Review Article
`
`Acute Myeloid Leukemia: Historical Perspective
`and Progress in Research and Therapy Over 5
`Decades
`3 Hagop M. Kantarjian, 1 Nicholas J. Short, 1 Amir T. Fathi, 2 Guido Marcucci,
`
`
`
`
`
`
`
`6 Farhad Ravandi, 1 Martin Tallman, 4 Eunice S. Wang, 5 Andrew H. Wei
`
`
`
`Abstract
`With the Food and Drug Administration approval of 9 agents for different acute myeloid leukemia (AML) indications, the
`prognosis and management of AML is evolving rapidly. Herein, we review the important milestones in the history of AML
`research and therapy, discuss insights regarding prognostic assessment and prediction of treatment outcome, detail
`practical supportive care measures, and summarize the current treatment landscape and areas of evolving research.
`
`Clinical Lymphoma, Myeloma and Leukemia, Vol. 21, No. 9, 580–597 © 2021 Elsevier Inc. All rights reserved.
`Keywords: AML, Therapy, Targeted therapy, Prognosis, Biology
`
`Introduction
`Understanding the pathophysiology of acute myeloid leukemia
`(AML) has translated into rapid clinical applications that are
`
`transforming its treatment and outcome. 1-3 Recent translational
`successes include the novel targeted therapies directed at BCL2
`(venetoclax), FMS-like tyrosine kinase 3 (FLT3), and isocitrate
`dehydrogenase (IDH). Such developments, and the highly effective
`novel combinations arising from them, raise the question of whether
`traditional intensive chemotherapy approaches, like the “3 + 7
`regimen” (3 days of daunorubicin plus 7 days of cytarabine), should
`remain as the optimal standard of care in the current era. In the early
`cooperative group trials of 3 + 7 in highly selected younger patients
`(usually 55 years or younger), the 5-year survival rates were 40% to
`
`50%. 4 Later trials included patients up to 60 to 65 years of age and
`
`reported lower long-term survival rates of 30% to 40%. 5 Patients
`older than 60 to 65 years receiving 3 + 7 experienced a higher risk of
`early mortality (4- to 8-week mortality rates greater than 10%-30%)
`6
`and poor long-term survival rates of less than 10% to 15%.
`
`1 Department of Leukemia, MD Anderson Cancer Center, Houston, TX, USA
`2 Leukemia Program, Massachusetts General Hospital, Harvard Medical School,
`Boston, MA, USA
`3 Gehr Family Center for Leukemia Research City of Hope, Duarte, CA, USA
`4 Leukemia Service, Department of Medicine, Memorial Sloan Kettering Cancer
`Center, Weill Cornell Medical College, New York, NY, USA
`5 Leukemia Service, Department of Medicine, Roswell Park Comprehensive Cancer
`Center, Buffalo, NY, USA
`6 Department of Clinical Hematology, The Alfred Hospital and Monash University,
`Melbourne, Australia
`
`Submitted: Apr 8, 2021; Revised: May 20, 2021; Accepted: May 22, 2021; Epub: 29
`May 2021
`
`Address for correspondence: Hagop M. Kantarjian, MD, MD Anderson Cancer Center,
`1400 Holcombe Blvd, Unit 428, Houston, TX, 77030 USA.
`E-mail contact: hkantarjian@mdanderson.org
`
`In community practice settings, patients treated with 3 + 7
`are more likely to be older and less selected, with multiple
`clinically impactful comorbidities (cardiac, pulmonary, hepatic,
`renal, diabetes, or hypertension), or with adverse-risk AML
`more frequently evolved as a result of prior genotoxic therapy or
`AML evolving from treated myelodysplastic syndrome (MDS),
`chronic myelomonocytic leukemia (CMML), or myeloproliferative
`neoplasm (MPN). All these factors contribute to worse outcomes
`with 3 + 7 or comparably intensive regimens, manifesting with
`
`higher rates of early mortality and lower rates of cure. 7-15 In the
`early publications from Swedish investigators, favorable results were
`
`reported with intensive chemotherapy in older patients, 10 , 11 but the
`subsequent updates indicated that only 60% of patients received
`intensive chemotherapy and that the early mortality rates were 6%
`
`to 9% in younger patients and 16% to 34% in older patients. 12 The
`5-year survival rates were 10% to 20% in patients 50 to 75 years old,
`12
`and the 2-year survival rate was 5% in patients older than 75 years.
`The Mayo Clinic detailed its experience in 1123 adults with AML,
`among whom 766 (68%) were treated with intensive chemotherapy,
`leading to a complete response (CR) rate of 44%. An additional
`33% of patients achieved CR with incomplete recovery of either
`
`neutrophils above 1 × 10 9 /L or platelets above 100 × 10 9 /L (CRi),
`
`resulting in an overall response rate of 77%. The 5-year survival
`rate was only 30% with intensive chemotherapy and less than 5%
`
`with lower intensity therapy or supportive care. 13 Similarly, poor
`results have been reported by others in younger cohorts of patients
`receiving intensive chemotherapy: 5-year survival rates of 20% in
`
`intermediate risk and only 10% in adverse risk. 15 Such findings have
`for some time highlighted the need to improve on the traditional
`cytarabine/anthracyclines regimens in younger or functionally fit
`
`580
`
`Clinical Lymphoma, Myeloma and Leukemia 2021
`
`CELGENE 2089
`APOTEX v. CELGENE
`IPR2023-00512
`
`2152-2650/$ - see front matter © 2021 Elsevier Inc. All rights reserved.
`https://doi.org/10.1016/j.clml.2021.05.016
`
`

`

`patients (younger/fit) and to develop novel lower-intensity targeted
`therapy combinations that would be efficacious in older or less fit
`patients (older/unfit). Figure 1 shows the outcomes in AML in
`younger and older patients from 1970 to 2020 in a single institution
`referral center.
`AML comprises several disease subsets that require different
`therapeutic approaches and carry very different prognoses. These
`include acute promyelocytic leukemia (APL) (now near univer-
`sally curable with therapies incorporating all- trans retinoic acid
`
`(ATRA) and arsenic trioxide) 16-19 and core binding factor (CBF)
`AML (remission rates approaching 90% with traditional intensive
`chemotherapy plus gemtuzumab ozogamicin [GO] and high rates
`
`of survival). 20-24 The remaining non-APL, non-CBF AMLs can
`be prognostically and functionally divided according to their
`cytogenetic-molecular profiles, as well as according to fitness for
`intensive chemotherapy, into AML in younger/fit patients and
`older/unfit patients. The latter are underrepresented in historical
`studies of intensive chemotherapy but comprise most AML cases
`managed in community practice (60%-70%). Finally, the prognosis
`of de novo AML is superior to AML that has evolved from MDS,
`CMML, or MPN, particularly after prior exposure to hypomethy-
`lating agents (HMAs) and to therapy-related AML after exposure to
`chemotherapy (eg, alkylating agents or topoisomerase-II inhibitors)
`7 , 13
`or irradiation as prior treatment of other malignancies.
`Herein we review important achievements in the history of
`AML research and therapy, discuss insights regarding prognostic
`assessment and prediction of treatment outcome, detail practical
`supportive care measures, and summarize the current treatment
`landscape and areas of evolving research.
`
`Acceptance and Evolution of Newly
`Introduced Therapies in AML
`The journey to the regulatory approval of various drugs and
`regimens for AML therapy in the United States, Europe, and
`elsewhere has historically followed at times convoluted paths.
`Many drugs approved by the Food and Drug Administration
`(FDA) for a particular indication have subsequently been used for a
`different purpose. This is logical because much of the research and
`progress in cancer and medicine often occurs after the regulatory
`drug approval. This has been true for the first 2 modalities approved
`for the treatment of AML, cytosine arabinoside (cytarabine or
`ara-C) and daunorubicin. Five decades after their introduction
`into AML therapy, researchers continue to explore different dose
`schedules of cytarabine and daunorubicin during induction and
`consolidation therapy.
`The research post–FDA approval is particulary vigorous for
`agents that have secure longer patent expiry terms, as this encour-
`ages greater pharmaceutical company investment to maximize
`commercial returns. For example, the FDA approved gilteritinib
`(FLT3 inhibitor) as monotherapy for refractory-relapsed FLT3 -
`mutated AML and enasidenib (IDH2 inhibitor) as monotherapy for
`refractory-relapsed IDH2 -mutated AML. However, it is likely that
`in a few years these targeted therapies will be used in combination
`with standard chemotherapy or in targeted therapy cocktails, both in
`salvage and frontline therapy, based on post–FDA approval research.
`
`Hagop M. Kantarjian et al
`
`Several agents were never approved for AML therapy but are
`now the most commonly used drugs in some AML subsets and
`are even accepted by the FDA as a control arm in investigational
`phase 3 trials. For example, neither of the 2 HMAs, decitabine
`and azacitidine, was approved for the treatment of older/unfit
`AML (European Medicines Agency [EMA] approval of decitabine
`in 2012 and of azacitidine [AML with ≥ 30% blasts] in 2015).
`Yet, both HMAs are now the most commonly used drugs in
`older/unfit AML, and the combination of HMAs plus venetoclax
`was recently FDA approved (October 2020) for the treatment of
`newly diagnosed older/unfit AML. The combination of ATRA
`and arsenic trioxide for APL was approved by the EMA in 2016
`and by the FDA only in 2018. Despite the lack of regulatory
`approvals, adenosine nucleoside analogues (fludarabine, cladrib-
`ine, clofarabine) and etoposide (topoisomerase 2 inhibitor) are
`commonly used in AML combination regimens around the world.
`Some drugs are limited to particular geographies: amsacrine
`in Europe (not produced anymore); homoharringtonine in
`25-30
`China.
`Finally, multiple agents have stumbled in their regulatory
`approval journey and, although potentially of value in AML
`therapy, may not be explored further. This is particularly true for
`“me-too” chemotherapy agents, in which interest waned in favor of
`rationally designed targeted therapies. Examples include topotecan,
`
`cloretazine, clofarabine, and vosaroxin. 31–34 But this also applies
`to targeted therapies such as sorafenib (first-generation FLT3
`inhibitor), vadastuximab talirine (SGN-CD33A; CD33-conjugated
`monoclonal antibody), and others.
`We are witnessing an ongoing “slow-motion” revolution in
`AML reseach and therapy, with the approval of 9 agents for
`different AML indications since 2017: GO; the 2 FLT3 inhibitors
`midostaurin and gilteritinib; the IDH1 inhibitor ivosidenib and
`IDH2 inhibitor enasidenib; the BCL2 inhibitor venetoclax; the
`oral HMA azacitidine; the liposomal formulation CPX351; and
`the hedgehog inhibitor glasdegib. Oral decitabine/cedazuridine was
`FDA approved for MDS and CMML and will likely be used as an
`alternative oral HMA in AML. An important challenge is how to
`promptly incorporate these drugs/modalities into effective and safe
`frontline and salvage combinations in AML.
`
`Determinants of Outcome in AML
`The outcomes in AML are very heterogeneous. With current
`therapies, some AML subsets are highly curable (eg, APL and
`CBF AML; cure rates, ≥ 75%), whereas others are highly adverse,
`with estimated 5-year survival rates of 10% or less: TP53 -mutated
`AML; MECOM (MDS1 and EVI1 complex locus)–rearranged
`AML [inversion of chromosome 3 or t(3;3)(q21q26)]; AML with
`complex kar yotype; secondar y AML after MDS or CMML treated
`with HMAs or after MPN; and therapy-related AML. The determi-
`nants of AML outcome include variables related to the patient, the
`leukemia, and the response to therapy (achievement of CR, CRi, or
`other responses; measurable residual disease [MRD] in remission).
`The therapeutic environment and choice of therapy (intensive vs.
`low intensity) are also important determinants of AML outcome.
`
`Clinical Lymphoma, Myeloma and Leukemia 2021 581
`
`

`

`Acute Myeloid Leukemia: Historical Perspective and Progress in Research
`
`Figure 1 Survival of Patients < 60 Years of Age (A) and ≥ 60 Years of Age (B) With De Novo AML Treated at MD Anderson Over 5
`Decades
`Abbreviations: AML = acute myeloid leukemia.
`
`582
`
`Clinical Lymphoma, Myeloma and Leukemia 2021
`
`

`

`Patient-Associated Factors
`Age is a consistent determinant of AML prognosis because
`older age is tightly linked to poor performance, increased comor-
`bidities, and poorer tolerance of and responsiveness to intensive
`chemotherapy. Older age is also associated with a higher frequency
`of AML-associated adverse factors such as non–de novo AML,
`complex karyotype, and unfavorable mutations (like TP53 ).
`Multivariate analyses account for some of these interactions.
`In discussing the choice of “intensive” versus “low-intensity”
`chemotherapy, some AML experts discourage such treatment labels
`as too dichotomous, suggesting that there is a fine line between
`the 2 approaches. They and the FDA advocate for a therapeutic
`choice based on the patient’s “biologic age” and comorbidities/organ
`dysfunctions, using tools such as the “Ferrara criteria” or the “Charl-
`son morbidity index.”35 , 36 However, older age is always selected
`as independently adverse in prognostic models of AML, even after
`
`meticulous accounting for all other adverse factors. 37-41 Moreover,
`despite achieving good CR rates of 40% to 50% with intensive
`chemotherapy in older AML, outcome is invariably poor, and the
`6 , 10-14 , 37 , 38
`early mortality rates often nullify any therapeutic benefit.
`In the Surveillance, Epidemiology, and End Results (SEER) data
`from the United States, the early mortality in older AML is ≥ 25%
`in patients 60 years and older and ≥ 40% in patients 70 to 75 years
`
`or older. 7 The emerging data with low-intensity chemotherapy
`in combination with venetoclax show high CR-CRi rates (almost
`equivalent to intensive chemotherapy) and low early mortality rates.
`Thus, it is perhaps time to consider low-intensity therapy in all
`older AML (aged 60-65 years or older) regardless of the “fitness”
`factors, if the aim of AML therapy is to achieve a marrow CR safely,
`to bridge to a curative option of allogeneic stem cell transplant
`(SCT) in first CR. It is also perhaps time to consider low-intensity
`therapy combinations even in younger patients, particularly if they
`have comorbidities or when intensive chemotherapy is known to
`have poor results. Regardless, the SEER data show that only about
`40% of older patients with AML receive any form of induction
`chemotherapy, highlighting the urgent need for a change in our
`42-44
`AML standard practice.
`
`Leukemia-Associated Factors: Cytogenetic and Molecular
`Abnormalities
`The National Comprehensive Cancer Network (NCCN)
`cytogenetic-molecular classification categorizes AML into “favor-
`
`able,” “intermediate,” and “poor/adverse” risk groups. 45 The
`NCCN classification has most relevance for younger and de novo
`patients with AML but has less discriminatory value for older AML,
`AML evolving from MDS/MPN, or therapy-related AML. Similar
`46
`observations apply to the European LeukemiaNet classification.
`The current NCCN classification is not dynamic and does not
`incorporate the modifying effect of novel therapies on AML
`outcome (eg, with FLT3 inhibitors plus intensive chemotherapy for
`FLT3 -mutant AML, and venetoclax-azacitidine for IDH mutant
`disease). It would be more relevant to categorize AML subsets
`based on estimated 3- to 5-year survival rates: favorable if rates
`are > 60%, intermediate if rates are 30%-60%, and unfavorable
`if rates are < 20% to 30%. For example, if we used absolute
`survival rates to guide risk stratification, nearly all older/unfit AMLs
`
`Hagop M. Kantarjian et al
`
`would be categorized as unfavorable, as would secondary and
`therapy-related AML. Such a dynamic risk stratification based on
`absolute (rather than relative) survival expectations would better
`reflect long-term clinical outcomes and comparisons across clinical
`47
`trials.
`Based on current practice, we consider a simpler classification
`of the AML karyoptypes as follows: (1) favorable, APL and CBF
`karyotypes; (2) intermediate, diploid karotype; (3) unfavorable, 3 or
`more chromosomal abnormalities, monosomy 5/5q −, monosomy
`7/7q −, translocation t(6;9), translocation t(9;22), all translocations
`involving 11q23, and translocations involving chromosome 3q26.2
`
`( EVI1 , location of the MECOM gene); 47 and (4) all others. Some
`studies consider certain cytogenetic abnormalities [eg, single trisomy
`
`8, or single translocation t(9;11)] as intermediate risk. 45 , 46 The
`intermediate risk classification of a single translocation t(9;11)(p22;
`
`q23)/ KMT2A-MLLT3 has been debated. 48-51 Although the NCCN
`classification puts it under intermediate risk, this may be true
`for only the small subset of younger patients with de novo
`15
`AML.
`Molecular studies have identified recurrent somatic mutations
`in more than 90% of patients with AML, the most frequent being
`FLT3, NPM1, DNMT3A, NRAS, TET2, IDH2, CEBPA, RUNX1,
`52 , 53
`PTPN11, IDH1, TP53 , and SRSF2 .
`Mutations may be prognostic and targetable. Their prognostic-
`predictive effect may depend on several factors: (1) the particular
`mutation; (2) the mutation burden or variant allelic frequency (ratio
`of mutated gene/total); (3) the cytogenetic risk group it associates
`with (favorable, normal, unfavorable, other); (4) the presence of
`other mutations; (5) the patient’s age; (6) whether the AML is de
`novo or evolving from MDS/MPN or therapy related; and (7) the
`treatment given.
`In normal karyotype AML, a biallelic CEBPA mutation (2% or
`less of AML) or a mutation of nucleophosmin-1 ( NPM1; 50% of
`AML with normal karyotype) are associated with better prognoses,
`45 , 46
`provided no other adverse concurrent mutations are present.
`A FLT3 internal tandem duplication (FLT3 -ITD) was traditionally
`associated with a poor prognosis; this is now changing with the
`use of newer and better FLT3 inhibitors with chemotherapy and
`as post-SCT maintenance. Prognosis was adverse particularly with
`a high allelic ratio (AR) (ratio of FLT3 -ITD/FLT3 wild type using
`
`a semiquantitative DNA fragment analysis) 46 and in the absence
`of NPM1 mutation. In normal karyotype AML, the prognosis
`with concurrent NPM1 and FLT3- ITD mutations depends on
`
`the FLT3 -ITD AR. 54-56 Other adverse mutations include RUNX1,
`ASXL1 , and TP53 ; when present, they categorize the AML as
`
`adverse risk. 57-62 In general, a greater number of adverse mutations
`in a patient with AML indicates worse prognosis.
`The prognostic effect of mutations is more relevant in
`
`diploid/intermediate karyotype AML. 54 Their impact in favor-
`able and unfavorable karyotypes is lessened and context dependent.
`Among the favorable karyotype AML, KIT mutations have been
`reported to be adverse in some studies using 3 + 7 63 , 64 but not in
`
`trials using fludarabine-high dose cytarabine and GO (FLAG-GO;
`FLAG-IDA + / − GO). 20 , 21 In unfavorable karyotype AML, a TP53
`
`mutation worsens prognosis further in an already poor prognosis
`57 , 58
`disease.
`
`Clinical Lymphoma, Myeloma and Leukemia 2021 583
`
`

`

`Acute Myeloid Leukemia: Historical Perspective and Progress in Research
`
`The significance of mutations had been defined primarily in
`younger patients with AML. The prognostic value of mutations
`is generally worse among older patients. The same is true for
`therapy-related AML and AML evolving from MDS or MPN.
`Because many of the molecular abnormalities are potentially
`targetable, the predictive value may change with the incorporation
`of effective targeted therapies into the standard chemotherapy
`regimens, simultaneously or sequentially. For example, the incorpo-
`ration of FLT3 inhibitors into AML chemotherapy and as post-SCT
`maintenance is already changing the previously poor outcome of
`
`FLT3 -mutated AML into a reasonably favorable one. 65-69 The
`IDH1/2 -mutated AML (20% of AML) can be effectively treated
`with combinations of chemotherapy and IDH inhibitors. The IDH
`mutations also generate BCL-2 dependence for survival, making
`IDH -mutated AML sensitive to venetoclax-based therapy and
`suggesting the potential for improved outcome with a triple-agent
`regimen (HMAs + venetoclax + IDH inhibitor, simultaneously or
`
`sequentially). 70 The TP53 -mutated AML responds poorly to inten-
`sive chemotherapy but may benefit from lower-intensity chemother-
`apy with HMAs and venetoclax and/or the addition of novel TP53 -
`
`directed strategies like magrolimab. 71-74 The cytogenetic-molecular
`subset of “mixed-lineage leukemia” (translocations involving 11q23;
`MLL1; KMT2A rearrangement) may benefit from novel menin
`
`inhibitors (SNDX-5613, KO-539, others). 75-77 The KIT -mutated
`CBF AML, associated with unfavorable outcome in 3 + 7 trials,
`may benefit from the addition of GO or potent c-KIT inhibitors
`
`(avapritinib, dasatinib). 78 , 79 Predicting the outcome in patients
`with AML, especially the impact of gene-gene interactions, is made
`easier with artificial intelligence modeling (“knowledge bank”)
`80 , 81
`based on annotated large cohorts of patients.
`
`Treatment Response-Associated Factors: Achievement of
`CR Versus Less-Than-CR Response; Measurable Residual
`Disease in Remission
`For decades, the achievement of CR with full hematologic recov-
`ery after intensive cytotoxic chemotherapy has been considered the
`only morphologic response associated with a significant survival
`
`benefit. 82 This is now challenged by multiple studies with intensive
`chemotherapy as well as with low-intensity therapy and targeted
`13 , 83
`agents (FLT3 inhibitors, IDH inhibitors, venetoclax, GO).
`More recently the term “CRh” has been used in a number of
`regulatory trials. The CRh requires the criteria for CR but with
`an absolute neutrophil count 0.5 to 1 × 10 9 /L and platelets 50
`
`to 100 × 10 9 /L. Recent studies suggest prognosis for CRh is
`
`intermediate between CR and CRi/CRp. It is likely that patients
`with marrow CR and MRD negative status (regardless of whether
`therapy is intensive or low intensity) will have an outcome close to
`a traditional morphologic CR.
`Measuring residual disease in AML in morphologic CR is now
`
`part of the standard of care in AML. 84-88 The detection of MRD
`at the time of morphologic CR or CRi is associated with a higher
`relapse rate and with worse survival. The MRD has been commonly
`investigated using 2 methodologies: multicolor flow-cytometric
`measurements , and molecular quantification of residual disease.
`Polymerase chain reaction (PCR) is used to monitor quantita-
`tively unique AML-specific translocations and mutations (eg, in
`
`APL, CBF AML, and NPM1- mutated AML) and is expanding
`to other molecular subsets ( IDH1/2 and FLT3 mutations). In
`APL and CBF AML, detection of MRD by quantitative PCR
`
`predicts for relapse. 89–91 Among patients with non-CBF/non-
`APL AML, monitoring MRD by next-generation sequencing of
`mutations is informative, for example in patients with NPM1
`
`mutations. 92 , 93 Combining multicolor flow cytometry and next-
`generation sequencing to detect molecular mutations in remission
`may further improve the capability of MRD studies to predict
`
`for relapse. 84 The persistence of some mutations like DNMT3A,
`TET2 , and ASXL1 (DTA mutations) does not predict for relapse
`and may be rather a feature of clonal hematopoiesis in some older
`84
`patients.
`The MRD status of patients with AML in CR may lead to
`
`consideration of therapeutic interventions. 89 , 94 Interventions
`that may eradicate MRD in CR now include allogeneic SCT,
`more intensified chemotherapy regimens, HMAs plus venetoclax,
`targeted therapy combinations when indicated for particular molec-
`ular abnormalities (FLT3 or IDH inhibitors), antibody therapies
`(eg, CD123 or CD33 antibody drug conjugates or bispecific T-cell
`engagers), or immune therapies (eg, checkpoint inhibitors).
`
`The Effect of the Environment Under Which the AML Is
`Treated, and the Impact of Supportive Care Measures
`Historically, in the United States it was assumed that the
`outcomes of AML are equivalent across National Cancer Institute
`(NCI)–designated cancer centers, other academic centers, and in
`community practice. This is not likely the case. In a National
`Cancer Database of 60,738 patients with AML, the 1-month
`mortality was 16% in academic centers and 29% in nonacademic
`centers ( P < .001). The estimated 5-year survival rates were 25%
`versus 15% ( P < .001). The center effect was identified by multi-
`variate analysis to be independently prognostic, with a hazard risk of
`
`1.52 in nonacademic centers ( P < .0001). 8 In another study from
`California, among 7007 patients with AML (1999-2014), the early
`4-week mortality was 12% in NCI-designated cancer centers versus
`24% in non–NCI-designated cancer centers. At MD Anderson, the
`early 4-week mortality is 5% or less with intensive chemotherapy
`in younger/fit AML and 2% to 3% in older/unfit AML (discussed
`and referenced later).
`The routine use of antibiotics (levofloxacin, cefpodoxime, others)
`and antifungal prophylaxis (posaconazole, voriconazole) in all newly
`diagnosed acute leukemias has reduced the incidences of infections
`and associated morbidities but importantly also early mortality rates
`95-98
`(5%-10%).
`AML is a rare cancer requiring cumulative expertise, constant
`vigilance, and prompt, readily available intensive supportive care
`(transfusion products, early recognition of sepsis, immediate imple-
`mentation of needed care in an emergency center and in intensive
`care units). Even perceived “insignificant delays” in implementing
`intravenous antibiotics in sepsis can result in increased mortality,
`particularly among older patients (most patients with AML) who
`have poorer organ reserve capacities during sepsis and may deteri-
`orate rapidly with multiorgan failure (pulmonary, cardiac, hepatic,
`renal) and have high complication and mortality rates.
`
`584
`
`Clinical Lymphoma, Myeloma and Leukemia 2021
`
`

`

`Translating the Biologic
`Information Into Therapy and
`Clinical Research
`The AML subsets are very heterogeneous and consequently
`benefit from selective therapies. Next, we discuss the treatment of
`different AML subsets using commercially available FDA approved
`agents, as well as approaches with investigational agents.
`
`APL
`APL (5%-10% of AML) is characterized by the cytoge-
`netic
`translocation between
`chromosomes 15
`and 17
`[t(15;17)(q22;q21)], which results in the PML-RAR alpha fusion
`oncogene and its encoded oncoprotein. The latter acts as a
`dominant negative inhibitor of wild-type RAR alpha , causing a
`maturation block and the clinical-pathologic picture of APL.
`In the 1970s, single-agent anthracyclines (daunorubicin) were
`99
`first shown to produce cure rates of 30% to 40% in APL.
`
`Single-agent cytarabine does not cure APL. 100 The addition of
`cytarabine to anthracyclines does not increase the APL cure rate
`substantially, nor does the addition of maintenance therapy with
`
`6-mercaptopurine-methotrexate combinations. 101 , 102 A “differ-
`entiation syndrome” was also reported for the first time with
`
`chemotherapy in APL. 103 The early mortality from the complex
`coagulopathy, which includes disseminated intravascular coagula-
`tion and bleeding, was significant (20%-30%).
`In the late 1980s and early 1990s, ATRA and arsenic trioxide
`were discovered to have major anti-APL activities. The curative
`effect of both agents is through induction of degradation of the
`PML-RAR alpha oncoprotein, thus reversing the maturation block
`and promoting differentiation of APL cells. Studies from China,
`India, and Iran of ATRA or arsenic trioxide as frontline APL
`monotherapy showed high CR rates and, with arsenic trioxide, 5-
`104-106
`year disease-free survival (DFS) rates exceeding 50% to 60%.
`
`Gemtuzumab ozogamicin was also highly active. 107 Both ATRA
`and arsenic trioxide, when added to chemotherapy during induc-
`tion and/or consolidation in comparative trials, improved outcome
`
`in APL. 108-111 The combination of idarubicin and ATRA (AIDA
`112
`regimen) became the standard of care in APL for a while.
`In the early 2000s, a nonchemotherapy regimen of ATRA plus
`arsenic trioxide was explored cautiously in APL salvage (2001),
`then as frontline APL therapy (2002). GO was added for high-risk
`APL. Following the demonstration of the high efficacy of this
`
`approach, 16 , 17 , 113 randomized studies confirmed the superiority of
`ATRA plus arsenic trioxide over AIDA in low- and intermediate-
`
`risk APL. 18 , 19 , 114 , 115 With ATRA plus arsenic trioxide, the CR rate
`is ≥ 95%, and the cure rate is ≥ 90%. Induction mortality from
`coagulopathy is low (about 5%), and resistant disease is rare, except
`in molecular variant APL (translocations between chromosome 11
`and 17 [ PLZF-RAR alpha ] or between chromosome 5 and 17).
`Patients with high-risk APL benefit from the addition of GO (or
`anthracylines). The details of the regimen have been previously
`
`published. 16-19 The Medical Research Council (MRC) comparative
`trial investigated a lower and less frequent dose schedule of arsenic
`
`trioxide. 114 Oral formulations of arsenic trioxide would make the
`treatment of APL more convenient, particularly in maintenance (80
`116 , 117
`doses).
`
`Hagop M. Kantarjian et al
`
`Figure 2 shows single-institution results in APL among younger
`and older patients and significant improvement in outcome in
`the era of ATRA and arsenic trioxide. Some important (not well-
`known) considerations in APL management are detailed in the
`118-120
`published literature.
`
`CBF AML
`The CBF AML includes the cytogenetic-molecular subsets
`of inversion 16 [inv16(p13;q22)] or t(16;16)(p13;q22)], and
`t(8;21)(q22;q22). Historically, CBF AML was treated with cytara-
`bine plus anthracycline induction chemotherapy followed by
`1 to 4 high-dose cytarabine consolidations. The cure rate was
`30% to 40% with 1 consolidation versus ≥ 50% with 3 to 4
`
`consolidations. 121 , 122 Optimizing the combinations of established
`chemotherapy drugs (fludarabine plus high-dose cytarabine for 5-6
`courses of induction consolidation; addition of GO to chemother-
`apy; monitoring and treatment of persistent MRD in CR) improved
`the cure rate in CBF AML to ≥ 75%. 20 - 24 A meta-analysis of 5
`
`randomized trials showed that the addition of GO to chemotherapy
`improved the estimated 5-year survival from 50% to 75% in CBF
`
`AML. 24 GO is now a standard component of CBF AML therapy.
`With fludarabine, high-dose cytarabine, and GO (FLAG-GO)
`during induction and consolidations (total, up to 6 courses), and
`modification of therapy (eg, allogeneic SCT, azacitidineazaciti-
`dine/venetoclax/GO) for persistent MRD in CR, the estimated
`5-year survival rates were ≥ 75% in both inversion 16 and t(8;21)
`
`AML ( Figure 3 ). 20 , 21 The results were better in younger patients.
`Patients who cannot tolerate FLAG-GO/IDA or who have persis-
`tent molecular disease may be offered HMA therapy (decitabine,
`azacitidine) in combination with venetoclax and GO, with the
`treatment duration adjusted according to the MRD results or for ≥
`12 months.
`Frequent mutations noted in CBF AML are FLT3 (15%-20%),
`KIT (25%-30%), N/KRAS (30%-50%) and others. Although
`some studies have reported worse outcomes with KIT or multiple
`
`mutations, 63 , 64 others have not, such as the experience with FLAG-
`GO/idarubicin. The improved efficacy of the regimen may have
`nullified the adverse effects of these mutations. Targeted therapies
`may also be considered (avapritinib or dasatinib for KIT mutations;
`78 , 79 , 123
`FLT3 inhibitors for FLT3 mutations).
`
`Choice of Intensive Chemotherapy in Younger/Fit AML
`Versus Low-Intensity Therapy in Older/Unfit AML
`
`The median age in AML is 68 to 70 years. 7 Still, most of the
`clinical research with 3 + 7 based regimens has been conducted
`in younger patients and proposed for older patients if considered
`fit enough for intensive chemotherapy. This is despite the poor
`outcome with 3 + 7 in older AML.
`7
`In a study of 813 selected patients 60 years and older (median
`age, 67 years) treated with 3 + 7 (randomization to higher vs. lower
`dose of daunorubicin), the early mortality rate was 11% to 12%,
`the median survival was 7 to 8 months, and the estimated 3-year
`
`survival rate was 20%/ 6 This and other experiences from carefully
`controlled studies in selected patients with good performance,
`normal organ functions, and few comorbidities translated poorly
`into community practice. Examination of 29,000 patients with
`
`Clinical Lymphoma, Myeloma and Leukemia 2021 585
`
`

`

`Acute Myeloid Leukemia: Historical Perspective and Progress in Research
`
`Figure 2 Survival in Patients < 60 Years of Age (A) and ≥ 60 Years of Age (B) W