Clinical Advances in Hematology & Oncology

January 2017 - Volume 15, Issue 1

The Evolution of Allogeneic Stem Cell Transplant for Children and Adolescents With Acute Myeloid Leukemia

Allyson Flower, MD, and Mitchell S. Cairo, MD

The authors are affiliated with New York Medical College in Valhalla, New York. Dr Flower is an assistant professor in the Department of Pediatrics and the Department of Microbiology and Immunology. Dr Cairo is the chief of Pediatric Hematology, Oncology, and Stem Cell Transplantation; director of the Children and Adolescent Cancer and Blood Diseases Center; medical and scientific director of the WMC Cellular and Tissue Engineering Laboratory; medical director of the WMC Hematotherapy Program, associate chairman of the Department of Pediatrics; and professor of Pediatrics, Medicine, Pathology, Microbiology & Immunology, and Cell Biology & Anatomy.

Corresponding author:
Allyson Flower, MD
Assistant Professor of Pediatrics and Microbiology and Immunology
40 Sunshine Cottage Road
Valhalla, NY 10595
Tel: (914) 594-2131


Abstract: Survival rates in subsets of pediatric patients who have acute myeloid leukemia (AML) with favorable risk features are now greater than 90%. However, outcomes for patients with high-risk (HR) features remain unacceptably poor. As novel technologies for the identification of HR biomarkers and the detection of residual disease are developed, risk stratification and the application of allogeneic hematopoietic stem cell transplant (HSCT) are evolving. HSCT has been shown to benefit subpopulations of pediatric patients with AML, including those with HR cytogenetic translocations, genetic mutations, and/or residual disease after induction. Targeted therapies have shown promise for improving outcomes, and their integration into standard therapy and HSCT regimens is a critical area of interest. Also, expansion of the donor pool has led to the successful use of alternative donor sources for those patients without a matched sibling. However, transplant-related morbidity and mortality and late effects are major limiting factors. Reduced-intensity conditioning regimens have resulted in outcomes equivalent to those achieved with myeloablative regimens among patients in complete remission. The limitation of transplant-related morbidity and mortality through reduced-intensity conditioning and supportive care, and improved survival through optimal alloreactivity in combination with targeted therapy, are steps toward advancing outcomes for pediatric patients who have AML with HR features.


Pediatric acute myeloid leukemia (AML) constitutes a minor portion of all pediatric cancers but remains a therapeutic challenge, with unacceptably low rates of event-free survival (EFS) and a high incidence of relapse when treated with traditional cytotoxic chemotherapy. The incidence of AML in children younger than 15 years is 7 per 1 million, and the disease accounts for one-fourth of cases of pediatric acute leukemia.1,2 Recent clinical trials addressing the treatment of pediatric AML have achieved 5-year EFS rates ranging from 49% to 63%, with a relapse rate of 30%.3,4 Importantly, outcomes in children with high-risk (HR) features remain significantly inferior to outcomes in those with favorable-risk disease despite intensification of therapy, implementation of hematopoietic stem cell transplant (HSCT), and advances in supportive care.

As cytogenetic markers of HR disease are identified, the stratification of pediatric patients with AML has been updated, and changes have been incorporated into the World Health Organization (WHO) classification of AML. The prevalence of certain cytogenetic mutations seen predominantly in pediatric AML highlights the need for pediatric-specific risk stratification based on known prognostic markers.2 With the advancement of technology for the detection of minimal residual disease (MRD), including multidimensional flow cytometry, cytogenetics, and next-generation sequencing, MRD positivity at critical points has been a major area of focus for the determination of prognostic significance. The discovery and application of clinically relevant HR features, the use of highly sensitive methods of MRD detection, and the increased sophistication of HSCT strategy—including expansion of the donor pool, human leukocyte antigen (HLA) matching, toxicity management, graft-versus-host disease (GVHD) prophylaxis, supportive care, and integration of targeted therapies—have led to an evolution of the role of HSCT for pediatric patients with AML. HSCT remains a critical area of interest in the effort to improve outcomes for pediatric patients with AML that has HR features. The focus of this article is identification of the subset of pediatric patients with AML who are likely to benefit from HSCT and current practices in HSCT for these patients, including timing, donor selection, conditioning regimens, and incorporation of novel agents.

Risk Stratification in Pediatric Acute Myeloid Leukemia

Clinically relevant cytogenetic markers and treatment response are strongly associated with survival in pediatric patients who have AML.2 A major factor limiting the interpretation of outcomes in pediatric AML is the variability in the classification of HR patients, which leads to the inconsistent use of HSCT in clinical trials. The risk stratification that has been used for recent clinical trials in pediatric AML reflects the 2008 WHO classification, which is based on prognostically significant cytogenetic markers and disease response according to morphologic evaluation. The role of MRD positivity as detected by flow cytometry, cytogenetics, and real-time polymerase chain reaction (RT-PCR) is under ongoing investigation.

Core binding factor mutations t(8;21) and inv(16) are well-established markers of favorable outcome in response to conventional chemotherapy for pediatric AML patients. Each is associated with an overall survival (OS) rate of greater than 90%.5,6 Although rare, biallelic mutations of CEBPA and NPM1 with a normal karyotype, and wild-type FLT3, also are associated with a favorable prognosis in pediatric AML.7-9 Other cytogenetic features that have been associated with favorable outcomes in pediatric AML include t(9;11)(p12;q23)/MLL-AF9 and t(1;11)(q21;q23)/MLL-AF1q.10

The FLT3 internal tandem duplication (FLT3/ITD) mutation, and to a greater extent FLT3/ITD with an allelic ratio greater than 0.4, is associated with a high risk for relapse and progression-free survival rates as low as 16%.7,11 The presence of monosomy 7 in pediatric patients with AML has been associated with a complete response (CR) rate of 67% and OS rate of 30%. Although HSCT did not improve outcomes for these patients, 30% of those with persistent disease at the time of HSCT achieved sustained remission.12 Other rare mutations that have been associated with an adverse prognosis in pediatric AML include monosomy 5 or del(5q), t(6;11)(q27;q23)/MLL-AF6, t(10;11)(p12;q23)/MLL-AF10, t(6;9)(p23q34)/DEK-NUP214, t(8;16)(p11;p13)/MYST3-CREBBP, and t(16;21)(q24;q22)/RUNX1-CBFA2T3.10 KIT, WT1, IDH1, TET2, and DNMT3A mutations currently are under investigation in an ongoing Children’s Oncology Group (COG) trial (AAML1031; A Phase III Randomized Trial for Patients With De Novo AML Using Bortezomib and Sorafenib for Patients With High Allelic Ratio FLT3/ITD).13 The implications of overlap of each of these mutations with independently validated HR markers are of special interest.

The identification of clinically relevant genomic and proteomic markers is of particular importance for the application of targeted therapeutics. A recent international expert panel recommends the evaluation of known prognostically significant genetic markers, including FLT3/ITD, WT1, c-KIT, CEBPA, and NPM1, and of specific mixed lineage leukemia (MLL) abnormalities for this purpose.2 The currently active AAML1031 trial from COG is evaluating the role of bortezomib (Velcade, Millennium/Takeda Oncology) in favorable-risk and HR groups, and of sorafenib (Nexavar, Bayer) for patients positive for FLT3/ITD with a high allelic ratio. The use of single-agent sorafenib has led to the successful induction of remission in adult patients with relapsed AML, and this agent can be safely administered to children and young adults in combination with standard chemotherapy.14,15 In a small cohort of pediatric patients with refractory AML, single-agent sorafenib induced a sustained CR.16 Sorafenib has a known association with cardiotoxicity, however, and has been shown to cause dose-limiting hematologic and nonhematologic toxicities in children.17,18 Ongoing research focuses on whole-genome analysis in addition to proteomic profiling of pediatric AML for the purpose of identifying prognostically significant and targetable biomarkers.19

Disease response, traditionally measured by morphology and more recently by flow cytometry, cytogenetics, and RT-PCR, is strongly associated with outcome. Among patients treated with standard chemotherapy, the 3-year relapse rate for those with residual disease after first induction (60%) was significantly higher than that seen in patients without residual disease (29%). Accordingly, relapse-free survival (RFS) was 30% in the residual disease group and 65% in those without residual disease. Similar results were found for patients with residual disease after second induction.20 These findings have been confirmed in an assessment of leukemic blast count after first induction for pediatric patients with AML, in whom an increasing level of residual disease was correlated with decreased 4-year OS.21 Patients with persistent disease after first induction who go on to achieve a CR after second induction remain at high risk for relapse, which demonstrates the vital significance of MRD.22 Patients with greater than 1% MRD after first induction or greater than 0.1% MRD after second induction are considered to be at high risk for relapse.23 Therefore, MRD can be used to identify HR patients among those without prognostically significant cytogenetic markers. Current methods for the detection of MRD include morphology, flow cytometry, cytogenetics, and RT-PCR.

Rationale for Allogeneic Hematopoietic Stem Cell Transplant During First Complete Remission

The disparity in long-term survival among risk groups justifies the application of alternative treatment strategies in patients with HR cytogenetic markers and MRD positivity after induction. In an attempt to improve survival, postconsolidation maintenance therapy with mercaptopurine and cytarabine was examined in pediatric patients with AML. At 5 years, disease-free survival (DFS) was not significantly increased, and OS was decreased in the maintenance therapy group (5-year OS, 58%) in comparison with the control group (5-year OS, 81%).24 Alternatively, increased dose intensity and intensively timed therapy have been associated with improved outcome in most AML subtypes and are the current standard of care.25 However, despite the value of current supportive care practices, these methods result in significant treatment-related toxicity.2

An evaluation of autologous HSCT for pediatric patients with AML demonstrated an OS equivalent to that achieved with chemotherapy alone. Although relapse rates were lower in the autologous HSCT group, the rate of treatment-related toxicity was significantly higher for these patients.26 A clinical trial of pediatric patients with AML that compared chemotherapy, autologous HSCT, and allogeneic HSCT demonstrated superior long-term OS for allogeneic HSCT (60%) vs autologous HSCT (48%) and chemotherapy (53%) for patients in first complete response (CR1) after 2 induction cycles (Figure 1).27 Therefore, autologous HSCT is not currently indicated for pediatric AML consolidation therapy.

The graft-versus-leukemia (GVL) effect of allogeneic HSCT has been proved. In a Children’s Cancer Group study of pediatric patients with AML undergoing allogeneic HSCT, grade 1 or 2 GVHD was associated with improved RFS.28 A pediatric clinical trial that attempted to improve outcomes through enhancement of the GVL effect with reduced dosing of alemtuzumab (Campath, Genzyme), early cessation of GVHD prophylaxis, post-transplant donor lymphocyte infusion, and treatment with interferon alfa demonstrated a 3-year RFS rate of 77%.29 Although a significant reduction in the relapse rate of pediatric patients with AML after HSCT consolidation therapy in comparison with intensive chemotherapy alone has been shown, OS is limited owing to transplant-related morbidity. For this reason, there is no significant benefit to HSCT consolidation therapy vs nonmyeloablative chemotherapy alone for pediatric patients with favorable-risk AML who are in first remission. However, this approach has been shown to confer a survival benefit for patients with intermediate-risk and HR features.

A high level of variability in risk assignment among recent clinical trials evaluating the survival benefit of HSCT has led to conflicting results in each individual subset. In a COG study, 8-year DFS and OS were significantly improved for intermediate-risk patients without favorable or HR cytogenetic markers who received allogeneic HSCT in CR1 vs chemotherapy alone.30 A retrospective review of pediatric patients with AML and HR cytogenetics found equivalent 5-year OS rates for those treated with chemotherapy (43%), matched family donor (MFD) HSCT (46%), or matched unrelated donor (MUD) HSCT (50%).31 In this analysis, assignment to poor risk was limited to those with monosomy 7/del(7q), monosomy 5/del(5q), abnormalities of 3q, t(6;9)(p23;q34), or complex karyotype. The AML02 trial (Treatment of Patients With Newly Diagnosed Acute Myeloid Leukemia or Myelodysplasia) from St Jude Children’s Research Hospital assessed MFD HSCT for patients without prognostically significant cytogenetic markers, and MFD or MUD HSCT for HR patients. This trial demonstrated that although outcomes in HR patients were not significantly different for those who underwent HSCT vs those who did not, patients with greater than 1% MRD after first induction in the HSCT group had improved OS (43%) compared with those in the chemotherapy-alone group (23%).23 The Nordic Society for Pediatric Hematology and Oncology (NOPHO)-AML 2004 clinical trial for pediatric AML (Response-Guided Induction Therapy in Pediatric Acute Myeloid Leukemia With Excellent Remission Rate) risk-stratified patients by response after induction and demonstrated that patients with more than 15% blasts who were ultimately treated with HSCT had improved EFS and OS vs patients with less than 5% blasts or with 5% to 15% blasts who were treated with chemotherapy alone.22 This evidence highlights the importance of cytogenetic and MRD-based risk stratification and enrollment in standardized clinical trials to further define the role of HSCT for pediatric patients with AML in CR1.

Among patients with FLT3/ITD mutations, the use of HSCT significantly improves the OS rate and decreased the relapse rate. OS after HSCT for patients with FLT3/ITD mutations nears that of patients with wild-type FLT3. This advantage demonstrates a clear indication for the use of HSCT in patients with FLT3/ITD mutations, particularly those with an allelic ratio higher than 0.4.32 Outcomes for pediatric myelodysplastic syndrome (MDS) and AML associated with monosomy 7 are poor. However, a 2-year EFS of 69% has been achieved with the use of HSCT.33

High levels of MRD at the time of HSCT have been associated with poor outcomes in adult patients.34 Among pediatric patients with AML, 5-year OS for those with MRD positivity (0.1%-5%) at the time of HSCT was 67%, compared with 80% for MRD-negative patients. MRD positivity, detected by multidimensional flow cytometry at the time of HSCT, is associated with a significantly increased relapse rate and decreased OS in comparison with MRD negativity (Figure 2).35

Donor Source, HLA Matching, and Alloreactivity

Over time, donor sources have expanded to include MUDs and umbilical cord blood (UCB) in addition to MFDs. Genomic typing and advances in supportive care have improved outcomes in MUD HSCT recipients to match those in MFD HSCT recipients, so that having a matched sibling is no longer a criterion for HSCT for pediatric patients in CR1 who have AML with HR markers, as it has been historically.36 In a recent clinical trial, 2-year leukemia-free survival rates in pediatric patients with AML in CR1, CR2, and no CR who received UCB donor HSCT with 1 or 2 HLA mismatches were 59%, 50%, and 21%, respectively. Although these results do not represent an improvement in outcome over MFD HSCT or MUD HSCT, UCB donor HSCT is an option for pediatric patients with HR AML and no alternative donor source. Of note, the incidence of GVHD in this population was 35%, a reduction from rates reported with MFD HSCT or MUD HSCT.37

A haploidentical or killer immunoglobulin receptor (KIR)-incompatible donor source confers the advantage of alloreactivity—and ultimately, GVL effect. The role of natural killer (NK) cells in the elimination of AML has been demonstrated in the setting of allogeneic HSCT. The mechanism of this effect is based on NK cell KIR phenotype. KIR-mismatched allogeneic HSCT creates an opportunity for donor NK cell alloreactivity, which contributes to a GVL effect.38 Certain KIR2DL1 alleles are known to be associated with stronger cytotoxicity signaling than others. A retrospective analysis of pediatric patients with leukemia undergoing HSCT revealed that patients receiving a KIR2DL1-R245 graft with an HLA-C receptor ligand mismatch had better survival and a lower risk for progression than did patients receiving KIR2DL1-C245 homozygous grafts.39 A recent study evaluating haploidentical HSCT with myeloablative conditioning (MAC) for pediatric AML demonstrated 5-year leukemia-free survival rates of 82%, 59%, and 42% for patients in CR1, in CR2, and with persistent disease at the time of HSCT, respectively. In this study, the 5-year leukemia-free survival for patients with MFDs was similar (at 71%), indicating that this approach is safe in children and may confer benefit through alloreactivity. This approach also widens the donor pool, providing ethnically diverse populations who lack representation in the international donor registries the opportunity for treatment with HSCT when indicated (Table).40

Conditioning Regimens for Hematopoietic Stem Cell Transplant in Pediatric Acute Myeloid Leukemia

An international study comparing reduced-intensity conditioning vs MAC HSCT in pediatric patients with AML showed no significant difference in relapse rates or 5-year EFS rates.41 However, any difference between the incidence of adverse late effects in these cohorts remains to be seen. Outcomes are not significantly different for patients undergoing nonmyeloablative HSCT from an MFD or MUD.42 The reduced-intensity approach may offer benefit to pediatric patients with AML who are ineligible for myeloablative therapy owing to heavy pretreatment and/or pretransplant comorbidities.

Total-body irradiation (TBI) for pretransplant conditioning in pediatric patients with AML in first remission is associated with an increased incidence of secondary malignancy and adverse late effects in comparison with busulfan/cyclophosphamide alone.43 Accordingly, outcomes are better with busulfan-based regimens than with TBI-based regimens regardless of donor source,44 so that chemotherapy-based conditioning regimens are preferred to TBI.45,46 In the setting of Fanconi anemia–associated pediatric AML, cross-linking chemotherapy and high-dose anthracyclines should be avoided to reduce the risk for secondary malignancy.47

The role of targeted therapy in combination with HSCT is under ongoing investigation. The benefits of this approach must be weighed against the potential for treatment-related mortality, and the ideal timing for such therapies is yet to be determined. Gemtuzumab ozogamicin (GO; Mylotarg, Pfizer) has shown promise in pediatric clinical trials. NOPHO-AML 2004, which compared GO as post-consolidation therapy with no further treatment, found no decrease in relapse rates or increase in OS in the GO group.48 However, a recent COG trial comparing the use of standard 5-course chemotherapy alone vs standard chemotherapy plus GO showed a significant reduction in relapse risk and improvement in EFS for patients with a high level of CD33 expression in all risk groups.49 In addition, MRD in pediatric patients with AML treated with GO was reduced in comparison with MRD in those treated with chemotherapy alone, suggesting a critical role for this targeted therapy in the peritransplant period. A recent phase 1 clinical trial that evaluated the safety and efficacy of GO before allogeneic HSCT with varying donor sources, in which reduced-intensity conditioning was used, demonstrated a 61% 5-year EFS for all patients and a 78% EFS for those in CR1 at the time of HSCT.50 These results suggest that GO plays a role in the clearance of residual MRD after HSCT. A phase 2 clinical trial based on these data is currently active.

In a retrospective review of pediatric patients with FLT3/ITD-mutated AML, in which the patients received sorafenib prophylactically or at the time of relapse after HSCT, all of 15 patients achieved a CR at a median of 48 months, and the incidence of GVHD was not increased.51 The COG AAML1031 study is evaluating the safety and efficacy of combination chemotherapy plus sorafenib followed by HSCT and a post-HSCT sorafenib maintenance phase for FLT3/ITD-mutated patients who have a high allelic ratio. As techniques in proteomic and genomic profiling advance and targetable pathways are identified, the development of a personalized approach to HSCT may be warranted. Panobinostat (Farydak, Novartis; NCT01451268), decitabine (NCT01277484), and azacitidine (NCT01995578) are currently under clinical investigation as maintenance therapy for adult patients with AML after HSCT.

The role of early withdrawal of immunosuppression and donor lymphocyte infusion after HSCT as relapse prophylaxis in pediatric patients with leukemia, including AML with mixed chimerism, was examined in limited numbers. A long-term CR was achieved in 7 of 12 patients; however, induction of GVHD and treatment-related mortality was a significant consideration.52 NK cell infusion has not been shown to increase the incidence of GVHD. Evaluation of a small cohort of adult patients with relapsed and refractory AML treated with third-party alloreactive haploidentical NK cell infusion before allogeneic HSCT demonstrated safe administration without outcome benefit, possibly owing to the small sample size. A phase 1/2 clinical trial based on these data is currently ongoing.53

Clinical advances in supportive care, prevention and early treatment of microbial infections, GVHD prophylaxis, veno-occlusive disease prophylaxis, and reduced-intensity conditioning all contribute to the reduction of transplant-related morbidity and mortality. As toxic outcomes of HSCT decrease and HR features are more precisely defined, HSCT for pediatric patients with AML in CR1 may become increasingly advantageous.

The Role of Allogeneic Hematopoietic Stem Cell Transplant in Relapsed and Refractory Disease

Between 3% and 19% of pediatric patients with AML do not achieve CR after induction therapy, and 30% to 40% of all pediatric patients with AML relapse.4 Although more than 60% of patients with relapsed disease achieve a second CR (CR2), only 29% achieve long-term remission. Key factors that are independently associated with survival include age younger than 10 years, favorable cytogenetics, duration of CR1, and no HSCT in CR1.54 OS for patients with relapse more than 12 months after diagnosis (48%) is superior to that for patients with relapse during the first 12 months (21%).55 For those achieving CR2 with cytarabine, etoposide, and an anthracycline, OS is significantly improved after HSCT (50%) vs chemotherapy alone (24%).56 An international study in which fludarabine, cytarabine, and granulocyte-colony stimulating factor (FLAG) or FLAG plus liposomal daunorubicin followed by HSCT was used for those achieving a CR reported a 64% CR rate and 38% OS.57 Others have reported OS of 62% for patients in CR2 before HSCT.55,58 A phase 1 study of pediatric patients with refractory or relapsed AML undergoing allogeneic HSCT with myeloablative conditioning consisting of GO, busulfan, and cyclophosphamide reported 1-year EFS and OS rates of 50%.59 There is a consensus that HSCT should be offered to all children with relapsed AML in CR2.2 In this setting, MFD transplant has not been associated with a long-term survival benefit vs alternative donor sources.60 Outcomes for patients with persistent disease before HSCT are poor. OS for patients not achieving CR2 before HSCT has been reported to be as low as 3%.54 Others have reported a 5-year leukemia-free survival rate for pediatric patients not achieving CR2 at the time of HSCT of 20%.61

Given the proven GVL effect of HSCT for pediatric AML, donor lymphocyte infusion for patients with relapse after MFD or MUD HSCT is a consideration but has a limited effect. Donor lymphocyte infusion as sole therapy for pediatric patients with hematologic malignancy relapsing after HSCT rarely induces remission and has not shown benefit over no donor lymphocyte infusion. However, donor lymphocyte infusion in combination with chemotherapy has resulted in 1-year DFS of 30%.62 Single-agent sorafenib and other novel targeted therapies, including antibody- and cell-based therapy, for patients with FLT3/ITD-mutated AML are under investigation for this purpose in the adult population.

AML cell surface antigens CD33 and CD123 have been investigated as immunotherapeutic targets. Chimeric antigen receptor (CAR)–modified cytokine-induced killer cells induced significant cytotoxicity against AML blasts and normal hematopoietic stem cells in preclinical models, although regenerative capacity was preserved.63,64 Virally transduced anti-CD123 CAR T cells have also shown efficacy in reducing CD123-positive AML blast and leukemia stem cell burden with a low hematopoietic toxicity profile, and they are currently under investigation in an open clinical trial (NCT02159495).65 Several novel targeted antibodies are under clinical investigation in the adult AML population, including an anti-CD33 pyrrolobenzodiazepine-conjugated monoclonal antibody that has shown increased apoptosis vs GO in preclinical models (NCT02326584, NCT01902329). Bispecific T-cell engager CD33 x CD3 (AMG 330), bispecific killer cell engager CD33 x CD16, and trispecific engager CD33 x CD16 x CD123 have shown evidence of successful clearance of AML blasts in preclinical models and are candidates for clinical investigation.66-70

Central Nervous System Therapy

Patients with central nervous system (CNS) involvement at the time of diagnosis are at increased risk for isolated CNS relapse and require intensified CNS-directed therapy.71 An evaluation of adult patients who had AML with CNS disease at diagnosis and underwent HSCT with intrathecal therapy (ITT) vs cranial or craniospinal irradiation boost demonstrated a significant advantage in 5-year relapse-free survival and OS in the radiation therapy group. However, a retrospective study of adults with CNS disease at diagnosis demonstrated successful clearance with or without radiation therapy, and no association between CNS disease at diagnosis and outcome after HSCT.72 The role of post-HSCT ITT and pre-HSCT irradiation boost in pediatric patients with acute lymphoblastic leukemia or AML has been evaluated. No significant difference in outcomes was found between those who received intrathecal chemotherapy and those who received no ITT. In addition, there was no significant difference in outcomes between patients who underwent pre-HSCT CNS irradiation boost and those who underwent ITT alone.73


Several reports have distinguished acute erythroid leukemia (FAB M6 in the French-American-British classification system) and acute megakaryoblastic leukemia (AMKL; FAB M7) from FAB M0-M5. A comparison of pediatric AML subtypes and MDS in patients without constitutional trisomy 21 (T21) found that those with MDS, M6 AML, or M7 AML had a higher frequency of monosomy 7 and –7q and EFS and OS significantly inferior to that of other AML subtypes.74 AMKL is the most common subtype of AML in children with T21 and accounts for up to 15% of cases of pediatric AML. In a single-center study, the remission induction rate for pediatric AMKL with or without T21 was reported at 60%, with a 48% relapse rate. The 2-year EFS for patients who had M7 AML without T21 (14%) was significantly inferior to the 2-year EFS for those who had M7 AML with T21 (83%). The 2-year EFS for patients without T21 was significantly improved after allogeneic HSCT, and those undergoing allogeneic HSCT in CR had a significantly higher 2-year EFS (46%) than did those not in CR at the time of HSCT (0%). Successful induction of remission was the prominent prognostic feature. The 5-year EFS was significantly lower for M7 AML without T21 than for other subtypes.75 These data clearly indicate that allogeneic HSCT in CR1 confers optimal survival for pediatric patients who have M7 AML without T21.

Acute promyelocytic leukemia, Down syndrome–associated AML, therapy-related secondary AML, and familial cancer syndrome–associated AML should be considered separately owing to their vastly different biology and sensitivity to therapy. Because of the excellent outcomes achieved in pediatric acute promyelocytic leukemia when all-trans retinoic acid is give in combination with cytotoxic chemotherapy, HSCT should be considered only for patients with relapse. In a comparison of allogeneic vs autologous HSCT for the treatment of relapsed or refractory childhood acute promyelocytic leukemia, 5-year EFS and OS were not significantly different. Although lower rates of relapse were observed in the allogeneic HSCT group, treatment-related mortality was significantly higher.76 Treatment-related mortality and morbidity are also a barrier to HSCT for patients with Down syndrome who have refractory or relapsed AML. The 3-year OS for pediatric patients who have Down syndrome–associated AML has been reported at 19%, and treatment-related mortality and relapse risk are significantly higher than in pediatric patients without Down syndrome who undergo HSCT for AML. Pediatric patients with chemotherapy-related and radiation therapy–related secondary AML can be salvaged. The 2-year OS for therapy-related secondary AML in pediatric patients undergoing HSCT while in CR after reinduction chemotherapy has been reported at 40%.77 There are few reported cases of familial cancer syndrome–associated pediatric AML; therefore, the role of HSCT in this population is unclear.

Late Effects

Late toxicity associated with HSCT for pediatric malignancies is an important consideration. A review of pediatric patients receiving HSCT for hematologic malignancy demonstrated a significantly increased relative risk for severe or life-threatening infections, 2 or more chronic health conditions, functional impairment, and activity limitation in comparison with matched sibling controls. Recipients of MUD HSCT were at greatest risk.78 The incidence of late adverse effects is significantly higher among pediatric patients with AML who receive HSCT (72%) than among those who receive chemotherapy alone (31%).79 Younger age and cranial irradiation or TBI are independent risk factors for the development of late effects.80 The risk for late effects and their impact on long-term survival, chronic comorbidities, and quality of life must be weighed against the significant risk for relapse in pediatric patients who have AML with HR features. All patients who have undergone HSCT in childhood are encouraged to participate when possible in organized survivorship programs developed for the early detection and prevention of late transplant-related morbidities, and for the improvement of long-term outcomes among survivors of pediatric AML.


The interpretation of outcomes for pediatric patients with AML based on previous clinical trials is limited by variations in the definition of high risk and the indications for allogeneic HSCT. The identification of genetic markers of HR disease and the detection of MRD are critical to risk stratification in pediatric AML and the identification of patients who might benefit from intensified therapy. The emergence of additional proteomic and genomic markers is expected, and these must be integrated into risk stratification and the treatment strategy to further define the HR population.

It is clear that patients with favorable features, such as core-binding factor mutations, biallelic CEBPA mutations, and NPM1 mutations without other adverse factors are unlikely to benefit from allogeneic HSCT. However, allogeneic HSCT has resulted in improved outcomes for specific subsets of pediatric patients with AML—including those who have ITD/FLT3-mutated disease with a high allelic ratio, monosomy 7, or AMKL and those who have MRD positivity at the end of induction—and may be of benefit for patients with newly identified HR features.

Outcomes for patients with refractory and relapsed disease are improved by allogeneic HSCT. Clearance of residual disease, however, is essential to long-term survival. The objective of a cytogenetic marker and an MRD-driven, pediatric-specific risk stratification system is to limit toxicity for patients without a high risk for relapse while intensifying therapy for those known to be at increased risk for relapse. The treatment strategy should be focused on MRD clearance and eradication of the preleukemic stem cell clone, which may be achieved through the GVL effect of allogeneic HSCT in combination with targeted therapies.

Advances in supportive care; expansion of the donor pool to include MFDs, MUDs, UCB, and haploidentical donors; high-resolution HLA typing; graft manipulation; and GVHD prophylaxis have all contributed to improved HSCT strategy. In addition, reductions in toxicity and conditioning and the limited use of TBI may result in a decrease in long-term toxicity. Ultimately, the identification of patients with known adverse features and the benefits of alloreactivity in combination with targeted therapy may reduce relapse rates and close the gap in OS for those pediatric patients with HR AML disease.


This manuscript was supported in part by a grant from the Pediatric Cancer Research Foundation. The authors have no disclosures to report. 


The authors would like to thank Erin Morris, RN, for her assistance in the preparation of this manuscript. 


1. Meshinchi S, Arceci RJ. Prognostic factors and risk-based therapy in pediatric acute myeloid leukemia. Oncologist. 2007;12(3):341-355.

2. Creutzig U, van den Heuvel-Eibrink MM, Gibson B, et al; AML Committee of the International BFM Study Group. Diagnosis and management of acute myeloid leukemia in children and adolescents: recommendations from an international expert panel. Blood. 2012;120(16):3187-3205.

3. Pui CH, Carroll WL, Meshinchi S, Arceci RJ. Biology, risk stratification, and therapy of pediatric acute leukemias: an update. J Clin Oncol. 2011;29(5):551-565.

4. Kaspers GJ, Creutzig U. Pediatric acute myeloid leukemia: international progress and future directions. Leukemia. 2005;19(12):2025-2029.

5. Harrison CJ, Hills RK, Moorman AV, et al. Cytogenetics of childhood acute myeloid leukemia: United Kingdom Medical Research Council Treatment trials AML 10 and 12. J Clin Oncol. 2010;28(16):2674-2681.

6. von Neuhoff C, Reinhardt D, Sander A, et al. Prognostic impact of specific chromosomal aberrations in a large group of pediatric patients with acute myeloid leukemia treated uniformly according to trial AML-BFM 98. J Clin Oncol. 2010;28(16):2682-2689.

7. Staffas A, Kanduri M, Hovland R, et al; Nordic Society of Pediatric Hematology and Oncology (NOPHO). Presence of FLT3-ITD and high BAALC expression are independent prognostic markers in childhood acute myeloid leukemia. Blood. 2011;118(22):5905-5913.

8. Ho PA, Alonzo TA, Gerbing RB, et al. Prevalence and prognostic implications of CEBPA mutations in pediatric acute myeloid leukemia (AML): a report from the Children’s Oncology Group. Blood. 2009;113(26):6558-6566.

9. Alonso CN, Rubio PL, Medina A, et al. FLT3, NPM1 and CEBPA mutations in 195 children with acute myeloid leukemia in Argentina [ASH abstract 5318]. Blood. 2014;124(21)(suppl).

10. Rubnitz JE. How I treat pediatric acute myeloid leukemia. Blood. 2012;119(25):5980-5988.

11. Meshinchi S, Alonzo TA, Stirewalt DL, et al. Clinical implications of FLT3 mutations in pediatric AML. Blood. 2006;108(12):3654-3661.

12. Hasle H, Alonzo TA, Auvrignon A, et al. Monosomy 7 and deletion 7q in children and adolescents with acute myeloid leukemia: an international retrospective study. Blood. 2007;109(11):4641-4647.

13. Delaney B, Zhang J, Carlson G, et al. A gene-shuffled glyphosate acetyltransferase protein from Bacillus licheniformis (GAT4601) shows no evidence of allergenicity or toxicity. Toxicol Sci. 2008;102(2):425-432.

14. Metzelder S, Wang Y, Wollmer E, et al. Compassionate use of sorafenib in FLT3-ITD-positive acute myeloid leukemia: sustained regression before and after allogeneic stem cell transplantation. Blood. 2009;113(26):6567-6571.

15. Ravandi F, Cortes JE, Jones D, et al. Phase I/II study of combination therapy with sorafenib, idarubicin, and cytarabine in younger patients with acute myeloid leukemia. J Clin Oncol. 2010;28(11):1856-1862.

16. Watt TC, Cooper T. Sorafenib as treatment for relapsed or refractory pediatric acute myelogenous leukemia. Pediatr Blood Cancer. 2012;59(4):756-757.

17. Widemann BC, Kim A, Fox E, et al. A phase I trial and pharmacokinetic study of sorafenib in children with refractory solid tumors or leukemias: a Children’s Oncology Group Phase I Consortium report. Clin Cancer Res. 2012;18(21):6011-6022.

18. Schmidinger M, Zielinski CC, Vogl UM, et al. Cardiac toxicity of sunitinib and sorafenib in patients with metastatic renal cell carcinoma. J Clin Oncol. 2008;26(32):5204-5212.

19. Braoudaki M, Tzortzatou-Stathopoulou F, Anagnostopoulos AK, et al. Proteomic analysis of childhood de novo acute myeloid leukemia and myelodysplastic syndrome/AML: correlation to molecular and cytogenetic analyses. Amino Acids. 2011;40(3):943-951.

20. Loken MR, Alonzo TA, Pardo L, et al. Residual disease detected by multidimensional flow cytometry signifies high relapse risk in patients with de novo acute myeloid leukemia: a report from Children’s Oncology Group. Blood. 2012;120(8):1581-1588.

21. Creutzig U, Zimmermann M, Dworzak MN, et al. The prognostic significance of early treatment response in pediatric relapsed acute myeloid leukemia: results of the international study Relapsed AML 2001/01. Haematologica. 2014;99(9):1472-1478.

22. Abrahamsson J, Forestier E, Heldrup J, et al. Response-guided induction therapy in pediatric acute myeloid leukemia with excellent remission rate. J Clin Oncol. 2011;29(3):310-315.

23. Rubnitz JE, Inaba H, Dahl G, et al. Minimal residual disease-directed therapy for childhood acute myeloid leukaemia: results of the AML02 multicentre trial. Lancet Oncol. 2010;11(6):543-552.

24. Perel Y, Auvrignon A, Leblanc T, et al; Group LAME of the French Society of Pediatric Hematology and Immunology. Impact of addition of maintenance therapy to intensive induction and consolidation chemotherapy for childhood acute myeloblastic leukemia: results of a prospective randomized trial, LAME 89/91. Leucámie Aiqüe Myéloïde Enfant. J Clin Oncol. 2002;20(12):2774-2782.

25. Smith FO, Alonzo TA, Gerbing RB, Woods WG, Arceci RJ; Children’s Cancer Group. Long-term results of children with acute myeloid leukemia: a report of three consecutive phase III trials by the Children’s Cancer Group: CCG 251, CCG 213 and CCG 2891. Leukemia. 2005;19(12):2054-2062.

26. Ravindranath Y, Yeager AM, Chang MN, et al; Pediatric Oncology Group. Autologous bone marrow transplantation versus intensive consolidation chemotherapy for acute myeloid leukemia in childhood. N Engl J Med. 1996;334(22):1428-1434.

27. Woods WG, Neudorf S, Gold S, et al; Children’s Cancer Group. A comparison of allogeneic bone marrow transplantation, autologous bone marrow transplantation, and aggressive chemotherapy in children with acute myeloid leukemia in remission. Blood. 2001;97(1):56-62.

28. Neudorf S, Sanders J, Kobrinsky N, et al. Allogeneic bone marrow transplantation for children with acute myelocytic leukemia in first remission demonstrates a role for graft versus leukemia in the maintenance of disease-free survival. Blood. 2004;103(10):3655-3661.

29. Bonanomi S, Connor P, Webb D, et al. Successful outcome of allo-SCT in high-risk pediatric AML using chemotherapy-only conditioning and post transplant immunotherapy. Bone Marrow Transplant. 2008;42(4):253-257.

30. Horan JT, Alonzo TA, Lyman GH, et al; Children’s Oncology Group. Impact of disease risk on efficacy of matched related bone marrow transplantation for pediatric acute myeloid leukemia: the Children’s Oncology Group. J Clin Oncol. 2008;26(35):5797-5801.

31. Kelly MJ, Horan JT, Alonzo TA, et al. Comparable survival for pediatric acute myeloid leukemia with poor-risk cytogenetics following chemotherapy, matched related donor, or unrelated donor transplantation. Pediatr Blood Cancer. 2014;61(2):269-275.

32. Meshinchi S, Arceci RJ, Sanders JE, et al. Role of allogeneic stem cell transplantation in FLT3/ITD-positive AML. Blood. 2006;108(1):400.

33. Trobaugh-Lotrario AD, Kletzel M, Quinones RR, et al. Monosomy 7 associated with pediatric acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS): successful management by allogeneic hematopoietic stem cell transplant (HSCT). Bone Marrow Transplant. 2005;35(2):143-149.

34. Walter RB, Othus M, Burnett AK, et al. Significance of FAB subclassification of “acute myeloid leukemia, NOS” in the 2008 WHO classification: analysis of 5848 newly diagnosed patients. Blood. 2013;121(13):2424-2431.

35. Leung W, Pui CH, Coustan-Smith E, et al. Detectable minimal residual disease before hematopoietic cell transplantation is prognostic but does not preclude cure for children with very-high-risk leukemia. Blood. 2012;120(2):468-472.

36. Moore J, Nivison-Smith I, Goh K, et al. Equivalent survival for sibling and unrelated donor allogeneic stem cell transplantation for acute myelogenous leukemia. Biol Blood Marrow Transplant. 2007;13(5):601-607.

37. Michel G, Rocha V, Chevret S, et al; Eurocord Group. Unrelated cord blood transplantation for childhood acute myeloid leukemia: a Eurocord Group analysis. Blood. 2003;102(13):4290-4297.

38. Ruggeri L, Capanni M, Casucci M, et al. Role of natural killer cell alloreactivity in HLA-mismatched hematopoietic stem cell transplantation. Blood. 1999;94(1):333-339.

39. Bari R, Rujkijyanont P, Sullivan E, et al. Effect of donor KIR2DL1 allelic polymorphism on the outcome of pediatric allogeneic hematopoietic stem-cell transplantation. J Clin Oncol. 2013;31(30):3782-3790.

40. Palma J, Salas L, Carrión F, et al. Haploidentical stem cell transplantation for children with high-risk leukemia. Pediatr Blood Cancer. 2012;59(5):895-901.

41. Bitan M, He W, Zhang MJ, et al. Transplantation for children with acute myeloid leukemia: a comparison of outcomes with reduced intensity and myeloablative regimens. Blood. 2014;123(10):1615-1620.

42. Peffault de Latour R, Brunstein CG, Porcher R, et al. Similar overall survival using sibling, unrelated donor, and cord blood grafts after reduced-intensity conditioning for older patients with acute myelogenous leukemia. Biol Blood Marrow Transplant. 2013;19(9):1355-1360.

43. Michel G, Socié G, Gebhard F, et al. Late effects of allogeneic bone marrow transplantation for children with acute myeloblastic leukemia in first complete remission: the impact of conditioning regimen without total-body irradiation—a report from the Société Française de Greffe de Moelle. J Clin Oncol. 1997;15(6):2238-2246.

44. Bresters D, van Gils IC, Kollen WJ, et al. High burden of late effects after haematopoietic stem cell transplantation in childhood: a single-centre study. Bone Marrow Transplant. 2010;45(1):79-85.

45. Bhatia S. Long-term health impacts of hematopoietic stem cell transplantation inform recommendations for follow-up. Expert Rev Hematol. 2011;4(4):437-452.

46. Gupta T, Kannan S, Dantkale V, Laskar S. Cyclophosphamide plus total body irradiation compared with busulfan plus cyclophosphamide as a conditioning regimen prior to hematopoietic stem cell transplantation in patients with leukemia: a systematic review and meta-analysis. Hematol Oncol Stem Cell Ther. 2011;4(1):

47. Mehta PA, Ileri T, Harris RE, et al. Chemotherapy for myeloid malignancy in children with Fanconi anemia. Pediatr Blood Cancer. 2007;48(7):668-672.

48. Hasle H, Abrahamsson J, Forestier E, et al; Nordic Society of Paediatric Haematology and Oncology (NOPHO). Gemtuzumab ozogamicin as postconsolidation therapy does not prevent relapse in children with AML: results from NOPHO-AML 2004. Blood. 2012;120(5):978-984.

49. Pollard JA, Loken M, Gerbing RB, et al. CD33 expression and its association with gemtuzumab ozogamicin response: results from the randomized phase III Children’s Oncology Group Trial AAML0531. J Clin Oncol. 2016;34(7):747-755.

50. Zahler S, Bhatia M, Ricci A, et al. A phase I study of reduced-intensity conditioning and allogeneic stem cell transplantation followed by dose escalation of targeted consolidation immunotherapy with gemtuzumab ozogamicin in children and adolescents with CD33+ acute myeloid leukemia. Biol Blood Marrow Transplant. 2016;22(4):698-704.

51. Tarlock K, Chang B, Cooper T, et al. Sorafenib treatment following hematopoietic stem cell transplant in pediatric FLT3/ITD acute myeloid leukemia. Pediatr Blood Cancer. 2015;62(6):1048-1054.

52. Bader P, Klingebiel T, Schaudt A, et al. Prevention of relapse in pediatric patients with acute leukemias and MDS after allogeneic SCT by early immunotherapy initiated on the basis of increasing mixed chimerism: a single center experience of 12 children. Leukemia. 1999;13(12):2079-2086.

53. Lee DA, Denman CJ, Rondon G, et al. Haploidentical natural killer cells infused before allogeneic stem cell transplantation for myeloid malignancies: a phase I trial. Biol Blood Marrow Transplant. 2016;22(7):1290-1298.

54. Sander A, Zimmermann M, Dworzak M, et al. Consequent and intensified relapse therapy improved survival in pediatric AML: results of relapse treatment in 379 patients of three consecutive AML-BFM trials. Leukemia. 2010;24(8):1422-1428.

55. Abrahamsson J, Clausen N, Gustafsson G, et al; Nordic Society for Paediatric Haematology and Oncology (NOPHO). Improved outcome after relapse in children with acute myeloid leukaemia. Br J Haematol. 2007;136(2):229-236.

56. Goemans BF, Tamminga RY, Corbijn CM, Hählen K, Kaspers GJ. Outcome for children with relapsed acute myeloid leukemia in the Netherlands following initial treatment between 1980 and 1998: survival after chemotherapy only? Haematologica. 2008;93(9):1418-1420.

57. Kaspers GJ, Zimmermann M, Reinhardt D, et al. Improved outcome in pediatric relapsed acute myeloid leukemia: results of a randomized trial on liposomal daunorubicin by the International BFM Study Group. J Clin Oncol. 2013;31(5):599-607.

58. Beier R, Albert MH, Bader P, et al. Allo-SCT using BU, CY and melphalan for children with AML in second CR. Bone Marrow Transplant. 2013;48(5):651-656.

59. Satwani P, Bhatia M, Garvin JH Jr, et al. A phase I study of gemtuzumab ozogamicin (GO) in combination with busulfan and cyclophosphamide (Bu/Cy) and allogeneic stem cell transplantation in children with poor-risk CD33+ AML: a new targeted immunochemotherapy myeloablative conditioning (MAC) regimen. Biol Blood Marrow Transplant. 2012;18(2):324-329.

60. Shaw PJ, Kan F, Woo Ahn K, et al. Outcomes of pediatric bone marrow transplantation for leukemia and myelodysplasia using matched sibling, mismatched related, or matched unrelated donors. Blood. 2010;116(19):4007-4015.

61. Bunin NJ, Davies SM, Aplenc R, et al. Unrelated donor bone marrow transplantation for children with acute myeloid leukemia beyond first remission or refractory to chemotherapy. J Clin Oncol. 2008;26(26):4326-4332.

62. Levine JE, Barrett AJ, Zhang MJ, et al. Donor leukocyte infusions to treat hematologic malignancy relapse following allo-SCT in a pediatric population. Bone Marrow Transplant. 2008;42(3):201-205.

63. Marin V, Pizzitola I, Agostoni V, et al. Cytokine-induced killer cells for cell therapy of acute myeloid leukemia: improvement of their immune activity by expression of CD33-specific chimeric receptors. Haematologica. 2010;95(12):2144-2152.

64. Tettamanti S, Marin V, Pizzitola I, et al. Targeting of acute myeloid leukaemia by cytokine-induced killer cells redirected with a novel CD123-specific chimeric antigen receptor. Br J Haematol. 2013;161(3):389-401.

65. Mardiros A, Dos Santos C, McDonald T, et al. T cells expressing CD123-specific chimeric antigen receptors exhibit specific cytolytic effector functions and antitumor effects against human acute myeloid leukemia. Blood. 2013;122(18):3138-3148.

66. Laszlo GS, Gudgeon CJ, Harrington KH, et al. Cellular determinants for preclinical activity of a novel CD33/CD3 bispecific T-cell engager (BiTE) antibody, AMG 330, against human AML. Blood. 2014;123(4):554-561.

67. Krupka C, Kufer P, Kischel R, et al. CD33 target validation and sustained depletion of AML blasts in long-term cultures by the bispecific T-cell-engaging antibody AMG 330. Blood. 2014;123(3):356-365.

68. Singer H, Kellner C, Lanig H, et al. Effective elimination of acute myeloid leukemic cells by recombinant bispecific antibody derivatives directed against CD33 and CD16. J Immunother. 2010;33(6):599-608.

69. Kügler M, Stein C, Kellner C, et al. A recombinant trispecific single-chain Fv derivative directed against CD123 and CD33 mediates effective elimination of acute myeloid leukaemia cells by dual targeting. Br J Haematol. 2010;150(5):

70. Gleason MK, Ross JA, Warlick ED, et al. CD16xCD33 bispecific killer cell engager (BiKE) activates NK cells against primary MDS and MDSC CD33+ targets. Blood. 2014;123(19):3016-3026.

71. Johnston DL, Alonzo TA, Gerbing RB, Lange BJ, Woods WG. The presence of central nervous system disease at diagnosis in pediatric acute myeloid leukemia does not affect survival: a Children’s Oncology Group study. Pediatr Blood Cancer. 2010;55(3):414-420.

72. Bar M, Tong W, Othus M, Loeb KR, Estey EH. Central nervous system involvement in acute myeloid leukemia patients undergoing hematopoietic cell transplantation. Biol Blood Marrow Transplant. 2015;21(3):546-551.

73. Rubin J, Vettenranta K, Vettenranta J, et al. Use of intrathecal chemoprophylaxis in children after SCT and the risk of central nervous system relapse. Bone Marrow Transplant. 2011;46(3):372-378.

74. Barnard DR, Alonzo TA, Gerbing RB, Lange B, Woods WG; Children’s Oncology Group. Comparison of childhood myelodysplastic syndrome, AML FAB M6 or M7, CCG 2891: report from the Children’s Oncology Group. Pediatr Blood Cancer. 2007;49(1):17-22.

75. Athale UH, Razzouk BI, Raimondi SC, et al. Biology and outcome of childhood acute megakaryoblastic leukemia: a single institution’s experience. Blood. 2001;97(12):3727-3732.

76. Dvorak CC, Agarwal R, Dahl GV, Gregory JJ, Feusner JH. Hematopoietic stem cell transplant for pediatric acute promyelocytic leukemia. Biol Blood Marrow Transplant. 2008;14(7):824-830.

77. Aguilera DG, Vaklavas C, Tsimberidou AM, Wen S, Medeiros LJ, Corey SJ. Pediatric therapy-related myelodysplastic syndrome/acute myeloid leukemia: the MD Anderson Cancer Center experience. J Pediatr Hematol Oncol. 2009;31(11):803-811.

78. Armenian SH, Sun CL, Kawashima T, et al. Long-term health-related outcomes in survivors of childhood cancer treated with HSCT versus conventional therapy: a report from the Bone Marrow Transplant Survivor Study (BMTSS) and Childhood Cancer Survivor Study (CCSS). Blood. 2011;118(5):1413-1420.

79. Klusmann JH, Reinhardt D, Zimmermann M, et al. The role of matched sibling donor allogeneic stem cell transplantation in pediatric high-risk acute myeloid leukemia: results from the AML-BFM 98 study. Haematologica. 2012;97(1):

80. Leung W, Hudson MM, Strickland DK, et al. Late effects of treatment in survivors of childhood acute myeloid leukemia. J Clin Oncol. 2000;18(18):3273-3279

81. Ravindranath Y, Yeager AM, Chang MN, et al. Autologous bone marrow transplantation versus intensive consolidation chemotherapy for acute myeloid leukemia in childhood. N Engl J Med. 1996;334(22):1428-1434.