|Home | About | Journals | Submit | Contact Us | Français|
We studied the role of HLA-matched sibling hematopoietic cell transplantation (HCT) in treating t(8;21) AML in first remission. Outcomes of 118 patients receiving HCT and reported to the Center for International Blood and Marrow Transplant Research were compared with 132 similar patients receiving chemotherapy selected from eight German AML Intergroup multicenter trials. Characteristics of the cohorts were similar except that chemotherapy recipients were significantly older. To adjust for time to treatment bias, outcomes were compared using left-truncated Cox regression models. Transplants were associated with higher treatment-related mortality [relative risk (RR) 6.76, 95% confidence interval (CI) 2.95–15.45, p<0.001], lower relapse (RR 0.47, 95% CI 0.25–0.85, p=0.01) and similar relapse-free survival (p=0.2). Loss of sex chromosomes (LOS) in addition to t(8;21) negatively impacted overall survival in patients receiving chemotherapy. Patients without LOS experienced shorter survival after HCT comparing to chemotherapy (RR 3.05, p=0.02), whereas patients with LOS had similar survival regardless of postremission therapy. In both cohorts, white blood cell count (WBC) at diagnosis > 25 x109/L was associated with a higher relapse risk (RR=2.09, p=0.03), lower relapse-free (RR=1.9, p=0.008) and overall survival (RR=1.91, p=0.01). In this cohort of patients with t(8;21) AML , HCT did not improve overall survival, since reduction of relapse was offset by high transplant-related mortality. In the group without LOS, survival after chemotherapy was far superior to HCT. These results suggest that patients with t(8;21) AML without poor prognostic factors have higher rates of survival after chemotherapy as a post remission therapy compared to HCT.
Translocation t(8;21)(q22;q22) [t(8;21)] is found in one third of karyotypically abnormal and approximately 8% of all patients with acute myeloid leukemia AML (1). This translocation is associated with secondary aberrations in 60 to 80 % of cases with loss of a sex chromosome (LOS) being the most frequent (2–4). At the molecular level this translocation involves RUNX1(AML1) on chromosome 21 and CBFA2T1 (ETO) on chromosome 8 (5). RUNX1 is a member of the core binding factor alpha family (CBFα) of transcriptional regulators and, through interaction with CBFβ, allows transcription of genes required for myeloid differentiation. The RUNX1/CBFA2T1 fusion protein acts as a dominant repressor of RUNX1-dependent transcription.
The role of post remission therapy in AML is to decrease relapse and prolong survival. Outcomes after postremission therapy vary considerably by specific AML subtypes and presence of poor prognostic factors. Slovak et al from the American Intergroup study (SWOG/ECOG) observed superior overall survival after transplantation compared to chemotherapy among patients with favorable risk AML [t(8;21), t(15;17) and inv(16)] (6). Conversely, both French and German AML-Intergroups showed no difference between hematopoietic stem cell transplantation (HCT) strategies and intensive chemotherapy for this group of AML patients (7, 8).
Relapse varies substantially according to additional risk factors within the same cytogentically risk categories. Leukocytosis at diagnosis is associated with worse outcomes (9)'(10), with the French and the German AML Intergroup analyses revealing nearly identical cut points for higher risk category (30x109/l and 25.4x109/l, respectively) (11) (7) (4). Additional cytogenetic abnormalities also predict treatment outcomes but less certainly than leukocytosis. LOS in men and del(9q) are associated with shorter survival outcomes in patients with t(8;21) (3,4,12,13).
In this study we examined two different postremission strategies for t(8;21) AML in first complete remission, namely, intensive chemotherapy with cytarabine-based regimes, using data from the German AML Intergroup, and HLA-matched sibling HCT using data from the Center for International Blood and Marrow Transplant Research (CIBMTR). Current practice does not suggest that t(8;21) AML in first complete remission is an indication for sibling donor HCT. We undertook this analysis to determine whether this currently held practice was evidence-based especially in the light of known prognostic factors in this specific type of AML.
The CIBMTR is a voluntary working group of more than 400 centers worldwide. Participating centers register basic information on all consecutive transplant recipients. Comprehensive demographic and clinical data are collected on a representative sample of registered patients selected using a weighted randomization scheme. Patients are followed longitudinally, with yearly follow up.
Computerized checks for errors, physician reviews of submitted data and on-site audits of participating centers ensure the quality of data. Studies utilizing CIBMTR data are approved by the Medical College of Wisconsin Internal Review Board.
The data from patients who received chemotherapy was obtained from the German AML Intergroup database using data from 8 multicenter prospective clinical trials (4): SHG-Hannover-AML-2/95 (14), SHG-Hannover-AML-1/99, SHG-Dresden-AML-96 (15), AMLSG ULM AMLHD93 (8), AMLSG ULM AML-HD 98A, AMLCG92 (16), AMLCG99 (17) and OSHO AML-96. All trials enrolled AML patients and included double induction chemotherapy followed by different postremission strategies including different dose levels of cytarabine-based chemotherapy, autologous and allogeneic HCT.
The study included patients with t(8;21) AML in morphologic first complete remission, ages 16 to 60 years, who underwent an HLA-identical sibling HCT with myeloablative conditioning regimen reported to the CIBMTR from 1990 to 2002 or who received chemotherapy in one of the German trials noted above from 1993 to 2002. One hundred and ninety seven patients with t(8;21) AML enrolled in the German trials. Thirty-one were excluded for not achieving remission, 7 for not receiving postremission therapy, and 27 for receiving an allogeneic (n=9) or an autologous (n=18) HCT. Thus 132 patients were eligible for this study. Among the 124 patients receiving HLA-identical sibling transplants for t(8;21) AML and reported to the CIBMTR, 2 were excluded for receiving reduced intensity conditioning regimens and 4 from inactive transplant centers for insufficient post transplant follow up data, leaving a total of 118 eligible patients, contributed by 66 centers in 26 countries.
All cytogenetic and molecular analyses in the chemotherapy cohort except for the OSHO trial were performed in central reference laboratories per protocol requirements. In the OSHO trial, cytogenetic analyses were required prior to study entry but there was no central review (n=9). The CIBMTR requests information on cytogenetic testing performed at diagnosis, prior to transplantation and at relapse, but not all patients have these analyses performed. Only cases reporting the presence of t(8;21) were included. Cytogenetic reports were available for review and confirmation for 69 of 118 patients in the transplant cohort. Descriptions of karyotypic abnormalities adhered to the International System for Human Cytogenetic Nomenclature (18). Five patients in the chemotherapy and one patient in the HCT group without evaluable metaphases had confirmed t(8;21) through AML1/ETO molecular analysis.
The most frequent additional cytogenetic change observed in both groups was loss of sex chromosomes (LOS) (-X or -Y). Three groups were analyzed: t(8;21) alone, t(8;21) and LOS with or without other cytogenetic changes and t(8;21) with other cytogenetic changes excluding LOS. Deletion 9q was present in only a small proportion of patients (chemotherapy, n= 21 and HCT, n= 9), precluding separate analysis.
Treatment-related deaths were defined as deaths occurring during continuous first remission as calculated by the cumulative incidence estimate with relapse as the competing risk. Patients were censored at time of last follow-up. Relapse was defined as morphologic leukemia recurrence at any site by the cumulative incidence estimate with death in remission as the competing risk. Patients were censored at death in continuous first remission or surviving in continuous complete remission at last contact. Relapse-free survival (RFS) and overall survival (OS) were defined, respectively, as survival in continuous complete remission and survival, with censoring at last follow up (19). For OS, death from any cause was considered an event. For RFS (i.e. treatment failure), relapse or death was considered an event.
Variables related to patient and disease characteristics were compared between groups using the chi-square statistic or Fishers exact test, if appropriate, for categorical variables and the Kruskal-Wallis test for continuous variables. The median follow up time was calculated utilizing the Kaplan Meier estimate for the censoring distribution.
When comparing outcomes of transplant versus non-transplant groups, differences in time to treatment and differences in patient baseline characteristics are potential sources of bias and require appropriate adjustments. As transplant recipients must survive long enough to undergo transplantation, they may represent population with inherently better outcome. In order to address this potential bias, left-truncated Cox regression models and left truncated cumulative incidence estimates were used. At each time point in this model, the risk set in the chemotherapy cohort consisted of all patients, while the risk set in the transplant cohort included only those whose waiting time to transplant was less than this time point (20).
To adjust for differences in baseline characteristics, multivariate Cox proportional hazards regression models were used (21). First, associations between each outcome and potential prognostic variables (listed in table 2) were evaluated using a stepwise approach. Variables significantly associated with each outcome event (p<0.05) were included as covariate factors in the subsequent comparisons. The proportionality assumption of the Cox model was tested by adding separately for each outcome event a time-dependent covariate for each of the covariates tested. Presence of LOS in the overall survival model was the only variable demonstrating significant interaction with type of postremission therapy. Adjusted probabilities of RFS and OS were then generated from the final Cox models stratified on postremission therapy. Results were expressed as relative risks (RR) of each outcome after transplantation versus after chemotherapy. P values are two-sided. Analyses were performed using SAS software, version 9.2 (SAS Institute).
Table 1 shows patient, disease and treatment-related characteristics by postremission therapy. Transplant recipients were younger than chemotherapy recipients (median age 32 vs. 42 years, p<0.001). We considered patients as having high risk leukocytosis, if their WBC count at diagnosis was >25×109/L (4). Cytogenetic abnormalities detected in addition to the t(8;21) were frequently observed in both groups. Fifty-two percent of HCT patients received TBI-based conditioning regimens and 88% received bone marrow grafts. None of the HCT recipients received donor lymphocyte infusions either as a planned therapy or for recurrent leukemia. Most transplants were performed prior to 1994; consequently, median follow up times of survivors were 48 and 94 months after chemotherapy and HCT, respectively. Nevertheless, 66% of chemotherapy and 87% of transplant recipients were followed for ≥ 5 years and the completeness of follow up (the ratio of the sum of the observed follow up time to the sum of the potential follow up time for all patients in the study (22) was 86% for both treatment groups. The rate of grade II-IV acute graft-versus host disease (GVHD) at day 100 after HCT was 9% [95% confidence intervals (CI) 5-15%] and chronic GVHD at one and three years were 27% (95% CI 19-35%) and 37% (95% CI 29-46%).
Risks of treatment related mortality (TRM) were significantly higher after HCT than after chemotherapy (Table 3). The 5-year cumulative incidences of TRM were 6% (95% CI 2–11%) and 32% (95% CI 22–44%) after chemotherapy and HCT, respectively (Figure 1).
Risks of leukemia relapse were significantly lower after HCT than after chemotherapy (Table 3). The 5-year probabilities of relapse was 29% (95% CI 21–37%) and 14% (95% CI 8-21%) after chemotherapy and HCT, respectively (p=0.005) (Figure 2). Risks of relapse were significantly higher in patients with WBC >25.4×109/L at diagnosis in both cohorts (Table 3).
Risks of treatment failure (relapse or death) were similar after chemotherapy and HCT (Table 3). The 5-year probabilities of relapse-free survival (RFS) were 64% (95% CI 53-73%) and 55% (95%CI 45-65%) after chemotherapy and HCT, respectively (p=0.2) (Figure 3). Risks of treatment failure were higher in patients with a diagnostic WBC >25×109/L in both cohorts (Table 3).
Overall survival after HCT versus chemotherapy differed in the presence or absence of LOS in addition to t(8;21). Risks of overall mortality were significantly higher after HCT than after chemotherapy in patients with t(8;21) and no LOS; risks of overall mortality were similar after HCT and chemotherapy in patients with t(8;21) and LOS (Figure 4). WBC >25.4×109/L was associated with higher risk of overall mortality after both chemotherapy and HCT (Table 3).
The most common cause of death in both treatment groups was recurrent leukemia, accounting for 70% of deaths after chemotherapy and 24% of deaths after HCT. Non-relapse causes of death after chemotherapy included infection (21%), organ failure (3%), secondary malignancy (3%) and other causes (3%). Non-relapse causes of death after HCT were infection (16%), interstitial pneumonitis (6%), acute respiratory distress syndrome (8%), acute graft-versus host disease (GVHD) (2%), chronic GVHD (10%), interstitial pneumonitis and GVHD (6%), organ failure (10%), secondary malignancy (6%), hemorrhage (4%) and other causes (2%). The cause of death was unknown for 3 patients (6%) in the HCT cohort.
This study shows that HLA-matched sibling HCT for t(8;21) AML in first complete remission for patients between 16 and 60 years is associated with lower relapse rates, higher TRM and similar RFS compared to cytarabine-based postremission chemotherapy. This is consistent with reports of unselected groups of AML patients treated in first complete remission (CR1). Reported relapse rates after allogeneic HCT for AML in CR1 range from 24% to 37% (23–26) compared to much higher rates after chemotherapy (36-60%) (23–27). This advantage (i.e. fewer relapses after HCT) does not always translate into longer relapse-free and overall survival due to higher rates of TRM after HCT. We observed 1-year TRM rates of 23% after HCT, compared to 4% after chemotherapy. Most transplants were performed in the early 1990's, however the year of transplantation did not significantly impact the rate of TRM. Despite changes in transplant practice in the last decade, TRM remains a major challenge in transplantation. The majority of patients in the HCT cohort received bone marrow grafts, which differs from the current practice of utilizing mobilized peripheral blood stem cells (28). The selection of graft source is likely to influence hematopoietic recovery and incidence of chronic GVHD. However, in patients with AML in CR1, rates of TRM are similar after bone marrow and peripheral blood stem cells transplants (29,30–32).
The MRC 10 prospectively compared allogeneic HCT with chemotherapy allocating patients with a matched sibling donor to HCT (biologic assignment) (33). Patients within the favorable risk category had similar rates of leukemia-free and overall survival, irrespective of donor availability. This observation and the overall favorable responses with standard dose chemotherapy led to recommendations against HCT for patients with favorable risk AML in first complete remission. In contrast, the American Intergroup (SWOG/ECOG) performed a similar subset analysis by postremission therapy and showed better survival in favorable risk patients undergoing either autologous or allogeneic HCT versus those receiving chemotherapy (6). Importantly, despite large numbers of patients in these landmark studies, the groups with specific cytogenetic changes were small and imbalances between comparison groups become significant. The French AML Intergroup analysis of t(8;21) AML demonstrated similar rates of leukemia-free survival with allogeneic HCT and chemotherapy. Thirty-seven of 154 patients in the French study underwent HLA-matched sibling HCT and 5-year probabilities of LFS were 56% and 52% for the transplant and chemotherapy groups, respectively (11).
Treatment outcomes after chemotherapy or HCT in patients with t(8;21) and additional cytogenetic abnormalities are mixed. The MRC and SWOG/ECOG cytogenetic classifications addressed this issue differently. The MRC included t(8;21) plus any additional changes in the favorable risk group since a separate analysis showed no significant impact of additional cytogenetic abnormalities on survival. In contrast, SWOG/ECOG excluded cases with cytogenetic abnormalities in addition to t(8;21) from the favorable risk group. The current study showed LOS to be an important predictor of overall mortality when present in addition to t(8;21) in patients receiving chemotherapy. Deletion of chromosome Y was the most frequent LOS and the main factor associated with poorer survival. Subset analysis on a group of patients with -X (Chemotherapy, n= 18 and HCT, n=16) did not show the same association. As the number of patients with -X was small, this interpretation deserves caution. Furthermore, the presence of additional cytogenetic abnormalities to t(8;21) other than LOS did not affect any outcomes analyzed and thus were combined to the group of patients with t(8;21) as a sole cytogenetic abnormality. Patients with t(8;21) alone or t(8;21) with additional cytogenetic abnormalities other than LOS had longer survival after chemotherapy than HCT, while among those with LOS, there was no significant difference in survival after chemotherapy or HCT. These findings contrast other reports (3,11). The French Intergroup trial, in which 43% of patients had LOS and 8% had del(9q), showed no significant impact of these abnormalities on survival. A CALGB analysis of CBF-AML identified 69% with cytogenetic abnormalities in addition to t(8;21), of which 90% were LOS and 17.4% del(9q). LOS did not impact survival, though in a subset of non-white patients, the presence of del(9q) was associated with longer overall survival. These discrepancies may represent inherent differences in study populations and methods of analyses. In the French study, the subset analysis did not specify additional cytogenetic changes among treatment groups. The majority of patients (61%) in the CALGB study presented with additional LOS and the 5-year probability of OS for t(8;21) AML patients was 46% (95% CI 37-55%). This survival probability is similar to the chemotherapy cohort with additional LOS (55%, 95% CI 41-69%) from our study. Small numbers of patients with normal sex chromosomes in the CALGB study may explain why LOS had no detectable impact on survival. Appelbaum et al also analyzed the effect of additional cytogenetic changes in the outcome of patients with t(8;21) AML(2). The most common additional abnormality was LOS and most patients received chemotherapy as postremission therapy. Neither LOS nor del(9q) were significantly associated with overall survival or any other outcomes. Patients with trisomy 8 or with three or more cytogenetic abnormalities had worse survival (2). The results on the impact of additional cytogenetic abnormalities to t(8;21) are conflicting and represent the heterogeneity of this patient population. More importantly this illustrates that background molecular factors such as c-KIT expression and others may play a greater prognostic role and presence or absence of certain cytogenetic markers may reveal to be more as confounders than biologically relevant.
In our study despite the association of LOS with lower survival in the chemotherapy cohort, it was not significantly associated with TRM, relapse or RFS. This might be explained by poor survival post relapse of patients with LOS. The CALGB study and others demonstrated shorter post relapse survival in patients with t(8;21) compared to those with inv(16) (2,3).A German meta-analysis also demonstrated shorter post relapse survival for patients with t(8;21) and LOS (4). Whether this represents an effect of LOS or an inherent characteristic of t(8;21) AML remains undetermined.
Extreme leukocytosis at diagnosis, although uncommon in t(8;21) AML is considered a poor prognostic factor. Several studies, including this analysis, confirmed the adverse impact of leukocytosis and higher marrow blast percentage on treatment outcome (2,9,10). Our study demonstrated that leukocytosis was associated with higher relapse rates, shorter relapse-free survival and shorter overall survival irrespective of postremission therapy. High percent of blasts in the bone marrow at diagnosis was associated with worse treatment related mortality. Both leukocytosis and percent marrow blasts reflect disease burden. The association between high disease burden and treatment related mortality is unclear. Other reports showed an association between percent of bone marrow blast at diagnosis and lower complete remissions rates after induction therapy in t(8;21) AML (3,11). The association of leukocytosis and survival in CBF AML is likely to be explained by overexpression of tyrosine kinase genes related to AML1/ETO. Gain of function mutations of cKIT are associated with high disease burden at diagnosis and significantly impact on survival due to high relapse rates (34–36).
In summary, we report a large comparison of postremission therapy in a homogenous population of AML with t(8;21). We acknowledge potential biases, which may affect the outcomes described. The chemotherapy group includes participants in several clinical trials and patient accrual was determined by trial specific eligibility criteria. Though all received cytarabine-based chemotherapy, it was administered in a spectrum of doses along with other anti-leukemic agents, the chemotherapy regimens considered in regression models did not influence any of the endpoints studied. Furthemore, the chemotherapy regimens used as post remission therapy differs from the four cycles of high dose cytarabine (HiDAC) consolidation commonly utilized in current clinical practice in North America. Despite the results from the CALGB study on HiDAC on this patient population being similar to the chemotherapy group (37), such differences in practice should be considered. For the HCT group, the decision to transplant and all aspects of transplant procedure (i.e. conditioning regimen, graft versus host disease prophylaxis, donor selection) were at the discretion of the transplant center and reflect routine clinical transplant practices. HCT-related variables were tested in the regression model and did not affect the comparisons with chemotherapy. Lastly, we combined leukocytosis and LOS to assess the impact of both significant covariates on survival, however the groups with both prognostic factor were too small for a meaningful comparison and thus it was omitted from the analysis.
AML with t(8;21) is regarded as favorable risk based on response to initial therapy and longer overall survival. This study confirms that cytarabine-based chemotherapy offers results similar or better than HLA-matched sibling HCT in first remission. LOS negatively affected survival in patients who received chemotherapy as postremission therapy but survival was still similar to that achieved with HCT. Patients with t(8;21) but without LOS, overall survival was longer after chemotherapy compared to HCT. Higher WBC counts at diagnosis were associated with worse outcomes after both chemotherapy and HCT. Selection of the best post-remission therapy for patients with t(8;21) and poor prognostic features would be best addressed on a risk adapted clinical trial.
The CIBMTR is supported by Public Health Service Grant U24-CA76518 from the National Cancer Institute, the National Institute of Allergy and Infectious Diseases, and the National Heart, Lung and Blood Institute; Office of Naval Research; Health Services Research Administration (DHHS); and grants from Abbott Laboratories; Aetna; American International Group, Inc.; American Red Cross; Amgen, Inc.; Anonymous donation to the Medical College of Wisconsin (MCW); AnorMED, Inc.; Astellas Pharma US, Inc.; Baxter International, Inc.; Berlex Laboratories, Inc.; Biogen IDEC, Inc.; BloodCenter of Wisconsin; Blue Cross and Blue Shield Association; Bristol-Myers Squibb Company; BRT Laboratories, Inc.; Cangene Corporation; Celgene Corporation; CellGenix, Inc.; Cell Therapeutics, Inc.; CelMed Biosciences; Cylex Inc.; Cytonome, Inc.; CytoTherm; DOR BioPharma, Inc.; Dynal Biotech, an Invitrogen Company; Enzon Pharmaceuticals, Inc.; Gambro BCT, Inc.; Gamida Cell, Ltd.; Genzyme Corporation; Gift of Life Bone Marrow Foundation; GlaxoSmithKline, Inc.; Histogenetics, Inc.; HKS Medical Information Systems; Kirin Brewery Co., Ltd.; Merck & Company; The Medical College of Wisconsin; Millennium Pharmaceuticals, Inc.; Miller Pharmacal Group; Milliman USA, Inc.; Miltenyi Biotec, Inc.; MultiPlan, Inc.; National Marrow Donor Program; Nature Publishing Group; Novartis Pharmaceuticals, Inc.; Osiris Therapeutics, Inc.; Pall Medical; Pfizer, Inc.; Pharmion Corporation; PDL BioPharma, Inc; Roche Laboratories; Sanofi-aventis; Schering Plough Corporation; StemCyte, Inc.; StemSoft Software, Inc.; SuperGen, Inc.; Sysmex; The Marrow Foundation; THERAKOS, Inc.; University of Colorado Cord Blood Bank; ViaCell, Inc.; ViraCor Laboratories; Wellpoint, Inc.; and Zelos Therapeutics, Inc. The German AML Intergroup is supported by grant 01GI9981 from the Bundesministerium für Bildung und Forschung (Kompetenznetz "Akute und chronische Leukämien"), Germany. The views expressed in this article do not reflect the official policy or position of the National Institute of Health, the Department of the Navy, the Department of Defense, or any other agency of the U.S. Government.
This study was presented as an abstract at the 47th annual American Society of Hematology in Atlanta, Georgia , December 2005.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.