Prognostic factors and treatment effects on survival in acute myeloid leukemia of M6 subtype: A retrospective study of 54 cases

Annals ofOncology 12 451-455. 2001 © 2001 Kluwer Academic Publishers Primed in the Netherlands

Original article Prognostic factors and treatment effects on survival in acute myeloid leukemia of M6 subtype: A retrospective study of 54 cases A. Colita,1 A. Belhabri,1 Y. Chelghoum,1 C. Charrin,2 D. Fiere1 & X.Thomas1 1

'Service d'Hemalologie. 2Laboraloirede Cylologie el de Cytogenetique, Hopital Edouard Hernol, Lyon, France

Summary Background Since 1976, erythroleukemia has been included within the FAB classification system of acute myeloid leukemia (AML) which designates it as M6 AML This report describes the data of 54 patients with newly diagnosed M6 AML, consecutively seen in our hospital between May 1976 and May 1999 Patients and methods There were 40 males and 14 females Median age was 59 years Pancytopenia was the most common feature at diagnosis Twenty-six percent of cases presented with secondary AML Karyotype was successfully performed in 35 cases Eleven patients presented with normal karyotype, nine with simple karyotypic abnormalities, and fifteen with major karyotypic abnormalities Fifty of the fifty-four patients received one or two courses of induction chemotherapy combining anthracyclines with cytarabine according to different successive protocols One elderly patient only received low-dose cytarabine, and three patients died before any chemotherapy could be given Results Complete remission (CR) was achieved in 29 cases (54%, 95% confidence interval (CI) 40%-67%) As postremission therapy, four patients could be allografted, and two underwent autologous transplantation All other treated patients received continuation chemotherapy Twenty-one patients have relapsed (72%) Median time to relapse was six

Introduction Erythroleukemia is a myeloprohferative disorder characterized by malignant proliferation of erythroid and myeloid precursors The disease has a long descriptive history which has gone through different stages of evolution. Although initially described in 1912 [1], the term of 'erythroleukemia' was first proposed by Di Guglielmo in 1917 [2] In the late 1950s and 1960s, Dameshek proposed the term 'Di Guglielmo's syndrome' to describe a myeloprohferative syndrome striking erythoblastic hyperplasia and a progressive increase of myeloblasts terminating in AML [3, 4]. After 1976, erythroleukemia was no longer considered as a chronic myeloprohferative disorder but rather as a form of AML and as such was included within the French-American-British (FAB) classification system of AML which designates it as AML-M6 [5]. Among the various types of acute myeloid leukemia (AML), erythroleukemia is regarded as a rela-

Downloaded from https://academic.oup.com/annonc/article-abstract/12/4/451/236782 by guest on 29 July 2018

months. Among those patients, only eight achieved a second CR (38%) The median disease-free survival (DFS) was eight months (95% CI 4-10 months) with a five-year survival rale of 17% Median overall survival (OS) was nine months (95% CI 5-12 months) with a five-year survival rate of 13% In univanate analysis, poor prognostic factors for DFS were secondary AML (P = 0 05) and initial platelet count < 50 x 109/1 (P = 0 02). Poor prognostic factors for OS were age > 60 years (P - 0.005), secondary A M L ( P = 0 05), initial 'blastic' fever (P - 0 0004), and initial haemoglobin level < 90 g/1 (P = 0.03). All factors, but haemoglobin level, remained significant in the multivanate analysis Although it was not statistically significant, there was a trent for a better prognosis of M6 patients presenting with normal karyotype as compared to those displaying chromosomal abnormality. Conclusions This retrospective analysis points to a somewhat heterogenous group of AML in terms of clinical and biological features, and outcome Distinctive subgroups can be identified according to prognostic factors related to survival A larger multicenter study with well-defined diagnostic criteria is warranted to further clarify treatment effects Key words acute myeloid leukemia, chemotherapy, M6 subtype, prognostic factors

tively rare disorder and considered to be a distinct entity with characteristic features such as complex chromosomal defects and poor prognosis [6-9]. Because of the rarity of this disorder, few investigators have accumulated sufficient numbers of cases to obtain data correlating various clinical and laboratory features with survival We therefore reviewed data from 54 patients with newly diagnosed M6 subtype AML, consecutively seen in our hospital over a 23-year period, in order to better assess characteristics and outcome of this disease Patients and methods Patient population All cases of adult ( > 15 years old) patients with AML of M6 subtype, diagnosed according to FAB criteria [5. 10] and referred to our center over a 23-year period, were included in this study This meant that all patients were required to have a bone marrow (BM) with an erythroid component of at least 50% and myeloid blasts comprising at least 30%

452 Table I Clinical and laboratory features of M6 AML at presentation (54 patients) Characteristic Age (years) Median Range Sex Male Female

Number of patients

59 19-84 40 14

Weight loss > 5% (yes/no) Tumoral syndrome (yes/no) Nodes enlargement (yes/no) Speen enlargement (yes/no) Liver enlargement (yes/no) Central nervous system involvement (yes/no) Clinical bleeding (yes/no)

3/51 17/37 5/49 8/46 10/44 0/54 8/46

Haemoglobin (g/l) Median Range

85 47-134

MCV(fl) Median Range White blood cell count (x 109/l) Median Range Platelets (x 109/l) Median Range

94 81-118 3 24 0 4-367 47 6-334

LDH(u/l) Median Range Uric acid (umol/l) Median Range Creatinine (umol/l) Median Range Fibnnogen (g/l) Median Range

612 167-4147 310 111-767 84 39-240 32 1 7-8 2

with complex abnormality Patients with simple abnormality had karyotypes consisted of one or at most two translocations, nondisjunctions, or numerical chromosomal changes In contrast, patients with complex abnormality had karyotypes involving three or more such cellular events Since 1996, cytogenetic analysis was occasionally completed by FISH technique using chromosome-specific paint probes for karyotypes with deletions or apparently not balanced translocations

Chemotherapy received Three patients with poor general status died before any chemotherapy could be given One elderly patient only received low-dose cytarabine by subcutaneous route The fifty other patients were treated according to ongoing first line intensive intercalating agent and cytosine arabinoside ± VP16 ± 6-mercaptopunne-contaimng chemotherapy protocols Once complete remission (CR) was achieved, patients received either no maintenance therapy, conventional maintenance therapy, additional courses of myelosuppressive consolidation, or autologous transplantation according to the protocol design in which they were included All protocols included one or two courses of intensive consolidation Patients received standard supportive care during the neutropenic period Approximately 1/3 of patients older than 60 years who achieved CR received maintenance using low-dose cytarabine after one course of intensive consolidation During this period of time, none of the therapeutic protocols differed according to patient karyotype Patients aged less than 45 years with an HLA-identical family donor were scheduled to receive allogeneic bone marrow transplantation (BMT) following standard conditioning regimen Four patients underwent allogeneic BMT, and two received autologous transplantation

Evaluation of therapy Complete remission was defined according to the CALGB criteria as less than 5% blasts in bone marrow aspirates with evidence of maturation of all cell lines and restoration of the peripheral blood counts [13] Failures to achieve CR were classified according to Preisler [14] Disease-free survival (DFS) was measured from the time of CR to the time of relapse or death, and overall survival (OS) from the time of initial diagnosis to the time of death Haematologic relapse was diagnosed when more than 5% blasts were seen in bone marrow aspirates

Statistical analysis of non-erythroid cells. Fifty-four patients with newly diagnosed untreated M6 AML and who fulfilled those criteria were retrospectively reviewed These patients represented about 3% of all AML cases seen during this time period Fourteen patients (26%) presented with secondary AML Indeed. M6 AML complicated chemotherapy given for malignant diseases in three cases (breast cancer in one patient, and Hodgkin's disease in two patients), and followed chronic haematologic myeloproliferative disorders in three cases (thrombocytopema in two patients, and chronic myeloid leukemia in one case) Five patients had a prior history ofmyelodysplastic syndrome Three patients have been previously exposed to benzene

Chromosomal analysis Since May 1980. cytogenetic analysis was performed at time of initial diagnosis on bone marrow leukemic cells or occasionally on short unstimulated peripheral blood blastic cell cultures, using standard RHG banding technique Karyotypes were analyzed according to the International System for Cytogenetic Nomenclature [II] A karyotype was considered normal when at least 15 normal mitoses were observed Karyotype was successfully performed in 35 cases (65%) According to previous reports [6. 8. 9, 12], the patients with chromosomal changes were divided into those with simple karyotypic abnormality and those

Downloaded from https://academic.oup.com/annonc/article-abstract/12/4/451/236782 by guest on 29 July 2018

Both groups were compared with regard to initial characteristics (past history of malignancy, toxic or radiation exposure, or myelodysplastic syndrome, organomegaly, fever, blood counts, percent blasts in bone marrow, age, karyotypic data, serum enzymes) using Yates corrected chi-square or two-tailed Fisher's exact test, when appropriate for discrete variables, and Mann Whitney non parametric test for continuous variables The same parameters were analyzed for potential prognostic significance for complete remission (CR) achievement, DFS and OS Quantitative variables were treated as dichotomous in the analysis Cutoff points used were the conventional value defining elderly AML for age, the approximate median value for haematological parameters, and the upper limit of normal value for serum parameters Complete remission rates were compared using Yate's corrected chi-square and 95% Cl on proportions of CR patients were calculated using the exact binomial formula Survival and DFS probabilities were calculated using the Kaplan and Meier product-limit estimate method and their 95% symmetrical Cl limit was calculated according to Greenwood's method Survival curves were compared using the logrank test For analysis of survival and DFS of chemotherapy-treated patients, patients undergoing allogeneic or autologous bone marrow transplantation while in first CR were censored at the time of transplantation Prognostic factors for CR were studied using stepwise multiple logistic regression, and prognostic factors for DFS and over-

453 all survival were studied using stepwise Cox's proportional hazard model All variables that were statistically (two-tailed P < 0 05) predictive for CR achievement or survival and their interactions were proposed for entry in the stepwise multivanate model for prediction of the same endpoint Goodness of fit of the models was tested using the likelihood ratio statistics All computations were made using BMDP software (BMDP Statistical Software, Los Angeles, California)

Results

Table 2 Efficacy of induction chemotherapy Results of induction chemotherapy

All patients (%)

Number of patients aged < 60 years (%)

Number of patients aged > 60 years (%)

CR (%) One course Two courses Failure (blastic death or resistant) Toxic death

29(54) 24 5

19(65) 17 2

10(40) 7 3

23 2

10 0

13 2

Initial characteristics of the patients From May 1976 to May 1999, 54 adults (40 males and 14 females, sex ratio' 2.8), seen in our institution at diagnosis, presented with AML of M6 subtype. Characteristics of the patients at diagnosis are indicated in Table 1. Median age was 59 years (range 19-84 years) Pancytopenia was the most common feature. Median haemoglobin level was 85 g/1 (range 47-134 g/1) White blood cell (WBC) count ranged from 0.4 to 367 x 109/l (median 3.24 x 109/l). Only 13 patients had WBC > 10 x 109/l. Median initial platelet count was 47 x 109/l (range 6-334 x 109/l). The incidence of tumoral syndrome was 31%. None of the patients presented with central nervous system (CNS) involvement Karyotypes were normal in 11 patients (31%). Simple karyotypes (9 patients, 26%) showed recurrent translocations such as t(9;22)(q34;qll) (1 case), t(3;12)(pl3;q26) (1 case), or partial deletion on the long arm of chromosome 20 (2 cases) or on the short arm of chromosome 1 (1 case), numerical changes such as tnsomy 8 (1 case), tnsomy 13 (1 case), or monosomy 7 (2 cases) as sole change. In complex karyotypes (15 patients; 43%), clonal changes combined partial or total deletion of chromosome 7 (12 cases; 34%), partial or total deletion of chromosome 5 (11 cases, 31%), monosomy 17 (8 patients; 23%), and miscellaneous changes such as deletions (del(4q) in three cases and del(12p) in 2 cases), translocations or chromosomal gains. Twelve patients (80%) displayed unidentified chromosomal rearrangements (markers) which could not be explained by the use of painting techniques. There were no significant differences between primary and secondary AMLs in terms of chromosomal rearrangements Treatment outcome CR was achieved in 29 (54%, 95% CI 40%-67%) patients after 1 (24 patients) or 2 (5 patients) induction courses CR rates were 65% and 40%, respectively, for patients aged less or more than 60 years (Table 2). Six patients died during the first induction course. Four patients died with relative drug resistance and two patients died during induction with severe marrow hypoplasia. Median follow-up for the entire cohort at the time of analysis was 138 months. Twenty-one patients (72%) had relapsed at the time of analysis, and seven were in continuing CR for one to more than one hundred twenty-eight months. Only eight (38%) relapsed patients

Downloaded from https://academic.oup.com/annonc/article-abstract/12/4/451/236782 by guest on 29 July 2018

achieved a second CR. The median time to relapse was 6 months with a range of < 1-28 months. Median DFS time for remitters was eight months (95% CI 4-10 months) with 17% of patients in continuing CR at five years Median OS of the entire cohort was nine months (95% CI: 5-12 months) with 13% of patients surviving five years from the time of diagnosis (Figure 1). Four patients underwent HLA-identical allogeneic bone marrow transplantation from a suitable familial donor while in first CR: two of these died from leukemic relapse, while the other two died from graft versus host disease. Two patients received autologous stem-cell transplantation in first CR: both patients died from toxicity. Prognostic/actors No individual characteristic of patients was significantly associated with CR rate. Factors predictive for a short DFS in the univanate analysis included secondary AML (p = 0 05), and initial platelet count <50 x 109/l (P = 0.02) Factors predictive for short OS included age ^ 6 0 years {P = 0.005), secondary AML (P - 0.05), initial 'blastic' fever (P = 0.0004), and initial haemoglobin <90 g/1 (P = 0.03). In a multivanate analysis taking into account potentially prognostic factors, no single factor remained significantly predictive for CR achievement. Secondary AML (P = 0 02, relative risk (RR). 0.16; CI (RR): 0.05-1.98) and thrombocytopenia (P = 0.0001; RR. 0.97, CI(RR): 0.95-1.00) remained the prognostic factors for DFS. Only age (P = 0.003; RR: 1.03, CI (RR). 1.00-105), secondary AML (P = 0.03, RR. 0 45; CI (RR). 0.22-109), and 'blastic' fever (P = 0.03; RR' 0 39; CI (RR). 0 19-124) remained significant for OS in the multivanate analysis. Regarding karyotypic data, there was a trend for a better survival in patients with normal karyotype (median OS 18 months) as compared to those with abnormal karyotypes (median OS 5 months) (Figure 2) Among patients with abnormal karyotypes, there were no significant differences between those with simple and those with complex abnormalities.

Discussion Because of its rare occurrence, M6 AML has not been extensively studied [15]. In our experience, M6 subtype

454 1.0 -i

'j

0.8 _ Normal karyotypej . . Abnormal karyotypes

W

0.6 -

o

median: 9 montht

S

0.4-1 ,[33X]

O

£

o.o

p - 006

0.2 H 0.0

0

2

4

6

8

10

12

Time from diagnosis (years)

0

2

4

6

8

10

12

Time from diagnosis (years)

Figure I Overall survival of the entire cohort (54 patients)

Figure 2 Overall survival according to karyotypic data

represented about 3% of all AML seen during this 23-year time period. This represented a bit less than that previously reported by others [9, 16] However, this could be an underestimation because of the retrospective character of the study reviewing only patients initially classified as M6 at diagnosis. M6 AML is distinctive in that it is defined by an oligoblastic proliferation of both erythroid and myeloid precursors, while most AMLs exhibit a uniform proliferation of single cell type The FAB M6 subtype, that we choose as the morphological standard to characterize our M6 cases, is defined by an excess of myeloblasts or proerythroblasts in the setting of dyserythropoiesis. However, the FAB criteria for M6 AML have been the subject of some controversy since they were put forth [17-19] Indeed, recent studies have shown identical data between FAB M6 AML and pure erythroid malignancies such as Di Gughelmo disease in terms of clinical and laboratory features, cytogenetic patterns, and poor survival data suggesting that the FAB classification scheme should be modified to include those malignancies. Erythroleukemia is distinguished by several features. Our data were in general in concordance with previous studies reported in the literature [16, 20] These patients present with pancytopenia. Almost all patients showed at presentation varying degrees of leukopenia The peripheral blast count in most of our patients was low. This review of cases collected over the past 23 years reveals a wide age range from 19-84 years, although most patients are over 50 years of age at time of diagnosis [16]. M6 is generally viewed as a disease of the elderly The disease was primarily one of the elderly, but a subset of younger patients also emerged in our cases, whose clinical outcome was significantly better than that of the elderly patients. In contrast to most previous reports [16, 20], there was a strong male sex predilection Like the other AMLs, M6 has been reported to complicate exposure to radiation or toxic drugs. M6 more commonly complicates treatment with alkalating agents and benzene exposure than any other types of AML [21-

23]. Twenty-six percent of our M6 patients had secondary AML This high percentage is in accordance with previous studies [20] and may indicate the necessity of investigating a detailed occupational history of M6 patients. Erythroleukemia tendency to transform to other types of AML late in its course has been previously reported [16], but was not assessed in our series because of the retrospective character of the study. However, several cases displayed this transformation (data not shown). Erythroleukemia is cytogenetically heterogeneous with high frequency of complex karyotypes [6, 8, 9, 24, 25] However, no specific cytogenetic abnormalities have been consistently identified in the relatively small series of M6 patients previously reported Most chromosomal abnormalities, described in M6 patients, have been reported in poor-risk AML. Loss of material from chromosomes 5, 7 and/or 17 was noted in 17 out of 35 cases studied. This points to the absence of specificity of —5/5q—, - 7 , or -17 abnormalities. Unfortunately, cytogenetic analysis was only performed successfully after 1980, yielding karyotypic data in only 35 cases. However, our results suggest that chromosomal findings might be a useful prognostic indicator of M6 AML There exist at least two cytogenetically distinct groups with different outcomes according to the presence or absence of cytogenetic abnormality. This confirms previously reported obvious differences in response to chemotherapy among the karyotypic groups [6, 8, 9, 12]. In contrast to a previous report [12], we did not find any difference between de novo AML and secondary AML in terms of karyotypic data. It has been suggested that M6 patients respond less well to chemotherapy than other AML patients [26-28]. This is confirmed by our retrospective study, but remains to be rigorously demonstrated in prospective studies using FAB criteria for patient selection and uniform treatment regimens. Previous retrospective studies have shown that survival was significantly better for patients treated with a regimen including daunorubicin alone or in combination with cytarabine, as compared with other

Downloaded from https://academic.oup.com/annonc/article-abstract/12/4/451/236782 by guest on 29 July 2018

455 therapeutic combinations [20, 28]. Too few patients received allogeneic or autologous transplantation to evaluate the effects of these approaches on survival We notice survival differences among certain subgroups. In contrast to a previous report [20], in which the group was subdivided according to age, a marked difference in survival was noted between the young subset and the older subset Another difference in survival emerged according to whether M6 was diagnosed de novo, or following chemotherapy, toxic exposure, or a chronic haematologic disorder, that has also been recognized as prognostic value by most previous studies [20, 29] In this retrospective study, we therefore confirmed the clinical aspects and the overall outcome of M6 AML. We also point to certain factors of prognostic value that would ultimately lead to a better understanding of this form of AML. However, larger series are still warranted in order to clarify treatment effects, particularly those of intensified therapy used as continuation treatment after CR achievement. Only a large prospective multicenter study with well-defined diagnostic criteria could break up those drawbacks inherent to a retrospective analysis.

12

13

14 15 16

17

18

19

20

21

References 1

Copelli M Di una emopatia sistemizzata rapresentdtd da una iperplasia enlroblastica (eritromatosi) Pathologica 1912, 4 460-5 2 Di Guglielmo G Richerche di ematologica Folia Medica 1917, 3 386-96 3 Dameshek W The Di Guglielmo syndrome Blood 1958, 13 192 4 Dameshek W The Di Guglielmo syndrome revisited J Hematol 1969,34 567-72 5 Bennetl JM, Catovsky D, Daniel MT et al Proposals for the classification of the acute leukemias Br J Haematol 1976. 33 451-8 6 Sakurai M, Sandberg AA Chromosomes and causation of human cancer and leukemia XIII An evaluation of karyotypic findings in erythroleukemia Cancer 1976, 37 790-804 7 Trent JM, Dune BGM, Davis JR et al Cytogenetic and clinical assessment of six patients with erythroleukemia Anti-Cancer Res 1983,3 111-6 8 Cuneo A, Van Orshoven A, Michaux JL et al Morphologic, immunologic and cytogenetic studies in erythroleukemia Evidence for multilineage involvement and identification of two distinct cytogenetic-clinicopathological types Br J Haematol 1990,75 346-54 9 Olopade OI, Thangavelu M, Larson RA et al Clinical, morphologic, and cytogenetic characteristics of 26 patients with erythroblastic leukemia Blood 1992, 80 2873-82 10 Bennett JM, Catovsky D, Daniel MT et al Proposed revised criteria for the classification of acute myeloid leukemia A report of the French-American-British Cooperative Group Ann Intern Med 1985, 103 620-5 11 ISCN (International System for Cytogenetic Nomenclature) Guidelines for cancer cytogenetics In Mitelman F (ed) Supple-

Downloaded from https://academic.oup.com/annonc/article-abstract/12/4/451/236782 by guest on 29 July 2018

22 23

24

25

26

27 28 29

ment to an International System for Human Cytogenetic Nomenclature Basel Karger 1991. 1-53 Nakamura H Cytogenetic heterogeneity in erythroleukemia defined as M6 by the French-American-British (FAB) Cooperative Group criteria Leukemia 1989, 3 305-9 Ellison RR, Holland JF. Weil M et al Arabinosyl cytosine A useful agent in the treatment of acute leukemia in adults Blood 1968,32 507-23 Preisler HD Failure of remission induction during the treatment of acute leukemia Med Pediatr Oncol 1990, 4 275-6 Sculy RE, Mark EJ, McNeely BU Case records of the Massachusetts General Hospital N Engl J Med 1983, 308 1081-91 Atkinson J, Hnsinko MA, Weil SC Erythroleukemia A review of 15 cases meeting 1985 FAB criteria and survey of the literature Blood Rev 1993, 6 204-14 Mazzella FM, Kowal-Vern A. Shnt MA et al Acute erylhroleukemia Evaluation of 48 cases with reference to classification, cell proliferation, cytogenelics, and prognosis Am J Clin Pathol 1998, 110 590-8 Goldberg SL, Noel P, KlumppTR et al The erythroid leukemias. a comparative study of erythroleukemia (FAB M6) and Di Guglielmo disease Am J Clin Oncol 1998, 21 42-7 Davey FR, Abraham N Jr, Brunetlo VL et al Morphologic characteristics of erythroleukemia (acute myeloid leukemia. FAB-M6) A CALGB study Am J Hematol 1995.49 29-38 Roggli VL, Subach J, Saleem A Prognostic factors and treatment effects on survival in erythroleukemia A retrospective study of 134 cases Cancer 1981, 1101-5 Form A, Moreo L Chromosome studies in a case of benzeneinduced erythroleukaemia EurJ Cancer 1969, 5 459-63 Rozman C, Woessner S. Saez-Serrania J Acute erythromyelosis after benzene poisoning Acta Haematol 1968, 40 234-7 Mantovani G, Del Giacco GS. Marongiu F el al Acute erythroblastic leukemia as a terminating event in a very long survivor (27 years) with Hodgkin's disease Haematologica 1982. 67 411-7 Kadam PR. Balsara BR. Zafaraullah KZ et al Cytogenetic features of erythroleukemia (EL) A study of II cases Cancer Genet Cytogenet 1990, 50 89-96 Torrabadella M, Vallespi T Chromosome analysis as an aid in differential diagnosis of erythroleukemia (M6) and myelodysplasia Cancer Genet Cytogenet 1990, 49 I39^»O Tamura K, Preisler HD Treatment of erythroleukemia with anthracychne antibiotics and cytosine arabinoside Cancer 1983. 51 1795-9 Sondergaard-Petersen H The Di Guglielmo syndrome A sludy of 17 cases Acta Medica Scand 1975, 198 165-74 Bloomfield CD Daunorubicin-prednisone treatment of erythroleukemia Ann Intern Med 1974. 81 746-50 Brandman J, Bukowski RM, Greenstreet R et al Prognostic factors affecting remission, remission duration, and survival in adult acute nonlymphocytic leukemia Cancer 1979. 44 1062-5

Received 18 September 2000. accepted 22 November 2000

Correspondence to X Thomas, MD, PhD Service d'Hematologie Hopital Edouard Hernot 69437 Lyon cedex 03 France E-mail xavier thomas(«)chu-lyon fr