Allogeneic transplantation as post-remission therapy for cytogenetically high-risk acute myeloid leukemia: landmark analysis from a single prospective multicenter trial

Abstract

Background Allogeneic hematopoietic cell transplantation is considered the preferred post-remission therapy in patients with acute myeloid leukemia cytogenetically defined as being at high risk. To substantiate evidence for allogeneic hematopoietic cell transplantation in first complete remission in these high-risk patients we performed a landmark analysis within a single prospective multicenter treatment trial.Design and Methods By the time of analysis, 2,347 patients had been accrued into the AMLCG 99 trial between 1999 – 2007. Out of this population, 243 patients under 60 years old fulfilled the criteria for high-risk cytogenetics. Landmark analyses were performed with a control cohort, who remained in first complete remission at least the median time from complete remission to transplantation in the intervention group.Results After standardized induction therapy, 111 patients under 60 years old achieved complete remission. A matched allogeneic donor was identified for 59 patients (30 sibling donors, 29 unrelated donors). Fifty-five patients received an allogeneic hematopoietic cell transplant after a median time of 88 days in first complete remission. Of the remaining 56 patients, 21 relapsed within 90 days after achieving first complete remission and for 7 patients with relevant comorbidities no donors search was initiated, leaving 28 patients given conventional post-remission therapy as the control cohort. The median follow-up of surviving patients was 60.4 months. Patients with an allogeneic donor had substantially better 5-year overall and relapse-free survival rates than the control group (48% versus 18%, P=0.004 and 39% versus 10%, P<0.001, respectively). A survival benefit from transplantation was evident regardless of donor type, age and monosomal karyotype.Conclusions Beyond evidence available for subgroups of high-risk patients, the findings of this study establish in a broader manner that allogeneic hematopoietic cell transplantation is a preferable consolidation treatment for patients with acute myeloid leukemia and high-risk cytogenetics.The study was registered at Clinicaltrials.gov as NCT00266136.

Introduction

High-risk acute myeloid leukemia (AML) is characterized mainly by cytogenetic features of the blast population, less often by immunophenotypic abnormalities, more regularly by secondary disease manifestation after myelodysplastic syndrome or cytotoxic treatment of another malignant disease and, finally, by response to induction therapy.1–5 Molecularly detectable mutations are currently being evaluated for their prognostic role and, consequently, as predictive markers guiding up-front treatment decisions.6,7 Evidence upon which treatment recommendations can be based must take into consideration this substantial heterogeneity.

Only about 20% of newly diagnosed AML patients are younger than 60 years and hence eligible for conventional allogeneic hematopoietic cell transplantation (HCT). It is obvious that any of the high-risk categories listed above represent rare condition and evidence for defining treatment standards is, therefore, unlikely to derive from single randomized trials. Indeed, most currently available evidence has been derived from cumulative analyses from sequential trials or meta-analyses covering trials conducted over substantial periods of time.8–11

From the body of evidence available it has emerged that, in high-risk AML patients, allogeneic HCT should be performed as soon as the patient, disease status and donor availability allow for it.2,9,10,12–15 Looking for support for this notion, we took the opportunity to analyze patients defined as at high risk according to cytogenetics (complex karyotype, −5/5q-, −7/7q-, abnormal 3q21/3q26, or abnormal 11q23) who underwent allogeneic HCT, comparing their outcomes with those of an appropriate landmark control cohort from one of the largest ongoing AML trials that had enrolled almost 2,400 patients at the time of this analysis.16,17 This large trial cohort allowed us to focus on patients in first complete remission (CR1) with balanced types of induction therapy, to address questions of donor types and conditioning intensities, to compare patients with different types of cytogenetic high risk including those with monosomal karyotypes (a recently established predictor for very poor prognosis of AML), and to perform multivariate analyses on sufficiently large subgroups. The benefit of allogeneic HCT in CR1 in cytogenetically defined high-risk AML patients has so far not been demonstrated under more stringent comparative conditions in the setting of a single prospective clinical trial.

Design and Methods

Study Population

Between May 1999 and May 2007 2,347 patients 16 years old or more (with no upper age limit) who had a previously untreated AML, except for acute promyelocytic leukemia, or suffered from high-risk myelodysplastic syndrome (refractory anemia with excess blasts in transformation) were included in the AMLCG 99 trial.16,17 At diagnosis, samples of bone marrow aspirates were examined for chromosomal abnormalities using standard banding techniques and classified according to the International System for Human Cytogenetic Nomenclature.2 A monosomal karyotype was defined as two or more monosomies, or a single monosomy in the presence of structural abnormalities.18,19 Two hundred and forty-three patients 21–59 years of age with a known unfavorable karyotype with or without monosomal karyotype, as detailed above, were considered in the present analysis (Online Supplementary Figure S1). According to the study protocol of the AMLCG99 trial, an unfavorable karyotype was defined as a complex karyotype, −5/5q-, −7/7q-, abnormal 3q21/3q26, or abnormal 11q23 throughout the whole study period. The trial was approved by the ethics committees of the participating centers and was conducted in accordance with the Declaration of Helsinki. Written informed consent was provided by all participants. The study was registered at Clinicaltrials.gov as NCT00266136.

Study design and treatment

Before treatment started, all patients were randomly assigned upfront to one of the two induction therapies: thioguanine, cytarabine, and daunorubicin (TAD) – high-dose cytarabine and mitox-anthrone (HAM) or HAM-HAM. In the same step patients were randomly assigned to post-remission therapy with either prolonged maintenance or autologous HCT. After achieving complete remission, patients received consolidation with one course identical to the TAD induction regimen. For maintenance treatment, patients were given monthly courses of cytarabine, and as a second agent from course to course, daunorubicin, thioguanine or cyclophosphamide, with the second agent changing in a rotating sequence. Maintenance was continued for 3 years. After recovery from TAD consolidation, patients younger than 60 years who were randomly assigned to autologous HCT received busulfan/cyclophosphamide conditioning before reinfusion of autologous stem cells. Independently of the random assignment it was recommended per protocol that younger patients with histo-compatible family donors should proceed to allogeneic HCT in CR1. In January 2002, the protocol was amended by recommending allogeneic HCT also from matched unrelated donors for patients under 60 years old with an unfavorable karyotype as consolidation therapy in CR1. The search for a related or unrelated stem cell donor was initiated immediately after the diagnosis of high-risk AML had been made. Until transplantation patients received conventional consolidation and maintenance treatment according to the protocol. Patients with a suitable donor and no contraindications (defined by the treating physicians) underwent allogeneic HCT according to the policies of the transplant center.

Post-induction/pretransplantation characteristics

Complete remission (CR) was defined according to standard morphological criteria; morphological CR with incomplete blood count recovery (CRi; absolute neutrophil count < 1x10/L and/or platelet count < 100x10/L) was included for this transplant analysis. Pre-transplantation comorbidities were assessed using the HCT comorbidity index (HCT-CI).20

HLA typing and matching

Patents and donors were typed at least for HLA-A, -B and –DRB1. After July 2001 all patients and donors were tested at least for HLA-A, and -B by at least intermediate-resolution DNA typing and HLA-DRB1 and DQB1 by high-resolution techniques. High-resolution typing of patients and donors for HLA-A, -B, -C, -DRB1 and -DQB1 was performed after April 2005.21

Statistical analyses

Data were analyzed as of April 1, 2010. The primary objective of this analysis was to compare allogeneic HCT and conventional chemotherapy as consolidation therapy in AML patients with an unfavorable karyotype in CR1. Overall survival was measured from the start of treatment until death. Relapse-free survival was counted from the achievement of CR1 until relapse or death in CR. Outcome comparisons were performed by landmark analyses, i.e. only patients with CR1 lasting 90 days or more were included in the control cohort in order to account for the median time from CR1 to transplantation (88 days) in the HCT comparison groups. Significances were calculated for overall survival and relapse-free survival by the log-rank test, and for multivariate analyses by logistic and Cox regression. Outcome measures were evaluated according to treatment received (as treated: allogeneic HCT in CR1 versus conventional treatment) and according to intended treatment (intention to treat: recommended allogeneic HCT, initiation of donor search and identification of a suitable allogeneic donor as defined per protocol versus recommended allogeneic HCT, initiation of donor search but without identification of a suitable donor as defined per protocol in patients who were otherwise eligible for an allogeneic transplant). The probabilities of relapse and non-relapse-related mortality were calculated using cumulative incidence estimates to accommodate competing risks (mortality in this analysis refers to number of deaths per number of patients involved in the evaluation). For analysis of non-relapse mortality, failure was defined as death during a continuous complete remission. For analysis of relapse, failure was defined as clinical or hematologic recurrence of AML at any site.

Results

Patients’ characteristics

The start of induction therapy of all prospectively enrolled 243 AML patients less than 60 years old with an unfavorable karyotype was between August 1999 and May 2007. After induction therapy with either TAD-HAM (123 patients) or HAM-HAM (120 patients), 111 patients achieved a CR1 (CR/CRi) within a median of 58 days (range, 12–113 days). As previously described, both induction therapies resulted in comparable outcomes.16,17 For landmark analyses, a minimum of 90 days of CR was required after achieving CR1, leaving a study cohort of 90 patients (21 patients relapsed within 90 days after achieving CR1 and were excluded from further analyses). A suitable allogeneic donor could be identified for 59 patients (a sibling donor for 30 and an unrelated donor for 29). Fifty-five patients underwent allogeneic HCT after a median of 88 days in documented CR1 (Table 1). The control cohort for landmark analyses of survival data consisted of 35 patients treated with conventional post-remission therapy (including autologous HCT in six cases). The median follow-up of all surviving patients was 60.4 months (range, 11–105 months), that for patients undergoing allogeneic HCT was 60.4 months, while that for patients receiving conventional consolidation therapy was 58.2 months.

Post-remission treatment

TAD consolidation therapy was given to 61/90 patients (30/55 patients undergoing allogeneic HCT and 31/35 patients receiving conventional treatment) and at least one cycle of maintenance therapy was administered to 19/90 patients (5/55 patients and 14/35 patients, respectively). Six out of 31 patients without a donor underwent autologous HCT according to the protocol.

Due to comorbidities/infections, four patients with an allogeneic donor did not proceed to transplantation in CR1 and were, therefore, included in the as treated control cohort. No suitable donor could by identified for 24 patients while in continued CR1 (control cohort for intent to treat analyses). For seven additional patients, only considered in the as treated analyses, no donor search was initiated because of comorbidities and/or severe infection associated with induction therapy.

Table 1.Characteristics of the patients, diseases and donors.

Among the subjects undergoing allogeneic HCT, 38 patients (69%) received standard intensity conditioning and 17 (31%) received reduced intensity conditioning, as previously defined.22 Total body irradiation (8–12Gy)-based standard intensity conditioning was used in 29 patients. Conditioning regimens and prophylaxis against graft-versus-host disease were chosen according to the policies of the individual centers. Six patients received an allogeneic bone marrow graft (11%) and 49 an allogeneic peripheral blood stem cell graft (89%), all without in vitro T-cell depletion. No graft failure was reported within the study population. The cumulative incidence of acute graft-versus-host disease grade II – IV was 33% and that of chronic graft-versus-host disease was 44% (limited disease in 11 patients and extensive disease in 13 patients).

Survival according to treatment received

At the time of last follow-up, 33 of 90 patients were alive (27/55 patients after allogeneic HCT in CR1 and 6/35 control patients). Patients who underwent allogeneic HCT in CR1 had substantially better 5-year overall survival and relapse-free survival rates than comparable patients given conventional post-remission therapy (48% versus 16%, P=0.001 and 40% versus 13%, P<0.001, respectively) (Figure 1). Both for the entire group and those patients transplanted, only age (≤40 years versus >40 years) had a significant impact on overall survival (54% versus 28%, P=0.001 and 69% versus 37%, P=0.011, respectively) (Table 2). Comparing allogeneic HCT in CR1 with conventional treatment in the subgroups of patients 40 years and less or over 40 years of age showed that the survival benefit of allogeneic HCT was independent of age (overall survival 69% versus 25%, P=0.026 and 37% versus 15%, P=0.032, respectively). Patients transplanted from a matched related or a matched unrelated donor had similar survival rates (hazard ratio for overall survival 0.77, 95% confidence interval 0.35–1.68, P=0.51) (Table 3).

Survival according to intended treatment

To eliminate factors such as pre-existing comorbidities, which might worsen the outcome of the control group, we performed an intent to treat analysis on the basis of donor availability. In consequence, four of the 35 patients receiving conventional therapy who had a matched related (n=3) or unrelated donor (n=1) were added to the transplant group (Online Supplementary Figure S1). In the remaining 31 patients, allogeneic HCT in CR1 was not intended because of a protocol recommendation/no suitable donor (24 patients) or pre-existing comorbidities (7 patients, all excluded from the intent to treat analysis). Comparing all 59 patients with identified allogeneic donors with the remaining 24 landmark control patients confirmed the benefit of allogeneic HCT with regard to overall survival (48% versus 18% P=0.004) and relapse-free survival (39% versus 10% P=0.001) also on the basis of intended treatment (Figure 2).

Table 2.Comparison of outcome in different subgroups of patients.

Relapse and non-relapse mortality

Of the 90 patients analyzed in this study, 47 had a relapse; 18/55 patients relapsed after allogeneic HCT and 29/35 patients after conventional post-remission therapy (Figure 3). Relapse was fatal in 41 patients (13/55 and 28/35 patients, respectively). The median time from CR1 to relapse after allogeneic HCT was 545 days (range, 177–1645 days) while that for patients receiving conventional therapy was 229 days (range, 105–1081 days). Interestingly, only 4/21 patients (19%) transplanted from matched unrelated donors relapsed in contrast to 13/27 patients (48%) transplanted from matched related donors (P=0.022). In contrast, transplantation from matched related donors resulted in three non-relapse-related deaths (11%) as compared to ten deaths (36%) in patients transplanted from matched unrelated donors (P=0.123). Of the 35 control patients, two died in CR (1 after an allogeneic HCT in second CR and 1 due to an infection).

Subset analyses

A monosomal karyotype was found in 27 of the 90 patients with high-risk cytogenetics at diagnosis of AML (33% of patients undergoing allogeneic HCT in CR1 and 26% of patients receiving conventional post-remission therapy). As shown by univariate and multivariate analyses, monosomal karyotype was an adverse prognostic factor in patients not transplanted from an allogeneic donor in CR1, but not in patients undergoing allogeneic HCT (Tables 2 and 3).

A small subset of ten patients with documented CR/CRi (median age 45 years, 5 transplanted from matched related donors and 5 from matched unrelated donors) received sequential reduced intensity conditioning therapy.23 These patients showed a tendency to a higher overall survival rate after 5 years (75%) compared to patients receiving other conditioning regimens (43%; P=0.068).

Table 3.Multivariate analyses of outcome.

Figure 1.Survival plots for comparison groups. HCT, hematopoietic stem cell transplantation; CR1, first complete remission. Patients transplanted in CR1 (—), landmark control group (—). MUD, matched unrelated donor (—); MRD, matched related donor (—).

Discussion

AML with poor-risk cytogenetics in patients less than 60 years of age is a rare disease and has a dismal prognosis. Accordingly, most evidence for treatment recommendations stems from meta-analyses or alternative cumulative evaluations of different, at best sequential, clinical trials.2,8–14 These limitations hold true specifically for the role of allogeneic HCT in front-line treatment. Only recently, data from two prospective analyses were reported which strongly favor allogeneic HCT from matched related or unrelated donors in the first-line treatment of high-risk AML patients defined by either cytogenetics and/or failure of induction treatment.11,15

Figure 2.Survival plots for corresponding intent to treat analyses (recommended allogeneic HCT, initiation of donor search and identification of a suitable allogeneic donor as defined per protocol (—); versus patients with recommended allogeneic HCT, initiation of donor search but without identification of a suitable donor as defined per protocol for patients who otherwise were eligible for an allogeneic transplant(—).

Figure 3.Cumulative incidences of relapse and non-relapse mortality. Transplanted patients are represented by blue and landmark controls by red.

The large data pool of the prospective AMLCG 99 trial, comprising almost 2400 patients enrolled at the time of this examination, enabled us to perform an analysis focusing on patients with high-risk cytogenetics aged less than 60 years, who achieved a CR1 after balanced induction treatment.17

The control cohort in our landmark comparisons certainly represents a positive selection as a minimal duration of CR1 was required in order to compensate for time from CR1 to transplantation in the intervention group. Landmark analysis was also used in a recent report by Schlenk et al. on the role of HCT in first-line treatment of high-risk AML. More than 80% of AML patients in this prospective cohort comparison were defined as being at high risk due to failure of induction treatment.15 This study also described a significant benefit of transplantation on the overall survival of patients with poor-risk cytogenetics (also including abn(12p), abn(17p) as high-risk cytogenetic abnormalities, in contrast to our study) in their multivariable regression model. The authors also ascertained an overall survival benefit for transplant recipients irrespective of whether the graft was from a matched related or unrelated donor. Their conclusions are, therefore, largely in line with the findings of the present study, although they refer to a mixed population of high-risk patients with either induction failure or poor cytogenetics.

In contrast, our prospective cohort comparisons are restricted to genetically high-risk patients in CR1 after balanced induction treatment, allowing us to compare outcomes of this specific high-risk population with an accordingly defined control group. The data clearly underscore that survival outcome after transplantation of a graft from a matched unrelated donor is certainly not worse than that after a related donor transplant. Since higher ranking evidence from randomized controlled trials will not be easily obtained in this rare disease entity, and with existing evidence some such trials may not even be ethically justifiable, there is little chance that data will appear that would question rather than corroborate our main finding. Consequently, together with existing data, we present here firm evidence that despite substantial non-relapse fatality rates of up to 20% with the transplants regimens used in this study, the benefit on survival end-points demonstrates that allogeneic HCT from matched related or unrelated donors in CR1 should be considered the standard of care for AML patients with high-risk cytogenetics.24 This conclusion also applies generally to patients aged less than 60 years, irrespectively of whether they belong to the age category less than or 40 years old, which benefits most, or to the category over 40 years old, which also shows a significant survival benefit. However, eligibility and donor matching criteria, transplant regimens and the extension of transplants to elderly patients require refinement in future studies and bear the potential for improving transplant outcome in this AML population.20,21,23,25–27

It might be asked whether allogeneic HCT in this setting can be postponed to first relapse. Although this question appears to be very reasonable for AML patients with intermediate cytogenetic risk, current consensus and the few data available suggest that an allogeneic HCT in CR1 should not be delayed in this population. Reports indicating a successful outcome after relapse with a curative potential of approximately 30% are highly selective and relate only to patients who have survived their relapse and remain fit enough to receive a transplant.28,29 The predicted overall survival of relapsed AML patients in this setting is exceedingly poor, being no more than about 10%.18,30

Finally, the comparatively large numbers of patients cytogenetically defined as high-risk accrued in this study also enabled us to address the role of rare subgroups such as patients with monosomal karyotype within the whole cohort of cytogenetically high-risk patients. In both univariable and multivariable analyses, monosomal karyotype was confirmed as a factor conferring the worst outcome in the control group.18,19 Surprisingly, monosomal karyotype was overcome as a prognostic factor in patients undergoing allogeneic HCT in CR1.

The main weaknesses of our conclusions depend on the fact that they were not derived from a randomized controlled trial specifically designed to address our key question. In consequence, imbalances of, for example, age groups or AML phenotypes and a remaining uncertainty of potential donor availability in the control group (the main limitation of all donor versus no donor comparisons) are also potential limitations of our study. On the other hand, randomized trials are unlikely to be feasible because of the rareness of the subentity under investigation and may even be of ethical concern. With these considerations in mind, it appears appropriate to conclude that allogeneic HCT in cytogenetically defined high-risk patients, with or without a monosomal karyotype, in CR1 who are eligible for a transplant and have a suitable donor is the current standard of care in first-line post-remission treatment. Our analyses further highlight the need to agree on standards on how to develop high-level evidence treatment guidelines for rare subentities.

Footnotes

• Funding: supported by Grants N. M17/92/Bü1 and 70-2839-Bü4 from Deutsche Krebshilfe, Grant N. 01 GI 9976 from Bundesministerium Bildung und Forschung Competence Network Acute and Chronic Leukemias, Grant N. LSH-2002-2.2.0-3 from the European Commission European LeukemiaNet, and an unrestricted grant from AMGEN.
• The online version of this article has a Supplementary Appendix.
• Authorship and Disclosures The information provided by the authors about contributions from persons listed as authors and in acknowledgments is available with the full text of this paper at www.haematologica.org.
• Financial and other disclosures provided by the authors using the ICMJE (www.icmje.org) Uniform Format for Disclosure of Competing Interests are also available at www.haematologica.org.

• Received January 19, 2011.
• Revision received March 15, 2011.
• Accepted March 21, 2011.