These are some ofthe specific areas where flow cytometry has advantages

over immunohistochemistry. Other areas where flowcytometry has advantages but remains under-utilised arethe analysis of fine needle aspirate samples (e.g., of lymphnodes) to assess a possible lymphoproliferative disorder,and bone marrow staging of lymphoma. For FNA samples,sufficient cells can be aspirated to assess clonality of B cells,and a basic panel of markers may be examined. Thistechnique is useful as an initial screening test for alymphoproliferative disorder, with a follow-up tissuebiopsy to confirm.5–7 In marrow staging of lymphoproli-ferative disorders, flow cytometry may be more sensitivethan histology in detecting low levels of clonal cells, butmay miss cases where there are small focal paratrabecularinfiltrates. The clinical relevance of small numbers of clonalcells in the bone marrow detected by flow cytometry aloneis yet to be determined.8–10

Fig. 2 Flow cytometry of chronic lymphocytic leukaemia demonstrating the expression of the prognostic markers CD38 and ZAP70. (A) Expression of CD5by the CD19-positive B cells, a gating strategy to isolate the CLL cells. (B) 78% of the cells are B cells that are CD38 positive. (C)(D) ZAP70 on two cases ofCLL. In (C) the CLL cells (identified with cytoplasmic CD79a) are predominantly ZAP70 negative, whereas in (D) the majority (64%) are ZAP70 positive.


Immunohistochemistry

Immunohistochemistry (IH) is an integral part of routinediagnostic histopathology and remains the phenotypingmethod of choice for tissue biopsies. IH is generallyperformed on routinely processed, formalin-fixed, paraf-fin-embedded tissue. The major advantage of IH over flowcytometry is that morphology and the antigen label(precipitated chromogenic substrate) can be visualisedsimultaneously by light microscopy, enabling accur-ate identification of the cells or region of interest(Table 1). Automated systems are available that can fullyautomate immunostaining for formalin-fixed, paraffin-embedded tissues, frozen sections, cytospins, cell smearsand FNA. Some will perform the entire process fromdewaxing paraffin-embedded tissues, performing antigenretrieval, application of the antibody and the chromogenicsubstrate through to nuclear counter-staining. These high-throughput devices can be programmed to optimise tissuepre-treatment and staining conditions for all types ofsamples and antibodies, with some having batch processingfacilities.

The majority of immunohistochemistry performed ontissue sections uses horseradish peroxidase as the enzymelabel, although some laboratories prefer alkaline phospha-tase. The immunological methods available include indirectantibody staining, unlabelled antibody bridge (e.g., perox-idase anti-peroxidase), avidin (or streptavidin)-biotinenzyme labelling and polymer-bound methods. These allvary in sensitivity but signal strength can be enhanced by anumber of approaches. For example, the avidin-biotinsystem can be enhanced with biotinyl tyramide, a phenoliccompound which binds covalently to electron-rich moietiesat the active sites of horseradish peroxidase.11–14 Thisincreases the intensity of the reaction product 100–10 000-fold over standard methods without loss of specificity,

thereby amplifying weak signals. This approach alsoenables antibodies that were previously unreactive informalin-fixed, paraffin-embedded tissue to be detectable.14

Other enhancement methods involve alteration of thecomposition of the chromogenic substrate (e.g., additionof cobalt chloride).

For many years, the number of antibodies for detectionof cellular antigens in paraffin-embedded material waslimited and diagnostic phenotyping was restricted to frozensections of fresh tissue. It was thought that antigens hadbeen destroyed during the formalin fixation and tissueprocessing; this is not the case. A range of antigen retrievalmethods are now available that expose or ‘unmask’previously undetectable antigens, thereby significantlyincreasing the range of antigens that can be detected informalin-fixed tissue.11,15–17 The retrieval methods includeproteolytic enzymes (e.g., pronase, trypsin), microwaveoven, autoclave or waterbath heating, or pressure cookingand use of buffers (e.g., citrate, EDTA) at pH 6.0–9.5.16,17

The choice of antigen retrieval method varies fromlaboratory to laboratory, in large part due to differencesin tissue preparation, and also by antibody clone. Hence,no advice can be given as to the optimal methodology to beused for a particular antibody.

Currently there is an extensive range of antibodies tolymphoid-associated antigens that can be used in routinelyprocessed tissue but this still remains more limited than forfresh cells (Table 2). This is most relevant in the assessmentof clonality, as light chains of immunoglobulin are not asreadily detected by IH on fixed material as on fresh cells.Other limitations of immunohistochemistry in the assess-ment of lymphoproliferative disorders are (Table 1):

1. The inability to assess more than one antigen on thesame cell (i.e., ‘multiple staining’).

TABLE 2 Antibodies of value for phenotyping lymphoproliferative disorders by flow cytometry (or fresh cell analysis) or immunohistochemistry on fixedtissue*

Disorder/cell typeFresh cells (e.g., flow cytometry,

cell smears or frozen sections) Fixed tissue (IH)

B-cell malignancies/Pan-B-cell antigens CD19, CD20, CD22, CD79b, Igm, Igd, Igk, Igl CD20, CD45RA, CD75, CD79a, PAX5, MUM1Precursor B-cells CD1d, CD10, CD34, CD45, CD58, TdT, NG2,

Myeloid markersCD10, CD34, TdT

Chronic lymphocytic leukaemia CD5, CD23, CD38, CD52{, ZAP70 CD5, CD23, p53Hairy cell leukaemia CD11c, CD25, CD103 CD11c, DBA-44Follicular lymphoma CD10 CD10, CD21, BCL2, BCL6Mantle cell lymphoma CD5 CD5, Cyclin-D1Diffuse large B-cell lymphoma CD5, CD10 CD5, CD10, CD30, BCL2, BCL6, MUM1Mediastinal B-cell lymphomaBurkitt’s lymphoma CD10, TdT CD10, BCL6, MIB-1

T-cell malignancies/Pan-T-cell antigens CD2, CD3, CD4, CD5, CD7, CD8, ab TCR, cd TCR CD2, CD3, CD4, CD5, CD7, CD8, CD43, CD45ROPrecursor T cells CD1a, TdT, Myeloid markers CD34, TdTT-prolymphocytic leukaemia CD26, CD52{T-large granular lymphocyte leukaemia CD16, CD56, CD57 TIA-1, granzyme BSezary syndrome/Mycosis fungoides CD26Adult T-cell leukaemia/lymphoma CD25, HLA-DR CD30Anaplastic large cell lymphoma CD15, CD30, CD45, ALK, TIA-1, perforin, EMA,

granzyme BHepatosplenic T-cell lymphoma TCRab, TCRcd, CD16, CD56 TIA-1

Natural killer cell antigens (and naturalkiller cell malignancies)

CD2, CD7, CD8, CD16, CD56, CD57 CD2, CD57

*Pan-T, -B and natural killer cell antigens are listed as well as those antibodies of particular value for individual lymphoproliferative disorders.{CD52 is not diagnostically useful but identifies a potentially useful therapeutic target.


2. Slow turnaround time as immunohistochemistry isperformed only after initial morphological assessmentof the tissue sections.

3. Only limited analysis is possible on small samples.

New techniques are being developed in IH and includethe analysis of tissue arrays and use of antibody arrays.Tissue arrays are a high throughput approach to the studyof antigen expression in tissue biopsies. The arrayed tissuebiopsies are stained using identical immunohistochemicalmethods to those used for routine diagnostic testing. Thisarray-based technology can be used to rapidly evaluate thediagnostic and prognostic value of newly produced anti-bodies. This approach is rapidly generating vast amounts ofinformation about the utility of these new antibodies indiagnostic pathology, some examples of which will be givenin this review.

Antibody arrays or solid-phase cytometry technologyhave also been developed.18,19 These are arrays of antibodiesimmobilised onto a solid phase (e.g., cellulose membrane orglass slide) with cell capture assessed by immunocytochem-ical staining or light transmission. The phenotype of cellsbound to the arrayed antibodies correlates with surfaceantigen expression. This methodology enables large num-bers of cellular antigens to be assessed simultaneously in asingle analysis. Both of these new techniques have potentialfor routine immunophenotyping but have yet to beincorporated into routine laboratory practice.

THE PHENOTYPE OF LYMPHOPROLIFERATIVEDISORDERS

Phenotyping is a routine component in the assessment ofthe majority of lymphoproliferative disorders. To date, nosingle antibody has been shown to be specific for amalignant lymphoid cell, so patterns of antigen expressionare used for the immunophenotypical classification oflymphoproliferative disorders. Hence, panels of antibodiesare used to determine the cell lineage, stage of differentia-tion and, especially for B cells, clonality. Using thisapproach, a number of the lymphoproliferative disordersexhibit characteristic phenotypes or signatures which assistin disease classification. The optimal number and range ofantibodies to evaluate lymphoproliferative disorders is amatter of debate. Table 2 lists many of the most usefulantibodies currently available. For most analyses, a limitedprimary panel of antibodies should be used. This wouldinclude pan-B (e.g., CD19, CD20, CD79a) and pan-T (e.g.,CD3, CD5) cell antibodies. If the morphology suggests aspecific lymphoproliferative disorder with a characteristicphenotype (e.g., hairy cell leukaemia) then these antibodies(e.g., CD11c, CD25, CD103) should be included (Table 2).

B-cell lymphoproliferative disorders are identified by theexpression of B-cell-associated antigens and light chain (kor l) restriction, indicating clonality. The precise diseasetype may also be determined from the pattern of antigenexpression. The immunophenotypical diagnosis of T-celllymphoproliferative disorders is more difficult than for theB-cell malignancies as T-cell clonality cannot be as readilyestablished. However, many T-cell malignancies showatypical T-cell phenotypes, such as aberrant loss of anexpected T-cell antigen (typically CD5 or CD7), and loss of

or co-expression of CD4 and CD8 antigens. Detection ofaberrant T-cell phenotypes can be used as an indicator ofneoplasia. For some B- and T-cell disorders, additionalnon-lineage-associated antibodies may be used to predictuse of immunotherapy (e.g., CD20), determining prolifera-tion rate (e.g., MIB1) or prognosis, or recognition ofantigens of use in residual disease monitoring during andfollowing treatment.

Initially, antibodies used in routine phenotyping were tocell surface antigens expressed by normal lymphoid cells,either during ontogeny or following activation, and thoserecognising heavy and light chains of immunoglobulin (Ig).More recently antibodies to gene products of chromosomaltranslocations have been introduced into routine pheno-typing, such as antibodies to cyclin-D1 from rearrangementof the CCND1 (BCL1) gene as a result of t(11;14) in mantlecell lymphoma. The protein products of other genesactivated by chromosomal rearrangements also have a roleas markers of either lineage (e.g., PAX-5 [B-cell-specificactivator protein; BSAP] for B cells) or maturation stage(e.g., BCL6 for germinal centre).20 Newer antibodies tomolecules whose genes are upregulated are also beingincorporated into phenotyping studies (e.g., FOXP1 andHGAL) for lymphoid malignancies. These antibodiesas well as lineage-specific or lineage-associated antibodieswill be discussed in relation to particular lymphoidmalignancies.21

Lymphoblastic leukaemia/lymphoma

Lymphoblastic leukaemias and lymphomas (ALL/LBL) areneoplasms derived from progenitor lymphoid cells orlymphoblasts of B or T lineage, and are overlappingpresentations of the same disease entity. They are classifiedaccording to phenotype, with both the cell lineage (i.e., B orT) and the stage of differentiation at which maturationarrest occurred being relevant. The lymphoblasts arecharacterised by expression of terminal deoxynucleotidyltransferase (TdT) within the nucleus. The B or T lineage ofALL/LBL is based on the expression of at least cytoplasmicCD79a for B-cell and cytoplasmic CD3 for T-cell disease.In 1995 the European Group for the ImmunologicalCharacterisation of Leukaemias (EGIL) proposed aclassification system that divided precursor B-ALL andprecursor T-ALL each into subgroups based on cellphenotype.22 This phenotypical subclassification providesuseful prognostic information. B-lineage ALL is morecommon than T-ALL, accounting for 85–90% of paedia-tric, and 75% of adult cases of ALL.21 In contrast, T-LBLis more common than B-LBL, comprising 85–90% ofcases.21

Minimal residual disease (MRD) detection is becomingincreasingly important in the management of ALL. Severalstudies have shown that combinations of antibodies andmulti-parameter flow cytometry has a sensitivity of 1024

for MRD assessment, which is comparable with molecularmethods.23–25 Clinical use of this approach necessitatesextensive phenotypical analysis at diagnosis to identifynovel phenotypical combinations expressed on the leukae-mic cells but not on normal haemopoietic cells.

Phenotyping is also performed in ALL/LBL to detectpotential immunotherapeutic targets, such as CD19, CD20,CD22, CD33 and CD52. A prerequisite for antibody


therapy is the presence of the target antigen on at least 20–30% of the blast cells.26,27 To date, most experience hasbeen with CD20 (rituximab) in precursor B-LBL.28,29

The role of immunophenotyping in lymphoblasticneoplasms therefore is not restricted to diagnosis and itcan be used to:

1. Confirm the neoplastic cells are primitive lymphoidcells and establish the B- or T-cell lineage of thelymphoblasts.

2. Determine the stage of differentiation of the lympho-blast and hence the precise classification of the ALL/LBL.

3. Identify markers which may have prognostic signifi-cance.

4. Identify unique antigens or patterns of antigenexpression that could be used to monitor residualdisease following therapy.

5. Assess the expression of molecules that may be targetsfor immunotherapy.

The phenotypical characteristics of B and T lympho-blastic leukaemias and lymphomas will be discussed. Mostof the data presented have been established by flowcytometric analysis of peripheral blood and/or bonemarrow. However, IH can also be used to detect many ofthe antigens required to assess these disorders (Table 2).

Precursor B-lymphoblastic leukaemia/lymphoma

The lymphoblasts in precursor B-ALL/LBL are CD19 andcytoplasmic CD79a positive, CD22 positive (initially in thecytoplasm and later on the cell surface), express TdT withinthe nucleus, and are HLA-DR positive. They have variableexpression of CD34 and weak CD45 expression (Fig. 1).Precursor B-ALL/LBL can be further subclassified basedon the degree of B cell differentiation and expression ofCD10, CD20 and cytoplasmic or surface Igm (heavy chainsof IgM) (Table 3). The three subtypes are:

1. Pro-B-ALL, the earliest stage in which the neoplasticcells are negative for CD10, CD20 and cytoplasmic (cy)Igm.

2. Common ALL, the intermediate stage, where the cellsexpress CD10, are negative for cy Igm and have variableCD20 expression (a pan-B-cell antigen expressed

slightly later in B-cell development than CD19, CD22and CD79a).

3. Pre-B-ALL, which is derived from the most matureprecursor B cell, with neoplastic cells positive forCD10, CD20 and cy Igm but negative for surface Igm.

Other antigens of note in B-cell ALL/LBL are CD45 andantibodies to myeloid associated antigens. CD45, theleucocyte common antigen, is characteristically expressedmore weakly on lymphoblasts than normal lymphocytes,and the intensity can vary between ALL/LBL cases. Onflow cytometry, this weak CD45 expression together withlow side scatter are useful characteristics for gating thelymphoblasts (Fig. 1). As ALL cases are commonlyanalysed with a panel of antibodies, aberrant expressionof myeloid antigens (e.g., CD33) may be detected.Approximately one-third of B-ALL aberrantly expressmyeloid associated antigens (e.g., CD13 or CD33). This isparticularly seen in cases with MLL gene rearrangements.30

These MLL rearranged B-ALL cases can also be identifiedby reactivity with antibody NG2 (which recogniseschondroitin sulfate proteoglycan neuron-glial antigen) witha specificity of 90–100%.31,32 Specific B-ALL/LBL pheno-types are also seen in association with other recurrentcytogenetic abnormalities, many of which are of diagnosticand prognostic significance (Table 4).21,33,34 Examplesinclude the pre-B-ALL phenotype with the t(1;19) andcommon ALL phenotype in BCR/ABL t(9;22).

Recent additions to antibody panels have indicated thatCD1d and CD58 may be valuable as prognostic markers orfor monitoring residual B-ALL. CD1d is expressed by 15%of B-ALL and these positive cases have been reported tohave a poorer prognosis. CD1d positivity has been detectedin infant leukaemia, leukaemias with a Pro-B phenotypeand mixed-lineage leukaemia with the MLL/AF4 generearrangement. It is unclear whether CD1d expression is anindependent poor prognostic marker, or its expression isassociated with B-ALL subgroups with poor prognosis.Detection of CD1d expression may also be useful to predictimmunotherapeutic response to glycolipids. As alpha-galactosylceramide is able to be presented via CD1d toCD1d-restricted T cells with cytotoxic potential, CD1d mayprove to be a useful therapeutic target.35

CD58, a member of the Ig superfamily, is over-expressedin B-ALL blasts when compared with normal precursor Bcells. This was first uncovered by gene expression analysis.Recent studies have suggested that CD58 expression is apowerful tool for disease monitoring and detection ofMRD;36,37 however, this has not gained widespreadacceptance. Gene expression profiling has also showndistinct patterns of gene expression in biologically distinctsubgroups of ALL. These genetic studies are likely toidentify additional novel biological groups not predicted byconventional morphological, phenotypical and cytogeneticvariables. Production of antibodies to gene products of theupregulated genes may provide further new diagnostictools, add significant prognostic information and may beable to be used to identify novel therapeutic targets.36,38

Precursor T-lymphoblastic leukaemia/lymphoma

T-ALL/LBL are malignancies derived from precursor Tcells or T lymphoblasts. They are TdT positive, expresscytoplasmic CD3 and surface CD7, and have variable

TABLE 3 Immunophenotype of B-lymphoblastic leukaemia/lymphoma(based on the EGIL Classification)22

Antibody Pro-B-ALL Common ALL Pre-B-ALL

TdT + + +Cy CD79a + + +CD19 + + +CD22 Cy+ + +CD10 2 + +CD20 2 +/2 +HLA-DR + + +CD34 + + +Cy Igm 2 2 +Sm Igm 2 2 2

ALL, acute lymphoblastic leukaemia; Cy, cytoplasmic; Sm, surfacemembrane.


expression of other T-cell-associated antigens and CD45.CD3 is lineage specific and is expressed within thecytoplasm early in T-cell development. CD7 is the earliestT-cell-associated antigen to be expressed on the cell surfacein T-cell differentiation. CD1 and CD5 are expressed later,followed by surface membrane CD3. T-ALL can beclassified therefore according to the stage of intra-thymicT-cell differentiation by antigen expression, as follows(Table 5):

1. Immature thymocyte T-ALL (early T-ALL) is derivedfrom the earliest T lymphoblast and does not expressCD1a or surface CD3. It may co-express CD34 andmyeloid-associated antigens (e.g., CD33).

2. Common thymocyte T-ALL (intermediate, thymic orcortical T-ALL) is defined by expression of CD1a, withfrequent co-expression of CD4 and CD8 antigens.Some cases may be surface CD3 positive.

3. Mature T-ALL is CD1 negative, expresses either CD4or CD8 and is surface CD3 positive.

This phenotypical subclassification is important, as thereis evidence that the immunological subtype is the mostimportant prognostic factor in T-ALL. The best disease-free survival is seen in common thymocyte T-ALL.39

Up to a quarter of T-ALL cases have aberrant myeloidantigen expression (e.g., CD13, CD33, CD117), thesignificance of which is uncertain. A small number of T-ALL cases express CD10 or cytoplasmic CD79a. BCL6 hasbeen reported to be positive in a significant number of T-LBL, suggesting that the expression of the anti-apoptoticgenes may contribute to the pathogenesis of the disease.40

Phenotypical assessment of minimal residual diseaseassessment for T-ALL can be performed on either blood orbone marrow as there is comparable dissemination of theleukaemic cells which are of thymic origin.41 TdT can beused in combination with one or more T-cell antigensknown to be expressed by the leukaemic cell (e.g., CD7 orCD2). The presence of a TdT-positive T cell in blood orbone marrow is indicative of leukaemia, as such doublepositive cells normally only reside in the thymus.

Microarray gene expression analysis of T-ALL/LBL hasprovided greater insight into the biological heterogeneity ofthe disease and revealed clinically relevant molecularsubtypes. Gene expression profiling has shown that over-expression of the HOX11 gene occurs in approximately 5–10% of childhood and 30% of adult T-ALL cases and isassociated with a favourable prognosis.42,43 Other significantgenes identified are LYL1 seen in immature thymocyte T-ALL, HOX11L2 in T-ALL of common thymocyte type, andTAL1 in mature T-ALL. The prognostic significance of thesegenomic features is yet to be determined and at presentphenotyping for the gene products is not routinely performed.

MATURE B-CELL NEOPLASMS

Mature B-cell malignancies are clonal expansions of matureB cells that are, by definition, TdT negative. They arisefrom B cells at different stages of differentiation which maybe pre- or post-antigenic stimulation in the germinal centre.As B-cell neoplasms arise from the clonal expansion of a B-cell population they express only one light chain, either Igkor Igl. Flow cytometry of fresh cells is particularly valuablefor the detection of light chain restriction or an over-whelming majority of one light chain. In some cases wherethere are residual normal B cells present, a skewed Igk orIgl ratio (i.e., Igk:Igl . 6:1 or Igk:Igl , 1:1) may beindicative of a clonal B-cell disorder. Many mature B-cellmalignancies have a characteristic immunophenotype thatcan easily be distinguished by immunohistochemistry ormultiparameter flow cytometry (Table 6). Phenotyping canalso be used to identify aberrant antigen expression by Bcells or down-regulation of immunoglobulin, both of whichmay be helpful in identifying a B-cell malignancy. Thephenotypes of the most common mature B-cell neoplasmsare presented.

Chronic lymphocytic leukaemia/Small lymphocyticlymphoma

Chronic lymphocytic leukaemia (CLL) and small lympho-cytic lymphoma (SLL), generally regarded as the blood and

TABLE 4 Phenotype/genotype correlations for paediatric precursor B-lymphoblastic leukaemia with non-random cytogenetic abnormalities21

Cytogenetic abnormality Genetic alteration Immunophenotype Frequency Prognosis

t(12;21)(p13;q22) TEL/AML1 CD10+/TdT+/CD19+/CD202/CD34+.Commonly CD33+ CD13+ (or both)

Up to 25% Good

t(1;19)(q23;p13.3) PBX1/E2A Most commonly Pre-B-ALL (CD10+/CD19+/CD20+). May have common ALL phenotype.

25% of Pre-B-ALL Poor or standardrisk

t(9;22)(q34;q11.2) BCR/ABL Common ALL phenotype (CD10+/CD19+/CD202/TdT+). CD33+ and/or CD13+

3–4% Poor

t(4;11)(q21;q23) AF4/MLL Pro-B-ALL (CD102/CD19+/CD202) and TdT2.Also CD34+, CD65+ and NG2+.

2–3% Poor

TABLE 5 Immunophenotype of T-lymphoblastic leukaemia/lymphomasubtypes

AntibodyImmature

thymocyte T-ALLCommon

thymocyte T-ALL Mature T-ALL

TdT + + +Cy CD3 + + +CD2 + + +CD7 + + +CD5 +/2 + +CD1a 2 + 2

Sm CD3 2 2/+ +CD42/ CD82 + 2 2

CD4+/CD8+ 2 + 2

CD4+ or CD8+ 2 2 +

ALL, acute lymphoblastic leukaemia; Cy, cytoplasmic; Sm, surfacemembrane.


tissue phases of the same disease, have the same phenotype,being CD5-positive B-cell disorders. The neoplastic B-cellsexpress pan-B-cell antigens with variable intensity, surfaceIg heavy chains (Igm) and show light chain restriction.Specifically, CD20, CD22 and Igm are weakly expressedwhilst CD79b is generally very weak or negative.44 CLLcells are characteristically positive for both CD5 and CD23antigens (by flow cytometry and IH). However, neitherCD5 nor CD23 alone is sufficient to diagnose CLL/SLL asthese antigens can also be detected on other B-cellmalignancies. Specifically, CD5 is expressed by mantle celllymphoma, however, CD23 is generally negative in thisdisorder. The phenotypical features (expression of CD5and CD23, absence of CD79b, weak Igm) form the basis ofthe CLL scoring system (Table 7).45,46 It is of note that thisscoring system was developed for cellular assessment byflow cytometry and is less applicable for IH. When CLL/SLL cases are analysed by IH the neoplastic cells can beshown to express CD5, CD20 and CD23 antigens; however,clonality (i.e., light chain restriction) and intensity ofexpression of Igm cannot usually be determined by IH.Up to 35% of cases of CLL have been reported to express(or lack) at least one aberrant marker and this may be

associated with more aggressive disease.47 These includelack of CD5, although this is uncommon, or CD23. Lesscommonly, CLL can express other B, T or myeloid-associated antigens, such as FMC7, CD11c, CD79b, CD2,CD7, CD10, CD13 or CD33.

Recent studies have shown that there are two subtypes ofCLL based on mutational status of Ig genes. The majorityof CLL cases are derived from a B cell that has undergonesomatic hypermutation of the Ig genes. A smaller numberare derived from pre-follicular B cells (i.e., lack somatichypermutation) and have more aggressive behaviour with apoorer prognosis. There is close (but not complete)correlation between mutational status and phenotypicalexpression of ZAP70, and to a lesser extent CD38antigen.48–50 A number of studies have now demonstratedthat immunophenotyping can provide prognostic informa-tion in CLL, based on expression of ZAP70 and CD38antigens. p53 expression is also reported to be associatedwith poor prognosis disease.

ZAP70 protein is an intracellular tyrosine kinaseessential for T-cell signalling, but not found in normal Bcells. Current data indicate that ZAP70 expression by morethan 20% CLL cells is reliably associated with a poorerprognosis. At present, flow cytometry for ZAP70 remainsto be standardised, in part because of the need formembrane permeabilisation to access this intracellularmolecule and also the difficulty in clear separation ofpositive from negative cells (Fig. 2). Few antibodies areavailable that enable assessment by IH on fixed material.

Many reports have described the poor prognosticsignificance of CD38 expression in CLL.49,51,52 CD38expression can be readily detected on B cells by multi-parameter flow cytometry on cells with the CLL phenotype(CD5, CD23 positive B cells); hence it is commonly used asa prognostic marker. Although there is debate as to thelevel of antigen expression that should be defined aspositive, with investigators using different criteria, a cut-offof more than 30% positive cells is commonly used. CD38has weak prognostic significance in univariate analysis, and

TABLE 6 Immunophenotype of mature B-cell lymphoproliferative disorders

Antibody CLL B-PLL HCL SLVL MCL FL DLBCL BL MALT

CD5 + 2/weak 2 2 + 2 2 2 2

CD10 2 2 2 2 2 + 2/+ + 2

CD19 + Strong Strong + Strong + + + +CD20 Weak Strong Strong + Strong Strong + + +FMC7 2 + + + + 2 2 2 2

CD22 Weak + Strong + + + + + +CD23 + 2 2 2 2 2 2 2 2

CD79b 2/weak + + + + + + + +Sm Igm Weak Strong Strong Strong Strong + + + +CD11c 2 2 Strong 2/+ 2 2 2 2 +/2

CD25 2 2 + 2 2 2 2 2 2

CD103 2 2 + 2 2 2 2 2 2

DBA-44 2 2 + + 2 2 2 2 2

Cyclin-D1 2 2 2 2 + 2 2 2 2

BCL2 2 2 2 2 2 + 2/+ 2 2

BCL6 2 2 2 2 2 2 + (GC) + 2

CLL, chronic lymphocytic leukaemia; PLL, polylymphocytic leukaemia; HCL, hairy cell leukaemia; SLVL, splenic lymphoma with villous lymphocytes;MCL, mantle cell lymphoma; FL, follicular lymphoma; DLBCL, diffuse large B-cell lymphoma; BL, Burkitt’s lymphoma; MALT, mucosal associatedlymphoid tissue lymphoma; GC, germinal centre type DLBCL.

TABLE 7 CLL scoring system from Moreau et al.46

Antibody CLL Score*Other B-cellneoplasms Score*

CD5 + 1 2{ 0CD23 + 1 2 0FMC7 2 1 + 0CD79b /

CD22Weak 1 Strong 0

Sm Igm Weak 1 Strong 0

*CLL usually scores 4–5, with other B-cell lymphoproliferative disordersscoring 0–2.{Except mantle cell lymphoma.CLL, chronic lymphocytic leukaemia.


does not improve predictive power of ZAP70 or Ig genemutational status in multivariate analysis. However, theantibody is readily available and flow cytometry simple toperform and detect positivity (Fig. 2). The predictive valueof CD38 expression can be increased by quantitativemeasurement of CD38 antigen density (antibody bindingcapacity) rather than as the percentage of CLL cellsexpressing the antigen. The antibody binding capacity ofCD38 closely correlates with the relative median fluores-cence intensity on flow cytometry. CD38 has the disadvan-tage of altered expression over the course of the disease.

Mutations in the p53 gene are also reported to beassociated with poor prognosis in CLL, as well as in severalother lymphoproliferative disorders (e.g., diffuse large B-cell lymphoma and mantle cell lymphoma). Mutations andover-expression of p53 are most commonly evaluated byfluorescence in situ hybridisation (FISH), but flow cyto-metry has also been studied using monoclonal antibodiesspecific for the N-terminus, Bp53 and DO-1, and thecentral region of the p53 protein, DO-11.53

The combination of ZAP70 expression, CD38 antigendensity, p53 function and the concentration of serumfactors (e.g., soluble CD23), is likely to provide extremelyaccurate prognostic information in future studies.52

Another antibody that may have prognostic value in CLLis MUM1. MUM1 (multiple myeloma oncogene 1) is aproto-oncogene that is deregulated as a result oft(6;14)(p25;q32) in multiple myeloma, and is also expressedin a variety of malignant lymphomas. The translocationjuxtaposes the Ig heavy-chain (IgH) locus to MUM1/IRF4gene, a member of the interferon regulatory factor (IRF)family. CLL with expression of MUM1/IRF4, indicatingpost-germinal centre origin, has been shown in one study tohave a more favourable clinical course.54 However, inanother it was shown that MUM1 expression was anindependent unfavourable prognostic factor in CLL/SLL.55

Therefore, more work is needed to clarify whether MUM1expression in CLL is a useful prognostic marker.

Another antibody of importance in the phenotypicalassessment of CLL is CD52. The CD52 antigen is aglycoprotein anchored on the cell membrane of mature Band T lymphocytes, monocytes and eosinophils. CD52 isthe antigenic target of the monoclonal antibody Campath-1H (alemtuzamab), which is used primarily in refractoryand relapsed CLL. Phenotypical detection of the antigen isimportant if this is to be a target for immunotherapy.

Immunophenotyping can also be used to monitorresidual disease following aggressive therapy for CLL.Approaches that have been used to detect small numbers ofresidual CLL cells include:

1. Flow cytometric quantitation of CD5, CD23 dual-positive B cells expressing the Ig light chain known tobe expressed by the neoplastic CLL clone.

2. Sensitive, quantitative flow cytometry utilising the weakCD79b expression of B cells.56 For example, multi-parameter flow cytometry for co-expression of CD19/CD5/CD20/CD79b can differentiate CLL cells fromnormal B cells and detect one CLL cell in 104 or 105.57

3. Determining CD22 and CD81 expression by CD19-gated B cells that express CD5, as there is differentialintensity of expression of these antigens compared withnormal B cells. This combination is currently being

assessed by flow cytometry for its value in detectingMRD following aggressive therapy.58

B-cell prolymphocytic leukaemia

B-cell prolymphocytic leukaemia (B-PLL) differs fromCLL by its morphology, phenotype and more aggressiveclinical course. Prolymphocytes of B-cell origin are large,round cells with abundant basophilic cytoplasm and asingle central nucleolus. In contrast to CLL, B-PLL cellshave high levels of surface Igm expression and pan-B cellmarkers (e.g., CD20, CD22), including CD79b. B-PLL isalso positive for the antibody FMC7, now known to be aweak antibody of the CD20 cluster. B-PLL cells arecommonly weakly CD5 positive and generally CD23negative. CD11c, CD25 and CD103, expressed by hairycells, are not expressed.

Hairy cell leukaemia and related disorders

Hairy cell leukaemia Hairy cell leukaemia (HCL) is adisorder of mature B cells that has a unique phenotype. Thecells strongly express pan-B-cell antigens (e.g., CD19, CD20,CD22) and surface membrane Igm. Most distinctivelythey are positive for CD11c, CD25, CD103 and HC2(Table 6).59–61 CD11c, normally expressed on neutrophilsand monocytes, is rare on B-cell disorders other than HCL.However, CD103, the human mucosal lymphocyte antigen,is the most specific marker of HCL (Table 6). CD52 isusually positive in HCL, leading to suggestions thatCampath-1H may be a useful therapeutic agent.62

Other markers that have recently been shown to be usefulin the diagnosis of HCL are CD123 (antibody to the a-chainof the human interleukin-3 receptor), TIA-1 and annexinA1.63 TIA-1 has been shown to be expressed by the majorityof HCL cases and none of 94 other B-cell lymphoproliferativedisorders tested, thereby suggesting that TIA-1 expressionmay be a useful confirmatory marker for HCL.64 By geneexpression profiling it has been shown that annexin A1(ANXA1) is upregulated in HCL. Immunohistochemicalstaining of 500 B-cell tumours with an anti-ANXA1monoclonal antibody has demonstrated that ANXA1 proteinexpression is specific for HCL. Further, ANXA1 candistinguish HCL from splenic lymphoma with villouslymphocytes (SLVL) and variant hairy cell leukaemia.65

Immunohistochemistry can be performed on paraffin-embedded biopsies, including decalcified bone marrowtrephines to phenotype HCL. To date, the most usefulantibodies for diagnosis have been the pan-B-cell markerCD20, DBA-44 and TRAcP (monoclonal antibody 9C5),an antibody that recognises tartrate-resistant acid phos-phatase.66,67 DBA-44, a monoclonal antibody that recog-nises a membrane antigen expressed by a subpopulation ofB-lymphoid cells has proved a useful marker of HCL but isnot specific. It is also positive in SLVL and therefore is nothelpful in differentiating between these two entities.Recently, a CD11c antibody that can be used in paraffin-embedded material and decalcified bone marrow trephineshas been produced (personal observation) (Fig. 3). Thisshould be of particular value in combination with CD20 forthe diagnosis and monitoring of HCL when the trephinebiopsy is the only available diagnostic material.68

Flow cytometry is particularly useful for identifying hairycells. On flow cytometry hairy cells have characteristic light


scatter properties, due to the fine cytoplasmic projectionsgiving increased side scatter, and stronger CD45 expression(mean cell fluorescence) than normal lymphocytes andmonocytes (Fig. 3).69 These unique characteristics providea sensitive way to ‘gate’ hairy cells, both at diagnosis andwhen monitoring residual disease following therapy.

Further, multi-parameter analysis enables B cells that co-express CD11c, CD25 and CD103 to be identified even whenonly present in small numbers (e.g., less than 1% of lymphoidcells). Therefore, flow cytometry is the immunophenotypicalmethod of choice for the diagnosis and monitoring of HCL,for the following reasons:

Fig. 3 CD11c expression by hairy cell leukaemia using different immunophenotyping methods. (A) Flow cytometry showing the distinctive strong CD45expression used to gate the cells and dual CD11c/CD19 positivity. (B) CD11c by immuno-alkaline phosphatase (APAAP) staining of peripheral blood. (C)CD11c positivity (immunoperoxidase stain) in formalin-fixed decalcified bone marrow trephine biopsy.


1. It has specific CD45 and side scatter properties.

2. There are limited specific antibodies of diagnosticutility that can be used by IH.

3. CD103 is the most specific marker of HCL and can beused in combination with CD11c and CD25 forsimultaneous assessment by flow cytometry.

4. Flow cytometric immunophenotyping of peripheralblood or bone marrow is capable of detecting low levelsof malignant cells, even when the patient is leukopenic.70

Hairy cell leukaemia variant HCL variant is a rare B-celldisorder which accounts for 10% of HCL cases.71,72

The morphology of the cells is intermediate betweenprolymphocytes and hairy cells, having a prominentnucleolus and cytoplasmic projections. The immuno-phenotype is similar to HCL. The cells are mature B cellsthat express pan-B-cell antigens (e.g., CD19, CD20, CD22),although more weakly than classical HCL, and CD11c. Incontrast to typical HCL, the neoplastic cells show weakexpression or are negative for CD103 and are negative forCD25, CD123 and HC2 antigens.73

Splenic marginal zone lymphoma Splenic marginal zonelymphoma is an indolent lymphoproliferative disease withmarked splenomegaly, and is frequently accompanied bycirculating villous lymphocytes. Hence it is also known assplenic lymphoma with villous lymphocytes (SLVL)and can be difficult to distinguish from HCL.74,75

Immunophenotypically, the tumour cells have a matureB-cell phenotype with moderately strong surface Igm andIgd. They express pan-B-cell antigens (e.g., CD19, CD20,CD22) and have moderate intensity CD79b. SLVL casesmay be CD11c and are usually DBA-44 positive, but arenegative for HCL-associated antigens CD25 and CD103.Other markers that are positive are CD24 and FMC7,which help to distinguish SLVL from other spleniclymphomas. In addition, SLVL are generally, but notuniversally, negative for CD5, CD10, CD23 and CD38,which are often expressed by other B-cell lymphomas, andHC2 and CD123 expressed by HCL.76 DBA-44 can be ofvalue in paraffin-embedded biopsies as its positivity inSLVL allows it to be distinguished from CLL but notHCL.66

In 1994, Matutes et al. proposed a scoring system todistinguish HCL from variant HCL and SLVL.76 Thisapproach, which has not been widely adopted, consideredthe reactivity with the four most common phenotypicalmarkers of HCL (CD11c, CD25, HC2 and CD103).Reactivity with each gives 1 point and 0 points if negative.Based on this system, 98% of HCL have scores of 3 or 4. Incontrast 88% of HCL variant and 77% of SLVL score amaximum of 2.76

Mantle cell lymphoma

The characteristic phenotype of mantle cell lymphoma(MCL) is expression of CD5 with strong pan-B-cell antigensand surface membrane Igm (Fig. 4). MCL has similarintensity of expression of CD5 to CLL but the CD20 andCD79b expression are significantly stronger. Cyclin-D1 isexpressed as a consequence of the BCL1-IgH fusionresulting from the t(11;14). At present, cyclin-D1 can only

be detected by IH; technical difficulties have resulted in therebeing no satisfactory reproducible method for detectingcyclin-D1 expression by flow cytometry. However, for thedirect detection of t(11;14)(q13;q32) FISH analysis hashigher sensitivity than IH and is the method of choice. Otheruseful markers to distinguish MCL from CLL (which is alsoCD5 positive) are the absence of CD23 and positivity withFMC7. These phenotypical differences result in a ‘CLLscore’ for MCL of below 3 in 96% of cases.77 MCL is alsonegative for CD10, enabling it to be distinguished fromfollicular lymphoma.

MCL has a relatively aggressive clinical course comparedwith other small cell B-cell lymphomas. CD5-negativeMCL cases have been reported and these may have a moreindolent clinical course.78 The blastoid variant behaves inan even more aggressive fashion. These blastoid cases arealso CD5 and cyclin-D1 positive and can express CD10.

Follicular lymphoma

The phenotype of follicular lymphoma (FL) reflects that ofnormal germinal centre B cells. The FL tumour cells arecharacterised by expression of CD10, BCL2, BCL6, pan-B-cell markers (CD19 and strong CD20), CD45RA, CD75and relatively strong surface membrane Igm. CD23 may beexpressed on a proportion of cells and Igd and Igc may bepositive. FL has tight meshworks of follicular dendriticcells which are CD21 positive, as seen on IH. Theneoplastic cells consistently lack CD5 and CD43 antigens.

CD10, BCL2 and BCL6 are the phenotypical hallmarksof germinal centre B cells from which FL originates. CD10,although characteristic of FL, is also present on Burkitt’slymphoma and some cases of diffuse large B-cell lym-phoma. Normal B-cell progenitors or haematogones, andprecursor B-ALL/LBL are also CD10 positive, but theseimmature B cells can be distinguished by their expression ofTdT and lack of surface Igm and light chain restriction.

The t(14;18)(q32;q21) chromosomal translocation, char-acteristic of FL, results in over-expression of BCL2 as thetranslocation places the BCL2 gene under the control of theIgH gene. BCL2, which is expressed intracellularly, can bedetected by IH and flow cytometry and distinguishes FLcells from their non-neoplastic counterparts in reactivegerminal centres. Approximately 10% of FL are reported tobe BCL2 negative by standard IH, and the majority ofthese cases lack the t(14;18) translocation. However, somecases with a t(14;18) have a mutation in the translocatedBCL2 gene, resulting in an amino acid substitution in theregion of the epitope recognised by the BCL2 antibody.79

Therefore, negative BCL2 by IH or flow cytometry doesnot exclude a diagnosis of FL. Paediatric FL, in contrast toadult FL, is generally BCL2 negative, suggesting thatdysregulated BCL2 expression does not play a significantpathogenetic role in most of these cases.80

Immunophenotypical detection of BCL6 is important todistinguish between FL, reactive lymphoid hyperplasia andother low-grade B-cell lymphomas.81 BCL6, a transcrip-tional repressor required for germinal centre formation, isover-expressed in the majority of FL. This expression islargely independent of chromosome 3q27 rearrange-ments.82,83 It is of note that BCL6 is over-expressed inthe majority of FL cases without t(14;18) but whichhave t(3;14)(q27;q32), implying a role for BCL6 in the


pathogenesis of these cases. Therefore, BCL6 is one of themost useful confirmatory immunophenotypical markers forthe detection of FL. However, there remains poorstandardisation for its detection by IH and it is rarelyperformed by flow cytometry.

The human germinal centre-associated lymphoma(HGAL) protein has recently been shown to be expressedin the majority of FL, when tested by immunohistochem-istry on tissue microarrays.84 HGAL protein is normallyexpressed in the cytoplasm of germinal centre lymphocytesand in lymphomas derived from germinal centre cells.Antibodies to HGAL protein are not yet being usedroutinely in diagnostic pathology but may prove to beanother useful phenotypical marker for FL.

Diffuse large B-cell lymphoma

Diffuse large B-cell lymphoma (DLBCL) includes allmalignant lymphomas characterised by the large size ofthe neoplastic cells, B-cell origin, aggressive clinical course,and the need for high dose cytotoxic chemotherapy

regimens. The phenotype of DLBCL is variable, reflectingthe heterogeneous nature of this group of tumours.

DLBCL cases express a number of pan-B-cell antigensincluding CD19, CD20, CD22 and CD79b. However, anyone of these may not be expressed in a specific case. Surfacemembrane Igm is detectable in up to 75% of cases inaddition to light chain restriction. CD10 is positive in up to50% of cases. CD5 expression is seen in approximately 10%of DLBCL, and has been shown to be associated with ahigher International Prognostic Index (IPI) score atdiagnosis, and a significantly shorter overall survival.85

However, CD23 is rarely expressed (approximately 15% ofcases). BCL2 protein is expressed in 30–50% of cases,usually in non-germinal centre type DLBCL, and generallyassociated with a poorer prognosis. CD21 expression maybe associated with improved prognosis disease, indepen-dent of IPI score.86 BCL6 protein is expressed in a highpercentage of cases of DLBCL. The BCL6 gene on 3q27 isrearranged as a result of chromosomal translocations, forexample t(3;14), in about one-third of cases of DLBCL.This disrupts the normal transcriptional regulation of the

Fig. 4 A case of mantle cell lymphoma. (A) Cytocentrifuge preparation of the lymphoma cells extracted from a lymph node. (B) Flow cytometry showingCD5 expression by the neoplastic B cells.


gene leading to constitutive over-expression of BCL6. Inaddition, about three-quarters of cases of DLBCL displaymultiple somatic mutations in the 59 non-coding region ofBCL6, which occur independently of chromosomal trans-locations and appear to be due to the IgV-associatedsomatic hypermutation process.87 Cyclin-D1 is negative inDLBCL, providing a useful marker for discriminating thisfrom blastoid variant of mantle cell lymphoma.

Gene expression profiling has demonstrated the extremegenetic heterogeneity of DLBCL. Data from cDNAmicroarray analysis have shown that DLBCL can bedivided into three prognostically distinct disease subgroupsbased on gene expression profiles:88

1. Germinal centre (GC)-like DLBCL which expressesgenes characteristic of normal germinal centre B cells(i.e., those encoding CD10 and BCL6 antigens) and hasvariable MUM1 expression.

2. Activated peripheral B-cell (ABC)-like DLBCL whichexpresses a subset of the genes that are characteristic ofplasma cells, particularly those encoding endoplasmicreticulum and golgi proteins involved in secretion.ABC-like DLBCL have lower BCL6 gene expressionthan GC-like DLBCL and express IRF-4. These casesdo not express CD10 and are usually BCL6 antigennegative and may express MUM1.

3. ‘Type 3’ which expresses CD5.

Immunohistochemical analysis of tissue microarrays hasshown some success in discriminating between the GC andnon-GC (ABC and ‘type 3’) subtypes of DLBCL.89 Theauthors suggest that GC and non-GC subtypes of DLBCLcan be determined and predict prognosis using antibodiesCD10, BCL6 and MUM1 and immunohistochemistry;however, this remains to be verified by other groups.Expression of GC-like phenotypical features (BCL6 andCD10) correlated with genotype and were reported to beassociated with better overall survival. In contrast, a non-GC phenotype, including expression of MUM1, cyclin-D2and BCL2 were associated with a poorer prognosis andworse overall survival. If these phenotypical features areconfirmed, these may be useful for determining DLBCLsubtypes and predicting survival, without the need forcDNA microarray. This could be of significant clinicalbenefit as the 5-year overall survival for the GC-likeDLBCL group is reported as 76%, compared with only 34%for the non-GC DLBCL group (p,0.001).89

A further marker that may prove useful in subclassifyingDLBCL is HGAL protein. Gene expression profiling hasshown that DLBCL cases with high levels of HGALmRNA are associated with significantly longer overallsurvival. IH staining of tissue microarrays of DLBCLshows that HGAL protein expression correlates withexpression of GC-associated proteins BCL6 and CD10.HGAL protein is also negative when MUM1 and BCL2,antigens associated with a non-GC phenotype, areexpressed. This HGAL protein expression pattern suggeststhat the antibody may have a role in identifying GC-typeDLBCL.84 It is of note that HGAL is not specific forGC-like DLBCL as it is also expressed by the majorityof follicular, Burkitt’s and mediastinal large B-celllymphomas.90

FOXP1, a forkhead transcription factor, is anothermolecule recently shown in gene expression studies to be ofvalue in identifying DLBCL subgroups. This has beenconfirmed by IH analysis on tissue microarrays.89 FOXP1is more commonly detected in non-GC type (MUM1 andBCL2 positive) than GC-like DLBCL. Its expression haspredictive value independent of the IPI subgrouping, withFOXP1-negative patients having better overall survival.91

It is likely that FOXP1 expression will become a standardpart of the phenotypical workup of DLBCL.

Other useful prognostic information in DLBCL can beobtained by assessing the infiltrating T cells in DLBCL.CD4-positive TH cells that comprise 20% or more of cellswithin the tumour are reported to be associated withprolonged overall survival and decreased likelihood ofrelapse.92 Another report showed increased overall survivalwhen T cells comprised 20% or more of tumour cells, andthere was a CD4 to CD8 ratio > 2.93

Mediastinal large B-cell lymphoma

Mediastinal large B-cell lymphoma (MLBCL) is a distinctsubtype of DLBCL seen predominantly in young females.It has a B-cell phenotype (CD20, CD79a, PAX5 positive),expresses CD30, BCL2 and usually BCL6 with variableexpression of surface Igm. MUM1/IRF4 is variably positiveand the tumour is negative for CD21 and CD10.94 Themajority of cases (approximately 70%) are also CD23positive, in contrast to other DLBCL cases which are rarelypositive.95 Microarray studies have shown that MLBCLhas a distinct gene expression profile which differs fromthat of DLBCL and shares some features with classicHodgkin lymphoma.96 IH has confirmed that the geneproducts of some of the upregulated genes can be detectedin MLBCL by IH (e.g., STAT1 and TNF receptor-associated factor-1). Preliminary data suggest that anumber of cell signalling molecules, including Syk, BLNKand PLC-gamma2 (absent from Reed-Sternberg cells) arealso expressed in the majority of MLBCL.97 Some of theseproteins (e.g., nuclear phosphorylated STAT6 and acti-vated c-Rel) can be used to distinguish MLBCL from bothDLBCL and classical Hodgkin lymphoma.98,99 However,these antibodies are not yet routinely available fordiagnostic IH analysis.

Burkitt’s lymphoma/leukaemia

Burkitt’s lymphoma is a malignancy derived from folliclecentre B cells and translocations involving MYC are aconstant feature. The malignant cells express surfacemembrane Igm, pan-B cell antigens (CD19, CD20,CD79a), CD10 and BCL6 and show light chain restriction.CD45 is more strongly expressed than on lymphoblasts ofALL/LBL and they are both CD34 and TdT negative. Theyare negative for CD23 and BCL2 and have variable MUM1expression. Occasional CD5-positive cases have beenreported. These must be distinguished from the blastoidvariant of mantle cell lymphoma, which is usually possibleby demonstrating rearrangement of MYC or IH for c-MYC.100,101

Burkitt’s lymphoma characteristically has a very high cellproliferation rate. Immunological assessment of cellproliferation with MIB1 (Ki-67) characteristically showspositivity approaching 100% of Burkitt’s lymphoma cells.


This is higher than in other follicle centre-derived tumoursand DLBCL, and can therefore be used to assist indiagnosis. This proliferation index can be performed byIH on fixed or fresh tissue, or by flow cytometry.102,103

Extranodal marginal zone lymphomas of mucosa-associatedlymphoid tissue

Extranodal marginal zone B-cell lymphomas of mucosal-associated lymphoid tissue, or MALT-type lymphomas,occur in a number of anatomical sites and are comprised ofmature B cells.104 These express pan-B-cell antigens (CD19,CD20), Igm and commonly Igd, and show light chainrestriction. The tumour shows CD21 and CD35 positivityand lacks expression of CD5, CD10 and CD23. There isno unique phenotypical marker of MALT-type lympho-mas, however, they are associated with chromosomaltranslocations t(11;18)(q21;q21), t(1;14)(p22;q32) andt(14;18)(q32;q21). High cytoplasmic expression ofMALT1 and BCL10 is seen in those MALT-lymphomacases with the t(14;18)(q32;q21).105 Despite the phenotypi-cal and morphological features of MALT-type lymphoma,the diagnosis can still be challenging as monoclonal B cellscan also be seen in Helicobacter pylori-associated gastritiswithout lymphoma.106

MATURE T-CELL LYMPHOPROLIFERATIVEDISORDERS

Mature T-cell lymphoproliferative disorders are uncom-mon, making up approximately 10% of all lymphoproli-ferative disorders, and can be very challenging diagnoses inhaematopathology. Unlike the more common B-celldisorders, in which clonality is often readily discernible bysurface immunoglobulin light chain restriction, there isno specific immunophenotypical signature that is diagnostic

of a clonal mature T-cell population. The definitivedemonstration of T-cell clonality is dependent onmolecular biological techniques using PCR or Southernblotting.

Mature T-cell disorders are TdT negative, and as such arealso referred to as post-thymic T malignancies. Antibodies toT-cell-associated antigens that are useful in establishing theT-cell origin of lymphoproliferations include CD3, CD5,CD4, CD7, CD8 and those to the T-cell receptors (TCR).CD43 and CD45RO are also used for the immunohisto-chemical determination of the T-cell nature of a prolifera-tion. Immunophenotypical criteria that are helpful in thediagnosis of T-cell neoplasms include:107,108

1. Aberrant loss of one or more pan-T-cell antigenrelative to normal T-cell populations (typically CD5or CD7).

2. T-cell subset antigen restriction (i.e., restricted CD4 orCD8) or predominance of either CD4 or CD8expressing T cells.

3. Dual-positive or dual-negative CD4 and CD8 expres-sion.

4. Loss of or reduced expression of CD45 (by flowcytometry).

5. Restricted cd T-cell receptor (rather than ab TCR)expression.

6. Restricted Vb expression. Flow cytometry using anti-bodies to the Vb repertoire antigens can be used as ascreening test for T-cell monoclonality; however, as thisis expensive it is rarely performed routinely.109,110

7. Expression of additional or aberrant antigens notexpressed by normal T cells (e.g., CD30, CD20).

In the following sections we will discuss those mature T-cell lymphoproliferative disorders with characteristic

TABLE 8 Immunophenotype of mature T-cell lymphoproliferative disorders

Antibody T-PLL T-LGL Sezary syndrome ATLL ALCL Hepatosplenic T-lymphoma

CD1 2 2 2 2 2 2

CD2 + + + + + +CD3 + + + + +/2 +CD4+/CD82 + 2 + + +/2 2

CD42/CD8+ 2 + 2 2 2 2

CD42/CD82 2 2 2 2 2/+ +CD4+/CD8+ 2/+ 2 2 2 2 2

CD5 + + + + 2/+ +CD7 ++ + 2 2 2 +CD16 2 +/2 2 2 2 2

CD25 2 2 2 + 2 2

CD26 + 2/+ 2 U U U

CD30 2 2 2 2/+ + 2

CD45 RO + + + + + +CD56 2 2 2 2 2 2

CD57 2 +/2 2 2 2 2

HLA-DR 2 2 2 + 2 2

ab TCR + + + + + 2

cd TCR 2 2 2 2 2 +TdT 2 2 2 2 2 2

T-PLL, T-cell prolymphocytic leukaemia; T-LGL, T-cell large granular lymphocytic leukaemia; ATLL, adult T-cell leukaemia/lymphoma; ALCL, anaplasticlarge cell lymphoma; U, not known.


phenotypes (Table 8). The large group of peripheral T-celllymphomas, unspecified, which have variable expression ofpan-T-cell antigens, will not be discussed.

T-cell prolymphocytic leukaemia

T-cell prolymphocytic leukaemia (T-PLL) is characterisedby small mature T lymphocytes that express pan-T-cellantigens CD2, CD3 and CD5 and have characteristicallystrong expression of CD7 (Fig. 5). The majority (approxi-mately 70%) of cases express CD4, whilst dual CD4 andCD8 expression (20% of cases), or CD8 only cases (10%) areoccasionally seen.111 The vast majority of T-PLL patientsexpress membrane ab TCR, however, case reports of cd T-PLL exist.112 T-PLL cells do not express HLA-DR or CD25antigens. CD52 is expressed, and good remissions have beendocumented with CD52 immunotherapy with Campath-1H.113 Immunostaining for the oncoprotein T-cell leukemia-1 (TCL-1) may aid diagnosis in rare cases of extra-medullaryT-PLL, where it is positive in 60–70% of cases.114,115

T-cell large granular lymphocytic leukaemia

A lymphocytosis of large granular lymphocytes can beof T-cell or natural killer (NK)-cell origin. T-cell largegranular lymphocytic leukaemia (T-LGL) is a clonaldisorder of mature post-thymic T-cells that expressCD3 and other pan-T-cell antigens (CD2, CD5, CD7)and CD8. Rare variant forms may be CD4-positive,co-express CD4 and CD8 antigens or express the cdT-cell receptor. They commonly express NK-cell associatedantigens CD16 and CD57, but rarely CD56. The T-LGLcells are TIA-1 (a cytotoxic granule-associated protein)and granzyme positive. CD26, a T-cell activation antigenand binding protein for adenine deaminase (ADA), hasbeen shown in flow cytometric analyses to identify caseswith more aggressive clinical behaviour and shortersurvival.116 Therefore, CD26 can be applied to identifythese poor prognostic T-LGL and to identify cases whichmay not respond to deoxycoformicin, a potent inhibitorof ADA.117 Research phenotyping studies and gene

Fig. 5 T-cell prolymphocytic leukaemia. (A) Strong CD7 expression by APPAP immunocytochemical staining. (B) Dual CD4 and CD8 expression by theneoplastic cells using flow cytometry.


expression analyses have demonstrated T-LGL leukaemiasto be derived from activated effector cytotoxic Tlymphocytes.118,119

An aggressive variant of T-LGL has been describedwhich presents with fever and hepatosplenomegaly and israpidly fatal. These cases are CD2 and CD3 positive,usually express HLA-DR and have aberrant CD56 expres-sion. CD5, CD7 and CD8 are variably expressed and CD16is negative. These cases are usually neoplasms of ab T cells.Other T-LGL leukaemia variants have been reported to bederived from cd T-LGL.

Sezary syndrome and mycosis fungoides

Sezary syndrome and mycosis fungoides are mature post-thymic T-cell malignancies with cutaneous involvement andwhich can have leukaemic spread. The malignant cellsexpress the pan-T-cell-associated antigens CD2, CD3, CD5and CD45RO and are usually of helper (CD4) phenotype.The majority of cases are negative for CD7, CD8 andCD25. Aberrant T-cell phenotypes are common; forexample, CD3 and CD4 expression levels can be lowerthan on normal T cells. This provides a useful means for

monitoring disease, especially leukaemic involvement.Sezary cells usually lack CD26, which is expressed by themajority of normal T cells.120 The phenotype CD4+/CD72/CD262 seen in Sezary cells has been successfully used todetect and monitor minimal circulating disease. Thespecificity of this phenotype can be further increased bydetecting a restricted TCR Vb repertoire.121

Adult T-cell leukaemia/lymphoma

Adult T-cell leukaemia/lymphoma (ATLL) is an aggressiveperipheral T-cell leukaemia/lymphoma associated withclonal HTLV-1 integration. Characteristically, the tumourcells have an activated T-helper phenotype expressing pan-T-cell-associated antigens (CD2, CD3 and CD5), CD4,CD25, CD95 (FAS) and HLA-DR. They are usuallynegative for CD7, CD8 and CD56 antigens and forcytotoxic-associated molecules granzyme and TIA-1.122

Variant ATLL phenotypes (e.g., CD42/CD8+ or dualCD4+/CD8+) have been described, as have rare CD30positive cases with anaplastic large cell lymphomamorphological features. These are ALK negative and donot express TIA-1 or granzyme B.123

Fig. 6 Anaplastic large cell lymphoma. (A) Anaplastic large cell lymphoma in a bone marrow aspirate (Romanowsky stain). (B) Bone marrow trephine(H&E stain). (C) Lymphoma cells are CD3 positive and (D) CD30 positive (immunoperoxidase stain).


Anaplastic large cell lymphoma

Anaplastic large cell lymphoma (ALCL) is a peripheral T-cell lymphoma that expresses pan-T-cell antigens (e.g.,CD3), CD30 antigen, anaplastic large cell lymphomakinase (ALK), epithelial membrane antigen, CD45, andis frequently CD4 positive (Fig. 6). Most cases expresscytotoxic granule-associated proteins TIA-1, granzymeB and perforin, but are CD8 negative. CD15, expressedby Hodgkin lymphoma, is negative.124 ALK-negative casesare also seen, usually in older patients, and theseare associated with poorer prognosis. In contrast, ALK-positive ALCL is the most common paediatric matureT-cell malignancy, and usually has a good prognosis.ALCL is most commonly diagnosed on tissue biopsies,including bone marrow trephine biopsies, with phenotyping

performed by IH (Fig. 6). Flow cytometry can give falsenegative results as the neoplastic cells may comprise only asmall proportion of the population analysed.

Hepatosplenic T-cell lymphoma

Hepatosplenic T-cell lymphoma is a rare clonal T-cellproliferation derived from cytotoxic T cells and usually ofcd T-cell type. It involves the sinusoids of the liver andspleen and can have blood involvement. Phenotypically, theneoplastic cells, being of T-cell origin, express CD3, CD43and CD45RO. Unlike the majority of other T-celllymphoproliferative disorders, the cd TCR receptor isusually expressed and not the ab T-cell receptor(Fig. 7).125,126 Characteristically, the neoplastic cells lackboth CD4 and CD8 antigens, although in approximately

Fig. 7 Hepatosplenic T-cell lymphoma of cd T-cell origin. (A) Morphology of the cells in peripheral blood (Romanowsky stain). Dual-colour flow cytometryin (B) cd TCR/CD3 and (C) ab TCR/CD3 shows the neoplastic cells are CD3 positive and express the cd T-cell receptor.


15% of cases the cells express CD8. Virtually all cases arepositive for TIA-1 but approximately 50% lack perforinand granzyme B, indicating a non-activated cytolytic T-cellphenotype. The neoplastic cells are commonly CD2 andCD16 positive and may express CD56.

NATURAL KILLER CELLLYMPHOPROLIFERATIVE DISORDERS

Lymphoproliferative disorders derived from natural killer(NK) cells are uncommon. These may have an acute clinicalpresentation with systemic symptoms or be a chronicdisorder and have minimal symptoms. The best charac-terised NK-cell disorders are the nasal and nasal-type NKlymphomas, the indolent NK-LGL leukaemias and aggres-sive NK-cell leukaemias. There is some overlap in theantigenic profile with T cells, as NK cells containcytoplasmic CD3 epsilon (CD3e) chain and can expressCD2, CD7 and CD8 antigens. However, they are negativefor surface membrane CD3 and CD5, do not express CD4or CD8, and are usually CD57 negative. The most commonNK lymphoproliferation is NK-LGL lymphocytosis. Inthis disorder the cells have the typical NK-cell phenotypeexpressing CD16, and commonly CD8, whilst lacking CD3.

Aggressive NK-cell leukaemia and extranodal NK/T-celllymphoma, nasal type

Aggressive natural killer cell leukaemia (ANKL), a rareform of LGL leukaemia with systemic proliferation of NKcells, and nasal type NK/T cell lymphoma have identicalphenotypes. The neoplastic cells express NK-cell antigensCD16 and CD56, but CD57 is usually negative.127,128

‘Surrogate’ NK-cell markers CD2 and CD7 are usuallypositive. Both CD4 and CD8 positive cases have beenreported. Cytoplasmic CD3e is expressed but not surfacemembrane CD3 or the T-cell receptor. CD11b and HLA-DR are commonly positive and other myeloid-associatedantigens may be expressed (e.g., CD13, CD33).

CONCLUSION

The lymphoproliferative disorders are a heterogeneousgroup of disorders and phenotyping has an essential role inboth the diagnosis and classification of these diseases. Thisis specifically to determine the lymphoid cell lineage (B, Tor NK cell) of the tumour, the stage of cell differentiation,clonality (especially of B cells) and to identify thosedisorders with characteristic disease-associated phenotypes.Recently the role of phenotyping has expanded to includethe demonstration of markers of prognostic significance,detect molecules which may be used as targets forimmunotherapy and to identify antigen combinations thatmay be used to monitor therapy. Results of gene expressionanalyses have identified genetic signatures for specificsubsets of diseases. The production of monoclonal anti-bodies to the products of upregulated genes is increasingthe repertoire of antibodies available. These are alreadyhaving impact in the phenotypical identification of prog-nostically distinct subgroups of these disorders. This shouldfurther enhance the capability of phenotyping in future.Optimisation of immunophenotyping methods is also

occurring. Most significant are the increasing range ofantibodies that can be applied to routinely processed tissuebiopsies, the use of flow cytometry for analysing cellsextracted from fresh tissue and the ability to use tissuearrays for the rapid assessment of new antibodies.

ACKNOWLEDGEMENTS We are grateful to Jane Smith forher excellent secretarial assistance. Thanks to DavidBloxham and Janine Davies for providing flow cytometryimages.

Address for correspondence: Dr W. N. Erber, Haematology Department

(Box 234), Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ,

United Kingdom. E-mail: [email protected]

References1. Greaves M, Janossy G. Patterns of gene expression and the cellular

origins of human leukaemias. Biochem Biophys Acta 1978; 516:193–230.

2. Nakane P, Pierce G. Enzyme-labelled antibodies: preparation andapplication for the localisation of antigens. J Histochem Cytochem1966; 14: 929–31.

3. Avrameas S. Coupling of enzymes to proteins with glutaraldehyde.Use of the conjugates for the detection of antigens and antibodies.Immunochemistry 1969; 6: 43–52.

4. Cordell JL, Falini B, Erber WN, et al. Immunoenzymatic labeling ofmonoclonal antibodies using immune complexes of alkaline phos-phatase and monoclonal anti-alkaline phosphatase (APAAP com-plexes). J Histochem Cytochem 1984; 32: 219–29.

5. Zeppa P, Marino G, Troncone G, et al. Fine-needle cytology and flowcytometry immunophenotyping and subclassification of non-Hodgkinlymphoma: a critical review of 307 cases with technical suggestions.Cancer 2004; 102: 55–65.

6. Nasuti JF, Yu G, Boudousquie A, et al. Diagnostic value of lymphnode fine needle aspiration cytology: an institutional experience of 387cases observed over a 5-year period. Cytopathology 2000; 11: 18–31.

7. Ravoet C, Demartin S, Gerard R, et al. Contribution of flowcytometry to the diagnosis of malignant and non malignantconditions in lymph node biopsies. Leuk Lymphoma 2004; 45:1587–93.

8. Mazur G, Halon A, Wrobel T, et al. Contribution of flow cytometricimmunophenotyping and bone marrow trephine biopsy in thedetection of lymphoid bone marrow infiltration in non-Hodgkinlymphomas. Neoplasma 2004; 51: 159–63.

9. Perea G, Altes A, Bellido M, et al. Clinical utility of bone marrowflow cytometry in B-cell non-Hodgkin lymphomas (B-NHL).Histopathology 2004; 45: 268–74.

10. Stacchini A, Demurtas A, Godio L, et al. Flow cytometry in the bonemarrow staging of mature B-cell neoplasms. Cytometry B Clin Cytom2003; 54: 10–8.

11. Merz H, Malisius R, Mannweiler S, et al. ImmunoMax. A maximizedimmunohistochemical method for the retrieval and enhancement ofhidden antigens. Lab Invest 1995; 73: 149–56.

12. Bobrow MN, Harris TD, Shaughnessy KJ, et al. Catalyzed reporterdeposition, a novel method of signal amplification. Application toimmunoassays. J Immunol Methods 1989; 125: 279–85.

13. Bobrow MN, Shaughnessy KJ, Litt GJ. Catalyzed reporterdeposition, a novel method of signal amplification. II. Applica-tion to membrane immunoassays. J Immunol Methods 1991; 137:103–12.

14. Key ME, Phillips T. Catalysed signal amplification (CSA): Anenhanced immunohistochemical staining method based on biotinyltyramide. J NIH Res 1996; 8: 72–3.

15. Erber WN, Gibbs TA, Ivey JG. Antigen retrieval by microwave ovenheating for immunohistochemical analysis of bone marrow trephinebiopsies. Pathology 1996; 28: 45–50.

16. Bankfalvi A, Navabi H, Bier B, et al. Wet autoclave pretreatment forantigen retrieval in diagnostic immunohistochemistry. J Pathol 1994;174: 223–8.

17. Miller K, Auld J, Jesseb E, et al. Antigen unmasking in formalin-fixedroutinely processed paraffin wax-embedded sections by pressurecooking: a comparision with microwave oven heating and traditionalmethods. Adv Anat Pathol 1995; 2: 60–4.

18. Ko IK, Kato K, Iwata H. Antibody microarray for correlating cellphenotype with surface marker. Biomaterials 2005; 26: 687–96.


19. Belov L, de la Vega O, dos Remedios CG, et al. Immunophenotypingof leukemias using a cluster of differentiation antibody microarray.Cancer Res 2001; 61: 4483–9.

20. Falini B, Mason DY. Proteins encoded by genes involved inchromosomal alterations in lymphoma and leukemia: clinical valueof their detection by immunocytochemistry. Blood 2002; 99: 409–26.

21. Jaffe ES, Harris NL, Stein H, et al., editors, Pathology and Genetics ofTumours of Haematopoietic and Lymphoid Tissues. Lyon: IARC Press,2001.

22. Bene MC, Castoldi G, Knapp W, et al. Proposals for theimmunological classification of acute leukemias. European Groupfor the Immunological Characterization of Leukemias (EGIL).Leukemia 1995; 9: 1783–6.

23. Campana D, Coustan-Smith E. Minimal residual disease studies byflow cytometry in acute leukemia. Acta Haematol 2004; 112: 8–15.

24. Neale GA, Coustan-Smith E, Stow P, et al. Comparative analysis offlow cytometry and polymerase chain reaction for the detection ofminimal residual disease in childhood acute lymphoblastic leukemia.Leukemia 2004; 18: 934–8.

25. Kerst G, Kreyenberg H, Roth C, et al. Concurrent detection ofminimal residual disease (MRD) in childhood acute lymphoblasticleukaemia by flow cytometry and real-time PCR. Br J Haematol 2005;128: 774–82.

26. Nemecek ER, Matthews DC. Antibody-based therapy of humanleukemia. Curr Opin Hematol 2002; 9: 316–21.

27. Szatrowski TP, Dodge RK, Reynolds C, et al. Lineage specifictreatment of adult patients with acute lymphoblastic leukemia in firstremission with anti-B4-blocked ricin or high-dose cytarabine: Cancerand Leukemia Group B Study 9311. Cancer 2003; 97: 1471–80.

28. Jandula BM, Nomdedeu J, Marin P, et al. Rituximab can be useful astreatment for minimal residual disease in bcr-abl-positive acutelymphoblastic leukemia. Bone Marrow Transplant 2001; 27: 225–7.

29. Gokbuget N, Hoelzer D. Treatment with monoclonal antibodies inacute lymphoblastic leukemia: current knowledge and future pro-spects. Ann Hematol 2004; 83: 201–5.

30. Szczepanski T, van der Velden VH, van Dongen JJ. Classificationsystems for acute and chronic leukaemias. Best Pract Res ClinHaematol 2003; 16: 561–82.

31. Smith FO, Rauch C, Williams DE, et al. The human homologue ofrat NG2, a chondroitin sulfate proteoglycan, is not expressed on thecell surface of normal hematopoietic cells but is expressed by acutemyeloid leukemia blasts from poor-prognosis patients with abnorm-alities of chromosome band 11q23. Blood 1996; 87: 1123–33.

32. Schwartz S, Rieder H, Schlager B, et al. Expression of the humanhomologue of rat NG2 in adult acute lymphoblastic leukemia: closeassociation with MLL rearrangement and a CD10(2)/CD24(2)/CD65s(+)/CD15(+) B-cell phenotype. Leukemia 2003; 17: 1589–95.

33. Czuczman MS, Dodge RK, Stewart CC, et al. Value of immunophe-notype in intensively treated adult acute lymphoblastic leukemia:cancer and leukemia Group B study 8364. Blood 1999; 93: 3931–9.

34. Hrusak O, Porwit-MacDonald A. Antigen expression patternsreflecting genotype of acute leukemias. Leukemia 2002; 16: 1233–58.

35. Fais F, Tenca C, Cimino G, et al. CD1d expression on B-precursoracute lymphoblastic leukemia subsets with poor prognosis. Leukemia2005; 19: 551–6.

36. Chen JS, Coustan-Smith E, Suzuki T, et al. Identification of novelmarkers for monitoring minimal residual disease in acute lympho-blastic leukemia. Blood 2001; 97: 2115–20.

37. Veltroni M, De Zen L, Sanzari MC, et al. Expression of CD58 innormal, regenerating and leukemic bone marrow B cells: implicationsfor the detection of minimal residual disease in acute lymphocyticleukemia. Haematologica 2003; 88: 1245–52.

38. Yeoh EJ, Ross ME, Shurtleff SA, et al. Classification, subtypediscovery, and prediction of outcome in pediatric acute lymphoblasticleukemia by gene expression profiling. Cancer Cell 2002; 1: 133–43.

39. Hoelzer D, Gokbuget N, Ottmann O, et al. Acute lymphoblasticleukemia. Hematology (Am Soc Hematol Educ Program) 2002:162–92.

40. Hyjek E, Chadburn A, Liu YF, et al. BCL6 protein is expressed inprecursor T-cell lymphoblastic lymphoma and in prenatal andpostnatal thymus. Blood 2001; 97: 270–6.

41. van der Velden VH, Jacobs DC, Wijkhuijs AJ, et al. Minimal residualdisease levels in bone marrow and peripheral blood are comparable inchildren with T cell acute lymphoblastic leukemia (ALL), but not inprecursor-B-ALL. Leukemia 2002; 16: 1432–6.

42. Pui CH, Schrappe M, Ribeiro RC, et al. Childhood and adolescentlymphoid and myeloid leukemia. Hematology (Am Soc Hematol EducProgram) 2004: 118–45.

43. Ferrando AA, Neuberg DS, Dodge RK, et al. Prognostic importanceof TLX1 (HOX11) oncogene expression in adults with T-cell acutelymphoblastic leukaemia. Lancet 2004; 363: 535–6.

44. Oscier D, Fegan C, Hillmen P, et al. Guidelines on the diagnosis andmanagement of chronic lymphocytic leukaemia. Br J Haematol 2004;125: 294–317.

45. Matutes E, Owusu-Ankomah K, Morilla R, et al. The immunologicalprofile of B-cell disorders and proposal of a scoring system for thediagnosis of CLL. Leukemia 1994; 8: 1640–5.

46. Moreau EJ, Matutes E, A’Hern RP, et al. Improvement of thechronic lymphocytic leukemia scoring system with the monoclonalantibody SN8 (CD79b). Am J Clin Pathol 1997; 108: 378–82.

47. Kampalath B, Barcos MP, Stewart C. Phenotypic heterogeneity of Bcells in patients with chronic lymphocytic leukemia/small lymphocyticlymphoma. Am J Clin Pathol 2003; 119: 824–32.

48. Chen L, Widhopf G, Huynh L, et al. Expression of ZAP-70 isassociated with increased B-cell receptor signaling in chroniclymphocytic leukemia. Blood 2002; 100: 4609–14.

49. Damle RN, Wasil T, Fais F, et al. Ig V gene mutation status andCD38 expression as novel prognostic indicators in chronic lympho-cytic leukemia. Blood 1999; 94: 1840–7.

50. Crespo M, Bosch F, Villamor N, et al. ZAP-70 expression as asurrogate for immunoglobulin-variable-region mutations in chroniclymphocytic leukemia. N Engl J Med 2003; 348: 1764–75.

51. Domingo-Domenech E, Domingo-Claros A, Gonzalez-Barca E, et al.CD38 expression in B-chronic lymphocytic leukemia: association withclinical presentation and outcome in 155 patients. Haematologica2002; 87: 1021–7.

52. Mainou-Fowler T, Dignum HM, Proctor SJ, et al. The prognosticvalue of CD38 expression and its quantification in B cell chroniclymphocytic leukemia (B-CLL). Leuk Lymphoma 2004; 45: 455–62.

53. Carter A, Lin K, Sherrington PD, et al. Detection of p53 dysfunctionby flow cytometry in chronic lymphocytic leukaemia. Br J Haematol2004; 127: 425–8.

54. Chang CC, Lorek J, Sabath DE, et al. Expression of MUM1/IRF4correlates with clinical outcome in patients with B-cell chroniclymphocytic leukemia. Blood 2002; 100: 4671–5.

55. Ito M, Iida S, Inagaki H, et al. MUM1/IRF4 expression is anunfavorable prognostic factor in B-cell chronic lymphocytic leukemia(CLL)/small lymphocytic lymphoma (SLL). Jpn J Cancer Res 2002;93: 685–94.

56. Garcia Vela J, Delgado I, Benito L, et al. CD79b expression in B cellchronic lymphocytic leukemia: its implication for minimal residualdisease detection. Leukemia 1999; 13: 1501–5.

57. Rawstron AC, Kennedy B, Evans PA, et al. Quantitation of minimaldisease levels in chronic lymphocytic leukemia using a sensitive flowcytometric assay improves the prediction of outcome and can be usedto optimize therapy. Blood 2001; 98: 29–35.

58. Moreton P, Kennedy B, Lucas G, et al. Eradication of minimalresidual disease in B-cell chronic lymphocytic leukemia afteralemtuzumab therapy is associated with prolonged survival. J ClinOncol 2005; 23: 2971–9.

59. Falini B, Pulford K, Erber WN, et al. Use of a panel of monoclonalantibodies for the diagnosis of hairy cell leukaemia. An immunocy-tochemical study of 36 cases. Histopathology 1986; 10: 671–87.

60. Robbins BA, Ellison DJ, Spinosa JC, et al. Diagnostic application oftwo-color flow cytometry in 161 cases of hairy cell leukemia. Blood1993; 82: 1277–87.

61. Matutes E, Morilla R, Owusu-Ankomah K, et al. The immunophe-notype of hairy cell leukemia (HCL). Proposal for a scoring system todistinguish HCL from B-cell disorders with hairy or villouslymphocytes. Leuk Lymphoma 1994; 14 (Suppl 1): 57–61.

62. Quigley MM, Bethel KJ, Sharpe RW, et al. CD52 expression in hairycell leukemia. Am J Hematol 2003; 74: 227–30.

63. Del Giudice I, Matutes E, Morilla R, et al. The diagnostic value ofCD123 in B-cell disorders with hairy or villous lymphocytes.Haematologica 2004; 89: 303–8.

64. Mori N, Murakami YI, Shimada S, et al. TIA-1 expression in hairycell leukemia. Mod Pathol 2004; 17: 840–6.

65. Falini B, Tiacci E, Liso A, et al. Simple diagnostic assay for hairy cellleukaemia by immunocytochemical detection of annexin A1(ANXA1). Lancet 2004; 363: 1869–70.

66. Salomon-Nguyen F, Valensi F, Troussard X, et al. The value of themonoclonal antibody, DBA44, in the diagnosis of B-lymphoiddisorders. Leuk Res 1996; 20: 909–13.

67. Janckila AJ, Cardwell EM, Yam LT, et al. Hairy cell identification byimmunohistochemistry of tartrate-resistant acid phosphatase. Blood1995; 85: 2839–44.

68. Tallman MS, Hakimian D, Kopecky KJ, et al. Minimal residualdisease in patients with hairy cell leukemia in complete remissiontreated with 2-chlorodeoxyadenosine or 2-deoxycoformycin andprediction of early relapse. Clin Cancer Res 1999; 5: 1665–70.

69. Tytherleigh L, Taparia M, Leahy MF. Detection of hairy cellleukaemia in blood and bone marrow using multidimensional flow


cytometry with CD45-PECy5 and SS gating. Clin Lab Haematol 2001;23: 385–90.

70. Cornfield DB, Mitchell Nelson DM, Rimsza LM, et al. The diagnosisof hairy cell leukemia can be established by flow cytometric analysisof peripheral blood, even in patients with low levels of circulatingmalignant cells. Am J Hematol 2001; 67: 223–6.

71. Matutes E, Wotherspoon A, Brito-Babapulle V, et al. The naturalhistory and clinico-pathological features of the variant form of hairycell leukemia. Leukemia 2001; 15: 184–6.

72. Matutes E, Wotherspoon A, Catovsky D. The variant form of hairy-cell leukaemia. Best Pract Res Clin Haematol 2003; 16: 41–56.

73. Cessna MH, Hartung L, Tripp S, et al. Hairy cell leukemia variant:fact or fiction. Am J Clin Pathol 2005; 123: 132–8.

74. Troussard X, Valensi F, Duchayne E, et al. Splenic lymphoma withvillous lymphocytes: clinical presentation, biology and prognosticfactors in a series of 100 patients. Groupe Francais d’HematologieCellulaire (GFHC). Br J Haematol 1996; 93: 731–6.

75. Dogan A, Isaacson PG. Splenic marginal zone lymphoma. SeminDiagn Pathol 2003; 20: 121–7.

76. Matutes E, Morilla R, Owusu-Ankomah K, et al. The immunophe-notype of splenic lymphoma with villous lymphocytes and itsrelevance to the differential diagnosis with other B-cell disorders.Blood 1994; 83: 1558–62.

77. Matutes E, Parry-Jones N, Brito-Babapulle V, et al. The leukemicpresentation of mantle-cell lymphoma: disease features and prognos-tic factors in 58 patients. Leuk Lymphoma 2004; 45: 2007–15.

78. Kaptain S, Zuckerberg LR, Ferry JA. BCL1 cyclin D1+CD5- mantlecell lymphoma. Mod Pathol 1998; 11: 133a.

79. Schraders M, de Jong D, Kluin P, et al. Lack of BCL2 expression infollicular lymphoma may be caused by mutations in the BCL2 gene orby absence of the t(14;18) translocation. J Pathol 2005; 205: 329–35.

80. Lorsbach RB, Shay-Seymore D, Moore J, et al. Clinicopathologicanalysis of follicular lymphoma occurring in children. Blood 2002; 99:1959–64.

81. Dogan A, Bagdi E, Munson P, et al. CD10 and BCL6 expression inparaffin sections of normal lymphoid tissue and B-cell lymphomas.Am J Surg Pathol 2000; 24: 846–52.

82. Flenghi L, Ye BH, Fizzotti M, et al. A specific monoclonal antibody(PG-B6) detects expression of the BCL6 protein in germinal center Bcells. Am J Pathol 1995; 147: 405–11.

83. Jardin F, Buchonnet G, Parmentier F, et al. Follicle center lymphomais associated with significantly elevated levels of BCL6 expressionamong lymphoma subtypes, independent of chromosome 3q27rearrangements. Leukemia 2002; 16: 2318–25.

84. Natkunam Y, Lossos IS, Taidi B, et al. Expression of the humangerminal center-associated lymphoma (HGAL) protein, a new markerof germinal center B-cell derivation. Blood 2005; 105: 3979–86.

85. Yamaguchi M, Seto M, Okamoto M, et al. De novo CD5+ diffuselarge B-cell lymphoma: a clinicopathologic study of 109 patients.Blood 2002; 99: 815–21.

86. Ogawa S, Yamaguchi M, Oka K, et al. CD21S antigen expression intumour cells of diffuse large B-cell lymphomas is an independentprognostic factor indicating better overall survival. Br J Haematol2004; 125: 180–6.

87. Pasqualucci L, Bereschenko O, Niu H, et al. Molecular pathogenesisof non-Hodgkin lymphoma: the role of BCL6. Leuk Lymphoma 2003;44 (Suppl 3): S5–12.

88. Alizadeh AA, Eisen MB, Davis RE, et al. Distinct types of diffuselarge B-cell lymphoma identified by gene expression profiling. Nature2000; 403: 503–11.

89. Hans CP, Weisenburger DD, Greiner TC, et al. Confirmation of themolecular classification of diffuse large B-cell lymphoma by immuno-histochemistry using a tissue microarray. Blood 2004; 103: 275–82.

90. de Leval L, Harris NL. Variability in immunophenotype in diffuselarge B-cell lymphoma and its clinical relevance. Histopathology 2003;43: 509–28.

91. Banham AH, Connors JM, Brown PJ, et al. Expression of the FOXP1transcription factor is strongly associated with inferior survival inpatients with diffuse large B-cell lymphoma. Clin Cancer Res 2005; 11:1065–72.

92. Ansell SM, Stenson M, Habermann TM, et al. CD4+ T-cell immuneresponse to large B-cell non-Hodgkin lymphoma predicts patientoutcome. J Clin Oncol 2001; 19: 720–6.

93. Xu Y, Kroft SH, McKenna RW, et al. Prognostic significance oftumour-infiltrating T lymphocytes and T-cell subsets in de novodiffuse large B-cell lymphoma: a multiparameter flow cytometrystudy. Br J Haematol 2001; 112: 945–9.

94. Pileri SA, Gaidano G, Zinzani PL, et al. Primary mediastinal B-celllymphoma: high frequency of BCL6 mutations and consistentexpression of the transcription factors OCT-2, BOB.1, and PU.1 inthe absence of immunoglobulins. Am J Pathol 2003; 162: 243–53.

95. Calaminici M, Piper K, Lee AM, et al. CD23 expression in mediastinallarge B-cell lymphomas. Histopathology 2004; 45: 619–24.

96. Savage KJ, Monti S, Kutok JL, et al. The molecular signature ofmediastinal large B-cell lymphoma differs from that of other diffuselarge B-cell lymphomas and shares features with classical Hodgkinlymphoma. Blood 2003; 102: 3871–9.

97. Marafioti T, Pozzobon M, Hansmann ML, et al. Expression patternof intracellular leukocyte-associated proteins in primary mediastinalB cell lymphoma. Leukemia 2005; 19: 856–61.

98. Guiter C, Dusanter-Fourt I, Copie-Bergman C, et al. ConstitutiveSTAT6 activation in primary mediastinal large B-cell lymphoma.Blood 2004; 104: 543–9.

99. Rodig SJ, Savage KJ, Nguyen V, et al. TRAF1 expression and c-Relactivation are useful adjuncts in distinguishing classical Hodgkinlymphoma from a subset of morphologically or immunophe-notypically similar lymphomas. Am J Surg Pathol 2005; 29: 196–203.

100. Lin CW, O’Brien S, Faber J, et al. De novo CD5+ Burkitt lymphoma/leukemia. Am J Clin Pathol 1999; 112: 828–35.

101. Frost M, Newell J, Lones MA, et al. Comparative immunohisto-chemical analysis of pediatric Burkitt lymphoma and diffuse large B-cell lymphoma. Am J Clin Pathol 2004; 121: 384–92.

102. Brown DC, Gatter KC, Mason DY. Proliferation in non-Hodgkinlymphoma: a comparison of Ki-67 staining on fine needle aspirationand cryostat sections. J Clin Pathol 1990; 43: 325–8.

103. Pich A, Ponti R, Valente G, et al. MIB-1, Ki67, and PCNA scores andDNA flow cytometry in intermediate grade malignant lymphomas. JClin Pathol 1994; 47: 18–22.

104. Isaacson PG. Update on MALT lymphomas. Best Pract Res ClinHaematol 2005; 18: 57–68.

105. Ye H, Gong L, Liu H, et al. MALT lymphoma witht(14;18)(q32;q21)/IGH-MALT1 is characterized by strong cytoplas-mic MALT1 and BCL10 expression. J Pathol 2005; 205: 293–301.

106. El-Zimaity HM, El-Zaatari FA, Dore MP, et al. The differentialdiagnosis of early gastric mucosa-associated lymphoma: polymerasechain reaction and paraffin section immunophenotyping. Mod Pathol1999; 12: 885–93.

107. Jamal S, Picker LJ, Aquino DB, et al. Immunophenotypic analysis ofperipheral T-cell neoplasms. A multiparameter flow cytometricapproach. Am J Clin Pathol 2001; 116: 512–26.

108. Gorczyca W, Weisberger J, Liu Z, et al. An approach to diagnosis ofT-cell lymphoproliferative disorders by flow cytometry. Cytometry2002; 50: 177–90.

109. Langerak AW, van Den Beemd R, Wolvers-Tettero IL, et al.Molecular and flow cytometric analysis of the Vbeta repertoire forclonality assessment in mature TCRalphabeta T-cell proliferations.Blood 2001; 98: 165–73.

110. Morice WG, Kimlinger T, Katzmann JA, et al. Flow cytometricassessment of TCR-Vbeta expression in the evaluation of peripheralblood involvement by T-cell lymphoproliferative disorders: a com-parison with conventional T-cell immunophenotyping and moleculargenetic techniques. Am J Clin Pathol 2004; 121: 373–83.

111. Matutes E. Chronic T-cell lymphoproliferative disorders. Rev ClinExp Hematol 2002; 6: 401–20.

112. Toyota S, Nakamura N, Dan K. Small cell variant of T-cellprolymphocytic leukemia with a gammadelta immunophenotype.Int J Hematol 2005; 81: 66–8.

113. Pawson R, Dyer MJ, Barge R, et al. Treatment of T-cellprolymphocytic leukemia with human CD52 antibody. J Clin Oncol1997; 15: 2667–72.

114. Valbuena JR, Herling M, Admirand JH, et al. T-cell prolymphocyticleukemia involving extramedullary sites. Am J Clin Pathol 2005; 123:456–64.

115. Herling M, Khoury JD, Washington LT, et al. A systematic approachto diagnosis of mature T-cell leukemias revealsheterogeneity among WHO categories. Blood 2004; 104: 328–35.

116. Langerak AW, Sandberg Y, van Dongen JJ. Spectrum of T-largegranular lymphocyte lymphoproliferations: ranging from expandedactivated effector T cells to T-cell leukaemia. Br J Haematol 2003;123: 561–2.

117. Aldinucci D, Poletto D, Lorenzon D, et al. CD26 expressioncorrelates with a reduced sensitivity to 29-deoxycoformycin-inducedgrowth inhibition and apoptosis in T-cell leukemia/lymphomas. ClinCancer Res 2004; 10: 508–20.

118. Greer JP, Kinney MC, Loughran TP Jr. T cell and NK celllymphoproliferative disorders. Hematology (Am Soc Hematol EducProgram) 2001: 259–81.

119. Kothapalli R, Bailey RD, Kusmartseva I, et al. Constitutiveexpression of cytotoxic proteases and down-regulation of proteaseinhibitors in LGL leukemia. Int J Oncol 2003; 22: 33–9.


120. Bernengo MG, Novelli M, Quaglino P, et al. The relevance of theCD4+ CD26- subset in the identification of circulating Sezary cells. BrJ Dermatol 2001; 144: 125–35.

121. Rappl G, Muche JM, Abken H, et al. CD4(+)CD7(–) T cells composethe dominant T-cell clone in the peripheral blood of patients withSezary syndrome. J Am Acad Dermatol 2001; 44: 456–61.

122. Ohshima K, Haraoka S, Suzumiya J, et al. Absence of cytotoxicmolecules in CD8- and/or CD56-positive adult T-cell leukaemia/lymphoma. Virchows Arch 1999; 435: 101–4.

123. Takeshita M. CD30-positive anaplastic large cell lymphoma withHTLV-I proviral integration: a unique histologic subgroup of adultT-cell leukemia/lymphoma. Intern Med 1999; 38: 757–8.

124. Pittaluga S, Wlodarska I, Pulford K, et al. The monoclonal antibodyALK1 identifies a distinct morphological subtype of anaplastic large

cell lymphoma associated with 2p23/ALK rearrangements. Am JPathol 1997; 151: 343–51.

125. Cooke CB, Krenacs L, Stetler-Stevenson M, et al. Hepatosplenic T-cell lymphoma: a distinct clinicopathologic entity of cytotoxic gammadelta T-cell origin. Blood 1996; 88: 4265–74.

126. Macon WR, Levy NB, Kurtin PJ, et al. Hepatosplenic alphabeta T-celllymphomas: a report of 14 cases and comparison with hepatosplenicgammadelta T-cell lymphomas. Am J Surg Pathol 2001; 25: 285–96.

127. Chan JK, Sin VC, Wong KF, et al. Nonnasal lymphoma expressingthe natural killer cell marker CD56: a clinicopathologic study of 49cases of an uncommon aggressive neoplasm. Blood 1997; 89: 4501–13.

128. Suzuki R, Suzumiya J, Nakamura S, et al. Aggressive natural killer-cell leukemia revisited: large granular lymphocyte leukemia ofcytotoxic NK cells. Leukemia 2004; 18: 763–70.


immunophenotyping of lymphoproliferative disorders: state...

Documents