British Iournal of Haernatologu. 1992, 81, 40-44

Surface marker expression in acute myeloid leukaemia at first relapse

XAVIER THOMAS,' '2 LYDIA CAMPOS, ' ERIC ARCHIMBAUD,' ZHI-HUA S H I , ' DANIBLE TREILLE-RITOUET,3

BRUNO ANCLARET' AND DENIS FIERE' 'Service de Cryobiologie, Centre de Transfusion Sanguine, Lyon, and 'Service d'Hematologie, and 3Laboratoire de Cy tologie, Hdpital Edouard Herriot, Lyon, France

Received 24 October 1991; accepted for publication 22 January 2992

Summary. Surface markers were studied at first relapse in 66 cases of acute myeloid leukaemia (AML), using a panel of five monoclonal antibodies directed to CD13. CD14. CDl5. CD33 and CD34 antigens. At time of relapse, there was increased expression ofCD33 (P=0*002) and CD34 (P=0*0001), and decreased expression of CD13 (P=0.004) and CD15 (P= 0.0001) antigens by comparison to initial diagnosis. There was no strict correlation with the FAB classification. However, CD13 and CD33 expression changes preferentially affected granulocytic leukaemias. At relapse, CD 14 and

CD34 were significantly more expressed in monocytic than in granulocytic AML (P=0.01 and 0.003 respectively). In a multivariate analysis, CD34 expression was associated with a low CR rate (P=0.001) and short survival (P=0.05), whereas CDl5 expression was associated with long survival (P= 0.0004). These results suggest that AML tends to relapse with a less differentiated phenotype than observed at diagno- sis and that AML with less differentiated phenotype is of poor prognosis after first relapse, as also observed at diagnosis.

Complete remission (CR) is achieved in 60-80% of previously untreated adult patients with acute myeloid leukaemia (AML) after intensive induction chemotherapy. However, up to 80% of the patients will ultimately relapse (Champlin & Gale, 1987). The prognostic value of initial characteristics has been often studied in patients with newly diagnosed AML, but little is known about the prognostic factors at relapse (Smits et al. 1987: Uhlman et al, 1990). A subset of patients in first relapse can still achieve long-term second CR when treated with autologous bone marrow transplantation. (Yeager et al, 1986: Ball et al, 1990) or chemotherapy (Archimbaud et al, 1991). Therefore the description of prognostic factors at relapse is important to determine which patients should be treated with these methods and which should be oriented to investigational therapies. Numerous differentiation antigens have been described in the myeloid lineage (Drexler & Minowada, 1986: Compos et al, 1987). The prognostic value of surface marker expression in AML at diagnosis is well known (Grif6n et al, 1986; Campos et al, 1989: Merle-Beral et al, 1989: San Miguel et al. 1989: Schwarzinger et al, 1990: Tucker et al. 1990). However, no study has been published so far on their clinical and biological relevance and potential prognostic value at relapse. The aim of this study was to compare cell surface markers analysis at

Correspondence: Dr Eric Archimbaud, Service d'Hematologie. HBpi- tal Edouard-Herriot, 69437 Lyon Cedex 0 3 , France.

diagnosis and at first relapse, and to investigate the potential relationships between surface myeloid antigen expression at relapse, and characteristics of AML and outcome of therapy at this time.

PATIENTS AND METHODS

Patient population Patient characteristics at initial diagnosis and at relapse.

Between March 1982 and November 1990,66 patients with AML diagnosed according to the revised French-American- British (FAB) Group criteria (Bennett et al. 198 5) had surface markers prospectively analysed at initial diagnosis and at first relapse. The main clinical and biological characteristics of this population at diagnosis and first relapse are shown in Table I. There were 34 males and 32 females. The median age of the patients was 57 years at presentation (range 17-81 years) and 58 years at relapse (range 18-82 years). Pretreat- ment parameters did not differ between diagnosis and first relapse except for a higher haemoglobin level (P< O-OOOl) , higher percentage of blood neutrophils (P=0.006) and decreased percentage of blood blast cells ( P = 0.05) at relapse. According to the French-American-British (FAB) criteria (Bennett et al, 1985), 12 patients were classified initially as M1, 14 as M2. eight as M3. 13 as M4, 15 as M5, one as M6 and three as undifferentiated acute leukaemia. At the time of first relapse, no obvious morphologic change of the blast cells

40

Surface Markers in Relapsed AML 41 Table 1. Characteristics of AML at diagnosis and at first relapse (66 patients)

Characteristic Diagnosis Relapse

Clinical parameters Age (years) Sex (male/female) Previous neoplasia (yes/no) Tumoural syndrome (yesJno)

Blood parameters Haemoglobin (g/l) WBC count ( 1 Oq/l) Granulocytes (x) Blasts ('X,) Platelets ( loy/l) LDH (iu/l) Creatinine (pmoI/l)

Bone marrow parameters Blasts (x) FAB subtype

( M l I M L I M 3/M4/MS/Mb/Other)

5 4 f 1 6 (66)* 5 5 f 1 6 (66) 3 5 / 3 1 9/57

18/47 10/56

9 2 f 1 9 (65) l l l f 2 4 (66) 4 2 f h 4 (66) 3 2 f 6 4 (66) 1 S f 1 5 (66) 2 2 f 2 0 (65) 4 5 f 36 (65) 3 8 f 3 7 (65) 7 5 f 6 4 (66) 61 f46 (66 )

5 3 7 f 364 (63) 5 9 4 f 8 3 6 (58) 8 6 f 2 3 (66) 8 9 f 2 9 (64)

7 3 f 2 0 ( 6 6 ) 6 9 f 2 7 (66) 12/14/8/13 / I 5/ 1 J 3 1211 4/8/13/1 511 /3

* Values are expressed as mean f SD (number of patients studied) unless otherwise indicated.

was noted in any of the patients by comparison with diagnosis.

Treatrnerrf received by tlic patients. At diagnosis, all patients received intensive cytarabine and daunorubicin-based induc- tion chemotherapy according to already described regimens (Archimbaud et al. 1987a; Zittoun et al. 1989) or EORTC- AML9 protocol, including daunorubicin 30 mg/m2/d or mitoxantrone (MTX) 8 mg/m'/d for 3 d. and cytarabine 100 mg/m2/d for 7 d. Following achievement of first CR. two patients aged less than 40 years received allogeneic bone marrow transplantation (BMT) in early first CR. Patients aged less than 60 years received either consolidation followed by maintenance (Archimbaud et a / , 1987a: Zittoun et al, 1989). or intensive consolidation with high-dose cytarabine. 3 g/m'/ 12 h from day 1 to day 4 and amsacrine, 9 0 mg/m*/d from day 5 to day 7, followed by repetition of induction therapy. Patients aged 60 years or more received mainte- nance therapy alone, generally using low-dose cytarabine (Archimbaud et 01. 1986). At the time of relapse, seven patients died before any therapeutic attempt and 59 received reinduction chemotherapy. Two patients received daunoru- bicin and cytarabine-based timed-sequential therapy (Fiere et a / , 1986). Forty-five patients were treated with mitoxan- trone, etoposide and cytarabine-based timed-sequential ther- apy (Archimbaud et al. 1991). Four patients aged more than 60 years received plicamycin and hydroxyurea (Archimbaud rt al, 1989). Since 1989, seven patients over 6 0 years were treated with daunorubicin 30 mg/mz/d from day 1 to day 4 and carboplatin 200 mg/m2/d from day 3 to day 7 followed by maintenance using reduced dosages of the same drugs. One patient relapsing after allogeneic bone marrow trans- plantation directly received a second transplantation. Four patients received allogeneic bone marrow transplantation while in second CR.

Results oltherupy afferfirst relapse. Complete remission (CR) and relapse were defined according to the CALGB criteria (Ellison et al, 1968). Treatment failures were classified as resistant disease and other failures according to Preisler (1978). Among the 66 patients included in this study, the median time to first relapse was 9 months with a range of 1- 3 8 months. Among the 59 patients receiving chemotherapy while in relapse, 30 ( 5 1 %) achieved a second CR. Six patients died from toxicity of reinduction chemotherapy and 2 3 had relative drug resistance. The median follow-up for the entire cohort after first relapse was 2 6 months. The median second CR duration was 6 months (range 0.5-46+ months). A second relapse occurred in 1 7 cases during the time of the study and four patients died while in second CR from post- remission chemotherapy-related toxicity. Median survival for the 6 6 patients after relapse was 4.5 months (range 0-47+ months). Fifteen patients are surviving at the time of last follow-up.

lmmunofluorescence analysis Bone marrow cell suspensions were obtained by centrifuga- tion of bone marrow samples on Ficoll-Hypaque gradients and resuspended in phosphate-buffered saline (PBS). In most cases the surface marker analysis was performed immedi- ately after cell separation. In other cases, the cells were cryopreserved in 20% fetal calf serum and 10% dimethylsul- foxide solution, and the immunological study was performed after quick thawing. Surface markers were analysed by indirect immunofluorescence as previously described (Cam- pos et a!, 1987, 1989). The monoclonal antibodies (MoAb) used were: MY7 (Coulter Immunology, Hialeah, Fla.) dir- ected to CDl3, CRIS-6 (gift from Dr Vilella. Barcelona. Spain) directed to CD14. VIMD5 (gift from Dr Knapp. Institut fur Immunologie, Vienna, Austria) directed to CDl5, MY9

42 Xavier Thomas et al Table 11. Expression of myeloid surface markers at first relapse compared to expression at diagnosis

No. of P value Marker cases Relapse Diagnosis t-test

CD13 65 30f22 (29/37)’ 43+28 (16/49) 0.004 CD14 66 19f18 (45/21) 16f2O (43/23) NS CD15 66 32+22 (27/39) 4 8 f 2 6 (9/57) 0.0001 CD33 6 5 3 3 f 1 9 (19/47) 2 2 f 2 3 (37/28) 0.002 CD34 66 34f25 (27/39) 21*22 (43/33) 0.0001

* Mean fSD of percentage of leukaemic cells stained (no. of

NS: non-significant (P value >0.05). negative/positive cases).

(Coulter Immunology) directed to CD33, and MY10 (Becton Dickinson. Mountain View, Calif.) directed to CD34. Human AB serum was added to the cell suspension to avoid non- specific fixation of the MoAb on Fc receptors. Cells were incubated with MoAb at 4OC for 30 min and washed with PBS. Fluorescein-labelled goat anti-mouse Fab’2 fragments (Bioart, Meudon, France) were added as a second layer (4°C 30 min). Cell fluorescence was assessed by using a Leitz microscope with epi-illumination. Controls were performed with a non-reactive antibody (MsIgG Kit or MsIgM Kit, Coulter Immunology). Results were considered positive when the percentage of stained cells exceeded that of the control by at least 20%. In tests performed after 1988. it was checked that this method gave results similar to flowcytometric immunophenotyping for the MoAbs used.

Statistical analysis Comparisons between immunophenotypes at initial presen- tation and at first relapse were made by the paired Student’s t-test. Correlations between the expression of different markers were studied using the non-parametric Spearman rank correlation test. The relationships between expression of surface antigens (considered as positive or negative) to quantitative characteristics at first relapse were studied by the Student’s t-test. A logarithmic transformation of the data was performed when necessary. The relationships to qualita- tive parameters were studied by the chi-square test or the two-tailed Fisher’s exact test. For analysis of overall survival and disease-free survival (DFS) following fist relapse, patients receiving allogeneic BMT in second CR were cen- sored at the time of transplantation. Survival and CR durations were estimated by the Kaplan-Meier method. Multivariate analysis of surface markers potentially related to the outcome of chemotherapy at relapse, and DFS and overall survival were performed using stepwise logistic regression and Cox’s proportional hazard regression model respectively. All computations were performed using BMDP statistical programs (BMDP Statistical Software, Los Angeles, Calif.)

RESULTS

Surface marker expression at first relapse The expression of myeloid surface markers at relapse com-

pared to expression at first diagnosis is shown in Table 11. Overall, there was a significant positive correlation between expressions at first relapse and at diagnosis of CD14 (P=0.002), CD33 (P=0.04) and CD34 (P=0.0005) while there was no significant correlation between expressions of CD13 and CD15. Quantitatively, CD14 had low antigenic expression in this AML population and showed no significant modification between diagnosis and relapse. The four other markers tested expressed significant changes in antigenic expression at first relapse (Table II), with decreased ex- pression of CD13 (P=0*004) and CD15 (P=O*OOOl) and increased expression of CD33 (P=0.002) and CD34 (P= 0.0001) by comparison with initial diagnosis. Changes in expression of surface markers were studied with respect to FAB classification. The granulocytic FAB subtypes ( M l , M2 and M3) showed modified expression of CD13 (P=0.04), CDl5 (P=0.0002), CD33 (P=O.OOl ) and CD34 (P=0.03) at relapse. The subtypes with a monocytic population (M4 and M5) showed modified expression of CD15 (P=O.Ol) and CD34 (P=0.0007) at relapse. There was a positive correla- tion between the expressions of CD13 and CDl5 (P = 0.0001) and a negative correlation between the expressions of CD15 and CD34 (P<O.OOOl) and CD15 and CD33 (P=0.03) at relapse.

Relationships between surface marker expression and biological and clinical features at first relapse Significant differences in surface marker expression were observed between granulocytic and monocytic FAB subtypes. The number of cases stained by CD14 and CD34 was significantly higher in the latter than in the former (P=O.Ol and P=0.003 respectively). CD15 positivity was associated with lower than average platelet counts (P=0.0007) and CD34 positivity with high platelet counts (P=0.02). CD15 positivity at relapse was signficantly associated with long previous CR duration (P=O.Ol).

Surface marker expression at first relapse and prognosis Factors that were significantly associated with a low second CR rate in the multivariate analysis taking into account immunophenotype and other patient characteristics at relapse included CD34 positivity (P= 0.001 ) and anaemia (P=O.Ol). Factors predictive of short survival after first relapse included CD15 negativity (P = 0.0004). elevated serum LDH (P=0-005) and CD34 positivity (P=O.O5). Factors predictive for short DFS in patients who achieved a second CR included short duration of the first CR ( P = 0.001), and high white blood cell count at relapse (P=O.Ol) .

DISCUSSION

We compared the expression of five surface markers on AML, cells at diagnosis and at fist relapse and we studied its relationships to clinical and biological characteristics and outcome of reinduction chemotherapy. Immunophenotypic changes at first relapse included increased expression of CD34 and CD33 and decreased expression of CD13 and CD15. indicating a loss of differentiation at relapse, which

Surface Markers in Relapsed AML 43 was not predictable by standard morphologic examination in this series. In other reports, the capacity to mature has been lost to some degree, concurrently with the usually observed resistance to therapy at relapse (van der Weide et a / . 1985; van Rhenen et a/, 1983). Immunophenotype modifications showing loss of cell maturation capacity were sometimes concomitant with morphologic changes (van Rhenen et a/, 198 3 1. Relapsed AML cells are also known to have decreased differentiation ability in vitro compared to cells studied at diagnosis (Archimbaud et a / . 1987b: Baer et al, 1987). Clinically. the association of CD14 positivity with extrame- dullary disease. high leucocyte count and elevated serum LDH level found at diagnosis (Griffin eta/. 1986: Campos et a/, 1989) was not found in our study at first relapse, perhaps because tumoural mass was lower in our relapsed patients than usually found at diagnosis. At time of first relapse, the only correlations noted were between CD15 and CD34 expressions and platelet count, and CD14 expression and haemoglobin value. In previous reports on prognostic factors at relapse for adults with AML, duration of first remission was the strongest prognostic factor and correlated with ability to achieve a second remission (Smits et a / . 1987: Kantarjian et al, 1988: Uhlman et al, 1990). second remission duration (Uhlman et a/. 1990), and overall survival from reinduction (Uhlman et al. 1990). Smits et a1 (1987) also described age under 40 years as predictive for high second CR rate. Our results using multivariate analysis of prognositic factors at first relapse confirm the prognostic importance of the duration of first CR and show the usefulness of myeloid marker studies to establish a prognosis at this stage of the disease. Expression of CD34 was associated with decreased CR rate. Lack of expression of CD15 and expression of CD34 were predictive of short survival. These results confirmed the poor outcome, previously noted in studies performed at diagnosis, of CD34-positive patients (Campos et a / , 1989: Vaughan et a / , 1988; Borowitz et al, 1989). We and others (Campos et al, 1989: Holowiecki et al, 1986; Schwarzinger et a/. 1990) have reported that expression of CDl5 at diagnosis was associated with a favourable outcome. Prognostic value of CD14 expression at diagnosis, described by others (Griffin et a / , 1986) and in one previous study from our group including most of the patients presented here (Campos et al. 1989). was not confirmed at first relapse. Overall, the present study confirms that several surface markers that are of prognostic value at diagnosis retain their prognostic value at first relapse. In addition, we show a rise in initial immunolo- gical risk factors at first relapse in the same patients. These results are consistent with reported lower CR rate in leukaemia showing evidence of undifferentiation (Swirsky et a / , 1986).

Our data suggest that, as already described in newly diagnosed AML, surface marker analysis is of clinical rele- vance in patients with relapsing AML. Assessment of several predictive factors might lead to the identification of patients who are likely to respond favourably to reinduction therapy and patients who will probably resist conventional chemo- therapy, and could consequently be helped by risk-adapted therapy.

REFERENCES

Archimbaud. E.. Guyotat. D.. Sebban. C.. Fiere. D. & Viala, J.J. (1986) Acute myelogenous leukemia (AML) in elderly patients: intensive induction chemotherapy and maintenance with low-dose cytara- bine. Blood, 68, (Suppl. l), 216a.

Archimbaud, E.. Fiere. D.. Treille-Ritouet, D.. Adeleine. P., Guyotat, D.. Sebban. C. & Viala. J.1. (1987a) Influenceofdaunorubicin and cytarabine sequencing on the outcome of therapy in acute myelogenous leukemia: a randomized trial. Cancer Treatment Reports, 71, 571-574.

Archimbaud. E. & Preisler. H.D. (1987b) Differentiation of myeloid cells in liquid culture. 2. Growth and differentiation of acute myelocytic leukemia blasts in primary liquid culture. Cell Biochem- istry and Function. 5, 15 7-1 66.

Archimbaud. E., Troncy. J.. Sebban, C., Guyotat, D.. Devaux. Y.. Ffrench. M.. Moriceau. M.. Viala, J.J. & Fiere. D. (1989) Phase I1 trial of plicamycin and hydroxyurea in acute myelogenous leukemia. Cancer Chemotherapy and Pharmacology, 2 5, 223-225.

Archimbaud. E., Leblond, V., Michallet. M.. Cordonnier, C.. Fenaux. P.. Travade. P.. Dreyfus. F.. Jaubert, J., Devaux. Y. & Fiere. D. (1 991) Sequential therapy with mitoxantrone and continuous infusion etoposide and cytarabine for previously treated acute myelogenous leukemia. Blood, 77, 1894-1 900.

Baer. M.R.. Dessypris. E.N.. Stein, R.S.. Cousar. J.B. & Krantz, S.B. (1987) Variable differentiation patterns of acute myelogenous leukemia blasts in liquid suspension cultures. British journal o/ Hematology. 66, 179-185.

Ball, E.D.. Mills, L.E.. Cornwell. G.G.. 111. Davis, B.H.. Coughlin, C.T., Howell, A.L.. Stukel. T.A.. Dain. B.j., McMillan. R., Spruce, W., Miller, W.E. & Thompson, L. (1990) Autologous bone marrow transplantation for acute myeloid leukemia using monoclonal antibody-purged bone marrow. Blood, 75, 1199-1206.

Bennett, J.M.. Catovsky, D.. Daniel, M.T.. Flandrin, G.. Galton, D.A.G.. Gralnick, H.R. & Sultan, C. (1985) Proposed revised criteria for the classification of acute myeloid leukemia. A report of the French-American-British Cooperative Group. Annals of Inter- nal Medicine. 103, 626-629.

Borowitz. M.J.. Gockerman. J.P. &Moore, J.O. (1 989) Clinicopatholo- gic and cytogenic features of CD34 (MY lO)-positive acute nonlym- phocytic leukemia. American Iournal oJCIinical Pathology. 91,26 5- 270.

Campos. L.. Guyotat. D.. Gentilhomme, O., Treille. D.. Archimbaud. E.. Fiere. D. & Germain. D. (1987) Surface markers in acute non- lymphoid leukemia: analysis with a panel of 36 monoclonal antibodies. Blut, 54, 65-71.

Campos, L., Guyotat. D.. Archimbaud, E.. Devaux. Y., Treille. D.. Larese. A,, Maupas, J.. Gentilhomme, 0.. Ehrsam. A. & Fiere. D. (1989) Surface marker expression in adult acute myeloid leukae- mia: correlations with initial characteristics, morphology and response to therapy. British journal ~(Haematology. 72, 161-166.

Champlin, R.E. & Gale, R.P. (1987) Acute myelogenous leukemia: recent advances in therapy. Blood. 69, 1551-1 562.

Drexler. H.G. & Minowada. j. (1986) The use of monoclonal antibodies for the identification and classification of acute myeloid leukemias. Leukemia Research, 10, 279-290.

Ellison, R.R., Holland, J.F.. Weil. M.. Jacquillat. C.. Boiron, M., Bernard, J.. Sawitsky. A.. Rosner. F., Gussoff. B.. Silver, R.T.. Karanas. A., Cuttner. J., Spurr, C.L., Hayes, D.M., Bloom, j., Leone, L.A., Haurani. F., Kyle, R., Hutchison, J.L., Forcier, R.J. & Moon. J.H. (1968) Arabinosyl cytosine: a useful agent in the treatment of acute leukemia in adults. Blood, 32, 507-523.

Fiere, D.. Campos, L.. Vu Van, H.. Guyotat. D.. Archimbaud. E., Extra,

44 Xavier Thomas et al J.M.. Coiffier. B.. Viala, J.J.. Bryon. P.A.. Milan, J.J. & Ritouet. D. (1986) Intensive timed chemotherapy protocol for 37 resistant or relapsing acute myeloid leukemias. Cancer Treatment Reports. 70,

Grfin. J.D., Davis, R., Nelson, D.A.. Davay, F.R., Mayer, R.J.. Schiffer. C.. McIntyre. O.R. & Bloomfield, C.D. (1986) Use of surface marker analysis to predict outcome of adult acute myeloblastic leukemia.

Holowiecki, 1.. Lutz. D.. Krzemien, S., Stella-Holowiecka, B., Graff, F.. Kelenyi. G . . Schranz. V., Callea. V.. Brugiatelli. M., Neri. A,. Magyarlaki, T.. Ihle. R.. Jagoda. K. & Rudzka, E. (1986) CD15 antigen detected by the VIM-D5 monoclonal antibody for predic- tion of ability to achieve complete remission in acute nonlympho- cytic leukemia. Acta Haematologica, 76, 16-1 9.

Kantarjian. H.M.. Keating. M.J., McCredie, K.B. & Freireich, E.J. (1988) The characteristics and outcome of patients with late relapse acute myelogenous leukemia. journal of Clinical Oncology.

Merle-Beral. H., Nguyen Cong Duc, L., Leblond. V.. Boucheix, C., Michel, A., Chastang. C. & Debre. P. (1989) Diagnostic and prognostic significance of myelomonocytic cell surface antigens in acute myeloid leukaemia. British journal ofHaematology. 73, 323- 3 30.

Preisler. H.D. (1978) Failure of remission induction in acute myelocytic leukemia. Medical and Pediatric Oncology. 4, 2 75-276.

San Miguel. J.F.. Ojeda. E.. Gonzalez. M.. Orfao. A.. Canizo, M.C.. Sanchez, J. & Lopex-Borasca. A. (1989) Prognostic value of immunological markers in acute myeloblastic leukemia. Leukemia,

Schwarzinger. I.. Valent, P., Koller. U.. Marosi, C.. Schneider. B.. Haas, 0.. Knapp. W.. Lechner. K. & Bettelheim. P. (1990) Prognostic significance of surface marker expression on blasts of patients with de novo acute myelobkdstic leukemia. Iournal of Clinical Oncology. 8, 423-430.

Smits. P., Schoots, L.. de Pauw. B.E.. de Witte. T.. Holdrinet, R.S.G., Janssen. J.T.P. & Haanen. C. (1987) Prognostic factors in adult patients with acute leukemia at first relapse. Cancer. 59, 1631- 1634.

285-286.

Blood. 68, 1232-1239.

6, 232-238.

3, 108-111.

Swirsky, D.M., De Bastos, M.. Parish, S.E.. Rees, J.K.H. & Hayhoe. F.G.H. (1986) Features affecting outcome during remission induc- tion of acute myeloid leukaemia in 619 adult patients. British journal of Haematology. 64, 435-453.

Tucker, J. , Dorey, E., Gregory, W.M.. Simpson, A.P.F., Arness. J.A.L., Lister, T.A. & Horton. M.A. (1990) Immunophenotype of blast cells in acute myeloid leukemia may be a useful predictive factor for outcome. Hematological Oncologu, 8, 47-58.

Uhlman. D.L.. Bloomfield, C.D.. Hurd. D.D. & Peterson, B.A. (1990) Prognostic factors at relapse for adults with acute myeloid leukemia. American Journal of Hematology, 33, 110-1 16.

van der Weide. M.. van Rhenen. D.J. & Langenhuijsen. M.M.A.C. ( 1 985) Maturation changes in relapsing acute myeloid leukaemia. Scandinavian journal ojHaematulog~, 35, 63-65.

van Rhenen. D.J.. van der Reijen. H.J., van’t Veer, M.B.. Engelfriet. C.P., vondemBorne, A.E.G.K. &Langenhuijsen. M.M.A.C. (1983) Changes in the surface marker pattern related to maturation in adult acute myeloid leukaemia. Leukemia Research. 7, 357-362.

Vaughan. W.P.. Civin, C.I., Weisenburger. D.D.. Karp. J.E.. Graham, M.L.,Sanger. W.G.. Grierson, H.L., Joshi. S.S.&Burke. P.J. (1988) Acute leukemia expressing the normal human hematopoeitic stem cell membrane glycoprotein CD34 (MY10). Leukemia. 2 .

Yeager. A.M., Kaizer. H., Santos. G.W.. Saral, R., Colvin. O.M.. Stuart. R.K.. Braine. H.G.. Burke. P.J., Ambinder. R.F.. Burns, W.H.. Fuller, D.J., Davis. J.M.. Karp, J.E.. Stratford. W.. Rowley. S.D., Sensenbrenner. L.L., Vogelsang. G.B. & Wingard. J.R. (1986) Autologous bone marrow transplantation in patients with acute nonlymphocytic leukemia. using ex vivo marrow treatment with 4-hydroperoxycyclophosphamide. New England journal of Medi- cine. 315, 141-146.

Zittoun. R.. Jehn, U.. Fiere. D.. Haanen. C.. Liiwenberg. B.. Willemze. R.. Abels. 1.. Bury. J.. Peetermans. M.. Hayat. M.. Suciu. S.. Solbu. G. & Stryckmans. P. (1989) Alternating versus repeated pos- tremission treatment in adult acute myelogenous leukemia: a randomized phase 111 study (Ah4L6) of the EORTC Leukemia Cooperative Group. Blood. 73, 896-906.

661-666.