Risk-adapted postremission therapy in acute myeloid leukemia: results of the german multicenter AML HD93 treatment trial

(2003) 17, 1521 1528 & 2003 Nature Publishing Group All rights reserved 0887-6924/03 $25.00 www.nature.com/leu Risk-adapted postremission therapy in acute myeloid leukemia: results of the german multicenter AML HD93 treatment trial RF Schlenk 1, A Benner 2, F Hartmann 3, F del Valle 4, C Weber 5, H Pralle 5, JTh Fischer 6, U Gunzer 7, A Pezzutto 8, W Weber 9, W Grimminger 10, J Prei 11, M Hensel 12, S Fröhling 1,KDöhner 1, R Haas 13 and H Döhner 1 for the AML Study Group Ulm (AMLSG ULM) 1 Department of Internal Medicine III, University of Ulm, Germany; 2 Central Unit of Biostatistics, German Cancer Research Center Heidelberg, Germany; 3 Department of Internal Medicine I, University of Homburg, Germany; 4 Städtische Kliniken, Oldenburg, Germany; 5 Department of Internal Medicine IV, University of Giessen, Germany; 6 Städtisches Klinikum, Karlsruhe, Germany; 7 Department of Internal Medicine I, University of Würzburg, Germany; 8 Robert Rössle Klinik, Charite, Humboldt-University of Berlin, Germany; 9 Krankenhaus der Barmherzigen Brüder, Trier, Germany; 10 Bürgerhospital, Stuttgart, Germany; 11 Caritas-Klinik St Theresia, Saarbrücken, Germany; 12 Department of Internal Medicine V, University of Heidelberg, Germany; and 13 Department of Hematology, University of Düsseldorf, Germany The objective of the AML HD93 treatment trial was to evaluate the outcome in young adults with acute myeloid leukemia (AML) after postremission therapy was stratified according to cytogenetically defined risk. The rationales for the study design were based (i) on previous favorable results with high-dose cytarabine in AML with t(8;21), inv/t(16q22) and in AML with normal karyotype, and ii) on encouraging results obtained in several phase II trials using autologous stem cell transplantation (SCT). Between July 1993 and January 1998, 223 eligible patients, 16 60 years of age with newly diagnosed AML other than French American British type M3/M3v, were entered into the trial. Risk groups were defined as follows: low risk: t(8;21) or inv/t(16q22); intermediate risk: normal karyotype; high risk: all other chromosomal abnormalities. Following intensive double induction therapy with idarubicin, cytarabine and etoposide, all patients in complete remission (CR) received a first consolidation therapy with high-dose cytarabine and mitoxantrone (HAM). A second consolidation therapy was stratified according to the risk group: low risk: HAM; intermediate risk: related allogeneic SCT or sequential HAM; high risk: related allogeneic or autologous SCT. Double induction therapy resulted in a high CR rate of 74.5%, and 90% of the responding patients were eligible for consolidation therapy. Survival for all 223 trial entrants was 40%, and for the 166 patients who entered CR, disease-free (DFS) and overall survival were 40 and 51% after 5 years, respectively. Within the low-, intermediate- and high-risk groups, DFS and survival after 5 years were 62.5 and 87, 40 and 49 and 17 and 26% respectively, without advantage for allogeneic transplantation in the intermediate- and high-risk groups. Postremission therapy-related mortality was 0, 7 and 14%, respectively. This study demonstrates the feasibility of cytogenetically defined risk-adapted consolidation therapy. The overall trial results are at least equivalent to those of published trials supporting the risk-adapted treatment strategy. (2003) 17, 1521 1528. doi:10.1038/sj.leu.2403009 Keywords: acute myeloid leukemia; risk-adapted postremission therapy Introduction Correspondence: Dr H Döhner, Department of Internal Medicine III, University of Ulm, Robert-Koch-Strasse 8, Ulm 89081, Germany; Fax: +49 731 5002 4493 In acute myeloid leukemia (AML), the karyotype at diagnosis is the most powerful factor predicting for response to therapy and survival. 1 5 This has opened the possibility to select therapy according to the biological risk profile. In the last decade, postremission strategies such as highdose cytarabine, allogeneic or autologous stem cell transplantation (SCT) have been shown to improve the outcome of adult patients with AML. In several randomized treatment trials, dose intensification of cytarabine significantly improved the disease-free (DFS) and overall survival (OS). 6,7 However, there is increasing evidence that not all patients benefit from the curative impact of cytarabine dose intensification, but only subgroups of patients who can be selected based on cytogenetic characteristics. In a pivotal study conducted by the Cancer and Group B (CALGB), it was shown that the remission duration was significantly improved by high-dose cytarabine in the group of patients exhibiting t(8;21) or inv/t(16q22) affecting the same core binding factor complex, and in the group of patients with a normal karyotype. 2 In contrast, the remission duration in this study was short for patients exhibiting other chromosomal abnormalities, irrespective of the dose of cytarabine given. For patients with poor risk features lacking an HLAcompatible donor, autologous bone marrow or peripheral blood SCT may be considered in first remission, since transplantation permits a further increase of dose intensity. In phase II studies, autologous SCT was shown to prolong DFS, in particular with mafosfamid purged bone marrow grafts. 8 Several randomized treatment trials addressed the question of whether autologous SCT is superior to either no further treatment or conventional chemotherapy. 9 12 In two studies, DFS was significantly improved for patients receiving autologous SCT compared to either no further treatment or a second course of intensive chemotherapy. 10,12 In contrast, in the other two studies, there was no difference in DFS between autologous SCT and intensive chemotherapy. 9,11 In the American intergroup study, OS was even significantly better after intensive chemotherapy compared to autologous and allogeneic SCT. 9 In 1993, we initiated the multicenter treatment trial AML HD93, in which late consolidation therapy was stratified according to the karyotype at diagnosis. Stratification was performed according to the criteria published by CALGB prior to initiation of our trial, that is, patients with the clonal chromosome abnormalities t(8;21) or inv/t(16q22) were considered as low-risk patients, those with normal karyotype were considered as intermediate-risk patients, and those with all other chromosomal abnormalities were assigned to the high-risk group. 13 The main objectives of this study were to evaluate the feasibility of a risk-adapted treatment approach in AML and to evaluate the predictive value of the karyotype for outcome after such a risk-adapted consolidation therapy.

1522 Methods Patients Patients 16 60 years of age with primary AML defined by the French American British (FAB) classification system 14,15 or secondary AML following treatment of a primary malignant disease were eligible for entry into the trial. Patients with acute promyelocytic leukemia FAB-type M3 or M3v were excluded and treated in the APL HD95 trial (Schlenk et al. Blood 2000; 96: 723; abstract). Patients with a history of myelodysplasia, concomitant liver, renal or cardiac disease were excluded. Written informed consent was obtained at study entry. The study was approved by the local Ethics Review Committee. Cytogenetics Chromosome banding analysis using standard methods was performed centrally for all patients in the Laboratory for Cytogenetic and Molecular Diagnosis of the AMLSG ULM. The criteria used to describe a cytogenetic clone and the description of the karyotype followed the recommendations of the International System for Human Cytogenetic Nomenclature. 16 All specimens were also analyzed by fluorescence in situ hybridization (FISH) as previously described, using a comprehensive set of DNA probes that allows one to detect the most recurrent structural and numerical chromosomal abnormalities in AML. 17,18 Study design All patients received a response-adapted double induction therapy consisting of a first induction therapy with ICE (idarubicin 12 mg/m 2 i.v., days 1, 3 and 5; cytarabine 100 mg/ m 2 continuously i.v., days 1 7; etoposide 100 mg/m 2 i.v., days 1 3), followed by a second cycle ICE in patients responding to the first cycle. Patients with refractory disease were assigned to a second induction therapy either with sequencial high-dose cytarabine and mitoxantrane (S-HAM) (cytarabine 3 g/m 2 /12 h i.v., days 1, 2, 8 and 9; mitoxantrone 10 mg/m 2 i.v., days 3, 4, 10 and 11) 19 for patients o55 years of age or HAM (cytarabine 3 g/m 2 /12 h i.v., days 1 3; mitoxantrone 12 mg/m 2 i.v., days 2 3) for patients X55 years of age. Bone marrow evaluation as well as start of the second cycle were scheduled between days 21 and 28. All patients in complete remission (CR) following double induction therapy were assigned to a first consolidation therapy with HAM. A peripheral blood stem cell collection or a bone marrow harvest was intended in all patients of the high-risk group lacking an HLA-compatible family donor. 20 Bone marrow was treated with mafosfamid in a dose-adapted manner prior to cryopreservation. 21 For a second consolidation therapy, patients o55 years of age were stratified according to the karyotype at diagnosis following criteria published by CALGB prior to initiation of our trial. 13 Low risk: AML exhibiting t(8;21) or inv/t(16q22) irrespective of additional chromosomal aberrations were assigned to a second consolidation with HAM. Intermediate risk: AML exhibiting a normal karyotype were assigned to an allogeneic SCT, if an HLA-identical family donor was available, or to a second consolidation with S-HAM. High risk: AML exhibiting all other chromosomal abnormalities were assigned to allogeneic SCT, if an HLA-identical family donor was available, or to autologous SCT. The preparative regimen for both autologous and allogeneic SCT consisted of either hyperfractioned total body irradiation (12.0 14.4 Gy) (TBI/Cy) or busulfan (16 mg/kg p.o.) (Bu/Cy) followed by cyclophosphamide (120 200 mg/kg i.v.). Patients X55 years of age were not stratified and assigned to a second consolidation with HAM. Criteria for response and definition of relapse Response assessment during response-adapted induction therapy was performed at two time points and was defined differently. The first time point was between days 21 and 28 after first induction therapy. Response to first induction therapy was defined by X50% reduction of the bone marrow blast count compared to the pretreatment count, reoccurrence of normal hematopoiesis not requiring full hematological recovery since the start of second induction therapy was scheduled directly after response evaluation according to double induction strategy. The second time point of response assessment was after double induction therapy. The definition of CR followed conventional criteria. 22 CR was defined by the presence of less than 5% blasts in the bone marrow and full hematological recovery with at least a granulocyte count of 1.5 10 9 /l, a platelet count of 100 10 9 /l, and no myeloid blasts in the peripheral blood. Causes of therapeutic failure were subdivided into two categories: (i) refractory disease (RD) and (ii) death less than 7 days after completion of the first induction therapy (early death: ED) or death during double induction therapy-induced bone marrow hypoplasia (hypoplastic death: HD). Relapse was defined as the occurrence of more than 10% blasts in two sequential bone marrow aspirates performed in a time interval of 2 weeks, or new extramedullary leukemia in patients with previously documented CR. Statistical analysis The primary end points of the study were OS and DFS. Survival end points, measured from study entry, were death (failure) and alive at last follow-up (censored). DFS end points, measured from the date of documented CR, were relapse (failure), death in CR (failure) and alive in CR at last follow-up (censored). The median duration of follow-up was calculated according to the method of Korn. 23 The Kaplan Meier method was used to estimate the distribution of OS and DFS, and probability of white blood cell (WBC) and platelet reconstitution times. 24 Confidence interval estimation was based on the cumulative hazard function using Greenwood s formula for the standard error estimation. 25 Survival distributions were compared using the (exact) log-rank test. 26 The proportional-hazard regression model of Cox was used to identify differences in survival due to prognostic factors. 27 Missing data were estimated using a multiple-imputation technique with 10 random draws. A limited backward-selection procedure was used to exclude redundant or unnecessary variables. 28 Testing and estimation of possible cutoff values for WBC was done by maximally selected log-rank statistics. 29 An effect was considered to be statistically significant if the P-value was 0.05 or less. To provide quantitative information on the relevance of statistically significant results, 95% confidence intervals of hazard ratios were computed. The comparison of the distribution of WBC according to the subtype of the core binding factor lesion was performed using a twosided Mann Whitney test. Pairwise correlations between WBC and LDH were estimated by Pearson s product moment correlation coefficient. The statistical analyses were performed

with the statistical software packages StatXact (Cytel Software Corp., Cambridge, MA, USA), SAS (SAS Institute Inc., Cary, NC, USA) and S-Plus (Insightful Corp., Seattle, WA, USA) together with the Design software library. 28 Results Accrual of patients Between July 1993 and January 1998, 256 patients were registered for the study. In all, 33 patients were considered to be ineligible for the following reasons: 20 patients had diagnosis other than AML, nine patients were excluded because of protocol violation, and four patients had concomitant disease (lung cancer, Wegener s granulomatosis with renal insufficiency, cardiac insufficiency, alcohol abuse). Table 1 shows the characteristics of the 223 eligible patients. Cytogenetic analysis as assessed by chromosome banding analysis and FISH was successful in 217 of the 223 patients (97%). 18 Risk assignment of the 223 patients was as follows: low risk, n ¼ 42 (19%); intermediate risk, n ¼ 101 (45%); high risk, n ¼ 74 (33%); not evaluable, n ¼ 6 (3%). Response to induction therapy Of the 223 patients, 158 (71%) responded to the first induction cycle with ICE. A total of 20 patients died from ED or HD during the first induction therapy, and five of the responding patients died due to HD after ICE II leading to an 11% ED/HD rate for double induction therapy. Of the 45 patients with RD after ICE I, 31 had high-risk cytogenetics, 12 had a normal karyotype and two had no evaluable metaphases. In all, 39 patients received a second induction therapy: 12 of 32 (37%) achieved CR after Table 1 Characteristics of the 223 eligible patients with AML Sex no. of patients (%) Male 110 (49) Female 113 (51) Age years (n=223) Median 46 Range 16 60 Type of AML no. of patients (%) De novo 207 (93) AML after cured cancer 16 (7) WBC count (10 9 /l) (n=208) Median 15.3 Range 0.54 369 LDH (U/l) (n=194) Median 403 Range 89 6430 FAB classification no. of patients (%) M0 8 (4) M1 34 (15) M2 52 (23.5) M4 65 (29) M5 34 (15) M6 6 (3) M7 3 (1) Missing 21 (9.5) Cytogenetic risk group a no. of patients (%) Low risk: inv/t(16q22); t(8;21) 42 (19) Intermediate risk: normal karyotype 101 (45) High risk: all other chromosomal aberrations 74 (33) Not evaluable 6 (3) a According to CALGB criteria. 13 Table 2 Results of response-adapted double induction therapy according to the karyotype No. of patients (%) CR RD ED/HD Low risk 39 (93) F 3 (6) abn(16q22) 23 (92) F 2 (8) t(8;21) 16 (94) F 1 (6) Intermediate risk 78 (77) 9 (9) 14 (14) High risk a 45 (61) 21 (28) 8 (11) abn(3q) 4 (40) 3 (30) 3 (30) 5/5q 7 (44) 7 (44) 2 (12) 7/7q 6 (32) 8 (42) 5 (26) abn(11q23) 16 (80) 1 (5) 3 (15) abn(12p) 1 (14) 5 (72) 1 (14) abn(17p) 1 (12.5) 5 (62.5) 2 (25) Various 15 (62.5) 9 (37.5) F Complex 5 (23) 12 (54) 5 (23) Not evaluable 4 (67) 2 (33) F Total 166 (74.5) 32 (14.5) 25 (11) a Some cases are listed in more than one category. S-HAM/HAM, and one of seven (14%) had CR after a second course of ICE. In total, response-adapted double induction therapy resulted in a CR rate of 74.5%. In Table 2, the results of response-adapted double induction therapy are summarized according to the karyotype. In the low-risk group CR rates were high without cases of RD, whereas in the high-risk group CR rates were heterogeneous, ranging from 80% in patients exhibiting abn(11q23) to less than 20% in patients exhibiting abn(12p) or abn(17p). Survival analysis after risk-adapted consolidation therapy Reason for patient exclusion: Of the 166 patients who entered CR on the protocol, 17 were not eligible for risk-adapted consolidation therapy for the following reasons: n ¼ 13, violation of the consolidation therapy protocol; n ¼ 4, no karyotype available. For these patients, DFS and survival were 56 and 53% at 5 years. The remaining 149 (90%) patients were eligible for risk-adapted consolidation therapy. Figure 1 illustrates the trial profile and the quantitative patient s distribution on the different treatment arms. Low risk (t(8;21) inv/t(16q22)) Of the 39 patients in CR after double induction therapy, 32 were eligible for consolidation therapy with two consecutive cycles of HAM. A total of 30 patients received the assigned two cycles of HAM. Two patients suffered from persistent thrombocytopenia after the first cycle of HAM and therefore did not receive the second cycle. Of the 32 patients, 12 relapsed between 6 and 22 months after first documented CR (Figure 2), resulting in an estimated DFS after 5 years of 62.5% (95% confidence interval (CI), 48 82%). An exploratory multiple Cox regression analysis was performed with DFS as end point, and included the following variables: subtype of the core binding factor lesion (t(8;21) vs inv/ t(16q22)), additional chromosomal abnormalities (except loss of sex chromosomes), WBC at diagnosis as binary variable with 1523

1524 Figure 1 Design of the study. The numbers represent the number of patients at each treatment step. HAM, high-dose cytarabine and mitoxantrone, S-HAM, sequential HAM, ALLO, allogeneic stem cell transplantation, AUTO, autologous stem cell transplantation. Patients X55 years of age were not stratified and assigned to a second consolidation with HAM. Figure 2 Estimated DFS of the 149 patients eligible for consolidation therapy according to the cytogenetically defined risk group. Figure 3 Estimated DFS of the 32 low-risk patients eligible for consolidation therapy according to the initial WBC count. The cutoff of 28.7 10 9 /l was defined by maximally selected log-rank statistics with P ¼ 0.002. a cut point at 28.7 10 9 /l defined by maximally selected logrank statistics, and age at diagnosis. Only WBC (P ¼ 0.005) as a binary variable was found to be a statistically significant prognostic variable in the Cox s regression analysis. The type of core binding factor lesion was not found to be informative for DFS, although patients with inv/t(16q22) showed significant higher WBC at diagnosis (median 35.1 10 9 /l) compared to patients with t(8;21) (median 10.8 10 9 /l) (P ¼ 0.03). Figure 3 shows the DFS according to the WBC at diagnosis with the cut point of 28.7 10 9 /l. Of the 12 relapsed patients, all received a reinduction therapy containing high-dose cytarabine and 11 achieved a second CR. Patients in second CR received the following consolidation therapies: allogeneic SCT from an HLA-matched family donor (n ¼ 6), autologous SCT (n ¼ 3), maintenance therapy with interleukin-2 (n ¼ 1), FLAG (n ¼ 1). Of the 32 patients, four died. The estimated survival after 5 years was 87% (95% confidence interval, 76 100%) (Figure 4). Intermediate risk (normal karyotype): Of the 78 patients in CR after double induction therapy, 73 were eligible for consolidation therapy and all received first consolidation with HAM. DFS for the entire group after 5 years was 40% (95% CI, 30 54%) (Figure 2).

1525 Figure 4 Estimated survival of the 149 patients eligible for consolidation therapy according to the cytogenetically defined risk group. The difference between the curves was significant by log-rank statistics (Po0.0001). Of the 16 patients with an HLA-compatible family donor, 13 received the assigned allogeneic SCT. Two patients received S- HAM and one patient received no further treatment. Of the 16 patients, five patients relapsed, and three patients died from transplantation-related causes resulting in a DFS after 5 years of 49% (95% CI, 30 82%). Of the 57 patients with no HLA-compatible family donor, 48 received the assigned S-HAM/HAM treatment. Five patients had one consolidation therapy only, and four patients relapsed before the start of the second consolidation therapy. The median times of unsupported platelet counts below 50 10 9 /l and of WBC below 1.0 10 9 /l were 44.5 days (range 21 85) and 31 days (range 23 49) in patients treated with S-HAM, and 18 days (range 9 37) and 12 days (range 6 24) in patients treated with HAM, respectively. In total, 34 patients relapsed, one died from invasive aspergillosis during hypoplasia and one patient died 1 year after completion of consolidation therapy because of sudden cardiac death resulting in a DFS after 5 years of 38% (95% CI, 27 53%). Comparison of the distribution of DFS and survival between intensive chemotherapy and allogeneic SCT within the intermediate-risk group showed no statistically significant difference (P ¼ 0.28, 0.7, respectively. Figure 5 for DFS). Notably, within the chemotherapy group there was no difference between the S- HAM and the HAM arm (P ¼ 0.87), although these arms represented two different age groups. The following variables were examined in an exploratory multiple Cox regression analysis: WBC at diagnosis, LDH at diagnosis and age at diagnosis. In Cox s regression model no variable was found to be statistically significant at the 5% level. Of the 39 relapsed patients, 27 received an intensive reinduction therapy and 13 achieved a second CR. Three patients received autologous SCT in second CR, 12 patients allogeneic SCT (eight in second CR, four with RD) and one patient FLAG. Of the 73 patients, 36 died; the estimated survival after 5 years was 50% (95% CI, 40 63%) (Figure 4). High risk (all other chromosomal aberrations): Of the 45 patients in CR after double induction therapy, 44 were eligible for consolidation therapy and all received first consolidation with HAM. The estimated DFS for the entire high-risk group after 5 years was 17% (95% CI, 9 34%) (Figure 2). Figure 5 Estimated DFS of the 73 intermediate-risk patients eligible for consolidation therapy according to the treatment arm. HAM/S-HAM, high-dose cytarabine and mitoxantrone/sequential HAM, ALLO, allogeneic stem cell transplantation. The difference between the curves was not significant by log-rank statistics (P ¼ 0.28). Of the 12 patients, 11 received the assigned allogeneic SCT, including one patient who had SCT from an unrelated donor. Five of the 12 patients had been refractory to the first induction therapy. Nine patients relapsed and two patients died from transplantation-related causes, resulting in a DFS after 5 years of 8% (95% CI, 1 54%). Of the 23 patients, 18 received the assigned autologous SCT. Three of the 23 patients had been refractory to the first induction therapy. In all, 12 patients were transplanted with mafosfamid purged bone marrow, and six patients received unmanipulated peripheral blood progenitor cells. Autologous SCT with mafosfamid purged bone marrow harvested after the first consolidation therapy with HAM was stopped in September 1996 because of unacceptable hematological toxicity; the median time from transplantation to an unsupported platelet count 450 10 9 /l and to a WBC count 41.0 10 9 /l was 125 days (range 42 330 days) and 71 days (range 17 108), respectively. It was replaced by transplantation with autologous peripheral blood progenitor cells collected during GCSF supported hematological recovery after the first consolidation therapy with HAM. The hematological toxicity was markedly reduced with a median time from transplantation to an unsupported platelet count 450 10 9 /l and to a WBC count 41.0 10 9 /l of 26.5 days (range 17 42) and 14.5 days (range 12 17 days), respectively. Of the 23 patients, 15 patients relapsed and three patients died from treatmentrelated causes, resulting in a DFS after 5 years of 22% (95% CI: 10 47%). Of the nine patients aged X55 years assigned to HAM, seven received the treatment. Seven patients relapsed and one died after first consolidation, resulting in a DFS after 5 years of 17% (95% CI, 3 88%). Of the 31 high-risk patients who relapsed, 18 received an intensive reinduction therapy and seven achieved a second CR. Of the 44 patients, 32 died and the survival after 5 years was 26% (95% CI, 15 43%) (Figure 4). Analyses were restricted to patients age o55 years to evaluate the influence of allogeneic and autologous transplantation on DFS in the presence of other risk factors. Cox s regression analysis for DFS examining WBC at diagnosis, LDH at diagnosis, age at diagnosis, response to first induction therapy and treatment group revealed the response to first induction as the only prognostic variable (P ¼ 0.02), with a hazard ratio for

1526 relapse of 2.7 for patients without response to first induction therapy (95% CI, 1.2 6.3). OS analysis: After a median follow-up time for survival of 64 months of all trial entrants, 134 of the 223 patients had died. The median survival time for the entire group was 21 months (95% CI, 16 41 months), and survival after 5 years was 40% (95% CI, 33 46%). For the 166 patients who entered CR on the protocol, DFS and survival after 5 years were 40% (95% CI, 33 48%) and 51% (95% CI, 44 59%), respectively. DFS and OS rates after 5 years were 62.5 and 87, 40 and 49, 17 and 26% for the low-, intermediate-, and high-risk groups, respectively. The distribution of treatment-related mortality during and after consolidation therapy was 0, 7 and 14% for the low-, intermediate- and highrisk groups, respectively. One additional analysis was performed on all eligible trial entrants. The patients were subdivided into three groups according to the response to first induction therapy with regard to the quality of response: (i) patients with response to first induction therapy and o5% blasts in the bone marrow; (ii) patients with response to first induction therapy and 5 25% blasts in the bone marrow; and (iii) no response to first induction therapy. The difference between patients with response compared to patients with no response to first induction therapy was highly significant (Po0.0001). However, there was no difference in survival for patients with response to first induction therapy regarding quality of response (BM blasts o5% vs 5 25%) (Figure 6). Discussion In this study, we used the karyotype at diagnosis for risk-adapted stratification of consolidation therapy in AML. As the study was initiated in 1993, risk assignment was performed using the classification scheme initially proposed by the CALGB, which considered t(8;21) and inv/t(16q22) as low risk, normal karyotype as intermediate risk, and all other chromosome abnormalities as high risk. 13 Using a response-adapted, intensive double induction therapy, we achieved a high overall CR rate of 74.5%, with a low ED/HD rate of 11% including hypoplastic deaths as late as 84 days from start of induction therapy. In agreement Figure 6 Estimated survival of 203 patients according to the response to first induction therapy. The difference between the curves was significant by log-rank statistics (Po0.0001). with previously published studies, 3 5 karyotype at diagnosis was the most powerful indicator for response to induction therapy with no refractory AML in the low-risk group and a high rate of refractoriness in AML exhibiting abn(3q), 5/5q, 7/7q, abn(12p), abn(17p) and/or complex karyotype. The incorporation of a high-dose cytarabine-based regimen in the response-adapted double induction therapy resulted in a 37% response rate in patients refractory to the first cycle ICE. Analogous to the data of the MRC AML 10 trial, 30 the prognosis of patients with no response to the first induction cycle was dismal in our study with an OS of 12% after 5 years, irrespective of the response to the second induction therapy. Whether further dose intensification of the induction therapy harboring the risk of an increased early death rate, 31 or rather early transplantation from a sibling or an unrelated donor after myeloablative, 32,33 or dose-reduced conditioning 34 may improve outcome for this subgroup of patients has to be evaluated in prospective studies. For OS, we did not find a difference with respect to quality of response to first induction therapy, that is, whether CR or PR (up to 25% bone marrow blasts) was achieved. According to the MRC data 30 this supports the view that response to first induction therapy should only be considered as a high-risk feature when there is resistant disease. In the low-risk group, outcome was excellent with a DFS and survival after 5 years of 62.5 and 87%, respectively. The conventional, high-dose cytarabine-based postremission therapy was well tolerated and there was no treatment-related mortality. Thus, our results are in agreement with those of the study by the CALGB that demonstrated the beneficial effect of high-dose cytarabine in consolidation therapy in this subgroup of patients. 2 The results compare favorably to those achieved with allogeneic SCT in first CR, 3,35 and our study supports the strategy to reserve allogeneic SCT for relapsing patients. As already discussed in the recent publication of the French AML Intergroup, 36 it remains unknown whether this beneficial effect of cytarabine in this subgroup is due to its specific mode of action or whether similar good results can also be achieved by dose intensification of other antileukemic drugs. It has been shown in several studies that the initial WBC is of prognostic value in AML exhibiting t(8;21) or inv/ t(16q22). 36 39 In our study, Cox s regression model revealed the WBC at diagnosis as a significant prognostic factor, with an estimated cutoff value of 28.7 10 9 /l, and this was confirmed in a meta-analysis of low-risk patients including patients of the present study and those treated in another German multicenter treatment trial (SHG 2/95) (Schlenk et al. Blood 2001; 98: 593; abstract). In the group of patients with normal karyotype, the DFS and survival after 5 years of 40 and 50%, respectively, are at least equivalent to those reported by other multicenter treatment trials. 2,3 In univariate analysis, there was no difference in DFS between the patients receiving conventional consolidation therapy and those who had allogeneic SCT from an HLAcompatible family donor. Only 69% of the relapsed patients received a reinduction therapy, and the CR rate of 48% was inferior compared to that of the low-risk group. Therefore, improvement of postremission therapy in first CR for this patient group is warranted. There are increasing data demonstrating that the subgroup of patients with a normal karyotype can be dissected based on molecular genetic findings. Patients with normal karyotype and mutations of, for example, the FLT3 or MLL gene have been shown to have significant inferior DFS and/ or survival. 40 44 In contrast, a recent study by a French group showed that patients with normal karyotype and dominant

negative mutations of the CEBPA gene may have an outcome similar to those with t(8;21). 45 Future treatment trials will likely have to consider these findings in this subgroup of patients, in particular with respect to the evaluation of novel drugs targeting the molecular defect and to the use of allogeneic SCT in all its facets. In the high-risk group we followed the approach of autologous SCT with mafosfamid purged bone marrow. 46 In contrast to previous studies that used perfosfamid 9 or mafosfamid 46 purged bone marrow collected after two induction cycles, in our trial the bone marrow was harvested after two intensive induction therapies and one consolidation therapy containing high-dose cytarabine. Due to the intensive in vivo and in vitro purging, the hematological reconstitution after autologous SCT was extremely long, with a median WBC reconstitution time of 71 days. In a pilot study, we showed the feasibility of collecting peripheral blood stem cells during the hematological reconstitution after two intensive induction therapies and first consolidation with HAM. 20 Using unpurged blood stem cell grafts, the median WBC reconstitution time was significantly shortened to 14.5 days. The DFS and survival after 5 years of 17 and 26%, respectively, of the entire high-risk group were discouraging but comparable to other multicenter treatment trials. 2,3 In univariate analysis restricted to patients o55 years of age, we did not identify a benefit in DFS for either allogeneic or autologous SCT. In multivariate analysis, the only prognostic factor was the response to first induction therapy, which was clearly correlated to cytogenetics. Recent studies evaluating the impact of dose intensification of the conditioning regimen with radio-labeled antibodies have shown promising results, and this experimental therapy should be explored prospectively in this very-high-risk group of patients. 47,48 Despite the various dose intensities used in the three risk strata, the karyotype retained its high impact on the clinical outcome. A critical issue related to the statistical analysis of our study is the question of whether the strategy of riskadjusted therapy improves outcome of the patients. The overall treatment results of our study with a survival of 40% after 5 years of all trial entrants are at least equivalent to those of published trials: in the MRC-10 trial, survival was 40% at 4.8 years in a clearly younger patient population, 10 in the ECOG study 35% at 4 years, 9 in the CALGB trial evaluating high-dose cytarabine 38% for patients under the age of 40 years and 27% for patients 40 60 years of age, 6 in the GOELAM trial 40% at 4 years also in a clearly younger patient population, 11 and in the AMLCG trial 31% after 4 years. 49 There seems to be a benefit when comparing survival of CR patients: in our study, survival of responding patients was 51% compared to 35 43% in the ECOG trial (depending on the treatment given), 9 40% in the GOELAM study 11 and 36% in the AMLCG trial. 49 Whether these differences are due to the treatment strategy or to other confounders remains open. To improve the comparability between different treatment strategies, the German AML Intergroup has recently initiated a common Intergroup study, in which patients are randomized upfront in a 1:10 ratio between a common standard treatment arm and the specific study that is active in the four AML study groups. 50 Our study demonstrates that karyotype-based stratified multicenter treatment trials are feasible in AML and provide an alternative to those trials in which all patients with AML are randomized either upfront or once a CR has been achieved. The overall treatment results compare favorably to those of published trials. Acknowledgements We gratefully acknowledge Brigitte Schreiter for technical assistance. References 1 Arthur DC, Berger R, Golomb HM, Swansbury GJ, Reeves BR, Alimena G et al. The clinical significance of karyotype in acute myelogenous leukemia. Cancer Genet Cytogenet 1989; 40: 203 216. 2 Bloomfield CD, Lawrence D, Byrd JC, Carroll A, Pettenati MJ, Tantravahi R et al. Frequency of prolonged remission duration after high-dose cytarabine intensification in acute myeloid leukemia varies by cytogenetic subtype. Cancer Res 1998; 58: 4173 4179. 3 Grimwade D, Walker H, Oliver F, Wheatley K, Harrison C, Harrison G et al. 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 Leukaemia Working Parties. Blood 1998; 92: 2322 2333. 4 Mrózek K, Heinonen K, Bloomfield CD. Clinical importance of cytogenetics in acute myeloid leukaemia. Baillieres Best Pract Res Clin Haematol 2001; 14: 19 47. 5 Slovak ML, Kopecky KJ, Cassileth PA, Harrington DH, Theil KS, Mohamed A et al. Karyotypic analysis predicts outcome of preremission and postremission therapy in adult acute myeloid leukemia: a Southwest Oncology Group/Eastern Cooperative Oncology Group Study. Blood 2000; 96: 4075 4083. 6 Mayer RJ, Davis RB, Schiffer CA, Berg DT, Powell BL, Schulman P et al. Intensive post remission chemotherapy in adults with acute myeloid leukemia. N Engl J Med 1994; 331: 896 903. 7 Weick JK, Kopecky KJ, Applebaum FR, Head DR, Kingsbury LL, Balcerzak SP et al. A randomized investigation of high-dose versus standard-dose cytosine arabinoside with daunorubicin in patients with previously untreated acute myeloid leukemia: A Southwest Oncology Group Study. Blood 1996; 88: 2841 2851. 8 Gorin NC. Autologous stem cell transplantation in acute myelocytic leukemia. Blood 1998; 92: 1073 1090. 9 Cassileth PA, Harrington DP, Appelbaum FR, Lazarus HM, Rowe JM, Paietta E et al. Chemotherapy compared with autologous or allogeneic bone marrow transplantation in the management of acute myeloid leukemia in first remission. N Engl J Med 1998; 339: 1649 1656. 10 Burnett AK, Goldstone AH, Stevens RM, Hann IM, Rees JK, Gray RG et al. Randomised comparison of addition of autologous bonemarrow transplantation to intensive chemotherapy for acute myeloid leukaemia in first remission: results of MRC AML 10 trial. UK Medical Research Council Adult and Children s Leukaemia Working Parties. Lancet 1998; 351: 700 708. 11 Harousseau JL, Cahn JY, Pignon B, Witz F, Milpied N, Delain M et al. Comparison of autologous bone marrow transplantation and intensive chemotherapy as postremission therapy in adult acute myeloid leukemia. The Groupe Ouest Est Leucemies Aigues Myeloblastiques (GOELAM). Blood 1997; 90: 2978 2986. 12 Zittoun RA, Mandelli F, Willemze R, de Witte T, Labar B, Resegotti L et al. Autologous or allogeneic bone marrow transplantation compared with intensive chemotherapy in acute myelogenous leukemia. European Organization for Research and Treatment of Cancer (EORTC) and the Gruppo Italiano Malattie Ematologiche Maligne dell Adulto (GIMEMA) Cooperative Groups. N Engl J Med 1995; 332: 217 223. 13 Bloomfield CD. Prognostic factors for selecting curative therapy for adult acute myeloid leukemia. 1992; 4 (Suppl 6): 65 67. 14 Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR et al. Proposals for the classification of the acute leukemias. Br J Haematol 1976; 33: 451. 15 Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DA, Gralnick HR et al. Proposed revised criteria for the classification of acute myeloid leukemia. Ann Intern Med 1985; 103: 626. 16 Mitelman F (ed) ISCN (1995). An International System for Human Cytogenetic Nomenclature. Basel, Switzerland: Karger, 1995. 17 Fischer K, Scholl C, Sàlat J, Fröhling S, Schlenk R, Bentz M et al. Design and validation of DNA probe sets for a comprehensive 1527

1528 interphase cytogenetic analysis of acute myeloid leukemia. Blood 1996; 88: 3962 3971. 18 Fröhling S, Skelin S, Liebisch C, Scholl C, Schlenk RF, Döhner H et al. Comparison of cytogenetic and molecular cytogenetic detection of chromosome abnormalities in 240 consecutive adult patients with acute myeloid leukemia. J Clin Oncol 2002; 20: 2480 2485. 19 Hiddemann W, Büchner T, Essink M, Koch O, Stenzinger W, van de Loo J. High-dose cytosine arabinoside and mitoxantrone: preliminary results of a pilot study with sequential application (S- HAM) indicating a high antileukemic activity in refractory acute leukemias. Onkologie 1988; 11: 10 12. 20 Schlenk RF, Döhner H, Pförsich M, Benner A, Fischer K, Hartmann F et al. Successful collection of peripheral blood progenitor cells in patients with acute myeloid leukaemia following early consolidation therapy with granulocyte colony-stimulating factor-supported high-dose cytarabine and mitoxantrone. Br J Haematol 1997; 99: 386 393. 21 Martin H, Bruecher J, Claude R, Hoelzer D. Cumulative chemotherapy increases mafosfamide toxicity for normal progenitor cells in AML patients: rationale for cryopreserving adapteddose purged marrow early in first complete remission. Bone Marrow Transplant 1993; 12: 495 499. 22 Cheson BD, Cassileth PA, Head DR, Schiffer CA, Bennett JM, Bloomfield CD et al. Report of the National Cancer Institutesponsored workshop on definitions of diagnosis and response in acute myeloid leukemia. J Clin Oncol 1990; 8: 813 819. 23 Korn EL. Censoring distributions as a measure of follow-up in survival analysis. Stat Med 1986; 5: 255 260. 24 Kaplan EL, Meier P. Non-parametric estimation form incomplete observation. J Am Stat Assoc 1958; 47: 457 481. 25 Therneau TM, Grambsch PM. Modeling Survival Data: Extending the Cox Model. New York: Springer-Verlag, 2000. 26 Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Nat Cancer Inst 1959; 22: 719 748. 27 Cox DR. Regression models and life tables. J R Stat Soc Ser B 1972; 34: 187 220. 28 Harrell FE. Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis. New York: Springer-Verlag, 2001. 29 Lausen B, Schumacher M. Maximally selected rank statistics. Biometrics 1992; 48: 73 85. 30 Wheatley K, Burnett AK, Goldstone AH, Gray RG, Hann IM, Harrison CJ et al. A simple, robust, validated and highly predictive index for the determination of risk-directed therapy in acute myeloid leukaemia derived from the MRC AML 10 trial. United Kingdom Medical Research Council s Adult and Childhood Leukaemia Working Parties. Br J Haematol 1999; 107: 69 79. 31 Kern W, Aul C, Maschmeyer G, Schonrock-Nabulsi R, Ludwig WD, Bartholomaus A et al. Superiority of high-dose over intermediate-dose cytosine arabinoside in the treatment of patients with high-risk acute myeloid leukemia: results of an age-adjusted prospective randomized comparison. 1998; 12: 1049 1055. 32 Biggs JC, Horowitz MM, Gale RP, Ash RC, Atkinson K, Helbig W et al. Bone marrow transplants may cure patients with acute leukemia never achieving remission with chemotherapy. Blood 1992; 80: 1090 1093. 33 Forman SJ, Schmidt GM, Nademanee AP, Amylon MD, Chao NJ, Fahey JL et al. Allogeneic bone marrow transplantation as therapy for primary induction failure for patients with acute leukemia. J Clin Oncol 1991; 9: 1570 1574. 34 Platzbecker U, Thiede C, Freiberg-Richter J, Rollig C, Helwig A, Schakel U et al. Early allogeneic blood stem cell transplantation after modified conditioning therapy during marrow aplasia: stable remission in high-risk acute myeloid leukemia. Bone Marrow Transplant 2001; 27: 543 546. 35 Ferrant A, Labopin M, Frassoni F, Prentice HG, Cahn JY, Blaise D et al. Karyotype in acute myeloblastic leukemia: prognostic significance for bone marrow transplantation in first remission: a European Group for Blood and Marrow Transplantation study. Acute Working Party of the European Group for Blood and Marrow Transplantation (EBMT). Blood 1997; 90: 2931 2938. 36 Nguyen S, Leblanc T, Fenaux P, Witz F, Blaise D, Pigneux A et al. A white blood cell index as the main prognostic factor in t(8;21) acute myeloid leukemia (AML): a survey of 161 cases from the French AML Intergroup. Blood 2002; 99: 3517 3523. 37 Billstrom R, Johansson B, Fioretos T, Garwicz S, Malm C, Zettervall O et al. Poor survival in t(8;21) (q22;q22)-associated acute myeloid leukaemia with leukocytosis. Eur J Haematol 1997; 59: 47 52. 38 Martin G, Barragan E, Bolufer P, Chillon C, Garcia-Sanz R, Gomez T et al. Relevance of presenting white blood cell count and kinetics of molecular remission in the prognosis of acute myeloid leukemia with CBFbeta/MYH11 rearrangement. Haematologica 2000; 85: 699 703. 39 O Brien S, Kantarjian HM, Keating M, Gagnon G, Cork A, Trujillo J et al. Association of granulocytosis with poor prognosis in patients with acute myelogenous leukemia and translocation of chromosomes 8 and 21. J Clin Oncol 1989; 7: 1081 1086. 40 Kiyoi H, Naoe T, Nakano Y, Yokota S, Minami S, Miyawaki S et al. Prognostic implication of FLT3 and N-RAS gene mutations in acute myeloid leukemia. Blood 1999; 93: 3074 3080. 41 Fröhling S, Schlenk RF, Breitruck J, Benner A, Kreitmeier S, Tobis K et al. Prognostic significance of activating FLT3 mutations in younger adults (16 to 60 years) with acute myeloid leukemia and normal cytogenetics: a study of the AML Study Group Ulm (AMLSG ULM). Blood 2002; 100: 4372 4380. 42 Thiede C, Steudel C, Mohr B, Schaich M, Schakel U, Platzbecker U et al. Analysis of FLT3-activating mutations in 979 patients with acute myelogenous leukemia: association with FAB subtypes and identification of subgroups with poor prognosis. Blood 2002; 99: 4326 4335. 43 Döhner K, Tobis K, Ulrich R, Fröhling S, Benner A, Schlenk RF et al. Prognostic significance of partial tandem duplications of the MLL gene in adult patients 16 to 60 years old with acute myeloid leukemia and normal cytogenetics: a study of the AML SG ULM. J Clin Oncol 2002; 20: 3254 3261. 44 Caligiuri MA, Strout MP, Lawrence D, Arthur DC, Baer MR, Yu F et al. Rearrangement of ALL1 (MLL) in acute myeloid leukemia with normal cytogenetics. Cancer Res 1998; 58: 55 59. 45 Preudhomme C, Sagot C, Boissel N, Cayuela JM, Tigaud I, de Botton S et al. Favorable prognostic significance of CEBPA mutations in patients with de novo acute myeloid leukemia: a study from the Acute French Association (ALFA). Blood 2002; 100: 2717 2723. 46 Körbling M, Fliedner TM, Holle R, Magrin S, Baumann M, Holdermann E et al. Autologous blood stem cell (ABSCT) versus purged bone marrow transplantation (pabmt) in standard risk AML: influence of source and cell composition of the autograft on hemopoietic reconstitution and disease-free survival. Bone Marrow Transplant 1991; 7: 343 349. 47 Bunjes D, Buchmann I, Duncker C, Seitz U, Kotzerke J, Wiesneth M et al. Rhenium 188-labeled anti-cd66 (a, b, c, e) monoclonal antibody to intensify the conditioning regimen prior to stem cell transplantation for patients with high-risk acute myeloid leukemia or myelodysplastic syndrome: results of a phase I II study. Blood 2001; 98: 565 572. 48 Matthews DC, Appelbaum FR, Eary JF, Fisher DR, Durack LD, Hui TE et al. Phase I study of (131)I-anti-CD45 antibody plus cyclophosphamide and total body irradiation for advanced acute leukemia and myelodysplastic syndrome. Blood 1999; 94: 1237 1247. 49 Büchner T, Hiddemann W, Wörmann B, Loffler H, Gassmann W, Haferlach T et al. Double induction strategy for acute myeloid leukemia: the effect of high-dose cytarabine with mitoxantrone instead of standard-dose cytarabine with daunorubicin and 6- thioguanine: a randomized trial by the German AML Cooperative Group. Blood 1999; 93: 4116 4124. 50 Büchner T, Döhner H, Ehninger G, Ganser A, Hasford J. Up-front randomization and common standard arm: a proposal for comparing AML treatment strategies between different studies. Res 2002; 26: 1073 1075.