Leukemia Research 25 (2001) 33–38

De novo acute myeloid leukemia in the elderly; a consistentfraction of long-term survivors by standard-dose chemotherapy

Shinichiro Yoshida a, Kazutaka Kuriyama a,*, Yasushi Miyazaki a, Jun Taguchi a,Takuya Fukushima a, Miyuki Honda a, Toshihisa Hayashibara b, Kazuhiro Nagai a,

Sunao Atogami c, Kazuhiro Toriya d, Hisashi Soda e, Hiroaki Nonaka f,Saburo Momita g, Itsuro Jinnai a, Tatsuhiko Amenomori h, Miyuki Kusano i,Yoshiharu Yoshida j, Shuichi Ikeda k, Tatsuki Matsuo l, Masao Tomonaga a

a Department of Hematology and Molecular Medicine Unit, Atomic Bomb Disease Institute, Nagasaki Uni6ersity School of Medicine,Sakamoto 1-12-4, Nagasaki 852-8523, Japan

b Nagasaki Prefectural Shimabara Onsen Hospital, Shimabara, Japanc Nagasaki Municipal Medical Center, Nagasaki, Japan

d Nagasaki Municipal Hospital, Nagasaki, Japane Isahaya Insurance General Hospital, Isahaya, Japan

f Nagasaki Rosai Hospital, Sasebo, Japang Nagasaki-Chuo National Hospital, Omura, Japan

h Japanese Red Cross Nagasaki Atomic Bomb Hospital, Nagasaki, Japani Sasebo Senjyu Hospital, Sasebo, Japan

j St Francis Hospital, Nagasaki, Japank Sasebo City General Hospital, Sasebo, Japan

l Blood Transfusion Ser6ice, Nagasaki Uni6ersity Hospital, Nagasaki, Japan

Received 26 January 225; accepted 16 June 2000

Abstract

To clarify the characteristics of de novo acute myeloid leukemia (AML) among the elderly, we reviewed 112 patients over 60years old (median age 72 years) who were treated at hospitals in Nagasaki Prefecture with a population of 1.5 million between1987 and 1994. Reclassification of morphological diagnosis revealed that the proportion of M3 was lower but that of M6 and theincidence of cases with trilineage dysplasia (TLD), known as poor prognostic features, were higher in the elderly than in patientsless than 60 years old. Similarly, chromosomal data showed a lower frequency of favorable karyotypes such as t(8;21) and t(15;17)in the elderly. The overall survival of all 112 patients was 10.3% at 5 years. Multivariate analysis indicated that good performancestatus (PS), low WBC at diagnosis, standard dose multi-drug chemotherapy and all-trans retinoic acid (ATRA) treatment for M3patients, and morphological findings without TLD were significantly correlated with longer survival. Most of the long-termsurvivors were found among those who received standard dose therapy in this series, although no consensus has been establishedhow to treat elderly AML patients. We propose that a prospective controlled trial is necessary to confirm the role of standard dosechemotherapy for elderly patients with de novo AML. © 2001 Elsevier Science Ltd. All rights reserved.

Keywords: De novo AML; Elderly patient; Clinical features; Prognostic factors; Standard chemotherapy

www.elsevier.com/locate/leukres

Abbre6iations: AML: acute myeloid leukemia; AMLL: acute mixed lineage leukemia; ALL: acute lymphoblastic leukemia; ATRA: all-transretinoic acid; CR: complete remission; MDS: myelodysplastic syndrome; MDS/AML: AML transformed from MDS; MPO: myeloperoxidase; PS:performance status; TLD: trilineage dysplasia.

* Corresponding author. Tel.: +81-95-8497111; fax: +81-95-8497113.E-mail address: kuriyama@net.nagasaki-u.ac.jp (K. Kuriyama).

0145-2126/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.PII: S 0 1 4 5 -2126 (00 )00089 -8

S. Yoshida et al. / Leukemia Research 25 (2001) 33–3834

1. Introduction

The incidence of acute myeloid leukemia (AML)increases with age. Brincker [1] showed that the medianage of AML patients was 62–64 years old, and approx-imately 60% were over 60 years old. Likewise, theproportion of elderly people is increasing among AMLpatients in many countries including Japan. The man-agement of elderly patients with AML is becomingincreasingly important for clinical hematologists.

However, clinical analysis for AML in the elderly isstill insufficient compared with that in younger popula-tions. Chemotherapy often cannot be applied to elderlypatients because of poor general condition. Elderlypatients tend to be treated in local hospitals, whileyoung or middle-aged patients receive care in centralhospitals for treatment of leukemia by specialists [2].

Heterogeneity of AML is another problem in theelderly. Myelodysplastic syndrome (MDS) is more com-mon among elderly patients, and 5–48% of MDS casestransform to leukemia [3]. Secondary leukemia due toanticancer drugs and/or radiation is also more commonamong the elderly [4]. As these types of leukemia arethought to be distinct in biology from de novo AML, itis important to separate them for analysis of treatmentoutcome.

To clarify the characteristics of de novo AML in theelderly, we reviewed AML patients over 60 years oldwho were treated at Nagasaki University hospital and12 affiliated hospitals.

2. Patients and methods

2.1. Patients

Between January 1987 and September 1994, 399 pa-tients with acute leukemia were admitted to the 13general hospitals in Nagasaki Prefecture. Among these,200 were 60 years or more at diagnosis. Bone marrowsmears stained with May–Grunwald–Giemsa and formyeloperoxidase (MPO) were available and reviewed in178 of 200 cases. Cases with acute lymphoblasticleukemia [5], hypoplastic leukemia [6] and acute mixedlineage leukemia [7] were excluded. Patients with his-

tory of antecedent hematological disorder at any timeup to 6 months before the diagnosis of AML or expo-sure to leukemogenic agents such as anti-cancer drugsor radiation were also excluded based on their clinicalrecords. One hundred and twelve cases were subse-quently diagnosed as having de novo AML. Table 1shows categories of diagnosis and numbers of patients.

FAB subtypes [8], percentage of bone marrow blastcells, evidence of co-existing trilineage dysplasia (TLD)[9] and MPO positivity of blast cells were furtherassessed. TLD was diagnosed according to Brito–Ba-bapulle’s criteria with some modifications [10]. Clinicaldata (age, sex, performance status [PS], blood cellcount, chromosomal findings), types of treatment andclinical outcome were analyzed. PS score was definedaccording to the Performance Status Scale of the East-ern Cooperative Oncology Group (ECOG) [11].

2.2. Treatment

Treatment regimens were categorized into threegroups according to the intensity of chemotherapy.Group A, standard dose induction chemotherapy (3-day infusion of daunorubicin 40 mg/m2 and 7-dayinfusion of cytosine-arabinoside 100 mg/m2) with orwithout the addition of daunorubicin, 6-MP and/orcytosine-arabinoside [12,13], and all-trans retinoic acid(ATRA) therapy for M3 combined with standard dosechemotherapy [14]; Group B, other inductionchemotherapy modified by dose reduction; Group C,supportive care only.

2.3. Response criteria and statistics

Statistical analyses were performed as of 30 Novem-ber, 1998. Complete remission (CR) was defined asnormalization of bone marrow cellularity with less than5% blast cells and normal peripheral blood count. Therelationships to qualitative and discrete parameterswere analyzed by x2 test. The duration of overall sur-vival was measured from the date of diagnosis to thedate of death. Survival curves of patients were preparedby the Kaplan–Meier method, and differences betweenthe survival curves were evaluated using the log-ranktest. Multivariate analysis of the prognostic factors forlong-term survival was performed using the Cox pro-portional hazard model.

3. Results

3.1. Patient characteristics and FAB classification

As shown in Table 2, 112 patients diagnosed ashaving de novo AML included 67 men and 45 womenwith a median age of 72 years (range, 60–92). Eighty

Table 1Diagnosis of 178 patients whose smears were reviewed

(62.9%)de novo AML 112(11.2)ALL 20

17 (9.6)Hypoplastic leukeima(1.7)3AMLL

18MDS/AML (10.1)(2.2)4Secondary leukemia

4 (2.2)Non-acute leukemia

S. Yoshida et al. / Leukemia Research 25 (2001) 33–38 35

Table 2Patient characteristics and FAB classification in de novo AML

112Number of patientsMale/Female 67/45

(60–92)72Median age (range)Performance status score

150381

2 2715394

Unknown 8FAB classification

M0 (3.6%)4M1 13 (11.6)

(42.9)48M2(6.3)M3 7

(17.0)19M4(5.4)M5 6

(10.7)12M61M7 (0.9)2Unclassified (1.8)

(36.6%)41Patients with TLD

ate risk or other adverse risk. Cytogenetic abnormalitiesnot classified as favorable or adverse karyotype andnormal karyotype were categorized into the intermedi-ate group. Seven of the 69 cases (10.1%) had favorablekaryotypes. Among 31 cases diagnosed as M2, only twocases (6.5%) were associated with t(8;21). None of thepatients had inv(16). Adverse karyotypes were found in11 cases (15.9%), and eight of these showed del(5q).There were no patients with −5, −7 or abnormal 3qin our study. The remaining 51 cases showed intermedi-ate karyotypes.

3.3. Treatment

Twenty-nine patients of Group A received conven-tional chemotherapy with standard dose or ATRAtherapy combined with chemotherapy (for M3). Fifty-eight patients were categorized into Group B with dosereduction, and 19 cases received only supportive care(Group C) for various reasons. Treatment data werenot available for six patients. Seven patients were diag-nosed as M3; four were treated with regimens includingATRA, and the remaining three received conventionalchemotherapy only. The median age of Group A pa-tients was 65 (range; 60–82), younger than those ofGroup B (median, 73; range 60–92) and Group C(median, 77; range 61–87). PS scores of patients in eachgroup were not significantly different. Other factorssuch as WBC, percentage of MPO-positive blast cellsand TLD were not different between Group A andGroup B. CR rate was 69% in Group A, whereas it was22.4% in Group B.

3.4. Sur6i6al and prognostic factors

The overall survival of the 112 patients is shown inFig. 1. Survival rate at 5 years was 10.3%. Patients witha PS score of 2 or below had significantly highersurvival rates than those with scores of 3 or 4 (11.8%vs. 8.3% at 5 years, P=0.007) (Fig. 2). Among the 24

patients (71.4%) had a PS score of 2 or below. Sevencases (6.3%) were M3 and 12 (10.7%) were M6. Forty-one patients (36.6%) were diagnosed as having AMLwith TLD.

3.2. Chromosomal findings

Cytogenetic analysis was performed in 74 patients.Analysis failed in five of these subjects, and availabledata for the remaining 69 cases (62%) are shown inTable 3. Karyotypes were classified into three groups,i.e. favorable, intermediate or adverse type according tothe criteria reported by Grimwade et al. [15]. Thefavorable group included all patients with t(8;21),t(15;17) and inv(16), regardless of whether other cytoge-netic abnormalities were present, and adverse groupincluded patients with −5, −7, del(5q), abnormal 3q,or complex abnormality in conjunction with intermedi-

Table 3Cytogenetic findings and FAB subtypesa

M3 M4 M5 M6 M7 others TotalM0 M1 M2

Fa6orable 755t(15;17)

t(8;21) 2 251Intermediate

no abnormality 1 3 19 1 6 4 2 1 37+8 4 1 5other numerical 4 1 1 6

31other structural 2Ad6erse 11

111complex 323del(5q) 3 8

a Cytogenetic data were available for 69 patients. These were classified according to the criteria of Grimwade et al. [15].

S. Yoshida et al. / Leukemia Research 25 (2001) 33–3836

Fig. 1. Overall survival curve of elderly patients with de novo AML.Overall survival was 10.3% at 5 years.

ment (Fig. 4), Group A patients showed a highersurvival rate of 29.3% at 5 years compared with thoseof Group B (3.9%) and Group C (0%) (P=0.0002).

The diagnosis of M3 was associated with a favorableoutcome. Survival rate of these patients at 5 years was41.7%, and none died before 1000 days after diagnosis.There were no differences in survival between patientstreated with regimens including ATRA and those givenconventional chemotherapy.

The prognostic factors for long-term survival byunivariate analysis were Group A treatment, lowerWBC (B10×109/l), favorable karyotype, high MPOpositivity (\50%), FAB subtype of M2 or M3, withoutTLD, good PS (0–2) and younger age (B70 years).However, multivariate analysis indicated only four ofthese factors as significant; i.e. good PS, low WBC,Group A treatment and without TLD, provided thatcytogenetic data were excluded from multivariate anal-ysis because of the small sample number (Table 4).

Fig. 2. Overall survival of patients with good performance status(0–2) and poor performance status (3, 4). Survival was 11.8% and8.3% at 5 years, respectively.

Fig. 4. Overall survival of patients of Group A, B and C (see Section2). Survival was 29.3%, 3.9% and 0% at 5 years, respectively.

Fig. 3. Overall survival of patients with favorable, intermediate andadverse karyotypes. Survival was 51.4%, 11.9% and 0% at 5 years,respectively.

Table 4Prognostic factors for long-term survival

Factors P-value

Uni6ariate analysisGroup Aa 0.0002WBCB10×109/l 0.0011

0.0016Favorable karyotypeMPO\50% 0.0026FAB: M2, M3 0.0033

0.0042without TLDPS: 0,1,2 0.0070

0.0072AgeB70 yrsMulti6ariate analysis

PS: 0,1,2 0.0001WBCB10×109/l 0.0020

0.0083Group Awithout TLD 0.0454

a Group A: Standard dose induction chemotherapy or all-transretinoic acid (ATRA) therapy for M3.

patients with a PS of 3 or 4, 10 died within 30 days and17 died within 100 days after diagnosis. As shown inFig. 3, four of the seven patients (57.1%) with favorablekaryotype were alive at the time of analysis. Thisnumber was higher than those obtained among patientswith intermediate or adverse karyotype (eight of 51,15.7% and zero of 10, respectively). In terms of treat-

S. Yoshida et al. / Leukemia Research 25 (2001) 33–38 37

4. Discussion

Patient selection has a marked impact on the treatmentoutcome of AML patients [2,16], and this is particularlytrue in the elderly [17,18]. We attempted to avoid patientselection in this study; the 13 hospitals participating inthis study included not only leukemia centers for adultAML but also community hospitals. Based on the TumorRegistry of Nagasaki Prefecture between 1987 and 1994,more than 80% of all adult acute leukemia patients weretreated in these hospitals (population of 1.5 million). Thissuggests the selection bias for the patients was minimalin this analysis.

We also attempted to examine elderly patients with denovo AML, as AML transformed from MDS andsecondary AML are common among elderly AMLpatients [17,19]. Therefore, we reviewed bone marrowsmears of each case to confirm diagnosis, and checkedclinical records including past history. Subsequently, wefocused on de novo AML in the elderly. This selectionprobably influenced the results of this study. Due to thiswide coverage of total elderly patients and re-evaluationof the clinical diagnosis, we believe that this studyrepresents the general features of de novo AML in theelderly.

In comparison with the data from a previous largescale clinical trial of AML therapy, the AML87 study(median age 48) [13] by the Japan Adult Leukemia StudyGroup (JALSG) in Japan, our series included less M3patients (6.3% vs. 17.8%) and more M6 patients (10.7%vs. 2.7%). TLD was more frequent among patients olderthan 60 (36.6% vs. 17.4%)[10]. Incidences of both M3 andnon-TLD phenotype, both good prognosis factors, werelower among AML in the elderly.

In this series, chromosomal data also had prognosticvalue and t(8;21) was correlated with a good response tochemotherapy. The frequency of t(8;21) was reported tobe about 40% among M2 subtype (8% of total de novoAML) [12,20,21], but we found only two patients witht(8;21) among 31 of M2 cases (6.5%). Such a low numberof cases with M2 with t(8;21) as well as M3 with t(15;17)suggested a lower frequency of AML with favorablekaryotypes for prognosis in the elderly as reportedpreviously [17,18].

Higher frequencies of less favorable karyotypes orTLD phenotype seemed to contribute to shorter survivalof de novo AML in cases over 60 years old.

We found that about 10% of the elderly patients in thisstudy lived long enough to be considered cured bychemotherapy. Multivariate analysis showed that factorssuch as PS 0–2, WBC B10×109/l at diagnosis, GroupA treatment and morphological findings without TLDwere correlated with long survival. Cytogenetic data wereexcluded from this multivariate analysis because of thesmall sample number, while favorable karyotype wasshown to be a significant factor by univariate analysis.

Although age was not significantly related to survival bymultivariate analysis, we should be careful to accept thisresult as age cannot be separated from the choice oftreatment (see Section 3).

Our study clearly demonstrated that most of thelong-term survivors in elderly AML patients had receivedstandard chemotherapy. Dose reduction (Group B) didnot increase the incidence of long-term survival. Fortreatment of AML in the elderly, some previous studiesdemonstrated the efficacy of chemotherapy. Immediateintensive induction was shown to provide longer survivalcompared with ‘wait and see’ policy [22]. Tilly et al.,demonstrated that low dose cytarabine was comparableto intensive chemotherapy in terms of survival; higherremission rate by intensive treatment was counterbal-anced by its toxicity and the stable disease induced bylow dose therapy [23]. Our analysis also showed theimportance of chemotherapy for long-term survival, butwe found no benefit of low dose therapy. This differencemay have been due to the length of the study period. Allpatients were followed more than 4 years (up to 10 years)in our study, but in that by Tilly et al., maximum followup time was almost 4 years. Our analysis was performedretrospectively, and we could not directly compare theresults of standard and low dose therapy. However, thislong-term observation suggested that only standardchemotherapy can improve clinical outcome in elderlypatients especially with good PS and favorable kary-otype. In this context, patients with good PS in GroupB (43 of 58) and Group C (14 of 19) could have beencandidates for standard chemotherapy, contributing tofurther improvement in outcome of the entire elderlyAML population. A single-arm prospective controlledstudy is required to confirm the role of standardchemotherapy in treating elderly patients with de novoAML who have a good general condition.

Acknowledgements

We thank Dr Daisuke Niino, Nagasaki-Chuo Na-tional Hospital, Dr Yuichi Danno, Izumikawa Hospitaland Dr Yuji Matsuo, Goto Central Hospital for theircooperation. S. Yoshida collected, analyzed the data,drafted the article and gave final approval. K. Kuriyamaprovided the concept, design, helped with the draftdocument and gave final approval. Y. Miyazaki providedtechnical support, critical revision and gave final ap-proval. J. Taguchi provided statistical expertise and gavefinal approval. T. Fukushima provided technical supportand gave final approval. M. Honda, T. Hayashibara, K.Nagai, S. Atogami, K. Toriya, H. Soda, H. Nonaka, S.Momita, I. Jinnai, T. Amenomori, M. Kusano and S.Ikeda provided study materials or patients and gave finalapproval. T. Matsuo provided critical revision and finalapproval. M. Tomonaga obtained the necessary funding,provided critical revision and gave final approval.

S. Yoshida et al. / Leukemia Research 25 (2001) 33–3838

References

[1] Brincker H. Estimate of overall treatment results in acute non-lymphocytic leukemia based on age-specific rates of incidenceand of complete remission. Cancer Treat Rep 1985;69:5.

[2] The Toronto Leukemia Study Group. Results of chemotherapyfor unselected patients with acute myeloblastic leukaemia: effectof exclusions on interpretation of results. The Toronto LeukemiaStudy Group. Lancet 1986;1:786.

[3] Hofmann WK, Ottmann OG, Ganser A, Hoelzer D. Myelodys-plastic syndromes: clinical features. Semin Hematol 1996;33:177.

[4] Kantarjian HM, Keating MJ, Walters RS, Smith TL, Cork A,McCredie KB, Freireich EJ. Therapy-related leukemia andmyelodysplastic syndrome: clinical, cytogenetic, and prognosticfeatures. J Clin Oncol 1986;4:1748.

[5] Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA,Gralnick HR, Sultan C. The morphological classification ofacute lymphoblastic leukaemia: concordance among observersand clinical correlations. Br J Haematol 1981;47:553.

[6] Nagai K, Kohno T, Chen YX, Tsushima H, Mori H, NakamuraH, Jinnai I, Matsuo T, Kuriyama K, Tomonaga M, Bennett JM.Diagnostic criteria for hypocellular acute leukemia: a clinicalentity distinct from overt acute leukemia and myelodysplasticsyndrome. Leuk Res 1996;20:563.

[7] Matutes E, Morilla R, Farahat N, Carbonell F, Swansbury J,Dyer M, Catovsky D. Definition of acute biphenotypicleukemia. Haematologica 1997;82:64.

[8] Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA,Gralnick HR, Sultan C. Proposed revised criteria for the classifi-cation of acute myeloid leukemia. A report of the French–Amer-ican–British Cooperative Group. Ann Intern Med 1985;103:620.

[9] Brito-Babapulle F, Catovsky D, Galton DA. Clinical and labo-ratory features of de novo acute myeloid leukaemia with trilin-eage myelodysplasia. Br J Haematol 1987;66:445.

[10] Kuriyama,K., Tomonaga, M., Matsuo, T., Kobayashi, T.,Miwa, H., Shirakawa, S., Tanimoto, M., Adachi, K., Emi, N.,Hiraoka, A., Tominaga, M., Imai, K., Asou, N., Tsubaki, K.,Takahashi, I., Minami, S., Yoshida, M., Murakami, H., Minato,K., Oshima, T., Furusawa, S., Ohno, R., Japan Adult LeukemiaStudy Group (JALSG). Poor response to intensive chemother-apy in de novo acute myeloid leukaemia with trilineagemyelodysplasia. Br J Haematol 1994;86:767.

[11] Oken MM, Creech RH, Tormey DC, Horton J, Davis TE,McFadden ET, Carbone PP. Toxicity and response criteria ofthe Eastern Cooperative Oncology Group. Am J Clin Oncol1982;5:649.

[12] Yates J, Glidewell O, Wiernik P, Cooper MR, Steinberg D,Dosik H, Levy R, Hoagland C, Henry P, Gottlieb A, Cornell C,Berenberg J, Hutchison JL, Raich P, Nissen N, Ellison RR,Frelick R, James GW, Falkson G, Silver RT, Haurani F, GreenM, Henderson E, Leone L, Holland JF. Cytosine arabinosidewith daunorubicin or adriamycin for therapy of acute myelocyticleukemia: a CALGB study. Blood 1982;60:454.

[13] Ohno R, Kobayashi T, Tanimoto M, Hiraoka A, Imai K, AsouN, Tomonaga M, Tsubaki K, Takahashi I, Kodera Y, YoshidaM, Murakami H, Naoe T, Shimoyama M, Tsukada T, Takeo T,

Teshima H, Onozawa Y, Fujimoto K, Kuiryama K, Horiuchi A,Kimura I, Minami S, Miura Y, Kageyama S, Tahara T,Masaoka T, Shirakawa S, Saito H. Randomized study of indi-vidualized induction therapy with or without vincristine, and ofmaintenance-intensification therapy between 4 or 12 courses inadult acute myeloid leukemia. AML-87 Study of the JapanAdult Leukemia Study Group. Cancer 1993;71:3888.

[14] Kanamaru A, Takemoto Y, Tanimoto M, Murakami H, AsouN, Kobayashi T, Kuriyama K, Ohmoto E, Sakamaki H, Tsub-aki K, Hiraoka A, Yamada O, Oh H, Saito K, Matsuda S,Minato K, Ohno R. All-trans retinoic acid for the treatment ofnewly diagnosed acute promyelocytic leukemia. Blood1995;85:1202.

[15] Grimwade D, Walker H, Oliver F, Wheatley K, Harrison C,Harrison G, Rees J, Hann I, Stevens R, Burnett A, Goldstone A.The importance of diagnostic cytogenetics on outcome in AML:analysis of 1,612 patients entered into the MRC AML 10 trial.The Medical Research Council Adult and Children’s LeukaemiaWorking Parties. Blood 1998;92:2322.

[16] Wahlin A, Hornsten P, Jonsson H. Remission rate and survivalin acute myeloid leukemia: impact of selection and chemother-apy. Eur J Haematol 1991;46:240.

[17] Baudard M, Marie JP, Cadiou M, Viguie F, Zittoun R. Acutemyelogenous leukaemia in the elderly: retrospective study of 235consecutive patients. Br J Haematol 1994;86:82.

[18] Taylor PR, Reid MM, Stark AN, Bown N, Hamilton PJ,Proctor SJ. De novo acute myeloid leukaemia in patients over55-years-old: a population-based study of incidence, treatmentand outcome. Northern Region Haematology Group. Leukemia1995;9:231.

[19] Leith CP, Kopecky KJ, Godwin J, McConnell T, Slovak ML,Chen IM, Head DR, Appelbaum FR, Willman CL. Acutemyeloid leukemia in the elderly: assessment of multidrug resis-tance (MDR1) and cytogenetics distinguishes biologic subgroupswith remarkably distinct responses to standard chemotherapy. ASouthwest Oncology Group study. Blood 1997;89:3323.

[20] Mitelman F, Heim S. Quantitative acute leukemia cytogenetics.Genes Chromosomes Cancer 1992;5:57.

[21] Nakamura H, Kuriyama K, Sadamori N, Mine M, Itoyama T,Sasagawa I, Matsumoto K, Tsuji Y, Asou N, Kageyama S,Sakamaki H, Emi N, Ohno R, Tomonaga M. Morphologicalsubtyping of acute myeloid leukemia with maturation (AML-M2): homogeneous pink-colored cytoplasm of mature neu-trophils is most characteristic of AML-M2 with t(8;21).Leukemia 1997;11:651.

[22] Lowenberg B, Zittoun R, Kerkhofs H, Jehn U, Abels J, Debuss-cher L, Cauchie C, Peetermans M, Solbu G, Suciu S, Stryck-mans P. On the value of intensive remission-inductionchemotherapy in elderly patients of 65+ years with acutemyeloid leukemia: a randomized phase III study of the EuropeanOrganization for Research and Treatment of Cancer LeukemiaGroup. J Clin Oncol 1989;7:1268.

[23] Tilly H, Castaigne S, Bordessoule D, Casassus P, Le Prise PY,Tertian G, Desablens B, Henry-Amar M, Degos L. Low-dosecytarabine versus intensive chemotherapy in the treatment ofacute nonlymphocytic leukemia in the elderly. J Clin Oncol1990;8:272.

.