CD79b expression in B cell chronic lymphocytic leukemia: its ... - Nature

Leukemia (1999) 13, 1501–1505  1999 Stockton Press All rights reserved 0887-6924/99 $15.00 http://www.stockton-press.co.uk/leu

CD79b expression in B cell chronic lymphocytic leukemia: its implication for minimal residual disease detection JA Garcia Vela, I Delgado, L Benito, MC Monteserin, L Garcia Alonso, N Somolinos, MA Andreu and F Onã Department of Hematology, Hospital Universitario de Getafe, Madrid, Spain

The surface expression of CD79b, using the monoclonal antibody (Mab) CB3–1, on B lymphocytes from normal individuals and patients with B cell chronic lymphocytic leukemia (CLL) has been analyzed using triple-staining cells for flow cytometry. In addition, the clinical significance of CD79b expression in CLL patients and its possible value for the evaluation of minimal residual disease (MRD) was explored. A total of 15 peripheral blood (PB) samples from healthy blood donors, five bone marrow (BM) samples from normal donors and 40 PB samples from CLL untreated patients were included in the study. In addition we studied the expression of CD79b in B lymphocytes from five CLL patients after fludarabine treatment in order to support our method. The expression of CD79b in B lymphocytes from PB was analyzed by flow cytometry, using simultaneous staining with the Mabs CD22, CD79b, CD19 and CD5, CD79b and CD19. Since normal immature bone marrow B cells are CD79b−/dim+ on their surface, in BM samples we used the combination CD45, CD79b and CD19 selecting mature B lymphocytes according to their bright CD45 intensity. Cell acquisition was performed in two consecutive steps using a live gate drawn on SSC/CD19+ cells. For data analysis, the PAINT-AGATE PRO software (Becton Dickinson) was used. Dilution experiments of CD79b− CLL cells and CD79bdim+ CLL cells with normal PB and BM cells were performed in order to assess the sensitivity level of the technique for detection of CD79b−/dim+ residual CLL cells. All B lymphocytes from normal samples showed reactivity for the CD79b antigen. In contrast, CD79b was absent in 18/40 CLL patients (42.5%) and 20/40 CLL cases (50%) exhibited a low CD79b expression. Therefore, CD79b− B lymphocytes would be restricted to the CLL population and thus could be considered a ‘tumor phenotype’ for monitoring MRD in CLL patients. Dilution experiments indicate that the detection limit with this marker almost reaches the levels obtained by molecular biology methods as the PCR technique. All cases studied after fludarabine presented leukemic cells in their PB or BM samples detected by flow cytometry. Upon comparing the clinical and morphological characteristics of CD79b− and CD79b+ cases, all atypical CLL cases included in the present study were CD79b+ and advanced clinical stage (B and C Binet stage) was most frequently observed in CD79b+ cases than in CD79b− cases. Keywords: CD79b; CLL; minimal residual disease; immunophenotype

Introduction It has been shown that immunophenotypic analysis is a useful approach for the investigation of minimal residual disease (MRD) in acute leukemia patients.1 At present this approach has not been used in B cell chronic lymphocytic leukemia (CLL) for MRD detection, with the exception of the kappa/lambda ratio but the sensitivity of this parameter is low (10−2).1 We and others2–4 have explored the incidence of phenotypic aberrances in CLL (maturational asynchronous

Correspondence: JA Garcia Vela, Department of Hematology, Hospital Universitario de Getafe, Carretera de Toledo KM 12,500, Getafe 28905, Madrid, Spain; Fax: 34 91 6833541 Received 7 July 1998; accepted 25 May 1999

antigen expression and antigen overexpression) in order to establish the applicability of immunophenotypic aberrances for monitoring MRD as in acute leukemias with a sensitivity level of 10−4 (one aberrant CLL cell among 10 000 normal cells). We have previously published that CD5 is overexpressed in most CLL cases.2 This aberrantly CD5high/CD19+ expression was present in 90% of our CLL. Dilutional experiments showed that CD5high/CD19+ were identified at frequencies as low as 10−4. Recently, different groups have communicated that CD79b (B cell receptor (BCR), B29 protein) is undetectable in the surface of most CLL cells.5,6 This aberrant phenotype (all CD19+ mature lymphocytes express CD79b) has not been studied for monitoring MRD in CLL. The aim of our group was to study the expression of CD79b in CLL, to estimate quantitatively the number of molecules per cell, and finally to explore if this antigen could be used in the detection of MRD by flow cytometry after treatment.

Materials and methods

Patients and samples A total of 20 healthy donors: 15 peripheral blood (PB) samples and five bone marrow (BM) samples and 40 PB samples from untreated CLL patients were included in the present study. PB and BM aspirates were available for review in all patients, bone marrow trephines in 33 patients and lymph node aspirates or biopsies in 13 cases. All slides were reviewed and classified according to both the REAL classification and the FAB cooperative group criteria.7,8 The diagnosis of CLL was confirmed by immunophenotype using a large panel of monoclonal antibodies (Mabs) with a direct immunofluorescence technique: FMC7, CD5, CD10, CD11c, CD19, CD20, CD22, CD23, CD25, CD43, CD103 and polyclonal anti-␬ and anti␭ surface light chains (slg) to determinate B cell clonal excess. A typical CLL immunophenotype was defined as CD5+/ CD23+/CD19+/CD22−/dim+/CD20dim+/CD10−/CD103−/FMC7−/dim+ and slgdim. CLL was further subdivided into typical and atypical CLL on morphology without considering the immunophenotype.9 Atypical CLL included cases with more than 10% circulating prolymphocytes (CLL/PL) and others with more than 15% lymphocytes with lymphoplasmacytic features and/or cleaved nuclei. According to the REAL classification and FAB criteria, all cases included were classified as CLL. There were no cases with mantle zone lymphoma, follicle center cell lymphoma, lymphoplasmacytoid lymphoma and splenic marginal zone B cell lymphoma. Six out of 40 CLL cases displayed atypical lymphocyte morphology in the blood film. The mean age of normal blood donors (eight males, seven females) was 49.5 ± 12.9 years and 68.2 ± 11.3 for CLL patients. According to the Binet clinical staging classi-

CD79b expression in CLL JA Garcia Vela et al

1502

fication,10 CLL patients were distributed as follows: A 67.5%, B 17.5% and C 15%. Five patients were studied prior to treatment and at the time of maximal response after fludarabine treatment which was defined as: CR, nodular PR (nPR) or PR. Criteria for CR were normal physical examination, absence of constitutional symptoms, PB lymphocyte count ⬍4 × 109/l, platelets ⬎100 × 109/l, Hb ⬎11 g/dl and ⬍30% lymphoid cells in the BM aspirate with normal histology. The presence of residual nodular or interstitial lymphoid aggregates in the BM was considered as nPR. The presence of more than 30% lymphocytes in the BM aspirate after normalization of all other parameters was defined as PR. Five ml of EDTA-anticoagulated PB or BM samples were obtained in all cases after informed consent according to the Hospital Ethical Committee.

Immunophenotypical studies In all cases unseparated PB samples (1–2 × 106/tube) were stained with a simultaneous triple labeling using the Mabs CD5 and CD22 conjugated with fluorescein isothiocyanate (FITC), CD79b conjugated with phycoerythrin (PE) and CD19 conjugated with the PE/Cyanin5 fluorochrome tandem (Tricolor). Since normal immature BM B cells are CD79b−/dim+ as CLL cells, in all BM samples we used the combination CD19-FITC/CD79b-PE/CD45-PerCP. Immature B cells (CD45−/dim+) were separate from mature B lymphocytes (CD45bright) according to the different CD45 expression.11 The CD5, CD19, CD22 and CD45 Mabs were purchased from Becton Dickinson (San Jose´, CA, USA), CD79b (CB3-1) from Immunotech (Marseille, France) and CD19-PE/Cy5 from Caltag Laboratories (San Francisco, CA, USA). Samples were incubated for 15 min at room temperature in the dark with saturating amounts of the fluorochrome-conjugated Mab reagents. Afterwards, 2 ml of FACS lysing solution (Becton Dickinson) was added and the samples incubated for 10 min in the same conditions. Then the samples were centrifuged for 15 min at 500 g and the cell pellets washed again with 5 ml of phosphate-buffered saline (PBS). Finally the cells were resuspended in 0.5 ml of PBS for analysis by flow cytometry. An isotype-matched negative control (Becton Dickinson) was used in all cases to assess background fluorescence intensity. Data acquisition was performed on a FACScan flow cytometer (Becton Dickinson) equipped with an argon ion laser tuned at 488 nm and 15 mW. Acquisition was performed in two consecutive steps: in the first step a total of 10 000 events/tube was acquired, and in a second step, cells were acquired through a ‘live gate’ drawn on the side-scatter (SSC)/CD19+ cells. The total number of nucleated cells acquired depended on the percentage of leukemic cells in the sample. We acquired 10 000, 100 000 and more than 500 000 nucleated cells when the percentages of leukemic cells were 1%, 0.1% and ⬍0.1%, respectively. In BM samples in the analysis of the ‘live gate’ SSC/CD19+ we selected those events with CD45bright intensity. For data analysis, the PAINT-A-GATE PRO software (Becton Dickinson) was used.

Assessment of fluorescence intensity (MESF) The mean fluorescence intensity was evaluated by the number of molecular equivalents of soluble fluorochrome (MESF), obtained by means of the QUICKCAL beads (Flow Cytometry,

San Juan, PR). Each marker was considered to be positive when the mean MESF obtained was at least double that observed in the isotype-negative control tube.

Dilution experiments Dilution experiments were performed in order to assess the sensitivity level of this technique for the detection of CD79b−/dim+/CD19+ leukemic cells. For this purpose, progressive dilutions of a CLL sample with normal PB and BM were carried out at ratios of up to one CLL cell in 106 normal cells prior to staining with Mabs as previously described.

Statistical analysis The Mann–Whitney test was used to compare the MESF value of CD79b among the two groups, CLL and normal donors, and to demonstrate the correlation between the expression of CD79b and the other markers analyzed in the present study. Results The mean frequency of B lymphocytes CD19+ in the normal PB samples analyzed was 8.4 ± 4.0% (range: 4.3–14%) of lymphocytes, whereas in CLL PB samples it was 74.09 ± 16.68% (range: 34–98%). The overall proportion of CD19+ cells in the normal BM samples was 3.14 ± 1.68% (range: 0.7–6.4%) and 70.8 ± 14.3% of them showed a bright expression of CD45. In normal PB, B lymphocytes expressing CD5 represented 2.5 ± 1.4% of lymphocytes (MESF 6.6 ± 0.4 × 103). In the CLL group, all cases were CD5+ and it was expressed in practically all CD19+ cells (MESF 78.4 ± 43.1 × 103); 24 out of 40 cases (60%) showed a weak expression of CD22 (MESF 7.5 ± 2.8 × 103). Regarding the expression of CD79b, normal mature B lymphocytes from the 15 PB and five BM samples obtained from healthy donors were positive for the CD79b antigen (MESF 36.5 ± 4.6 × 103). In BM samples, B cells CD45−/dim+ (immature B cells) were CD79−/dim+ (Figure 1). When we analyzed the expression of CD79b in CD19+ cells regarding the expression of CD5 we did not find any difference between CD5+ and CD5−B lymphocytes (Table 1). On the other hand, 17 out of 40 CLL samples studied (42.5%) were CD79b negative (MESF ⬍1.5 × 103); 20 out of 40 cases (50%) exhibited low CD79b expression (MESF 8.4 ± 2.8 × 103) (P ⬍ 0.001) and only three cases (7.5%) showed a normal expression of CD79b (MESF 31.9 ± 6.4 × 103) as in normal B lymphocytes (Figure 2). These three cases were CD20dim+/CD22dim+/CD23+/CD10−/CD103− and FMC7−/dim+. When considering a cut-off point of 15 × 103 molecules per cell, 92.5% of the CLL samples analyzed in the present study had CD79b values below this level. Advanced clinical stage (B and C Binet stage) was observed most frequently in CD79b+ cases (43.4%) than in CD79b− cases (17.6%) (P ⬍ 0.01). All cases with atypical morphology were CD79b+. A comparison of the CD79b and other B cell markers showed a correlation between CD79b positivity and strong expression of sIg (P = 0.004) but there was no correlation between the levels of CD5 and CD79b nor of CD22 and CD79b. Dilution experiments showed that CD79b− CLL cells were recognized and distinguished from normal PB CD19+ lymphocytes and normal BM mature B cells (CD45bright), even when


1503

Figure 1 CD79b expression in B cells from a normal BM sample. B precursors (blue color) showed a lower SSC and CD45 expression than mature B lymphocytes (red color) and were CD79b−/dim+. Table 1 CD79b-PE MESF value in normal B lymphocytes CD19+ regarding the expression of CD5. No differences were seen between CD19+CD5+ and CD19+CD5− B lymphocytes

CD19+

CD19+/CD5+

CD19+/CD5−

% of lymphocytes

8.4 ± 4.0

2.5 ± 1.4

6.1 ± 3.9

CD79b MESF (PE)

32.8 ± 3.0

36.5 ± 5.5

34.7 ± 4.7

MESF, molecular equivalents of soluble fluorochrome expressed in 103.

diluted at frequencies as low as 10−4 (Figure 3). Dilution experiments with CLL cells with normal CD79b expression were not able to differentiate it from normal B lymphocytes. We analyzed the expression of CD79b in B lymphocytes from PB and BM in five CLL patients after six courses of fludarabine to support this analysis. Three of five cases were in PR; 1/5 in nPR and 1/5 in CR at the moment of the flow cytometry study. All cases, including the patient considered in CR by bone marrow biopsy, did have leukemic cells (CD19+ with low or undetectable expression of CD79b) detectable by flow cytometry (17.3 ± 7.2% of lymphocytes). With a mean followup of 17 months, all patients relapsed.

Figure 2 Immunocytometric analysis of surface CD79b expression on normal and CLL samples. Surface CD79b expression was determined on PB from a normal individual (plot a, analysis of CD79b in a lymphocyte gated region) with normal CD79b reactivity and from three representative CLL patients. (b) CLL patient with B lymphocytes CD5+ and CD79b negative; (c) CLL patient with low level of CD79b and (d) CLL case with apparently normal surface CD79b expression (plots b to d represent ungated events).


1504

Figure 3 Dilution experiments were able to detect one tumoral cell (CD79b negative/CD22 negative/CD19+) in 104 normal peripheral blood cells (0.01%). Acquisition was performed using a ‘live gate’ in SSC/CD19+ cells as described in the text.

Discussion Surface Ig on B cells is complexed with heterodimers of two critical accessory proteins called B29 (Ig␤ and CD79b) and mb-1 (Ig␣ and CD79a) in the B cell antigen receptor (BCR).12 The presence of perturbation or alterations of either of these genes could prevent surface Ig expression characteristic of CLL cells.5 In the present study, using a triple-staining procedure and a highly sensitive method for the detection of antigen expression (flow cytometry combined with PE-conjugated monoclonal antibody and a double-step acquisition procedure), we have demonstrated that 92.5% of the CLL samples analyzed in our study had low or negative surface B29 protein that directly correlated with the level of surface Ig expression. Strikingly different results of B29 expression were reported recently in chronic B cell diseases using the SN8 Mab.6 SN8 was produced by immunizing mice with an antigen preparation from human B-prolymphocytic leukemia cells. It identifies an epitope on the extracellular part of the B29 protein. Because only 5% of their CLL cases had detectable surface expression of B29, they concluded that SN8 could be a useful marker for non-CLL B cell lymphoproliferative disorders.13 The CB3-1 Mab used by us and Thompson et al5 also recognizes an extracellular epitope. Its clear specificity for human B29 that can be detected throughout B-lymphocyte differentiation would perhaps make it a more selective antibody for studies of B29 expression.14 42.5% of our CLL cases were CD79b− and 50% had a low level of expression of this marker (less than 15 × 103 molecules per cell). In contrast mature B lymphocytes from normal blood and BM donors were constantly CD79b+ with a level of expression higher than 25 × 103 molecules per cell. Therefore this marker would appear to be absent or underexpressed in CLL cells, and thus could be considered as a ‘tumor

phenotype’ associated with CLL. In this sense, it would be a useful phenotype for monitoring residual disease in CLL patients. Dilution experiments, in which CD79b− CLL cells were diluted with normal PB and BM cells, enabled us to detect tumoral cells with a sensitivity level of 10−4. Since normal immature BM B cells are CD79b−/dim+ in the analysis of the ‘live gate’ drawn on the SSC/CD19+ cells we selected those events with a bright expression of CD45 (mature B lymphocytes with an expression of CD79b identical to PB B lymphocytes). The sensitivity level obtained by flow cytometry to detect residual CLL cells is favorable to the morphology approach (1%) and almost reaches the levels obtained by molecular biology methods, such as the PCR technique.15 In addition quantitative studies may be useful in samples with a low percentage of CD19+ lymphocytes, and the different intensity of expression of CD79b in normal B lymphocytes and CLL cells can be exploited to identify both populations separately. There have been reports investigating quantitatively the expression of several antigens in CLL3,16,17 but this is the first report investigating quantitatively the expression of CD79b in normal B lymphocytes and CLL cells. This study demonstrated that flow cytometry contributes to increase the sensitivity of the clinicohematological criteria to detect residual malignant cells in CLL patients and may be useful to monitor disease status and therefore to assess the effectiveness of different therapeutic strategies as allogeneic, autologous bone marrow transplantation or purine analogues. This method may also be useful to evaluate the presence of MRD in the peripheral blood stem cell harvests of CLL patients and to evaluate the different methods of purging. In our patients, we were able to detect leukemic cells after fludarabine treatment, including the case considered in CR. All patients with persistence of MRD assessed by flow cytometry relapsed. Cabezudo et al3 showed that the persistence of MRD detected by flow cytometry in their patients could be corre-


Figure 4 PB lymphocytes in a CLL patient 2 months after the sixth course of fludarabine. Note an increased number of non-clonal CD5+ B cells. The ratio CD5+CD19+/CD19+ was ⬎0.25 but all circulating B lymphocytes were polyclonal.

lated with progression. They considered patients in immunological remission when the ratio of CD5+/CD19+/CD19+ cells was ⬍0.25 in PB and ⬍0.15 in BM and the density of expression of both antigens was normal. Recently, we and others18 have observed an increased number of non-clonal CD5+ B cells in PB of CLL patients after purine analogue therapy (Figure 4). These cells may represent normal B-1a cells and are seen in the PB as the BM regenerates after the chemotherapy. The presence of these cells can complicate the monitoring of residual disease by analysis of the ratio CD5+CD19+/CD19+. It would be interesting to compare this method with that proposed in the present study. Finally, Thompson et al5 hypothesized that mutations or other alterations affecting B29 gene expression or function produce cells with diminished surface BCR that are unresponsive to antigen binding and therefore unable to initiate either growth stimulatory or apoptotic signal cascades. A recent study showed that BCR-signaled apoptosis of B cells required both B29 and mb-1.19 The B29 gene alterations that result in reduced CD79b and Ig expression may have prognostic significance because normal surface Ig is correlated with poor survival in CLL.20 In this context, although the follow-up of our patients is still relatively short for a definitive evaluation of the prognostic impact of CD79b expression in CLL, our CD79b+ cases had more frequently an advanced clinical stage and an atypical morphology than the CD79b− group. Acknowledgements This work was in part supported by a grant from the Fundacioń Espanõla de Hematologia y Hemoterapia, premio Ernst Schering. References 1 Jennings CD, Foon KA. Recent advances in flow cytometry: application to the diagnosis of hematologic malignancy. Blood 1997; 90: 2863–2892. 2 Garcıá Vela JA, Delgado I, Garcıá Alonso L, Monteserin MC, Benito L, Onã F, Lastra AM. Detection of minimal residual disease in B cell chronic lymphocytic leukemia by flow cytometry. Br J Haematol 1997; 99: 464–465. 3 Cabezudo E, Matutes E, Ramrattan M, Morilla R, Catovsky D. Analysis of residual disease in chronic lymphocytic leukemia by flow cytometry. Leukemia 1997; 11: 1909–1914.

4 Sańchez-Guijo FM, Sańchez ML, Almeida J, Vallejo JA, Lo´pezBerges MC, Fuentes M, Fernańdez Calvo J, Martıńez MA, Ba´rez A, Casanueva F, Orfao A, San Miguel JF. Incidencia de aberraciones fenotı´picas en sıńdromes linfoproliferativos B leucemizados: ana´lisis de 239 casos. Haematologica 1997; 82 (Suppl. 2): 25. 5 Thompson AA, Talley JA, Do HN, Kagah HL, Kunkel L, Berenson J, Cooper MD, Sazon A, Wall R. Aberrations of the B cell receptor B29 (CD79b) gene in chronic lymphocytic leukemia. Blood 1997; 90: 1387–1394. 6 Zomas AP, Matutes E, Morilla R, Owusu-Ankomah K, Seon BK, Catovsky D. Expression of the immunoglobulin-associated protein B29 in B cell disorders with the monoclonal antibody SN8 (CD79b). Leukemia 1996; 10: 1966–1970. 7 Harris NL, Jaffe ES, Stein H, Banks PM, Chan JKC, Cleary ML, Delsol G, De Wolf-Peeters C, Falini B, Gatter KC, Grogan TM, Isaacson PG, Knowles DM, Mason DY, Muller-Hermelink H-K, Pileri SA, Piris MA, Ralfkiaer E, Warnke RA. A revised European– American classification of lymphoid neoplasm: a proposal from the International Lymphoma Study Group. Blood 1994; 84: 1361–1392. 8 Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DAG, Gralnick HRT, Sultan C. Proposals for the classification of chronic (mature) B and T lymphoid leukemias. J Clin Pathol 1989; 42: 567–584. 9 Matutes E, Oscier DG, Garcıá Marco J, Ellis J, Copplestone A, Gillinghan R, Hamblin T, Lens D, Swansbury GJ, Catovsky D. Trisomy 12 defines a group of CLL with atypical morphology: correlation between cytogenetics, clinical and laboratory features in 544 patients. Br J Haematol 1996; 92: 382–388. 10 Binet J, Auquier A, Dighiero G. A new prognostic classification of chronic lymphocytic leukemia derived from a multivariance survival analysis. Cancer 1981; 48: 198–206. 11 Ciudad J, Orfao A, Vidriales B, Macedo A, Martıńez A, Gonzalez M, Lo´pez-Berges MC, Valverde B, San Miguel JF. Immunophenotypic analysis of CD19+ precursors in normal human adult bone marrow: implications for minimal residual disease detection. Haematologica 1998; 83: 1069–1075. 12 Clark MR, Campbell KS, Kazlauskas A, Johnson SA, Hertz M, Potter TA, Pleiman C, Cambier JC. The B cell antigen receptor complex: association of Ig-alpha and Ig-beta with distinct cytoplasmic effectors. Science 1992; 258: 123–126. 13 Moreau EJ, Matutes E, A’hern RP, Morilla AM, Morilla R, OwusuAnkomah KA, Seon BK, Catovsky D. Improvement of the chronic lymphocytic leukemia scoring system with the monoclonal antibody SN8 (CD79b). Am J Clin Pathol 1997; 108: 378–382. 14 Nakamura T, Kubagawa H, Cooper MD. Heterogeneity of immunoglobulin associated molecules on human B cells identified by monoclonal antibodies. Proc Natl Acad Sci USA 1992; 89: 8522–8526. 15 Provan D, Bartlett-Pandite L, Zwicky C, Neuberg D, Maddocks A, Corradini P, Soiffer R, Ritz J, Nadler LM, Gribben JG. Eradication of polymerase chain reaction-detectable chronic lymphocytic leukemia cells is associated with improved outcome after bone marrow transplantation. Blood 1996; 88: 2228–2235. 16 Witzig TE, Li CY, Tefferi A, Katzmann JA. Measurement of the intensity of cell surface antigen expression in B cell chronic lymphocytic leukemia. Am J Clin Pathol 1994; 101: 312–317. 17 Poncelet P, Lavabre-Bertrand T, Carayon P. Quantitative phenotypes of B chronic lymphocytic leukemia B cells established with monoclonal antibodies from the B cell protocol. In: Reinhertz EL (ed.). Leukocyte Typing II. Springer-Verlag: New York, 1986, pp 329–341. 18 Howe D, Bromidge T, Johnson S, Rule S. Increased numbers of non-clonal CD5 positive B-cells in the peripheral blood of lymphoma patients after purine analogue therapy. Br J Haematol 1997; 97 (Suppl. 1): 29. 19 Tseng J, Eisfelder BJ, Clark MR. B-cell antigen receptor-induced apoptosis requires Ig-␣ and Ig-␤. Blood 1997; 89: 1513–1520. 20 Geisler CH, Larsen JK, Hansen EN, Hansen MM, Christensen BE, Lund B, Nielsen H, Plesner T, Thorling K, Andersen E, Andersen PK. Phenotypic importance of flow cytometric immunophenotyping of 540 consecutive patients with B cell chronic lymphocytic leukemia. Blood 1991; 78: 1795–1802.

1505