Clinical Manifestations and Treatment of Acute Myeloid Leukemia

Epidemiology

In 2019, 21,450 estimated new cases of patients with AML were diagnosed in the United States and an estimated 10,920 patients with AML died from it. The age-adjusted incidence rate of AML is 3.6/100,000 population. However, with a median age at diagnosis of 68 years, incidence rates are as high as greater than 15/100,000 in the older age group, and about 70% of all diagnoses of AML are in patients over 55 years of age. The 5-year survival rate declines with advancing age for many reasons, including declining overall health, higher incidence of poor risk cytogenetic and mutational features, adverse socioeconomic factors such as absence of adequate caregiver support, and inability to tolerate treatments associated with lower risk of relapse, such as high-dose cytarabine consolidation therapy and allogeneic hematopoietic stem cell transplantation (allo HSCT). AML is slightly more frequent in men than women (lifetime risk of acquiring AML: 1 in 227 [male] to 1 in 278 [female]). Based on Surveillance, Epidemiology, and End Results (SEER) data from 2005 to 2009, small differences in frequency also exist by race; AML is more common in whites than other ethnic groups. African Americans have a worse overall survival (OS) than Whites in National Cancer Institute-sponsored cooperative group clinical trials. According to one study of 27,525 patients with AML, based on the SEER database from 1999 to 2008, African Americans and Hispanics had a 12% and 6% increased hazard of death, respectively. This unfavorable hazard rate pertained despite the higher rate in these two ethnic groups of AML with translocations t(8;21) and t(15;17), both associated with a better outcome.

AML has been associated with exposure to benzene, cigarette smoke, chemotherapy, ionizing radiation, and immunosuppressive agents. AML may develop following immunosuppressive therapy for aplastic anemia ( Chapter 31, Chapter 32 ) or evolve from an antecedent myeloid neoplasm (myelodysplastic syndrome [MDS]; Chapter 73 ), myeloproliferative neoplasms, and myelodysplastic/myeloproliferative neoplasms) ( Chapter 70, Chapter 71, Chapter 72, Chapter 73 ). AML may also be due to an inherited germline mutation and indicate a familial cancer syndrome ( Chapter 30 ). AML patients with two or more first- or second-degree relatives with hematologic neoplasms, prolonged cytopenias, or bleeding disorders should be screened for germline mutations which predispose to myeloid neoplasms. AML patients may have physical stigmata associated with known clinical syndromes. For example, mucosal leukoplakia, reticular hyperpigmentation, nail dystrophy, and interstitial lung disease are all suggestive of short telomere syndromes, such as dyskeratosis congenita ( Chapter 30 ). The European Leukemia Network (ELN) 2017 guidelines suggests testing for germline mutations of RUNX1 , CEBPA , DDX41 , TERC , TERT , TP53 , and others, when appropriate. Some of these variants will be detected by NGS panels at the time of AML diagnosis. The patient should be referred for genetic counseling, especially if present at 50% variant allelic frequency at the time of remission. Identification of other affected family members is an important consideration for selection of family members as hematopoietic stem cell donors.

Clinical and Laboratory Manifestations

The initial presentation of patients with AML may vary from the discovery of relatively asymptomatic cytopenias at the time of routine health maintenance exams to life-threatening complications, requiring urgent intervention. Signs and symptoms of AML mostly reflect the effect of cytopenias. Patients typically present with a short history (few weeks to 3 months) of constitutional symptoms (fatigue, lack of energy, malaise, fever, profuse sweats), infections, and/or manifestations of bleeding (e.g., gingival bleeding, bruising, epistaxis, menorrhagia). Fevers should always be presumed to be secondary to infections even in the absence of an identifiable focus and lead to rapid institution of antibiotic therapy. “Tumor fever” due to AML does occur but remains a diagnosis of exclusion. Extramedullary leukemic infiltrates in the gingiva, skin, lymph nodes, or other organs occur occasionally, and are more often associated with monocytic differentiation. Signs and symptoms referable to central nervous system (CNS) involvement (cranial neuropathies and other focal neurologic abnormalities, encephalopathy, seizures) are rare, except for AML with monocytic/monoblastic differentiation or in any AML with considerable leukocytosis (>100,000/μL). The life-threatening complications requiring emergent intervention include clinical leukostasis, spontaneous tumor lysis syndrome (TLS) with renal failure, and DIC with hemorrhagic or occasionally thrombotic complications (see Initial Therapy).

Findings after physical examination are often nonspecific. Patients may demonstrate pallor, tachycardia, oral purpura, ecchymoses, and petechiae due to anemia and thrombocytopenia. There may be signs of extramedullary leukemia such as gingival hypertrophy, enlargement of lymph nodes, cutaneous nodules (leukemia cutis), or hepatosplenomegaly. Examination of the lungs may reveal signs of an infectious process. Many patients with AML have no abnormal findings on physical examination at presentation.

The laboratory evaluation should include blood counts with evaluation of the blood smear, a standard chemistry panel (electrolytes, urea nitrogen, creatinine, total bilirubin, transaminases, uric acid, lactate dehydrogenase [LDH]), and coagulation studies, including prothrombin time (PT), partial thromboplastin time (PTT), and fibrinogen levels. Anemia and thrombocytopenia are almost always present. The white blood cell (WBC) count can vary from low to extremely high (>100,000/μL). Spurious hypoxia and hypoglycemia may occur in the setting of hyperleukocytosis due to continued metabolism by the leukemic blasts prior to processing of a blood sample for analysis; spurious hyperkalemia can occur in the setting of hyperleukocytosis due to cell lysis during clot formation in the preparation of serum. On the other hand, hypokalemia is not infrequently observed with monoblastic AML. DIC may be present at diagnosis and is often seen in patients with monocytic AML and APL ( Chapter 137 ). Subclinical DIC, evident by elevated D-dimers, is common in many forms of AML and can worsen with the institution of therapy. Abnormalities of renal and hepatic values may represent infiltration of these organs; renal dysfunction is commonly observed at presentation due to spontaneous TLS or hypoperfusion due to capillary leak syndrome.

Imaging studies are of little help in diagnosis but allow assessment of complications (pneumonia, cerebral hemorrhage). If patients present with any neurologic deficit, the threshold for computed tomography (CT) scan (non-contrast if bleeding is of concern) or any other imaging modality of the brain should be low. Further evaluations should be based on the clinical assessment of the patients.

Diagnosis and Classification

The diagnosis of AML relies on morphologic evaluation (cytochemical stains), immunophenotyping by flow cytometry, and assessment of karyotype and molecular studies ( Fig. 60.1 ). Blast percentage is best determined by a 500-cell differential of the marrow aspirate. The presence of ≥20% myeloid blasts in the peripheral blood or bone marrow is required for the diagnosis of AML, except in AML with t(8;21), inv(16), or t(15;17), in which case, these specific cytogenetic changes/gene fusions are sufficient for the diagnosis of AML regardless of the blast percentage. Three broad types of myeloblasts are described based on the granular content and nuclear features of the blasts (type 1: agranular basophilic cytoplasm, nucleus with fine chromatin, and two to four distinct nucleoli; type 2: basophilic cytoplasm with 20 or fewer azurophilic granules and similar nuclear features as type 1 blasts; type 3: basophilic cytoplasm with more than 20 azurophilic granules), although the morphologic variety of blasts exceeds the defined categories ( Fig. 60.2 ). Auer rods are linear aggregates of primary granules composed predominantly of myeloperoxidase (MPO) that are found in 30% to 50% of newly diagnosed patients with AML and, if present, are one of the hallmark morphologic features to establish a diagnosis of AML (see Fig. 60.2C ). Malignant promyelocytes have moderately basophilic cytoplasm with numerous azurophilic granules, frequent Auer rods including cells with multiple Auer rods (“faggot cells”), and bilobed or sliding-plate nuclei. Monoblasts and promonocytes usually exhibit folded or convoluted nuclei and may contain prominent vacuoles. Promyelocytes, promonocytes, and atypical pronormoblasts are all considered blast equivalents. Micromegakaryocytes and pronormoblasts are not considered blasts (see Fig. 60.2K and L ). AML marrows are typically hypercellular with decreased or absent megakaryocytes. Exceptions are marrows of older patients or those with t-AML, which may be hypocellular with dysplastic changes of one or more hematopoietic lineages. Prominent dysplasia may suggest a previous diagnosis of MDS but can also be found in patients with de novo AML, where the prognostic significance of dysplastic changes is less clear. For example, multilineage dysplasia in natural killer (NK) AML with NPM1 or biallelic CEBPA mutations does not appear to affect the otherwise more favorable outcome of AML patients with these mutations. Cases with extensive fibrosis may be due to a preceding myeloproliferative neoplasm or acute megakaryocytic leukemia.

Several cytochemical reactions further highlight morphologic characteristics (MPO, periodic acid–Schiff, Sudan black B, naphthol AS-D chloroacetate esterase [specific esterase], α-naphthyl acetate/butyrate esterase [nonspecific esterase], acid phosphatase; Fig. 60.3 ). MPO is the most specific granulocytic marker, and MPO positivity in at least 3% of the blasts is consistent with a diagnosis of AML. On the other hand, lack of MPO staining does not rule out AML because it is often not present in AML with minimal differentiation, acute monoblastic leukemia, and acute megakaryocytic leukemia. Monoblastic leukemias are stained by nonspecific esterase. The diagnostic utility of cytochemical reactions is limited by the more rapid availability of immunophenotypic data by flow cytometry. Nonetheless, cytochemical stains of a touch imprint of the core biopsy can be very useful in the event of a “dry tap.” Likewise, immunohistochemical staining for CD34 and CD117 can be used to estimate the blast percentage in a bone marrow core biopsy.

The first systematic attempt to classify AML was made by the French–American–British (FAB) Cooperative Group in 1976 and updated in 1985, and was based solely on morphology (blast percentage, degree of differentiation, lineage involvement). Because of its limited scope, the FAB system is now considered inadequate. Rapidly growing insights from genetic mutation analyses, their association with prognosis, and, in some cases, prediction of response to therapy triggered a revision of the old system and led to the changes in the 2008 edition of the World Health Organization (WHO) classification of AML and updated in 2016. The focus has shifted to identification of recurrent cytogenetic and molecular abnormalities, information regarding exposure to prior leukemogenic therapy, and morphologic features related to dysplasia-related changes as well as morphologic remnants of the FAB system ( Table 60.1 ). Several categories have been defined:

• 1.

  AML with recurrent genetic abnormalities. This includes AML with relatively common cytogenetic changes: AML with t(8;21)(q22;q22), RUNX1-RUNX1T1 ( Fig. 60.4 ); AML with inv(16)(p13.1q22) or t(16;16)(p13.1;q22), CBFB-MYH11 ( Fig. 60.5 ); AML with t(15;17)(q22;q12); PML-RARA ( Fig. 60.6 ); and AML with t(9;11)(p22;q23), MLLT3-KMT2A ( Fig. 60.7 ). Less common subtypes are AML with t(6;9)(p23;q34), DEK-NUP214 ; AML with inv(3)(q21q26.2) or t(3;3)(q21;q26.2), GATA1 , MECOM ; and AML with t(1;22) (p13;q13), RBM15-MKL1 ( Fig. 60.8 ). Normal karyotype AML with nucleophosmin 1 mutation ( NPM1 ; see Fig. 60.8G and H ), and biallelic CEBPA mutations (and in the absence of additional mutations) were initially considered as provisional, but are now incorporated in this subtype, with other molecular subsets expected to be added in subsequent editions.

  
  
  
  
  

• 2.

  AML with myelodysplasia-related changes (AML-MRC) is defined in one of three ways: (a) AML in patients with a preceding diagnosis of MDS or MDS/MPN; (b) AML with multilineage dysplasia recognized morphologically by the presence of dysplastic features in 50% or more of cells in two separate lineages ( Fig. 60.9 ); and (c) AML associated with a myelodysplasia-related cytogenetic abnormality. Patients with AML-MRC appear to have a worse outcome compared with AML-not otherwise specified (AML-NOS). However, the OS of patients treated with intensive chemotherapy for AML-MRC as defined solely by multilineage dysplasia is similar to that of patients with AML-NOS.

  

• 3.

  The category of t-AML was retained from the 2001 WHO classification, although the distinctions between AML associated with alkylating agents or radiation and those with topoisomerase II inhibitors were abandoned. Many patients receive complex therapeutic regimens, making this distinction difficult and impractical. Nonetheless, AML following exposure to topoisomerase II inhibitors typically have a shorter latency (less than 2 years) and frequently have monoblastic morphology with fusions of the KMT2A (previously MLL ) gene. CBF AML and APL have also been associated with topoisomerase II inhibitors exposure. AML following alkylating agents typically have a longer latency, are preceded by MDS, and have complex cytogenetic changes including monosomy and deletions of chromosomes 5 and/or 7 as well as TP53 mutations ( Chapter 59 ). Immunosuppressive therapies are now considered potentially leukemogenic.

• 4.

  NOS subtypes are AML cases that do not fit the categories already described and for the most part are classified using the FAB scheme (AML with minimal differentiation, AML without maturation, AML with maturation, acute myelomonocytic leukemia, acute monocytic or monoblastic leukemia, acute erythroleukemia, acute megakaryoblastic leukemia, acute basophilic leukemia). The AML-like disease of acute panmyelosis with myelofibrosis is included here ( Fig. 60.10 ).

  

• 5.

  Lastly, the classification now includes myeloid sarcoma ( Fig. 60.11 ), myeloid proliferations associated with Down syndrome, and the rather rare entity of blastic plasmacytoid dendritic cell neoplasm (BPDCN) ( Fig. 60.12 ) ( Chapter 64 ).

  
  

Prognosis

Prognosis in AML is dependent on patient characteristics and disease-related factors. The former determines the likelihood of surviving initial induction chemotherapy as well as eligibility for potentially curative post-remission therapy. Age and performance status are key determinants of early mortality as well as long-term outcome ( Chapter 157 ). Comorbidity scores such as the Charlson comorbidity index and the hematopoietic cell transplantation comorbidity index have been incorporated into predictive models of early mortality with intensive therapy ( Chapter 157 ). Comorbidity indices (ideally as part of a more global geriatric assessment strategy) are particularly useful in older patients where the question of intensive chemotherapy versus lower-intensity interventions assumes more significance. AML prognostic models often include both factors that determine early treatment-related mortality as well as OS, since the impact of these features on early death and long-term outcome are difficult to separate. More recently, the approval of drugs for AML patients unfit for intensive chemotherapy have used some of the criteria identified by Ferrara and colleagues to define this population, including age 75 years or older, performance status greater than 2, left ventricular ejection fraction ≤50%, congestive heart failure, pulmonary DLCO or FEV1 less than 65% predicted, and hepatic disease.

Disease-related factors determine resistance to therapy. Secondary AML, herein defined as AML following an antecedent hematologic neoplasm (MDS, MPN, MDS/MPN) or prior exposure to leukemogenic chemotherapy or ionizing radiation, has been associated with a worse prognosis. The response rates following intensive therapy are lower in patients with an antecedent hematologic neoplasm. Patients with therapy-related AML (t-AML) are typically older and have reduced survival due to other cancers or comorbid illnesses. The cumulative toxicity of prior therapy likely contributes to the higher rate of treatment-related mortality in t-AML patients. Regardless of age, patients with t-AML have lower remission rates, higher relapse rates, and increased early therapy-related mortality compared with de novo AML patients. The multi-drug resistance (MDR) phenotype due to expression of transmembrane drug efflux proteins negatively impacts outcomes of AML patients. Anthracyclines, etoposide, and calicheamicin (a component of Gemtuzumab Ozogamicin [GO]) are substrates for these drug efflux proteins. The MDR phenotype is commonly observed in secondary AML.

Among the disease-related factors, karyotype ( Chapter 57 ) has historically been considered as the most important prognostic factor dividing patients into those with favorable (CBF leukemias associated with t[8;21] and inv[16] or t[16;16]; APL with t[15;17]), unfavorable (complex abnormalities and monosomies, among others), and intermediate prognosis (mainly diploid; Fig. 60.13 ). Fluorescence in situ hybridization (FISH) can identify some of the more common AML gene deletions (deletions of chromosomes 5 and/or 7) and AML gene rearrangements in AML, including RUNX1/RUNX1T1 , CBFB/MYH11 , PML/RARA , and KMT2A fusions. FISH analysis can provide more rapid identification of these common changes and can help guide initial therapeutic choices. DNA based assays may replace classical cytogenetic and FISH analyses in the future.

Karyotype is supplemented by analyses of commonly mutated genes ( Chapter 59 ). Whereas prognosis is hardly or not at all augmented by gene mutation testing in cytogenetically unfavorable groups, the situation is different in the intermediate cytogenetic risk group. Identification of NPM1 mutation or biallelic CEBPA mutations in the absence of a FLT3 ITD mutation has been associated with a more favorable outcome than other patients with intermediate risk karyotype. Normal diploid karyotype, associated with intermediate risk, is the largest subset of both younger and older AML patients. Even in older, normal karyotype AML patients treated with intensive therapy, the presence of NPM1 or CEBPA mutations, and absence of the FLT3 ITD mutation, along with age, performance status, and WBC count, determine prognosis. On the other hand, even in the absence of an adverse risk karyotype, identification of the FLT3 ITD (especially high allelic burden), ASXL1, RUNX1, and TP53 variants are considered poor risk.

Core binding factor (CBF) leukemias are characterized by translocation t(8;21) and inv(16)/t(16;16) and, at the molecular level, by a disruption of the genes for the CBFα and CBFβ subunits, respectively. The CBF transcription factor plays a crucial role in hematopoietic differentiation. CBF AML comprises about 15% to 20% of patients with AML, but only 5% in AML patients over age 60 to 65 years. Although generally considered favorable risk, the prognosis of AML patients with CBF translocations is adversely affected by other clinical, cytogenetic, and mutational features. Poor prognostic factors in t(8;21) AML include high WBC count at diagnosis, presentation with granulocytic sarcoma, and expression of CD56. In general, cytogenetic abnormalities in addition to CBF translocations do not appear to influence outcome; these more complex karyotypes with CBF translocations are still considered favorable risk by ELN 2017 and NCCN risk stratification. However, del(9q) and loss of Y chromosome in t(8;21) AML have been associated with an inferior OS. The presence of a KIT tyrosine kinase activating mutation has been considered to have a negative impact on OS in CBF AML due to a higher rate of relapse. More recently, the prognostic impact of KIT has been shown to vary according to the variant allelic frequency as well as the concomitant mutations in t(8;21). Mutations in genes for chromatin modifying enzymes and cohesin complex proteins increase the risk of relapse in patients with t(8;21) and KIT mutations. The impact of KIT mutation is much less clear in inv(16) compared with t(8;21).

The first cancer genome to be completely sequenced was a cytogenetically normal AML. Whole-genome sequencing identifies a finite number of genes that are recurrently mutated in AML samples ( Chapter. 3, Chapter. 59 ). These repetitively mutated genes encode proteins/enzymes that generally fall into just a few classes including DNA methylation, epigenetic/chromatin modifiers, transcription factors, RNA splicing, activating signals, and the cohesin complex. These mutations are considered in risk assessment, but often with incongruous results (e.g., NPM1 , FLT3 , CEBPA , ASXL1 , RUNX1 , TP53, KIT , IDH1 , IDH2 , TET2 , DNMT3A ). Nonetheless, a comprehensive cytogenetic and mutational analysis of samples from 1540 AML patients treated with intensive chemotherapy identified 5234 driver mutations across 76 genes. At least one pathogenic mutation was found in 96% of the samples. Eleven prognostic groups were identified, allowing classification of 80% of patients into mutually exclusive prognostic subsets ( Fig. 60.14 ). The ELN currently recommends testing for mutations of NPM1 , CEBPA , FLT3, RUNX1, ASXL1, TP53 in all patients with a new presentation of AML. Evaluation for specific germline mutations associated with familial AML (e.g., RUNX1, CEBPA, TP53, BRCA1/2, GATA2, DDX41, DKC, TERC, TERT ) should be performed in those with a family history of AML or patients with physical stigmata of congenital marrow failure syndromes; this information may guide genetic counseling and donor selection for allo HSCT ( Chapter 30 ). The ELN has published a revised risk model based on integration of cytogenetic and gene mutation information ( Table 60.2 ). Mutations of the chromatin modifying proteins ( ASXL1 , EZH2 , BCOR ), spliceosome enzymes ( SRSF2 , U2AF1 , SF3B1 , ZRSR2 ), and the cohesin complex protein STAG2 are more commonly identified in AML evolving from MDS compared with de novo AML. In fact, patients with apparently de novo AML with one of these mutations have inferior response to and survival following intensive therapy compared with those who do not have this genotype.

Assessment of post-therapy response provides additional valuable information. Achievement of morphologic complete remission (CR) after one or two courses of induction therapy is associated with improvement in survival. The most powerful post-treatment prognostic factor is the degree of minimal (or measurable) residual disease (MRD). The deeper the response, the better the long-term outcome. MRD is defined as disease that is detected below the arbitrary definition of persistent AML, 5% blasts by morphology and/or immunohistochemical assays. It is estimated that patients in morphologic CR still harbor up to 10 ⁹ leukemic cells, which constitute an important reservoir for later recurrences. Assessing MRD has several advantages: (1) better determination of the quality of response, (2) more timely diagnosis of an impending relapse and possibly improvement of outcome by early intervention, and (3) potential indication to intensify post-remission therapy (allo HSCT) or not.

Persistence of karyotypic abnormalities by cytogenetic analysis at the time of morphologic CR is associated with increased relapse risk and inferior OS. However, the cytogenetic analysis only evaluates 20 metaphases. FISH analysis can detect lower levels of persistent karyotypic abnormalities. MRD, as detected by reverse-transcriptase quantitative polymerase chain reaction (RT-qPCR) or multiparameter flow cytometry (MFC), is even more sensitive. RT-PCR requires a suitable molecular target such as fusion transcripts (e.g., PML-RARA , RUNX1-RUNX1T1 , CBFB-MYH11 , DEK-CAN, BCR-ABL ), NPM1 insertion mutations, or overexpressed genes (e.g., WT1 ). Persistence of these transcripts following intensive chemotherapy has been associated with higher risk of relapse and inferior survival. For example, detection of the clone-specific NPM1 mutant transcript in the peripheral blood by RT-PCR following 2 cycles of intensive chemotherapy was associated with increased risk of relapse at 3 years (82% if MRD positive vs. 30% if MRD negative) and inferior survival at 3 years (24% if MRD positive vs. 75% if MRD negative). An expert panel has recommended the use of RT-PCR for MRD assessment at informative clinical timepoints only in AML patients with PML-RARA , RUNX1-RUNX1T1 , CBFB-MYH11 , and NPM1 insertion mutations.

Since only the minority of AML patients harbor a leukemia-specific marker that can be identified by RT-PCR, MFC is more widely applicable to the assessment of MRD in AML patients. MRD is based on the identification of a leukemia-associated immunophenotype, which can be detected in the majority of patients with AML in various ways: (a) “different-from normal” antigen expression in leukemic blasts; (b) detection of lineage-foreign markers; (c) asynchronous expression of markers; (d) altered density of surface antigens. Commonly, a level of 0.1% has been determined to distinguish patients with MRD from those without. MRD levels following induction or early in consolidation have been associated with relapse and survival in all age groups regardless of cytogenetic and mutational status. Even though MRD levels tend to be higher in patients with unfavorable pretreatment characteristics, the prognostic impact of MRD extends to all risk groups and has shown to improve standard risk classification independently of cytogenetic–molecular markers.

Despite the clear prognostic significance of persistent MRD even in the setting of CR, there are many barriers to fully adopting MRD assessments into the routine care of AML patients. The list of these obstacles includes lack of assay standardization, absence of a universal definition of most appropriate cut-point for reporting positive MRD, choice of peripheral blood versus bone marrow for MRD analysis, quality of the bone marrow aspirate sample, timing of the assay following therapy, and variation in the kinetics of MRD clearance. The persistence of specific mutations by NGS may not actually affect prognosis. For example, persistence of mutations in DNMT3A , TET2 , and ASXL1 following intensive chemotherapy has not been shown to affect prognosis. Intensification of therapy by means of allo HSCT is associated with the lowest risk of relapse in AML patients with MRD persistence after initial therapy. However, even with a myeloablative preparative regimen, risk of relapse and OS are negatively impacted by persistence of disease by MFC at the time of the allo. Reduced intensity preparative regimens seem to provide little protection from relapse in patients with genomic evidence of persistent AML using highly sensitive duplex DNA sequencing. Effective strategies to eradicate MRD following intensive chemotherapy are not currently available, but novel agents are being assessed for this purpose (e.g., antibody-drug conjugates, Bispecific T-cell engagers [BiTE], targeted therapies, and cellular therapy).

Therapeutic Considerations at Initial Presentation

Leukocytosis with 100,000 WBC/μL or higher is considered an emergency and requires immediate efforts to reduce the disease burden. Hyperleukocytosis may result in leukostasis. The clinical features of leukostasis are due to the rheologic effects of the large blasts in the microvasculature as well as the presence of endothelial adhesion molecules on the leukemic blasts. Leukostasis is a multisystem syndrome due to compromise of blood flow in the lungs (dyspnea, hypoxemia, diffuse alveolar hemorrhage), CNS (encephalopathy, coma, tinnitus, dizziness, visual disturbance, retinal hemorrhages), and heart (myocardial ischemia). Patients with leukostasis require urgent cytoreduction with chemotherapy (immediate start of intensive induction therapy, high-dose hydroxyurea, single dose of cytarabine) and/or leukapheresis. Retrospective studies have not confirmed any survival benefit with leukapheresis compared with chemotherapy alone. Furthermore, leukapheresis is contraindicated in the setting of significant DIC and hypotension; extracorporeal leukapheresis can be associated with respiratory failure due to cytokine release and exacerbation of pulmonary capillary leak. Leukapheresis also requires placement of large-bore central venous catheters which may result in traumatic hemorrhage.

Other complications of AML requiring urgent intervention are TLS and DIC. Either may be present at the time of diagnosis, and both may worsen as therapy is started. Therefore, frequent laboratory assessments (2 to 4 times daily) may be needed, especially as therapy is initiated in patients with high tumor burden, e.g., hyperleukocytosis. DIC is managed with aggressive transfusion of cryoprecipitate, plasma, and platelet transfusion based on close monitoring of the fibrinogen, PT/PTT, and platelet count, respectively ( Chapter 137 ). The use of heparin and/or anti-fibrinolytic agents have not been shown to decrease the risk of serious hemorrhagic complications associated with DIC. However, prospective evaluation of these interventions versus transfusion support alone is not available. TLS should be anticipated, and prophylactic therapy with xanthine oxidase inhibitors (allopurinol, febuxostat) and intravenous hydration are recommended with initial therapy. Urine alkalinization to increase urinary excretion of uric acid is not recommended, since alkalinization may exacerbate metastatic calcinosis due to hyperphosphatemia. Instead, hyperuricemia, especially if accompanied by poor renal function, can be corrected rapidly with intravenous rasburicase. Rasburicase metabolized uric acid to the more soluble allantoin. However, patients with G6PD deficiency are at risk of fulminant oxidative hemolysis with rasburicase due to the concomitant production of hydrogen peroxide. Phosphate binders are given for the hyperphosphatemia. Insulin with glucose, sodium bicarbonate, and sodium polystyrene sulfonate in sorbitol (a potassium exchange resin) are used to control life-threatening hyperkalemia. Patiromer may also be used but has not been evaluated for hyperkalemia associated with TLS. Finally, urgent renal replacement therapy (dialysis) should be considered for management of acute renal failure due to TLS that is not adequately controlled by the management outlined above.

Initial Therapy: Intensive Versus Less Intensive Treatment Algorithm

Until 2017, the general therapeutic algorithm following the diagnosis of AML had not changed substantially in the preceding four decades. The physician first assessed the patient’s ability to tolerate intensive chemotherapy approaches, which may ultimately include allo HSCT. Therapeutic options for “unfit” AML patients were quite limited and included DNA methyltransferase inhibitors (also known as DNA hypomethylating agents [HMA]) such as azacitidine and decitabine, Low-Dose Cytarabine (LoDAC), and Best Supportive Care (BSC) alone. Cytarabine 20 mg by subcutaneous injection twice daily for 10 days every 4 weeks (LoDAC) was compared with hydroxyurea and BSC (with or without all trans-retinoic acid) in older, unfit AML patients. Both the median survival and CR rate (18% vs. 1%; P < .001) were superior with LoDAC, except for those patients with adverse cytogenetics, who did poorly on either treatment arm. The survival benefit was only observed in those achieving remission.

Azacitidine 75 mg/m ² /day for 7 days improved the survival of AML patients with 20% to 30% marrow blasts compared with conventional care regimens (BSC, LoDAC, and intensive chemotherapy) in a phase III study enrolling high-risk, high-grade MDS patients (median OS 24.5 vs. 16 months, P = .005). Subsequently, azacitidine was compared with conventional care regimens (predominantly, LoDAC) in AML patients ages 60 years and older. There was a trend for improvement in the median survival with azacitidine (10.4 vs. 6.5 months, P = .1009); however, subsequent therapies confounded the assessment of any survival benefit. The rate of CR was similar with both azacitidine and LoDAC (19.5% vs. 21.9%). Decitabine 20 mg/m ² IV daily for 5 days every 4 weeks was evaluated in a phase II study of treatment-naïve older AML patients and subsequently compared to either LoDAC (in most subjects) or BSC alone in a large randomized multicenter study. The decitabine arm achieved a higher CR rate (17.8% vs. 7.8%; P = .001). In a post hoc sensitivity analysis of the mature data a survival advantage was observed with decitabine at fixed time points over 2 years, and median survival improved from 5.0 to 7.7 months. A phase II study suggested that an extended schedule of decitabine 20 mg/m ² /day for 10 days may result in better response (CR 47%) and survival (median OS 55 weeks). However, a randomized phase II study of 5 versus 10 days did not detect any benefit in terms of CR/CRi rate (43% vs. 40%), median relapse-free survival (RFS) (5.7 vs. 4.6 months), median OS (5.5 vs. 6.0 months), or 1-year survival (25% both arms). With both azacitidine and decitabine, several monthly cycles may be required to achieve the response. Although some analyses suggested a higher response rate in MDS patients with TET2 mutations, there are no clear predictors of response to HMA therapy in AML patients.

In AML patients over the age of 65 years, there has even been conflicting data regarding the benefit of intensive versus less intensive chemotherapy from randomized trials conducted in the 1980s. A randomized trial conducted by the European Organization for Research and Treatment of Cancer (EORTC) compared immediate induction therapy with daunorubicin, cytarabine, and vincristine with supportive care until disease progression (leukocytosis, thrombocytopenia, pain) and then therapy with hydroxyurea 3 g on days 1 and 4 with cytarabine 100 mg/m ² twice daily on day 2 to 3 and 5 to 6. The remission rate was higher (58% vs. 0%), median survival longer (21 vs. 1 weeks), and survival at 2.5 years higher (17% vs. 0%) with the immediate induction therapy; early mortality was higher with the more palliative approach. On the other hand, another randomized trial comparing an anthracycline and cytarabine induction therapy with LoDAC demonstrated higher CR rate with intensive therapy, but lower early mortality with LoDAC, resulting in no significant difference in the OS. In a retrospective analysis of AML patients older than 65 years treated at the M.D. Anderson Cancer Center from 2000 to 2010, OS was similar regardless of initial therapy, either intensive chemotherapy or epigenetic-directed therapy. Given the low response rate with less intensive therapy, greater toxicity with intensive therapy, and minimal (if any) survival benefit, it is not surprising that the majority of AML patients age ≥65 years chose not to receive any form of chemotherapy for AML. In an analysis of SEER and Medicare data between 2000 and 2007, only 38.6% of U.S. Medicare recipients with AML received any form of chemotherapy. Increasing age, higher Charlson Comorbidity Index, and prior MDS diagnosis were associated with a lower rate of treatment. However, those who were selected for therapy had a longer median survival than those who were not (6 vs. 2 months).

With the recent availability of more effective, less intensive treatment regimens (see section Less Intensive Therapy), more older AML patients are now eligible for therapy. On the other hand, younger, fit AML patients may potentially benefit from less intensive therapies that have demonstrated efficacy in AML subtypes typically associated with chemotherapy resistance. Conversely, AML patients may become candidates for allo HSCT in first remission even if initially treated with less intensive therapy. The current AML treatment algorithm is summarized in Fig. 60.15 .

When selecting initial therapy now, the clinician needs to consider a more global assessment of fitness (performance status, comorbidities, geriatric assessment scores), disease biology, as well as the goal of therapy (see box on Selection of Initial Therapy for the Patient With Acute Myeloid Leukemia ). Age alone is considered inadequate for determination of initial and subsequent therapy. On the other hand, comorbidities are frequent in older patients, who are more likely to have a poor performance status. Hence, tolerance to the myelosuppressive and immunosuppressive consequences of intensive chemotherapy is diminished and there is greater risk of early therapy-related mortality. There are also differences intrinsic to the blast biology between older and younger patients. Patients over 60 years of age are more likely to have secondary AML, demonstrate unfavorable cytogenetics, and more often express multidrug-resistant phenotypes. Nevertheless, a select group of patients may still benefit from more intensive interventions. Older patients are not a homogeneous group and several models have been devised to identify variables that predict which patients may do well with intensive therapy versus those who will not. These models are comprised of patient characteristics, easily accessible clinical variables, and tumor characteristics. Efforts are therefore underway to capture multiple patient characteristics (physical and cognitive function, nutritional status, comorbidity, psychological state, and social support) as part of a comprehensive geriatric assessment and to better distinguish patients into fit, vulnerable, and frail (indicating a significantly increased risk of treatment complications). Despite thorough discussions with their oncologist regarding the goals of therapy, patients often misunderstand terms, e.g., equating remission with cure; patients may overestimate the benefit of therapy in terms of long-term survival and cure.

Selection of Initial Therapy for the Patient With Acute Myeloid Leukemia

The selection of either an intensive or less intensive chemotherapeutic regimen for AML patients remains a matter of significant debate and part of the art of medicine. The choice has been made more difficult by the recent approval of less intensive regimens with high initial response rates, even in patients with adverse AML biology. Fitness for intensive chemotherapy has never been adequately defined. I weigh multiple factors that will determine tolerance to chemotherapy including age, functional or performance status, comorbid illnesses, organ dysfunction, tobacco use, and nutritional status. However, I no longer consider only fitness; I also consider whether intensive chemotherapy is appropriate. The disease biology is critical to decision making. AML with adverse features based on cytogenetic and mutational analysis has a low response rate with intensive chemotherapy, and the only potentially curative option will ultimately be allogeneic hematopoietic stem cell transplantation once a remission is achieved.

For patients appropriate for intensive therapy, I base my choice of initial therapy on the following features: documented history of antecedent myeloid neoplasm, prior leukemogenic therapy for another disease, presence of a FLT3 ITD or TKD mutation, presence of RUNX1/RUNX1T1 or CBFB/MYH11 fusion, and presence of the PML/RARA fusion. I have partnered with our pathology colleagues to expedite PCR for the FLT3 ITD and TKD mutations and FISH analysis for PML/RARA , RUNX1/RUNX1T1 , CBFB/MYH11 , as well as deletions of chromosomes 5q, 7q, and 17p. These results are available within 72 hours. If there are 20% circulating myeloblasts, the diagnosis of AML is confirmed, and these assays can be done on a peripheral blood sample without waiting for a bone marrow biopsy to be performed.

If there is any suggestion of APL at initial presentation, I start ATRA immediately, before the diagnosis is confirmed. A myeloid immunophenotype lacking the stem cell markers CD34 and HLA-DR, blasts with bilobed nuclei without cytoplasmic granules, or even leukopenia with DIC, may suggest the diagnosis of APL. The ATRA can be discontinued, after the diagnosis is excluded. If APL is confirmed, I treat with ATRA and ATO. I add GO 9 mg/m ² , if the WBC count is over 10,000/μL. If the patient has CBF AML, I treat with cytarabine and daunorubicin (7+3) with GO 3 mg/m ² . If the FLT3 ITD or TKD is present, I treat with daunorubicin 60 mg/m ² /day for 3 days and cytarabine 200 mg/m ² /day for 7 days with midostaurin 50 mg twice daily on days 8–21. Finally, if the patient has t-AML or AML following MDS or MDS/MPN, I treat with liposomal daunorubicin/cytarabine (CPX-351). If the patient apparently has a de novo AML but with multiple abnormalities on the FISH panel including del(5q) and del(7q), I will also treat with CPX-351. However, I will also consider a HMA and venetoclax (VEN) for these patients, especially if there is del(17p), and the patient had not been previously treated with an HMA. If none of these markers are present, I often select 7+3 with idarubicin 12 mg/m ² /day for 3 days, to avoid the controversy surrounding the optimal dose of daunorubicin (90 vs. 60 mg/m ² ).

For patients who are truly unfit for intensive chemotherapy, or decline intensive chemotherapy, I treat with azacitidine 75 mg/m ² /day for 7 days plus VEN once daily for up to 28 days. I do not wait for any cytogenetic or mutational data since this information would not change the decision to use AZA/VEN in this approved population. Although certain subsets of patients did not have a survival benefit with AZA/VEN compared with AZA/placebo (e.g., FLT3 ITD- or TP53 -mutated), the initial response rate is over 50%. The one possible exception are those patients with IDH1 - or IDH2 -mutated AML. If these patients had previously received an HMA for MDS, or if they decline any form of parenteral therapy, ivosidenib and enasidenib would be acceptable alternatives for IDH1 - and IDH2 -mutated AML, respectively. Therefore, our pathologists also provide the IDH mutation status within 72 hours as well.

I will treat patients with HMA/VEN as an outpatient if they live close to our center, have a caregiver at home, have normal renal function, do not have leukocytosis, and do not already have complications from active AML such as infection. Otherwise, I will admit patients for the first week of therapy. I generally give posaconazole to all patients during initial induction due to the high rates of invasive fungal infection seen in our population. I give venetoclax 100 mg once daily in combination with posaconazole. If there is clearance of the peripheral blood blasts by day 21, I repeat a bone marrow biopsy at that time. If morphologic leukemia-free state (MLFS) or marrow aplasia has been achieved, I hold venetoclax at that time and allow for neutrophil recovery. In that case, I only prescribe 21 days of venetoclax during cycle #2. If the patient has not achieved MLFS, I proceed with cycle #2 as planned at day 29. Generally, remission is achieved with the first 2 cycles, but I will continue AZA/VEN if there appears to be clinical benefit even if CR or CRi has not been achieved.

Of course, I consider every patient for a clinical trial if available and appropriate.

The initial goal of either intensive or less intensive therapy is the prolongation of survival through the achievement of a CR. Following CR, further post-remission therapy is required to maintain remission by eliminating or reducing residual disease. CR is defined as achievement of less than 5% marrow blasts (even if these are known to be leukemic blasts by immunophenotypic, cytogenetic, or mutational analyses), absence of Auer rods, and clearance of the peripheral blood and all extramedullary deposits of leukemic blasts. CR also requires restitution of relatively normal marrow function, indicated by an absolute neutrophil count (ANC) greater than 1000/μL, a platelet count greater than 100,000/μL, and independence of transfusion support. CR defines a landmark point in time because patients who achieve CR after beginning therapy have longer survival than patients who are resistant to therapy. Lesser response criteria (CR with incomplete platelet recovery [CRp], CR with incomplete recovery of neutrophils or platelets [CRi], CR with partial hematologic recovery [CRh], partial remission [PR], and morphologic leukemia-free state [MLFS]) may serve as useful end points in the context of early phase clinical trials, but they do not necessarily carry the same significance for survival as does CR. Each of these (except CRh) are defined by International Working Group and ELN criteria. The definition of CRh is identical to that of CR, except that the ANC must only be greater than 500/μL and the platelet count only greater than 50,000/μL.

After achievement of an initial response, the continued therapeutic path is different for patients treated with intensive versus less intensive therapy. Early clinical trials demonstrated the need to provide some post-remission therapy following achievement of CR with intensive chemotherapy. For example, in an ECOG trial AML patients ≤ age 65 years in first CR were assigned to allo HSCT (if family member donor available and age ≤40 years), or randomly assigned to observation, 2 years of oral mercaptopurine (MP) maintenance, or 1 cycle of high-dose cytarabine. The observation arm was closed early after 28 subjects assigned to this arm were noted to have inferior remission duration compared with continuing maintenance therapy. Post-remission therapy following intensive remission induction therapy consists of chemotherapy (typically 1 to 4 cycles), and/or HSCT, and/or maintenance therapy. Patients may receive one of these post-remission treatments or all of them, depending on initial risk stratification, eligibility for allo HSCT over time, as well as MRD assessments. For patients receiving less intensive regimens, the same therapy is continued in the absence of disease progression, disease relapse, or unacceptable toxicity (discussed later). However, even patients achieving CR after less intensive therapy may become candidates for potentially curative allo HSCT.

Intensive Therapy: Induction

Since first published by Yates and colleagues in 1973, intensive AML therapy is still built on the same two drugs as 50 years ago: cytarabine (also known as cytosine arabinoside or Ara-C) and an anthracycline. Cytarabine, an analogue of a physiologic pyrimidine nucleoside, is a cell cycle-specific antimetabolite that requires intracellular conversion to its triphosphate compound, ara-CTP, and incorporation into DNA to become active ( Chapter 58 ). When given by itself at standard doses of 100 to 200 mg/m ² intravenously (IV) daily for 5 to 7 days, it produces CR rates of around 40%. The anthracyclines, daunorubicin and idarubicin, and the anthraquinone, mitoxantrone, have been the most extensively evaluated agents in combination with cytarabine. The anthracycline doxorubicin was associated with more mucositis and gastrointestinal toxicity compared with daunorubicin and is not used in AML therapy. These agents intercalate into DNA and stabilize the complex between DNA and the enzyme topoisomerase II, thereby leading to apoptotic cell death ( Chapter 58 ). Daunorubicin achieves similar CR rates as does cytarabine alone. The combination of cytarabine 100 to 200 mg/m ² as a continuous IV infusion (its serum half-life is only 15 minutes) daily on days 1 to 7 and daunorubicin 45 to 90 mg/m ² IV daily on days 1 to 3 has become known as the “7+3” regimen and, for more than 40 years, has been the standard induction combination for AML patients selected for intensive therapy. Remission rates range from 50% to 80% (depending on patient age and cytogenetic risk), and long-term disease-free survival (DFS) is 35% to 40%.

Early clearance of the peripheral blood blasts (e.g., less than 3 days by flow cytometry) is associated with superior long-term survival. In clinical practice, patients undergo a repeat bone marrow examination between 14 and 21 days from the start of treatment ( Fig. 60.16 ). If the marrow continues to show blasts and is cellular, a reinduction is usually given. However, there is little consensus on what degree of persistent disease at day 14 to 21 can be tolerated. It is not uncommon for a patient to be too ill 2 weeks after intensive therapy to receive re-induction therapy for persistent disease, and yet ultimately achieve remission without further therapy. Clinical trials often recommend re-induction, if there are more than 10% blasts, and the marrow cellularity is ≥15%. The reinduction may be an attenuated repetition of the induction (“5+2”) or an intensification with intermediate-dose cytarabine (IDAC) or high-dose cytarabine (HiDAC) or fludarabine, HiDAC and granulocyte colony stimulating factor (G-CSF) (FLAG). There has been debate regarding the prognostic significance of requiring 2 versus 1 cycle of induction therapy to achieve first CR. Earlier studies reported a negative impact on RFS whereas others showed no impact on outcome. However, younger AML patients treated on SWOG S0106 and S1203 studies with 7+3 containing daunorubicin 60 and 90 mg/m ² , respectively, the requirement of 2 induction cycles to achieve remission had a significant negative impact on RFS and OS.

On the other hand, if the day 14 or 21 marrow is hypoplastic and appropriately ablated, supportive care continues until bone marrow recovery. If blood counts are not recovering, the bone marrow studies are repeated every 2 weeks to determine if there is recurrent AML causing the pancytopenia or delayed marrow recovery. If in CR, post-remission therapy starts shortly thereafter, once the patient has recovered from the toxicities of induction therapy. The degree of neutrophil and platelet recovery at the time of remission has prognostic significance. Higher neutrophil and platelet count at the time of remission are predictive of better RFS. In some cases, a regenerating marrow may have an increased number of blasts, which may look like persistent leukemia. Immunophenotyping by flow cytometry often helps to make that distinction. If due to regeneration, then follow-up marrow studies will show reduction in blasts concomitant with a rise of neutrophils and platelets. In case of no response to one or two courses of induction therapy (“primary refractory”), treatment is often changed.

Many modifications to the 7+3 regimen have been evaluated over the years. Until recently, modifications of 7+3 have not improved outcomes for AML patients (see below). Two different standard cytarabine dose regimens were compared: 100 versus 200 mg/m ² continuous IV infusion once daily days 1 to 7 with daunorubicin. The rates of remission, median DFS, median OS, and 5-year survival were not statistically different. However, there was a higher mortality rate related to chemotherapy with 200 mg/m ² (21% vs. 13%, P = .05). Likewise, 10 days instead of 7 days of cytarabine (10+3) did not improve outcomes. Other cytotoxic agents (e.g., vincristine, etoposide, thioguanine, topotecan, fludarabine) have been added to 7+3 without proven benefit, but often with increased toxicity. Inhibitors of multi-drug efflux protein MDR1 (also known as P-glycoprotein or P-gp) could potentiate the cytotoxicity of P-gp substrates, such as the anthracyclines and etoposide, by allowing intracellular accumulation of these drugs. However, P-gp inhibitors also affect clearance of these drugs. Therefore, the doses of anthracyclines and etoposide needed to be reduced due to increased toxicity in the presence of the P-gp inhibitor. Neither the response rates nor OS of AML patients was improved by the addition of the P-gp inhibitors, valspodar or zosuquidar, to induction chemotherapy in phase III randomized trials. The B-cell lymphoma/leukemia-2 (BCL2) anti-sense oligonucleotide, oblimersen, did not improve OS when administered with 7+3 induction therapy. Since cytarabine is a cell-cycle specific cytotoxic agent, priming with hematopoietic growth factors (G-CSF and granulocyte-macrophage colony stimulating factor [GM-CSF]) has been performed during the week of chemotherapy to force leukemic blasts into S phase. However, priming has not been shown to improve OS. The greatest risk of intensive chemotherapy is morbidity and mortality due to neutropenic infection. The addition of either G-CSF or GM-CSF following chemotherapy shortens the duration of absolute neutropenia by a few days, but did not improve CR rate, early mortality, or OS.

The intracellular concentration of ara-CTP is increased by culturing leukemic cell lines in the presence of both a purine analog and cytarabine, due to increased expression of cytidine kinase. Based on this pharmacokinetic interaction and other effects of the purine analogs, many studies have evaluated the addition of purine analogs to cytarabine-based chemotherapy. The Polish Adult Leukemia Group published the outcome of an open-label, randomized study where addition of cladribine 5 mg/m ² once daily days 1 to 5, but not fludarabine, to a daunorubicin 60 mg/m ² for three days and cytarabine 200 mg/m ² for 7 days resulted in significantly increased rates for remission (68% vs. 56%; P =.01) and 3-year OS (OS; 45% vs. 33%; P =.02). Interestingly, leukemia-free survival was not improved by the addition of cladribine. The survival benefit was less significant when subjects were censored at the time of allo HSCT. The addition of the second-generation purine analog, clofarabine at 20 mg/m ² /day for five days, to idarubicin 10 mg/m ² /day on days 1 to 3 and cytarabine 1 g/m ² /day on days 1 to 5 (IA) led to significantly better event-free survival (EFS) and OS in the three-drug combination when compared with a historical IA-treated group, especially in patients younger than 40 years of age. The effect of ludarabinee on cytarabine accumulation in leukemic blasts was also demonstrated in patients with relapsed/refractory AML receiving high-dose cytarabine salvage therapy with or without fludarabine; however, there was no apparent clinical benefit in this phase III study. The use of fludarabine with idarubicin and high-dose cytarabine in the United Kingdom Medical Research Council (UK MRC) trials is discussed later.

The dose of daunorubicin has been the subject of debate. A study by the Eastern Cooperative Oncology Group (ECOG 1900) randomized patients between the ages of 17 and 60 years with untreated AML to a 7+3 combination with either standard-dose daunorubicin (45 mg/m ² daily for 3 days) or high-dose daunorubicin (90 mg/m ² daily for 3 days). The higher dose of daunorubicin achieved higher CR rates (70.6% vs. 57.3%, P < .001) and improved the median OS (23.7 vs. 15.7 months, P = .003). This improvement was limited to patients younger than 50 years and with intermediate-risk cytogenetics. The daunorubicin dose of 90 mg/m ² /day did not lead to a higher incidence of adverse events (particularly cardiomyopathy and infectious complications). For patients with FLT3 ITD mutations, CR rates (70% vs. 48%), median survival (15.2 vs. 10.1 months), and 4-year survival estimates (28% vs. 17%) were also superior with 90 versus 45 mg/m ² . The study can be criticized on the basis that the dose of the comparator arm, daunorubicin 45 mg/m ² /dose, is no longer considered the standard of care for younger AML patients, and instead 60 mg/m ² /dose should be considered standard. In older AML patients ages 60 to 83 years (median 67 years), daunorubicin 90 mg/m ² /day for 3 days was compared with the standard dose for this population, 45 mg/m ² /day for 3 days both in combination with cytarabine 200 mg/m ² /day for 7 days. The CR rate was greater with the high-dose daunorubicin (64% vs. 54%, P = .002) without an increase in induction mortality. There was no difference in EFS or OS for the entire population. In the patients ages 60 to 65 years, 2-year survival was superior with high-dose daunorubicin (38% vs. 23%). The UK MRC AML 17 trial compared daunorubicin 60 versus 90 mg/m ² during induction and showed equivalent CR and relapse rates as well as survival. However, in the patients with FLT3 ITD mutation, the 90 mg/m ² daunorubicin resulted in higher RFS and OS. The UK MRC AML 17 trial did not settle the debate, because patients received a second cycle of daunorubicin-containing chemotherapy. The cumulative dose of daunorubicin was higher in both arms of the UK MRC AML17 trial compared with the two arms of the ECOG 1900 study.

Idarubicin is a 4-demethoxy anthracycline analogue of daunorubicin, which results in increased lipophilicity and better cellular uptake compared with daunorubicin. Idarubicin, and its metabolite idarubicinol, are not substrates for the multidrug resistance efflux pump. Idarubicin 12 to 13 mg/m ² for 3 days appeared to be superior to daunorubicin 45 to 50 mg/m ² for 3 days in several phase III studies. However, the doses did not appear to be biologically equivalent, since idarubicin was associated with greater myelosuppression in those studies continuing the comparison into the post-remission setting. In younger adults with treatment naïve AML receiving cytarabine 200 mg/m ² for 7 days with either daunorubicin 90 mg/m ² /day (AD) for 3 days or idarubicin 12 mg/m ² /day for 3 days (AI), there was no difference in CR rates (AD 74.5% vs. AI 80.5%), cumulative incidence of relapse (AD 25.1% vs. AI 35.2%), and 4-year survival (AD 54.7% vs. AI 51.1%). However, EFS and OS were significantly better with IA in patients with FLT3 ITD positive AML.

A meta-analysis of 29 randomized controlled trials compared the efficacy of different anthracyclines and dosing schedules during induction therapy for AML patients less than age 60 years. Idarubicin compared with daunorubicin improved remission rates, although this effect was limited to studies with a daunorubicin/idarubicin ratio of less than 5. Likewise, higher-dose compared with lower-dose daunorubicin improved remission rates. Survival estimates suggest that both high-dose daunorubicin (90 mg/m ² /dose × 3 days or 50 mg/m ² /dose × 5 days) and idarubicin (12 mg/m ² /dose × 3) can achieve 5-year survival rates of between 40% and 50%. On the other hand, there was no difference in CR rates or OS in AML patients older than age 55 years treated with cytarabine 100 mg/m ² /day for 7 days with either idarubicin 12 mg/m ² /day, daunorubicin 45 mg/m ² /day, or mitoxantrone 12 mg/m ² /day for 3 days. However, the induction mortality was greater in the idarubicin treatment arm.

A meta-analysis of 3 randomized clinical trials using HiDAC during induction encompassed 1691 patients and arrived at the following conclusions: there were no differences between HiDAC and standard-dose cytarabine (SDAC) with respect to percent CR and rates of persisting leukemia and early death. On the other hand, 4-year OS and RFS were significantly better with HiDAC but at the cost of more toxicities (infections, nausea/vomiting, CNS). HiDAC benefited mostly patients who already were more likely to achieve CR (i.e., patients with intermediate and favorable karyotype). The major drawback of HiDAC during induction is that it may make further consolidation therapy difficult. It is also not clear if HiDAC should be used during initial induction or is best used as a post-remission therapy after marrow function has been restored. The UK MRC AML 15 trial randomized 1268 treatment-naïve younger AML patients to receive fludarabine, HiDAC, G-CSF, and idarubicin (FLAG-Ida; fludarabine 30 mg/m ² IV on days 2 to 6, cytarabine 2 g/m ² IV on days 2 to 6, idarubicin 10 mg/m ² IV on days 4 to 6, and G-CSF on days 1 to 7) and 1983 to receive cytarabine (ara-C), daunorubicin, and etoposide (ADE; daunorubicin 50 mg/m ² IV on days 1, 3, 5; cytarabine 100 mg/m ² IV on days 1 to 10; and etoposide 100 mg/m ² IV on days 1 to 5), the latter representing the SDAC arm. FLAG-Ida resulted in higher response rates following the first induction and fewer patients on FLAG-Ida relapsed. However, more myelosuppression and deaths in CR offset the survival benefit except in the FLAG-Ida treated patients who were able to receive four courses (FLAG-Ida × 2, HiDAC × 2) where 8-year survival was 63% versus 47% with ADE with intermediate-risk patients and 95% for those with favorable risk.

The EORTC and the Italian Cooperative Group for Hematologic Diseases in Adults (GIMEMA) AML-12 trial randomized 1942 newly diagnosed AML patients between ages 15 and 60 years to an induction with SDAC or high-dose cytarabine (3 g/m ² every 12 hours on days 1, 3, 5, and 7). The study demonstrated significantly higher CR rates in all patients and significantly improved EFS (43.6% vs. 35.1%; P =.003) and OS (51.9% vs. 43.3%; P =.009) for patients between 15 and 45 years of age. The impact on survival was particularly evident in younger patients with secondary AML, high-risk cytogenetics, and FLT3 mutations, emphasizing that HiDAC may overcome poor-risk features during induction that SDAC cannot.

Increasing the efficacy of induction therapy has been attempted with the use of additional cytotoxic chemotherapy shortly after completion of the first course and prior to recovery of normal marrow function. This approach is known as timed-sequential chemotherapy and is based on the observation that leukemic blasts will enter cell cycle shortly after completion of chemotherapy, theoretically increasing the sensitivity to cytarabine-based chemotherapy. OS was improved in pediatric patients treated with multi-agent chemotherapy (dexamethasone, cytarabine, thioguanine, etoposide, daunorubicin, and intrathecal cytarabine) followed by the same regimen at day 10 to 14 regardless of initial response compared to standard sequencing of these two courses based on initial response. On the other hand, the German AML Cooperative Group did not detect any difference in CR rate or OS with a dose dense regimen of high-dose cytarabine and mitoxantrone on days 1 to 4 and 8 to 11 compared with their standard double induction regimen with thioguanine, cytarabine, and daunorubicin (TAD) followed by high dose cytarabine and mitoxantrone (HAM) 21 days later. The duration of neutropenia was shorter with the dose dense sequential HiDAC and mitoxantrone.