standardization of bone marrow features - does it work in ... · neoplastic conditions. this fact...

Summary. Standardized bone marrow (BM) featuresdetermined by semiquantitative scoring are valuabletools for the recognition and easily reproducibleinterpretation of histological patterns inhematopathology. This procedure may help tocharacterize various disease entities, but especially todifferentiate chronic myeloproliferative disorders(MPDs) with increased platelet counts from reactivethrombocytosis (RTh). A clear-cut separation of theseconditions continues to present a major problem inhematology. Therefore MPDs are a most suitable modelto test the diagnostic relevance of this procedure. Byregarding the literature and based on archive materialthat involved BM biopsies of 319 patients, asemiquantitative grading of histological parameters wasperformed. Standardized features were applied for astepwise discriminant analysis to establish different setsof variables exerting a diagnostic impact. A distinctioninto five histological patterns was achieved that showeda correctly predicted group membership of about 94 %.These were consistent with the clinicopathologicaldiagnosis of polycythemia vera, essentialthrombocythemia (ET), prefibrotic or early fibroticchronic idiopathic myelofibrosis (CIMF) and finallyRTh. Variables of discriminating potency according totheir ranking included megakaryopoiesis (maturationdefects, nuclear lobulation, naked and bulbous nuclei,small and giant size), reticulin fibers, erythro- andgranulopoiesis (left shifting and quantity) and cellularity.These findings are in keeping with the assumption thatcharacteristic patterns of BM histopathology can beassigned to different subtypes of MDPs mimicking ET.Discrimination between ET and especially early stageCIMF with thrombocythemia is warranted because ofsignificant implications concerning therapeuticstrategies, follow-up examinations and survival.

Regarding these results, a schematic procedure isproposed to be used for daily routine diagnosisconcerning the discrimination of MPDs.

Key words: Bone marrow biopsies, Polycythemia vera,Essential thrombocythemia, Idiopathic myelofibrosis,Thrombocytosis, Histological patterns, Discriminantanalysis, Prognosis

Introduction

Regarding complex structures that are composed ofwidely varying constituents like bone marrow (BM),histopathological evaluations aimed at achieving adefinite diagnosis are significantly impaired bysubjective aspects of description. These obviousshortcomings are especially important for the diagnosisof hematological neoplasias and the ensuingconsequences for a successful therapeutic approach. Forthis reason, there is a general demand to standardizemorphological features and to create characteristicpatterns that may be readily assigned to specific diseaseentities. Clinicians are aware of this unfavorablesituation and therefore insist on reproducible BMinterpretations for a clear-cut distinction of conditionsthat require elaborate up-to-date treatment regimens. Asuitable model to test the impact of standardized BMfeatures for the discrimination of disease entities are thePhiladelphia chromosome-negative (Ph1-) chronicmyeloproliferative disorders (MPDs) that comprise threemain subtypes: polycythemia vera (PV), essentialthrombocythemia (ET) and chronic idiopathicmyelofibrosis (CIMF)(Vardiman et al., 2001). In

Review

Standardization of bone marrow features - does it work in hematopathology forhistological discrimination of different disease patterns?J. Thiele1, H.M. Kvasnicka1 and V. Diehl21Institute of Pathology and 2First Clinic of Medicine, University of Cologne, Cologne, Germany

Histol Histopathol (2005) 20: 633-644

Offprint requests to: Juergen Thiele, M.D., Institute of Pathology,University of Cologne, Joseph-Stelzmannstr. 9, D-50924 Cologne,Germany. Fax: +49-0221-4786360. e-mail: [email protected]

http://www.hh.um.es

Histology andHistopathology

Cellular and Molecular Biology

Abbreviations: MPDs: myeloproliferative disorders; BM; bone marrow;Th, thrombocytosis; PV: polycythemia vera; ET: essentialthrombocythemia; CIMF: chronic idiopathic myelofibrosis; RTh: reactivethrombocytosis; PAS: periodic acid Schiff reagent; EDTA: ethylene-diamine tetra-acetic acid; H&E: hematoxylin and eosin

particular controversy and discussion still continuesconcerning the differentiation of ET from the othersubtypes of the MPDs that may present an elevatedplatelet count (Michiels and Thiele, 2002; Thiele andKvasnicka, 2003a; Florena et al., 2004). Although itseems reasonable to define thrombocytosis (Th) as aplatelet count exceeding 500x109/l it is important to notethat clinically significant complications (thrombosis,hemorrhage) may occur in a patient at this low range,even in ET below the generally acknowledged referencevalue of 600x109/l. (Lengfelder et al., 1998; Sacchi etal., 2000; Michiels and Thiele, 2002). Therefore, thedegree of Th is a poor discriminator concerning reactivestates (RTh) or defined entities of MPDs. In this contextit is noteworthy that in routine clinical practice RTh maybe responsible for more than 80% of patients presentinga substantial elevation of the platelet count, however,mostly without important consequences (Griesshammeret al., 1999). In this situation, in order to follow areasonable cost-effective approach for evaluating Th, anumber of suggestions have been proposed (Tefferi andMurphy, 2001). In addition to a scrutinized study ofpatient history, previous laboratory records and basiclaboratory tests, BM morphology was apparently not

considered as a routine method capable of providingappropriate means for distinction between reactive andneoplastic conditions. This fact may be due todifficulties in the recognition of distinctive histologicalpatterns charaterizing each entity and is especiallyfocused on an unequivocal differentiation of ET from theother subtypes of Ph1--MPDs. In the context of a reviewof the pertinent literature a retrospective analysis of alarge series of patients with specimens from our archivematerial was performed to elucidate whetherstandardized morphological features and/or specifichistological patterns facilitate such a separation ofentities that may clinically present Th. Moreover, toenhance diagnostic validity a comparative re-evaluationof clinical data and survival in combination with thesemorphological features of discriminative value iswarranted to control the predictive relevance of ourassessment.

Standardization of bone marrow (BM) features

Until now, a semiquantitative, systematicallyconducted evaluation with standardization of certain BMconstituents has been infrequently performed for thedifferentiation of MPDs associated with an elevatedplatelet count (Iland et al., 1987; Annaloro et al., 1999;Thiele and Kvasnicka, 2003b; Florena et al., 2004) or inpatients presenting erythrocytosis (Thiele et al., 2001c;Thiele and Kvasnicka, 2005). Moreover, this methodseems to be suitable to determine therapy-related effects,especially concerning BM cellularity and fibrosis orother features to characterize the efficacy of newlyintroduced drugs (Thiele et al., 2004a,b). Forpracticability it seems to be reasonble to use no morethan 20 BM features (Table 1) with a focus onmegakaryopoiesis, considering those entities presentingTh under study (Fig. 1a,b). Moreover, in this context a

634

Standardization of bone marrow features

Fig. 1. Prominent standardized features of megakaryopoiesis. Differentiation (a) and clustering (b) of megakaryocytes.

Table 1. Standardized major features of hematopoiesis and myeloidstroma.

CONTITUENTS PARAMETERS

Cellularity age-matched quantity

Granulopoiesis and quantity, left shifting and Erythropoiesis maturation defects

Megakaryopoiesis (see Fig. 1 a, b)

Myeloid stroma fibers (see Table 2)lymphoid nodulesiron deposits

proper quantification of the fiber content is essential(Table 2) by considering previously reported gradingsystems (Bauermeister, 1971; Manoharan et al., 1979;Beckman and Brown, 1990). The same applies to theassessment of cellularity that has to regard age-related

changes (Hartsock et al., 1965; Frisch et al., 1982). Asshown in Table 3 semiquantitative scoring should berestricted to three grades for decrease or increase,respectively, in addition and comparison to the normalstate. This procedure is recommended in order to avoid

635


Table 2. Grading of Myelofibrosis in MPDs adapted from semiquantitative and morphometric findings (Georgii et al., 1996; Thiele et al., 1996) andcompared with the Bauermeister score (Bauermeister, 1971) modified by Manoharan (Manoharan et al., 1979).

GRADING MORPHOMETRY FIBER DENSITY PER HEMATOPOIESIS(normal density: 16 i x 102/mm2) - Bauermeister score in brackets -

MF-0 < than twice (upper limit of the normal) Scattered linear reticulin with no inter-sections/cross-over corresponding with the normal BM (score N/+1)

MF-1 > three-fold Loose network of reticulin especially around the vessel (sinus wall sclerosis) with many intersections (cross-over) in most areas of the sectinon (score +2)

MF-2 > six-fold Diffuse and marked increase in reticulin with extensive intersections (dense network) and some bundles of collagen throughout the section, occasionally associated with initial (plaque-like) osteosclerosis (score +3)

MF-3 > six-fold Diffuse and dense increase in reticulin and coarse bundles of collagen throughout the section, often associated with osteosclerosis (score +4)

Table 3. Descriptive analysis of 18 standardized histological features comprising a semiquantitative evaluation with reference to age-related normalvalues (control group) in 319 patients presenting the presumptive diagnosis of Ph1-MPDs and accompanying thrombocytosis.

PARAMETER MARKED MODERATE SLIGHT NORMAL OR SLIGHT MODERATE MARKED DECREASE DECREASE DECREASE ABSENT INCREASE INCREASE INCREASE

(-3) (-2) (-1) (+1) (+2) (+3)

1. Cellularity - - 1 52 202 61 3

2. Megakaryopoiesis

quantity - - - 3 145 152 19

clusters (loose and dense) - - - 87 159 60 13

size:small - - - 30 161 114 14giant - - - 18 131 148 22

nuclear lobulation 5 54 91 47 76 44 2

maturation defects - - - 162 85 63 9

bulbous nuclei - - - 164 77 69 9

naked nuclei - - - 134 165 20 -

3. Granulopoiesis

quantity - - 6 94 169 50 -

left shifting - - - 177 142 - -

maturation defect - - - 307 12 - -

4. Erythropoiesis

quantity - 1 68 192 36 22 -

left shifting - - - 260 69 - -

maturation defects - - - 299 19 1 -

5. Myeloid stroma

reticulin fibres - - - 275 44 - -

lymphoid nodules - - - 248 48 18 5

iron deposits 57 44 42 158 13 3 2

much overlappings and thus enhance interobserverconsensus. Because this review included our own datathat were partially derived from a previous study (Thieleand Kvasnicka, 2003b) we performed a retrospectivereview of BM trephine biopsies and clinical records in319 patients (122 men, 197 women) with a median ageof 66 years (range 28 to 87 years) and the presumptiveclinical diagnosis of a Ph1--MPD that was associatedwith a Th in excess of 600x109/l sustained for more than3 months. Initially it was not clear, whether and to whichextent this cohort included only patients with a putativePh1--MPD in particular ET according to the PVSG-criteria (Murphy et al., 1997; Pearson, 1998; Streiff etal., 2002) or also ill-defined conditions suggesting areactive raise in the platelet count (RTh). BM samples of40 age-matched patients without known hematologicaldisorders served as controls. It has to be explicitlyemphasized that retrospective evaluation of clinicalrecords and analysis of BM biopsies were carried outindependently and in a blinded fashion.

Discriminant analysis and statistical evaluation

The diagnostic importance of standardized BMfeatures that are gained by semiquantitative evaluation ispreferably tested by means of stepwise discriminantanalysis. The purpose of this method is to obtain a smallset and ranking of useful discriminating variables thatprovide a reliable prediction of group classification(Everitt and Dunn, 2001) which greatly facilitates thedistinction of new cases in a prospective trial. Bycomparing predicted group memberships (test samples -standardized BM features) with actual memberships(training samples - with a definitive clinicopathologicaldiagnosis according to the WHO classification), we wereable to measure empirically the success in discriminatingpotency by observing the proportion of correctclassifications. Thus, by tabulation of the predictedgroup memberships one can validate the efficacy ofthese variables (Everitt and Dunn, 2001). To enhancesensitivity of this calculation it should be mentioned thatevaluation of the histological parameters was based onenzyme-histochemical staining techniques, likechloroacetate reaction (Leder stain) and also on PAS(Thiele and Kvasnicka, 2003b, 2005) contrasting thesemiquantitative evaluations in previous studies thatwere limited to routine staining techniques includingHematoxylin-Eosin (H&E) (Iland et al., 1987; Annaloroet al., 1999).

Problems of data interpretation

One of the most controversial issues inhematopathology is the discriminating impact of BMhistology as a characteristic feature and independentparameter to differentiate the various subtypes of MPDs.(Dickstein and Vardiman, 1993; Michiels and Thiele,2002). In this context the question arises whetherstandardized and easy to reproduce morphological

parameters may be generated that alone, or within acertain pattern, are specific for clinically defined entities.In this regard, a conflict of opinion continues to persistas to whether ET may be definitely separated from RTh(Iland et al., 1983; Kutti and Wadenvik, 1996), as well asfrom PV and early (hypercellular) stages of CIMF inpatients presenting a high platelet count (Iland et al.,1987; Murphy et al., 1997; Pearson, 1998). In therecently published WHO classification on MPDs,histopathology was included as a relevant parameter inthe corresponding sets of diagnostic criteria (Imbert etal., 2001). Nevertheless, one has to realize that until nowthe majority of clinical trials fail to recognize a BMbiopsy as an essential tool for differentiation between ETand the other subtypes of MPDs (Thiele and Kvasnicka,2003a). In addition to time-consuming cell culturestudies (Eridani et al., 1987; Juvonen et al., 1993; Shihand Lee, 1994; Florensa et al., 1995; Westwood andPearson, 1996; Yan et al., 1996), a promising approachto solve this problem seems to focus firstly on aretrospective evaluation of prominent BM features.Regarding this point, application of subsequentdiscriminant analysis (Everitt and Dunn, 2001)performed in a blinded and independent fashion onstandardized BM features should be considered as auseful procedure to establish certain patterns ofhistopathlogy (Thiele and Kvasnicka, 2003b, 2005). Thediagnostic importance of these have to be validated by aproperly conducted correlation with clinical dataincluding follow-up examinations and survival.Furthermore, as a stringent consequence of this preludethe accordingly established histological parameters needto be tested and approved in the context of randomizedprospective clinical trials. Finally, regarding otherdiagnostic features, intriguing data on the identificationof the so-called PV gene (PRV-1) and its assumedexclusive expression in corresponding patients has notonly renewed interest in finding mutations leading to thedevelopement of a MPD, but raised considerable hopesfor a valuable molecular marker (Klippel et al., 2002;Pahl, 2002, 2003, 2004) However, a recently publishedstudy reported that PRV-1 is constitutively expressed byBM cells and does not discriminate PV from reactivestates or other Ph1--MPDs (Bock et al., 2003).Contrasting erythropoietin-independent erythroid colony(EEC) formation in a few patients with the clinical androutine laboratory characteristics of PV normal PRV-1mRNA levels were found and also overlappings betweenerythropoietin (EPO) expression and other biomarkers inso-called ET (Kralovics et al., 2003; Liu et al., 2003;Griesshammer et al., 2004). Fluorescence in-situanalysis of PRV-1 provided persuasive evidence that thisgene is neither amplificated, deleted nor structurallyrearranged in PV patients (Najfeld et al., 2003)Additionally, real-time PCR-based assays showed aPRV-1 expression across the MPDs and also in reactivestates (Tefferi et al., 2004). Finally, this obviousconfusion concerning the diagnostic validity ofbiomarkers, but especially the PRV-1 gene, originally

636


637


Fig. 2. Heterogeneityof bone marrowfeatures in Ph1-- MPDsassociated withthrombocytosis. PVwith markedproliferation of largeand small-sizedmegakaryocytes(pleomorphic aspect)besides increase inerythro- and neutrophilgranulopoiesis(panmyelosis) (a). ETwith predominance oflarge to giantmegakaryocytes, butno significant growthof the other celllineages (b). CIMF - 0consistent with aprefibrotic stageshowing clusters ofabnormalmegakaryocytes andmarked granulocyticproliferation (c). CIMF- 1 with early fibrosischaracterized bybundles of reticulin inclose association withlarge megakaryocytes(d). a, b, c,chloroacetate-esterase; d, silverimpregnation afterGomori). x 180

638


Fig. 3. Heterogeneityof megakaryocytefeatures in Ph1-- MPDsassociated withthrombocytosis incomparison with RTh.PV revealingsignificant differencesin size (pleomorphusaspect) ranging fromsmall to giant, yet,maturemegakaryocytes (a).ET shows giantmegakaryocytes withhyperlobulated(staghorn-like) nuclei(b). CIMF ischaracterized by aloose clustering ofsmall to largemegakaryocytescontaininghyperlobulated(bulbous) cloud-likenuclei and a fewalmost denuded nuclei(arrow) (c). Nakednuclei of CIMF may belarge and of bizarreshape (arrows) (d). InRTh megakaryocytesare dispersedthroughout the marrowand of medium size(compare with Fig.2a,b) and do notexhibit maturationdefects (compare withFig. 2c,d) (e). a, b, c,d, e, PAS reaction; x 380

postulated to exert a major discriminating power for thedifferentiation of these disorders has been clarified.Recently, PRV-1 has been shown to represent a non-

specific marker for an abnormal leukocyte productionand/or release in reactive as well as various neoplastichematological conditions (Passamonti et al., 2004). Forthis reason, compared to histopathology, until now nobiological indicator has been shown to exert by itself adefinite diagnostic or discriminating impact concerningthe Ph1--MPDs.

Heterogeneity of bone marrow histopathology

Contrasting the clinical diagnosis of sustained Thand presumptive ET in the majority of patients,histological features of the BM reveal conspicuouslyexpressed varieties of patterns. Consequentlysemiquantitative analysis of the 18 standardizedprominent BM features show wide ranges concerningthese parameters (Table 3). This striking heterogeneity inhistological appearance is properly reflected by theamount and ratio of erythro- and granulopoiesisincluding their left-shifting (Fig. 3a-c). Moreover,density of reticulin fibers ranges from no increase in themajority of patients to a slight and moderate fibrosis in asmall cohort (Fig. 3d). However, most remarkable arechanges involving the megakaryocytes (Fig. 3a-e).Besides differences in quantity, features like loose ordense clusters (Figs. 2c, 3c), varieties in size (Fig. 2a,b,e) and presence of naked nuclei (Fig. 3c,d) emerge asparameters of distinctive value (Table 3). Concerning thematuration defects of this cell lineage, anomalies ofnuclear lobulation including bulbous (hyperchromatic,cloud-like) nuclei and an obvious deviation of normalnuclear-cytoplasmic maturation (Fig. 3c) are apparent.

Opposed to those findings derived from our filedmaterial of BM specimens, a previous study on BMcharacteristics in 50 patients with ET and 27 patientswith PV presenting a platelet count exceeding1,000x109/l failed to demonstrate significant differences(Iland et al., 1987) Although semiquantitative evaluationwas performed by employing four grades comparable to

639


Table 4. Diagnostic impact of certain standardized histological bonemarrow features (ranking with relative incidence) in 319 patientsfollowing discriminant analysis and clinical diagnosis according to theWHO classification (Vardiman et al., 2001).

CLINICAL DIAGNOSIS PV ET CIMF-0 CIMF-1 RTh(WHO classification)

No. of patients 52 90 118 37 22

Megakaryopoiesis1. maturation defects - - + + -2. nuclear lobulation + + - - -

Myeloid stroma3. reticulin fibers - - - + -

Erythropoiesis4. left shifting + - - - -

Megakaryopoiesis5. naked nuclei - - + + -6. small forms + - + + +

Erythropoiesis7. quantity + - - - -

Granulopoiesis8. left shifting + - + - +

Megakaryopoiesis9. giant forms + + + + -10. Cellularity + - + + -

Megakaryopoiesis11. bulbous nuclei - - + + -12. clusters + + + + -

Relative incidence of histological features displaying a discriminatingeffect: -, ≤ 10%; +, ≥ 80%. Statistical analysis: Predicted groupmembership: 299 of 319 of originally grouped cases correctly classified(93.6%).

Table 5. Stepwise discriminant analysis of standardized BM features(relative incidence and ranking - see also Table 3) showing distinctivepatterns that enable separation of PV from thrombocytosis and ETaccording to the diagnostic criteria of the PVSG (Murphy et al., 1997;Pearson, 1998; Murphy, 1999).

VARIABLE (%) PV versus ET

1. Erythropoiesis (quantity) 68 62. Granulopoiesis (left shifting) 66 233. Megakaryocytes /naked nuclei) 25 754. Cellularity 81 185. Erythropoiesis (left shifting) 88 126. Megakaryocytes (giant forms) 19 997. Megakaryocytes (nuclear lobulation) 37 648. Reticulin fibers 15 85No. of patients 52 245

Statistical analysis: Wilks’ lambda statistics = 0.260; p < 0.0001.Predicted group membership: PV 49/52 and ET (PVSG) 241/245 oforiginally grouped cases correctly classified (97.6 %).

Table 6. Stepwise discriminant analysis of standardized BM features(relative incidence and ranking - see also Table 3) showing distinctivepatterns that enable separation of PV from thrombocytosis and ETaccording to the diagnostic criteria of the WHO (Vardiman et al., 2001).

VARIABLE (%) PV versus ET

1. Erythropoiesis (quantity) 89 92. Granulopoiesis (left shifting) 78 213. Erythropoiesis (left shifting) 75 294. Megakaryocytes (naked nuclei) 25 985. Cellularity 100 06. Megakaryocytes (giant forms) 20 987. Reticulin fibers 14 0No. of patients 52 90

Statistical analysis: Wilks’ lambda statistics = 0.217; p < 0.0001.Predicted group membership: PV 48/52 and ET (WHO) 88/90 oforiginally grouped cases correctly classified (95.8 %).

our scoring system, histopathological features were mostprobably determined by assessment of H&E stainedspecimens not involving more refined methods likechloroacetate esterase (Leder stain). In particular thelatter staining technique is very useful to determine veryeasily the exact amount of erythropoiesis versusneutrophil granulopoiesis in patients with PV (Thiele etal., 1999b, 2001c) In both groups of patients a moderateincrease in cellularity, erythroid precursors and myeloidelements was reported and some of the PV patientsshowed even a slight collagen fibrosis (Iland et al., 1983,1987). For this reason, it was concluded that theseresults highlight the similarities of apparently closelyrelated conditions and therefore, additional methods forachieving the correct diagnosis were suggested. Thisadverse impression of an insignificant discriminatingimpact of morphology has been ameliorated by our re-evaluation of corresponding specimens (Tables 4-6)including up-to-date staining techniques to characterizemore accurately the different cell lineages.

On the other hand, the impact of BM morphologyconcerning the diagnosis of PV has been recentlyquestioned (Spivak, 2002). It has been stated thatabnormalities such as an increase in megakaryocytes andcellular hyperplasia with a loss of fat spaces (Ellis et al.,1975, 1986; Ellis and Peterson, 1979) can never be alonediagnostic for PV. However, one should be aware thatthis argument reflects a rather historical view, becausesignificant advances in handling, processing andevaluation of BM specimens achieved over the last twodecades have repeatedly emphasized the existence ofcharacteristic BM features in PV (Bartl et al., 1993;Georgii et al., 1998; Thiele et al., 1999a,b, 2001a,c;Thiele and Kvasnicka, 2005) The argument that resultsof morphological studies are generally ambiguousbecause discriminating features are not standardized andtherefore easily reproducible (Pearson, 1998) has beencontested by our data (Table 3). Finally, the statementthat until now there was no investigation testing thevalidity of histopathology in a blinded and independentfashion is not supported by previous studies (Thiele etal., 2001b,c).

Discriminate analysis of bone marrow features

Discriminate analysis revealed that 12 of the 18variables allowed a corresponding ranking andseparation into five distinctive histological patterns.When applying the WHO classification (Vardiman et al.,2001) these specific sets of morphological parametersare compatible with PV, ET, prefibrotic CIMF (CIMF-0),early fibrotic CIMF (CIMF-1) and finally RTh (Table 4).It is noteworthy that the predicted group membershipreveals a correct overall classification of 93.6 %.Regarding the diagnosis of ET significant differences arepresent when comparing the PVSG criteria (Murphy etal., 1997; Pearson, 1998) with the WHO classification(Vardiman et al., 2001). To investigate this importantissue, a comparative evaluation was performed between

ET and PV accompanied by Th (Tables 5, 6).Discriminant analysis between these two groups andclassifications yields almost identical results except forthe degree of nuclear lobulation of megakaryocytes(Table 5). The latter result is in keeping with theassumption of a highly discriminating power of BMfeatures in PV (Georgii et al., 1996, 1998; Thiele et al.,2001b,c) compared to the other subtypes of Ph1--MPD.Finally, there is still the problem of how to differentiateET from reactive lesions. With respect to this salientpoint of diagnosis, a corresponding calculation displaysa number of histological features that exert a clear-cutpotential (more than 98%) to separate ET from RTh(Table 7).

Regarding the parameters of discriminating powerbetween ET and PV it is noticable that the differences inTables 5 and 6 are obviously limited to maturationdefects of erythropoiesis and quantity ofmegakaryocytes. These differences are dependent on theclassification applied; because by following the PVSGcriteria (Pearson, 1998; Murphy, 1999) a considerablenumber of patients with CIMF-0 and CIMF-1 areincluded and therefore, consistent with false ET (Thieleand Kvasnicka, 2003a). In this regard it is noticable thatseveral reports on histopathology are in line with theassumption that distinctive features exist allowing theseparation of the different entities of Ph1--MPDsassociated with a significant excess of platelets. (Thieleet al., 1988; Georgii et al., 1990; Buhr et al., 1992;Michiels and Juvonen, 1997; Georgii et al., 1998; Thieleet al., 1999a; Michiels and Thiele, 2002; Florena et al.,2004)

Clinical impact

Contrasting the discriminating impact ofhistopathology, some doubts have been expressed aboutthe usefulness to differentiate more thoroughly MPDswith an associated raise in the platelet count, which

640


Table 7. Stepwise discriminant analysis of standardized BM features(relative incidence and ranking - see also Table 3) revealing distinctivepatterns that enable separation of RTh from ET (according to the WHOclassification) with special focus on megakaryocyte morphology.

VARIABLE (%) RTh versus ET

1.Megakaryopoiesisgiant forms 21 99small forms 86 37nuclear lobulation 0 79

2. Granulopoiesis (left shifting) 81 22

3. Megakaryocytes (naked nuclei) 0 45

No. of patients 22 90

Statistical analysis: Wilks’ lambda statistics = 0.250; p < 0.0001.Predicted group membership: RTh 22/22 and ET 88/90 of originallygrouped cases correctly classified (98.2 %).

clinically often mimic ET (Thiele et al., 2002). Anindependently performed determination of clinical datashows that PV with Th is characterized by a number ofparameters facilitating an easy distinction from the otherentities (Table 8). It has to be mentioned that themajority of leading clinical trials of ET fail to enter a(pretreatment, representative) BM biopsy as one of thecriteria for diagnosis (Bellucci et al., 1986; Chistolini etal., 1990; Fenaux et al., 1990; Colombi et al., 1991;Jantunen et al., 1998). Furthermore, one has to bear inmind that the platelet limit of 600x109/l is arbitrary andsome patients may present relevant complications likethrombosis, easy bruising and transitional ischemicattacks associated with a far lower thrombocyte count(Lengfelder et al., 1998; Sacchi et al., 2000; Michielsand Thiele, 2002). In these cases a BM biopsy may offera major clue to the nature of the disease process byrevealing features that are in keeping with ET accordingto the specific histologial pattern described in thisreview. Therefore the need for a clear-cut definition ofcharacteristic BM features and more postitive diagnosticguidelines has been recognized by the WHOclassification, where megakaryocyte morphology wasespecially regarded (Imbert et al., 2001). By followingthese postulates initial (prefibrotic) and early CIMF thatfrequently present Th (false ET) were distinguished(Thiele et al., 2001a, 2002; Michiels and Thiele, 2002;Thiele and Kvasnicka, 2003a; Florena et al., 2004). Thisconcept implicates a more stringent definition ofdiagnostic criteria than has been exerted by the PVSGguidelines (Murphy et al., 1997; Pearson, 1998; Murphy,1999). For this reason, it is not suprising thatsignificantly wide ranges of complications like theoccurrence of myelofibrosis or acute leukemia werereported in so-called ET patients, particularly in the highrisk groups (Yonemitsu and Okuda, 1985; Bellucci et al.,

1986; Van de Pette et al., 1986; Liberato et al., 1989;Fenaux et al., 1990; Kimura et al., 1990; Emilia et al.,1993; Murphy et al., 1997; Tefferi and Murphy, 2001;Cervantes et al., 2002; Radaelli et al., 2002). Therefore,by following the PVSG criteria, a considerableheterogeneity of disease features may be expectedaccounting for these adverse findings (Thiele andKvasnicka, 2003a). To underline this argument, acontroversy has recently emerged concerning thepresumptive varieties of BM patterns in patients with ETderived from the Italian Study Group where differencesof histological features were associated with survival.(Annaloro et al., 1999) Moreover, this finding of avariable morphology in so-called ET patients wasvalidated by a longitudonal study involving sequentialBM biopsies and a long-time follow-up. (Thiele et al.,2002). When applying the WHO classification (Imbert etal., 2001) only about one third of the 120 patientsrevealed (true) ET, while the other presented with initial-early stage CIMF of whom the majority developedmyelofibrosis - osteosclerosis during the lengthy courseof disease (Thiele et al., 2002). Follow-up data exhibiteda strikingly different prognosis in the various groups thatare discriminated by their histological features. Incontrast with patients with RTh, a less favorableobserved (median) survival was calculable in PV (148months) and ET (193 months), whereas early CIMF ischaracterized by a worsening regarding outcome (111

641


Table 8. Clinical findings (mean ± SD) following discriminant analysis ofBM features into different categories in the 297 patients with Ph1--MPDsand accompanying thrombocytosis according to the updated WHOcriteria (Vardiman et al., 2001).

PV ET CIMF-0 CIMF-1

No. of patients 52 90 118 37Erythrocytes (x1012/l) 6.5±0.9* 4.9±0.5 5.0±0.6 4.7±0.7Hemoglobin (g/dl) 17.3±2.9* 14.2±1.5 14.1±2.0 12.9±1.7Hematocrit (%) 54.2±9.1* 42.7±3.4 43.0±4.7 39.3±5.8Leukocytes (x109/l) 14.4±4.9 10.4±3.5 12.6±0.9 12.1±4.5Basophils (%) 1.0±1.2 1.1±1.3 0.9±1.2 1.2±1.6Myeloblasts (%) 0 0 0 0.2±0.8*Erythroblasts (%) 0 0 0 0.2±0.5*Thrombocytes (x109/l) 882±232 1,043±337 1,068±311 1,049±287LAP** 191±85* 55±33 100±76 113±76LDH (U/l) 343±147 365±216 275±95 617±22*Spleen size*** 2.1±1.9* 0.5±0.7 0.7±1.1 1.0±1.3

*: parameters of discriminating value; **: leukocyte alkalinephosphatase: normal score 10-80; ***: cm below costal margin onpalpation Wilks’ lambda statistics = 0.240; p< .0000.

Fig. 4. Schematic description (flow-sheet) of prominent histologicalfeatures exerting diagnostic relevance to differentiate chronicmyeloproliferative disorders associated with an evaluated platelet countfrom reactive lesions in bone marrow biopsy specimens.

months) and also a higher risk of developing thrombosisand hemorrhage. All these data support the notion ofcharacteristic laboratory (Table 8) and especiallymorphological features that are of invaluable aid toestablish and validate the diagnostic classification ofPh1--MPDs accompanied by Th.

In conclusion, because of the significant differencesin clinical findings, therapeutic strategies and prognosis,a clear-cut distinction between ET, CIMF and PVaccompanied by an excess in platelets seems to bewarranted in addition to a separation from RTh. One ofthe major points of differentiation is characteristichistological BM patterns as shown by discriminantanalysis of standardized morphological featuresfollowing a semiqunatitative scoring. Principally basedon these results, a schematic procedure (in Figure 4)regarding these diagnostic BM parameters may bepursued in daily routine to differentiate Ph1--MPDs.

References

Annaloro C., Lambertenghi Delil iers G., Oriani A., Pozzoli E.,Lambertenghi Deliliers D., Radaelli F. and Faccini P. (1999).Prognostic significance of bone marrow biopsy in essentialthrombocythemia. Haematologica 84, 17-21.

Bartl R., Frisch B. and Wilmanns W. (1993). Potential of bone marrowbiopsy in chronic myeloproliferative disorders (MPD). Eur. J.Haematol. 50, 41-52.

Bauermeister D.E. (1971). Quantitation of bone marrow reticulin--anormal range. Am. J. Clin. Pathol. 56, 24-31.

Beckman E.N. and Brown A.W. Jr. (1990). Normal reticulin level in iliacbone marrow. Arch. Pathol. Lab. Med. 114, 1241-1243.

Bellucci S., Janvier M., Tobelem G., Flandrin G., Charpak Y., Berger R.and Boiron M. (1986). Essential thrombocythemias. Clinicalevolutionary and biological data. Cancer 58, 2440-2447.

Bock O., Serinsöz E., Neusch M., Schlué J. and Kreipe H. (2003). Thepolycythaemia rubra vera-1 gene is constitutively expressed by bonemarrow cells and does not discriminate polycythaemia vera fromreactive and other chronic myeloproliferative disorders. Br. J.Haematol. 123, 472-474.

Buhr T., Georgii A., Schuppan O., Amor A. and Kaloutsi V. (1992).Histologic findings in bone marrow biopsies of patients withthrombocythemic cell counts. Ann. Hematol. 64, 286-291.

Cervantes F., Alvarez-Larran A., Talarn C., Gomez M. and MontserratE. (2002). Myelofibrosis with myeloid metaplasia following essentialthrombocythaemia: actuarial probability, presenting characteristicsand evolution in a series of 195 patients. Br. J. Haematol. 118, 786-790.

Chistolini A., Mazzucconi M.G., Ferrari A., la Verde G., Ferrazza G.,Dragoni F., Vitale A., Arcieri R. and Mandelli F. (1990). Essentialthrombocythemia: a retrospective study on the clinical course of 100patients. Haematologica 75, 537-540.

Colombi M., Radaelli F., Zocchi L. and Maiolo A.T. (1991). Thromboticand hemorrhagic complications in essential thrombocythemia. Aretrospective study of 103 patients. Cancer 67, 2926-2930.

Dickstein J.I. and Vardiman J.W. (1993). Issues in the pathology anddiagnosis of the chronic myeloproliferative disorders and themyelodysplastic syndromes. Am. J. Clin. Pathol. 99, 513-525.

Ellis J.T. and Peterson P. (1979). The bone marrow in polycythemia

vera. Pathol. Annu. 14, 383-403.Ellis J.T., Silver R.T., Coleman M. and Geller S.A. (1975). The bone

marrow in polycythemia vera. Semin. Hematol. 12, 433-444.Ellis J.T., Peterson P., Geller S.A. and Rappaport H. (1986). Studies of

the bone marrow in polycythemia vera and the evolution ofmyelofibrosis and second hematologic malignancies. Semin.Hematol. 23, 144-155.

Emilia G., Sacchi S., Temperani P., Longo R. and Vecchi A. (1993).Progression of essential thrombocythemia to blastic crisis viaidiopathic myelofibrosis. Leuk. Lymphoma 9, 423-426.

Eridani S., Sawyer B. and Pearson T.C. (1987). Patterns of in vitro BFU-E proliferation in different forms of polycythaemia and inthrombocythaemia. Eur. J. Haematol. 38, 363-369.

Everitt B.S. and Dunn G. (2001). Applied multivariate data analysis. 2ndedn. Arnold. London.

Fenaux P., Simon M., Caulier M.T., Lai J.L., Goudemand J. and BautersF. (1990). Clinical course of essential thrombocythemia in 147cases. Cancer 66, 549-556.

Florena A.M., Tripodo C., Iannitto E., Porcasi R., Ingrao S. and FrancoV. (2004). Value of bone marrow biopsy in the diagnosis of essentialthrombocythemia. Haematologica 89, 911-919.

Florensa L., Besses C., Woessner S., Sole F., Acin P., Pedro C. andSans-Sabrafen J. (1995). Endogenous megakaryocyte and erythroidcolony formation from blood in essential thrombocythaemia.Leukemia 9, 271-273.

Frisch B., Bartl R. and Burkhardt R. (1982). Bone marrow biopsy inclinical medicine: an overview. Haematologia (Budap). 15, 245-285.

Georgii A., Vykoupil K.F., Buhr T., Choritz H., Dohler U., Kaloutsi V. andWerner M. (1990). Chronic myeloproliferative disorders in bonemarrow biopsies. Pathol. Res. Pract. 186, 3-27.

Georgii A., Buhr T., Buesche G., Kreft A. and Choritz H. (1996).Classification and staging of Ph-negative myeloproliferativedisorders by histopathology from bone marrow biopsies. Leuk.Lymphoma 22 Suppl 1, 15-29.

Georgii A., Buesche G. and Kreft A. (1998). The histopathology ofchronic myeloproliferative diseases. Baillieres Clin. Haematol. 11,721-749.

Griesshammer M., Bangerter M., Sauer T., Wennauer R., Bergmann L.and Heimpel H. (1999). Aetiology and clinical significance ofthrombocytosis: analysis of 732 patients with an elevated plateletcount. J. Intern. Med. 245, 295-300.

Griesshammer M., Klippel S., Strunck E., Temerinac S., Mohr U.,Heimpel H. and Pahl H.L. (2004). PRV-1 mRNA expressiondiscriminates two types of essential thrombocythemia. Ann.Hematol. 83, 364-370.

Hartsock R.J., Smith E.B. and Petty C.S. (1965). Normal Variations withaging of the amount of hematopoietic tissue in bone marrow fromthe anterior iliac crest. A study made from 177 cases of suddendeath examined by necropsy. Am. J. Clin. Pathol. 43, 326-331.

Iland H.J., Laszlo J., Peterson P., Murphy S., Briere J., Weinfeld A.,Rosenthal D.S., Landaw S.A., Ellis J.T. and Silverstein M.N. (1983).Essential thrombocythemia: clinical and laboratory characteristics atpresentation. Trans. Assoc. Am. Physicians 96, 165-174.

Iland H.J., Laszlo J., Case D.C. Jr., Murphy S., Reichert T.A., Tso C.Y.and Wasserman L.R. (1987). Differentiation between essentialthrombocythemia and polycythemia vera with markedthrombocytosis. Am. J. Hematol. 25, 191-201.

Imbert M., Pierre R., Thiele J., Vardiman J.W., Brunning R.D. andFlandrin G. (2001). Essential thrombocythaemia. In WHO

642


Classification of tumours: Tumours of haematopoietic and lymphoidtissues. Jaffe E.S., Harris N.L., Stein H. and Vardiman J.W. (eds).IARC Press. Lyon. pp 39-41.

Jantunen R., Juvonen E., Ikkala E., Oksanen K., Anttila P., Hormila P.,Jansson S.E., Kekomaki R. and Ruutu T. (1998). Essentialthrombocythemia at diagnosis: causes of diagnostic evaluation andpresence of positive diagnostic findings. Ann. Hematol. 77, 101-106.

Juvonen E., Ikkala E., Oksanen K. and Ruutu T. (1993). Megakaryocyteand erythroid colony formation in essential thrombocythaemia andreactive thrombocytosis: diagnostic value and correlation tocomplications. Br. J. Haematol. 83, 192-197.

Kimura A., Fujimoto T., Inada T., Imamura N., Oguma N., Kajihara H.,Mtasiwa D.M., Katoh O., Fujimura K. and Kuramoto A. (1990).Blastic transformation in essential thrombocythemia. In vitrodifferentiation of blast cells into granulocytic, erythroid, andmegakaryocytic lineages. Cancer 65, 1538-1544.

Klippel S., Strunck E., Busse C.E., Behringer D. and Pahl H.L. (2002).Biochemical characterization of PRV-1, a novel hematopoietic cellsurface receptor, which is overexpressed in polycythemia rubravera. Blood 100, 2441-2448.

Kralovics R., Buser A.S., Teo S.S., Coers J., Tichelli A., van der MaasA.P. and Skoda R.C. (2003). Comparison of molecular markers in acohort of patients with chronic myeloproliferative disorders. Blood102, 1869-1871.

Kutti J. and Wadenvik H. (1996). Diagnostic and differential criteria ofessential thrombocythemia and reactive thrombocytosis. Leuk.Lymphoma 22 Suppl 1, 41-45.

Lengfelder E., Hochhaus A., Kronawitter U., Hoche D., Queisser W.,Jahn-Eder M., Burkhardt R., Reiter A., Ansari H. and Hehlmann R.(1998). Should a platelet limit of 600 x 109/l be used as a diagnosticcriterion in essential thrombocythaemia? An analysis of the naturalcourse including early stages. Br. J. Haematol. 100, 15-23.

Liberato N.L., Barosi G., Costa A., D'Elia P. and Boccaccio P. (1989).Myelofibrosis with myeloid metaplasia fol lowing essentialthrombocythaemia. Acta Haematol. 82, 150-153.

Liu E., Jelinek J., Pastore Y.D., Guan Y., Prchal J.F. and Prchal J.T.(2003). Discrimination of polycythemias and thrombocytoses bynovel, simple, accurate clonality assays and comparison with PRV-1expression and BFU-E response to erythropoietin. Blood 101, 3294-3301.

Manoharan A., Horsley R. and Pitney W.R. (1979). The reticulin contentof bone marrow in acute leukaemia in adults. Br. J. Haematol. 43,185-190.

Michiels J.J. and Juvonen E. (1997). Proposal for revised diagnosticcriteria of essential thrombocythemia and polycythemia vera by theThrombocythemia Vera Study Group. Semin. Thromb. Hemost. 23,339-347.

Michiels J.J. and Thiele J. (2002). Clinical and pathological criteria forthe diagnosis of essential thrombocythemia, polycythemia vera, andidiopathic myelofibrosis (agnogenic myeloid metaplasia). Int. J.Hematol. 76, 133-145.

Murphy S. (1999). Diagnostic criteria and prognosis in polycythemiavera and essential thrombocythemia. Semin. Hematol. 36, 9-13.

Murphy S., Peterson P., Iland H. and Laszlo J. (1997). Experience of thepolycythemia vera study group with essential thrombocythemia: afinal report on diagnostic criteria, survival, and leukemic transition bytreatment. Semin. Hematol. 34, 29-39.

Najfeld V., Fuchs S., Merando P., Lezon-Geyda K. and Fruchtman S.(2003). Fluorescence in situ hybridization analysis of the PRV-1

gene in polycythemia vera: implications for its role in diagnosis andpathogenesis. Exp. Hematol. 31, 118-121.

Pahl H.L. (2002). Polycythaemia vera: will new markers help us answerold questions? Acta Haematol. 108, 120-131.

Pahl H.L. (2003). Molecular markers in myeloproliferative disorders:from classification to prognosis? Hematology 8, 199-209.

Pahl H.L. (2004). Diagnostic approaches to polycythemia vera in 2004.Expert. Rev. Mol. Diagn. 4, 495-502.

Passamonti F., Pietra D., Malabarba L., Rumi E., Porta M.G., MalcovatiL., Bonfichi M., Pascutto C., Lazzarino M. and Cazzola M. (2004).Clinical significance of neutrophil CD177 mRNA expression in Ph-negative chronic myeloproliferative disorders. Br. J. Haematol. 126,650-656.

Pearson T.C. (1998). Diagnosis and classification of erythrocytoses andthrombocytoses. Baillieres Clin. Haematol. 11, 695-720.

Radaelli F., Mazza R., Curioni E., Ciani A., Pomati M. and Maiolo A.T.(2002). Acute megakaryocytic leukemia in essentialthrombocythemia: an unusual evolution? Eur. J. Haematol. 69, 108-111.

Sacchi S., Vinci G., Gugliotta L., Rupoli S., Gargantini L., Martinelli V.,Baravelli S., Lazzarino M. and Finazzi G. (2000). Diagnosis ofessential thrombocythemia at platelet counts between 400 and600x109/L. Gruppo Ital iano Malatt ie MieloproliferativeCroniche(GIMMC). Haematologica 85, 492-495.

Shih L.Y. and Lee C.T. (1994). Identification of masked polycythemiavera from patients with idiopathic marked thrombocytosis byendogenous erythroid colony assay. Blood 83, 744-748.

Spivak J.L. (2002). Polycythemia vera: myths, mechanisms, andmanagement. Blood 100, 4272-4290.

Streiff M.B., Smith B. and Spivak J.L. (2002). The diagnosis andmanagement of polycythemia vera in the era since the PolycythemiaVera Study Group: a survey of American Society of Hematologymembers' practice patterns. Blood 99, 1144-1149.

Tefferi A. and Murphy S. (2001). Current opinion in essentialthrombocythemia: pathogenesis, diagnosis, and management.Blood Rev. 15, 121-131.

Tefferi A., Lasho T.L., Wolanskyj A.P. and Mesa R.A. (2004). NeutrophilPRV-1 expression across the chronic myeloproliferative disordersand in secondary or spurious polycythemia. Blood 103, 3547-3548.

Thiele J. and Kvasnicka H.M. (2003a). Chronic myeloproliferativedisorders with thrombocythemia: a comparative study of twoclassification systems (PVSG, WHO) on 839 patients. Ann. Hematol.82, 148-152.

Thiele J. and Kvasnicka H.M. (2003b). Diagnostic differentiation ofessential thrombocythaemia from thrombocythaemias associatedwith chronic idiopathic myelofibrosis by discriminate analysis of bonemarrow features--a clinicopathological study on 272 patients. Histol.Histopathol. 18, 93-102.

Thiele J. and Kvasnicka H.M. (2005). Diagnostic impact of bone marrowhistopathology in polycythemia vera (PV). Histol. Histopathol. 20,317-328.

Thiele J., Schneider G., Hoeppner B., Wienhold S., Zankovich R. andFischer R. (1988). Histomorphometry of bone marrow biopsies inchronic myeloproliferative disorders with associated thrombocytosis-features of significance for the diagnosis of primary (essential)thrombocythaemia. Virchows Arch. (A) 413, 407-417.

Thiele J., Kvasnicka H.M., Werden C., Zankovich R., Diehl V. andFischer R. (1996). Idiopathic primary osteo-myelofibrosis: a clinico-pathological study on 208 patients with special emphasis on

643


evolution of disease features, differentiation from essentialthrombocythemia and variables of prognostic impact. Leuk.Lymphoma 22, 303-317.

Thiele J., Kvasnicka H.M., Diehl V., Fischer R. and Michiels J. (1999a).Clinicopathological diagnosis and differential criteria ofthrombocythemias in various myeloproliferative disorders byhistopathology, histochemistry and immunostaining from bonemarrow biopsies. Leuk. Lymphoma 33, 207-218.

Thiele J., Kvasnicka H.M. and Fischer R. (1999b). Histochemistry andmorphometry on bone marrow biopsies in chronic myeloproliferativedisorders - aids to diagnosis and classification. Ann. Hematol. 78,495-506.

Thiele J., Imbert M., Pierre R., Vardiman J.W., Brunning R.D. andFlandrin G. (2001a). Chronic idiopathic myelofibrosis. In WHOClassification of Tumours: tumours of haematopoietic and lymphoidtissues. Jaffe E.S., Harris N.L., Stein H. and Vardiman J.W. (eds).IARC Press. Lyon. pp 35-38.

Thiele J., Kvasnicka H.M., Muehlhausen K., Walter S., Zankovich R.and Diehl V. (2001b). Polycythemia rubra vera versus secondarypolycythemias. A clinicopathological evaluation of distinctivefeatures in 199 patients. Pathol. Res. Pract. 197, 77-84.

Thiele J., Kvasnicka H.M., Zankovich R. and Diehl V. (2001c). The valueof bone marrow histology in differentiating between early stagepolycythemia vera and secondary (reactive) polycythemias.Haematologica 86, 368-374.

Thiele J., Kvasnicka H.M., Schmitt-Graeff A., Zankovich R. and Diehl V.(2002). Follow-up examinations including sequential bone marrowbiopsies in essential thrombocythemia (ET): a retrospectiveclinicopathological study of 120 patients. Am. J. Hematol. 70, 283-

291.Thiele J., Kvasnicka H.M., Schmitt-Graeff A., Kriener S., Engels K.,

Staib P., Griesshammer M., Waller C.F., Ottmann O.G. andHansmann M.L. (2004a). Effects of the tyrosine kinase inhibitorImatinib mesylate (STI571) on bone marrow features in patients withchronic myelogenous leukemia. Histol. Histopathol. 19, 1277-1288.

Thiele J., Kvasnicka H.M., Schmitt-Graff A., Hulsemann R. and Diehl V.(2004b). Therapy-related changes of the bone marrow in chronicidiopathic myelofibrosis. Histol Histopathol 19, 239-250.

Van de Pette J.E., Prochazka A.V., Pearson T.C., Singh A.K., DicksonE.R. and Wetherley-Mein G. (1986). Primary thrombocythaemiatreated with busulphan. Br. J. Haematol. 62, 229-237.

Vardiman J.W., Pierre R., Brunning R.D., Harris N.L., Thiele J., ImbertM., Flandrin G. and Bain B. (2001). Chronic myeloproliferativediseases. In: WHO classif ication of tumours: Tumours ofhaematopoietic and lymphoid tissues. Jaffe E.S., Harris N.L., SteinH. and Vardiman J.W. (eds). IARC Press. Lyon. pp 9-38.

Westwood N.B. and Pearson T.C. (1996). Diagnostic applications ofhaemopoietic progenitor culture techniques in polycythaemias andthrombocythaemias. Leuk. Lymphoma 22 Suppl 1, 95-103.

Yan L., Elkassar N., Gardin C. and Briere J. (1996). Clonality assaysand megakaryocyte culture techniques in essentialthrombocythemia. Leuk. Lymphoma 22 Suppl 1, 31-40.

Yonemitsu H. and Okuda K. (1985). Essential thrombocythemiaterminating in myelofibrosis and myeloblastic transformation. Jpn. J.Med. 24, 57-61.

Accepted December 20, 2004

644


standardization of bone marrow features - does it work in ... · neoplastic conditions. this fact...

Documents