review
`
`How we treat older patients with acute myeloid leukaemia
`
`Avraham Frisch,1 Jacob M. Rowe1,2,3
`
`and Yishai Ofran1,3
`
`1Department of Hematology and Bone Marrow Transplantation, Rambam Health Care Campus, Haifa, 2Department of Hematology,
`Shaare Zedek Medical Center, Jerusalem, and 3The Ruth and Bruce Rappaport Faculty of Medicine, Technion, Israel Institute of
`Technology, Haifa, Israel
`
`Summary
`
`After decades when intensive chemotherapy remained the
`only effective anti-acute myeloid leukaemia (AML) treat-
`ment, a torrent of novel, less toxic agents are about to revo-
`lutionise AML therapy. Prolonged remissions with good
`quality of life become achievable for many patients previ-
`ously considered only for palliative care because they could
`not tolerate intensive therapy. As treatment options multiply,
`the importance of genetic profile is recognised, even for
`advanced-age patients for whom cure is unlikely. With lack
`of randomised comparative trials for most treatment regi-
`mens, one can only extrapolate data from existing studies to
`make evidence-based decisions. We herein present seven
`common clinical
`scenarios
`illustrating the complexity of
`treating older AML patients and describe our approach to
`their management. In each case, up-to-date data on relevant
`agents to be offered to a particular patient are discussed. The
`current review is limited to the drugs, available and approved
`in the Western world and many promising agents, still under
`investigation, are not discussed.
`
`Keywords: acute myeloid leukaemia, older adults, molecular
`profile, intensive chemotherapy, novel agents.
`
`Acute myeloid leukaemia (AML) is the most prevalent acute
`leukaemia in adults, with a median age at diagnosis of
`68 years. For decades, the main treatment option for newly-
`diagnosed AML patients has been intensive chemotherapy,
`which optimally offers a complete remission (CR) rate of
`70% and a long-term survival of about 40%.1 Specifically for
`patients with advanced age, not only is the remission rate
`significantly inferior, but also toxicity and treatment-related
`mortality are excessive compared to younger patients. Fortu-
`nately, in recent years, novel, effective and less toxic drugs
`
`Correspondence: Yishai Ofran, Department of Hematology and Bone
`
`Marrow Transplantation, Rambam Health Care Campus, 8 Ha’Aliya
`
`st., Haifa 3109601, Israel.
`
`E-mail: y_ofran@rambam.health.gov.il
`
`First published online 30 April 2020
`doi: 10.1111/bjh.16701
`
`CELGENE 2082
`APOTEX v. CELGENE
`IPR2023-00512
`ª 2020 British Society for Haematology and John Wiley & Sons Ltd
`British Journal of Haematology, 2020, 191, 682–691
`
`have become available. Moreover, improvements in support-
`ive care, and better tools for risk stratification and patient
`selection for allogeneic stem cell transplantation (allo-SCT),
`have all contributed to an improved outcome. Choosing
`wisely from the many treatment options available for AML
`patients with advanced age is currently a common clinical
`dilemma with no simple textbook answer. It is challenging to
`translate the accumulating but limited data regarding the
`effectiveness of each drug into a reliable, effective and evi-
`dence-based approach. We herein, with the use of six vign-
`ettes, present common treatment options and discuss a few
`case scenarios, aiming to apply principles that can be used in
`common clinical practice.
`
`Case 1: Is there still a role for intensive
`chemotherapy for elderly AML patients?
`
`A previously healthy 72-year-old woman is referred because
`of increased fatigue. Her complete blood count (CBC) shows
`a white blood cell (WBC) count of 1,300/ll with 6% blasts,
`32,000/ll platelets and haemoglobin of 6 5 g/dl. Her marrow
`is infiltrated with myeloid blasts, and molecular mutation
`analysis is negative for NPM1, FLT3-ITD mutations and for
`core binding factor (CBF) aberrations. Her echocardiogram
`demonstrates normal
`left systolic and diastolic functions
`without pulmonary hypertension or wall motion abnormali-
`ties. She has normal renal and liver functions. Cytogenetic
`results will be available within a week. What is the optimal
`approach for this patient? Should cytogenetics impact the
`chosen strategy?
`This is a case of an apparently fit patient who, had she
`been 10 years younger, would have been considered for
`intensive chemotherapy by most physicians. The question
`therefore
`is whether
`intermediate-
`and/or
`adverse-risk
`patients should be treated differently in an advanced age. To
`answer this question,
`it is prudent to consider both the
`immediate treatment plan and post-remission therapy. It has
`been well-known for decades that achieving CR in the end of
`induction is a prerequisite but
`insufficient condition for
`long-term remission. A substantial number of patients will
`relapse early after achieving CR, if no consolidation therapy
`
`

`

`is prescribed.2,3 Thus, an immediate/induction treatment
`plan should be such as to allow optimal post-remission ther-
`apy, particularly in fit patients, preserving the option for
`allo-SCT.
`Following intensive chemotherapy, fit patients aged 70–
`80 years with intermediate-risk AML have about 60–65%
`probability to achieve CR, with a 10–15% early death risk.1,4
`It should be noted that in the large prospective study by the
`Eastern Cooperative Oncology Group-American College of
`Radiology Imaging Network Cancer Research Group (ECOG-
`ACRIN, E2906) exploring intensive chemotherapy in patients
`with advanced age, minimal residual disease (MRD) negativ-
`ity was achieved in 58/147 (39 5%) of patients who achieved
`remission. These MRD-negative patients have an excellent
`long-term survival rate.5 Further improvement of intensive
`regimes in older patients is challenging. The addition of gem-
`tuzumab ozogamicin (GO) in induction, known to improve
`CR and survival in younger patients, is probably too toxic
`for patients over 70 years.6,7 As GO increases induction toxi-
`city by prolonging the time to bone marrow recovery8 and
`increasing early mortality,9
`its use in combination with
`intensive chemotherapy for older patients cannot be recom-
`mended.
`For long-term remission in fit older patients, allo-SCT
`remains the most effective post-remission strategy, with 38%
`2-year overall survival (OS) in CR1 for patients older than
`69 years and a 3-year OS of 49% for patients older than
`60 years.10,11
`A novel alternative for this patient, considered by some, is
`using venetoclax with hypomethylating agents (HMA) or
`low-dose cytarabine (LDAC; discussed later
`in case 2).
`Although data from phase Ib/II studies reported CR + CR
`with incomplete blood count recovery (CRi) rates as high as
`73% and low early mortality rates, early results from the
`phase III trial studying the LDAC and venetoclax combina-
`tion found a much lower response rate of 47% and a median
`OS of 7 2 months only. Furthermore, updated results of
`using an HMA + venetoclax combination in both NPM1 and
`IDH2 wildtype AML patients demonstrated a 2-year OS of
`26 6%.12 Of note, given SCT ineligibility of most patients
`included in the phase Ib/II trials, data about the outcome of
`patients fit to be transplanted are insufficient. Interestingly,
`in 21 patients (out of 145) initially treated with HMAs and
`venetoclax combination and then transplanted,13,14 the med-
`ian OS is reported to be as high as 24 4 months. However,
`the required depth of response and optimal timing for SCT
`after induction with a venetoclax combination still need to
`be assessed.
`Therefore, as long-term data are lacking, and in the
`absence of prospective
`comparisons between intensive
`chemotherapy and venetoclax combinations, it seems that for
`fit intermediate-risk AML patients, outcomes are better when
`intensive chemotherapy rather than venetoclax combinations
`is used.
`
`Treatment of Older Patients with AML
`
`Assuming that our patient has AML with adverse-risk
`cytogenetics would lead to complicated decision-making.
`Available data in patients with advanced age suggest that, for
`patients presenting with a monosomal/complex karyotype,
`the CR rates with intensive chemotherapy are as low as 30–
`35% and become even lower if TP53 is mutated/deleted.4,15,16
`Moreover, early death rates are higher, presumably because
`of prolonged cytopenia
`and the need for
`additional
`chemotherapy.
`Novel agents (e.g., venetoclax + azacytidine) offer at least
`the same CR with reduced early death rates. However, as dis-
`cussed earlier, to achieve a long-term outcome, the feasibility
`and efficacy of allo-SCT after these novel therapies need to
`be demonstrated in prospective studies.
`Older patients who are not candidates for allo-SCT but
`are in remission after chemotherapy are commonly offered at
`least one cycle of intermediate-dose cytarabine. The overall
`results, however, remain disappointing, with a relapse-free
`survival (RFS) of only 16% after 4 years.2
`Historically, maintenance therapy with various agents con-
`sistently failed to demonstrate OS advantage. Exceptions are
`three early studies showing an event-free survival (EFS) bene-
`fit with LDAC17,18 and with IL-2 and histamine combina-
`tion,19 but this has not been adopted in clinical practice,
`particularly in the U.S. Recently,
`the Haemato-Oncology
`Foundation for Adults in the Netherlands (HOVON) pub-
`lished a phase III trial where patients >60 years who received
`intensive chemotherapy and achieved remission were ran-
`domised to azacytidine maintenance or observation alone.
`This study reported an EFS benefit without OS prolonga-
`tion.20 Other studies established an OS benefit for azacytidine
`maintenance, but with some limitations. Oliva et al. pre-
`sented at the 23rd European Hematology Association Con-
`gress a phase III
`trial where azacytidine maintenance,
`preceded by intermediate-dose cytarabine consolidation, led
`to improved OS compared to non-maintenance.21 Notably,
`the number of patients participating in that study was rela-
`tively small. A phase II ECOG-ACRIN trial as an ancillary to
`the large prospective E2906 study included a randomisation
`for patients who achieved CR1 with intensive chemotherapy.
`One year of decitabine maintenance was shown to improve
`OS in the FLT3-ITD-negative subgroup.22 Finally, the most
`exciting breakthrough confirming the value of HMA mainte-
`nance therapy is in the results of the phase III randomised
`placebo-control QUAZAR AML-001 study. In this trial, CC-
`486, an oral formulation of azacytidine, was tested as mainte-
`nance therapy for both de novo and secondary AML patients
`aged 55 and older. The drug was given until relapse or allo-
`SCT, and was reported to double EFS and OS.23,24
`Our suggestion is to treat this patient with a curative
`intent, comprising intensive chemotherapy and, if possible,
`an allo-SCT. If a decision to consolidate her with chemother-
`apy is taken, at least 1 year of post-remission maintenance
`with HMAs should also be considered.
`
`ª 2020 British Society for Haematology and John Wiley & Sons Ltd
`British Journal of Haematology, 2020, 191, 682–691
`
`683
`
`

`

`Frisch et al.
`
`Case 2: Initial therapy in an unfit older patient
`
`An active 74-year-old male presents with AML. His ECOG
`performance status (PS) is 1. He is known to have type II
`diabetes and hypertension, a prior cerebrovascular accident
`(CVA) from which he recovered with no sequelae and
`chronic obstructive pulmonary disease (COPD) with moder-
`ate pulmonary hypertension. His WBC count is 60 000/ll,
`he has no disseminated intravascular coagulation (DIC) and
`no evidence of end-organ injury on admission. Two months
`previously, his blood counts were within normal limits. He
`has a normal karyotype and molecularly is FLT3-ITD-nega-
`tive/NPM1-positive. Is there a preferred targeted approach
`for this patient?
`Older patients with comorbidities have lower CR rates
`and higher mortality with intensive chemotherapy than
`younger patients. Assessing fitness for intensive induction is
`problematic and biased by considerable subjectivity; thus,
`published data need to be carefully scrutinised. In general,
`truly unfit patients have an early mortality rate of approxi-
`mately 30% with standard chemotherapy, as reported by
`real-world data from the Swedish registry.1
`Despite the good PS, his comorbidities, which can be eval-
`uated using
`either Hematopoietic Cell Transplantation
`Comorbidities Index or Charlson Comorbidities Index, are
`predictive of early mortality.25,26 Thus, treatment with inten-
`sive chemotherapy should to be discouraged.
`Few low-intensity options are currently approved for unfit
`newly-diagnosed AML patients. These regimens have not
`been directly compared;
`thus, recommending a particular
`regimen for this patient is based on a careful review of the
`available data.
`
`Venetoclax
`
`BCL2 superfamily proteins regulate mitochondrial apoptotic
`signalling, with BCL2, MCL1 and BCLXL proteins being the
`key anti-apoptotic members. Venetoclax is a potent oral
`specific BCL2-inhibitor which has only a modest activity
`against AML blasts as a single agent, possibly due to MCL1-
`dependent apoptosis evasion.27 However, based on phase I-II
`trials, the combination of venetoclax with either HMA or
`LDAC was approved by the FDA (though not the European
`Medicines Agency) for newly-diagnosed older or unfit AML
`patients.
`DiNardo et al.13,14 demonstrated that a combination of
`venetoclax with HMA in newly-diagnosed patients (median
`age 74 years) who were HMA-na€ıve induced CR/CRi rates of
`67%, compared with the historic rates of 20–25% CR/CRi
`for azacytidine or decitabine when used as monotherapy.28,29
`Moreover, high response rates were reported even in patients
`with poor cytogenetics or P53 mutations, with 60% and 47%
`CR/CRi rates, respectively. With a median follow-up of
`15 months,
`the median OS was not reached in NPM1-
`
`mutated AML patients. With the same duration of follow-
`up, the median OS was not reached if isocitrate dehydroge-
`nase (IDH) 1/2 was also mutated. Early mortality rates were
`as low as 3% in the first 30 days and 8% in 60 days. Inter-
`estingly, MRD-negativity was defined using a low-sensitivity
`3 and was reached in 30% of responsive
`threshold of 10
`patients. An optimal venetoclax dose, when prescribed in
`combination with HMA, was determined to be 400 mg/day,
`with no additive value for 800 mg/day and excessive toxicity
`in the 1200 mg/day dose.
`Another phase II trial examined the combination of LDAC
`(20 mg/day for 10 days) with 600 mg venetoclax. This study
`also included patients who failed previous treatment with
`HMA (about a third of the patients). The CR/CRi rates were
`26 and 28% respectively, but were dramatically influenced by
`presence of secondary AML or prior HMA treatment. For
`patients without prior HMA treatment
`the CR/CRi rates
`were 62%, and the median OS was 13 5 months. Of note,
`patients with NPM1 mutation had favourable response to
`this combination, with 89% CR/CRi rates.24
`Retrospective observations report encouraging results of
`venetoclax combinations even in the relapse setting. Studies
`demonstrate a response spectrum of 21–64%, emphasising
`the heterogeneity of
`these
`select
`reports, with higher
`responses in IDH1/2-mutated AML, intermediate-risk cytoge-
`netics and RUX1-mutated AML.30–32
`The kinetics of response in the newly-diagnosed AML and
`relapse settings are similar. The vast majority of the patients
`who respond to venetoclax-containing regimens do so in the
`two cycles.13,14,24,30 It
`is therefore rational
`to assess
`first
`response after two cycles.
`Venetoclax-containing regimens have a few important lim-
`itations. The major haematologic toxicity is neutropenia,
`which may often be profound. If severe neutropenia is evolv-
`ing while the patient
`is in remission, granulocyte-colony
`stimulating factor (GCSF) support provides a short and min-
`imal benefit;
`therefore, reducing or withholding doses is
`advised. Detailed dose modification varies and relies on min-
`imal data. Published recommendations include withholding
`venetoclax administration until neutrophil recovery and then
`resuming treatment in lower dose (50–75% full dose), reduc-
`ing the HMA/LDAC doses or extending intervals between
`treatment cycles.33 The two latter might be advised when
`marrow hypocellularity or pancytopenia is evident. Such rec-
`ommendations are based on clinical experience/observation
`of dose modifications that lead to peripheral count recovery
`in some patients. Experience in the practical use of veneto-
`clax is crucial, and precise data regarding the long-term effect
`of any modifications on survival and relapse are lacking.
`Venetoclax is metabolised via CYP3A4, and thus strong
`CYP3A4 inhibitors, commonly used for other indications in
`AML patients, may increase venetoclax blood levels. Com-
`monly used azoles, particularly voriconazole and posacona-
`zole, and quinolone antibiotics may have a strong interaction
`
`684
`
`ª 2020 British Society for Haematology and John Wiley & Sons Ltd
`British Journal of Haematology, 2020, 191, 682–691
`
`

`

`with venetoclax. It is recommended to reduce the venetoclax
`dose by 75% when co-administering strong CYP3A4 inhibi-
`tors, like the aforementioned azoles, and by 50% for moder-
`ate CYP3A4 inhibitors.24,34 However, as previously noted,
`these recommendations are based on limited pharmacoki-
`netic data with no clinical evidence supporting the long-term
`effect of these dose modifications.
`A major limitation for venetoclax-containing regimens is
`venetoclax resistance, as lack of response to venetoclax-based
`treatment is usually associated with prolonged cytopenias.
`Moreover, as other options for first-line treatment are avail-
`able, prediction of response to venetoclax-based therapy will
`become very important. Venetoclax resistance is driven
`mainly by overexpression or mutations in MCL1 and BCL-
`XL, the two other anti-apoptotic proteins of the BCL-2 fam-
`ily.35 Recent studies suggested that high BCL2 expression or
`BCL2/MCL1 ratio can predict for good response to veneto-
`clax. It was also shown that HMA and chemotherapy sup-
`press MCL1
`expression
`and
`thus
`synergise with
`venetoclax.36,37 A few recently-published articles demonstrate
`that AML cells with monocytic differentiation respond poorly
`to venetoclax, possibly because of low BCL2/MCL1 ratio, and
`that CD14 can serve as a marker for venetoclax sensitivity.37–
`40 Other pathways/mechanisms which confer venetoclax
`resistance are currently being actively investigated, as are
`other venetoclax combinations, such as those with MEK
`inhibitor, MDM2 inhibitor and FLT3-ITD inhibitors.41–44
`
`Glasdegib
`
`Glasdegib is a specific inhibitor of Smoothened (SMO), a
`receptor regulating the hedgehog pathway. SMO inhibition
`in cell lines is found to reduce the percentage of G0 cells,
`especially in leukaemic stem cells, and to abrogate cytarabine
`resistance. In mouse model, SMO inhibition is demonstrated
`to attenuate leukaemia-initiation potential of AML cells, pre-
`sumably by targeting leukaemic stem cells.45 Based on these
`results, a phase II study was conducted, comparing LDAC to
`LDAC + glasdegib in AML patients and high-risk myelodys-
`plastic syndrome (MDS) patients. The glasdegib + LDAC
`combination was found to have an OS benefit
`in AML
`patients (median OS 8 8 months, compared to 4 9 months
`with LDAC alone), which was consistent in all cytogenetic
`risk groups, albeit less pronounced in high-risk cytogenetics.
`Despite the survival advantage, only 27% patients achieved
`CR/CRi. Notably, 17% of patients who received the combi-
`nation were previously treated with HMAs. Most adverse
`events more common with glasdegib were gastrointestinal
`symptoms, mostly grade I/II.46 Based on these results, glas-
`degib was approved by the FDA for relapsed/refractory
`patients over 75 years or those with comorbidities, and a
`phase III trial for newly-diagnosed patients combining glas-
`degib with either intensive chemotherapy or HMAs was initi-
`ated in 2018.
`
`Treatment of Older Patients with AML
`
`Our priority for this patient is a combination of veneto-
`clax with azacytidine. While LDAC is a perfectly reasonable
`option, we prefer the combination with HMA due to the
`lower dose of venetoclax, used with less risk of drug interac-
`tions.
`
`Case 3: Treatment of an unfit AML patient with
`IDH1/2 mutation
`
`A 77-year-old female presents with normal karyotype,
`NPM1wt/FLT3-ITD-negative/IDH1-positive de novo AML.
`She is known to have hypertension, treated with valsartan/
`amlodipine. She had a hip fracture 4 months ago, for which
`she underwent total hip replacement. She is assisted with a
`walker, and her ECOG PS is 2. On CBC, she is pancytopenic
`with WBC of 2500/ll, absolute neutrophil count (ANC) of
`1200/ll, 72 000/ll platelets, haemoglobin 7 9 g/dl. The
`patient is a widow, and her children live abroad. She is afraid
`that she would not be able to attend frequent visits to the
`clinic.
`Is a total ambulatory care possible in this patient?
`
`IDH1/2 inhibitors
`
`IDH1 and IDH2 mutations are present in about 20% of
`AML cases, predominantly in normal karyotype AML. The
`IDH1 inhibitor ivosidenib and the IDH2 inhibitor enasidenib
`will be discussed together, since their safety profiles and effi-
`cacies are very similar. In relapsed/refractory AML, each of
`these IDH inhibitors is effective as a single agent, with as
`much as 30% CR + CRi rates and a median survival of over
`1 year for CR/CRi responders. Also, many patients who
`achieved only partial response (PR) had prolonged survival
`and became transfusion-independent.47,48 Based on these
`results,
`the
`IDH inhibitors gained FDA approval
`for
`relapsed/refractory AML. Interestingly, patients with FLT3-
`ITD co-mutation and RAS co-mutations respond poorly to
`IDH inhibition. For newly-diagnosed older patients, both
`IDH inhibitors were tested in phase I studies as single agents.
`Ivosidenib treatment achieved 55% ORR and 42% CR/CRh
`rates, with more
`than 60% of
`responses
`lasting for
`12 months. Based on these data, ivosidenib was approved by
`the FDA for this indication.49 Enasidenib achieved 31%
`ORR, with median duration response not reached.50 Adverse
`effects with these drugs are minor. Leukocytosis associated
`with differentiation syndrome is unique but uncommon, and
`possible QT prolongation, particularly with co-administra-
`tion of CYP3A4 inhibitors such as azoles, may also occur.
`IDH inhibitors were also tested in combination with aza-
`cytidine, demonstrating relatively high CR rates of 50–57%
`and ORR of 68–78%. Grade 3–4 neutropenia occurred in
`about 30%, whereas grade 3–4 febrile neutropenia occurred
`in 12% of patients treated with enasidenib and 39% of
`patients treated with ivosedenib combinations.13,51
`
`ª 2020 British Society for Haematology and John Wiley & Sons Ltd
`British Journal of Haematology, 2020, 191, 682–691
`
`685
`
`

`

`Frisch et al.
`
`Given that IDH inhibitors are oral anti-leukaemic agents
`with a remarkable safety profile, this would easily be our
`preference for older unfit patients who clearly elect to avoid
`hospitalisations. The sustained response and easy administra-
`tion of single agent ivosidenib make this an exciting thera-
`peutic strategy. However, in cases where immobilisation is
`less of a problem, venetoclax combinations, very effective
`also in IDH1/2-mutated AML, would be offered to this
`patient, given the overall greater likelihood of achieving a
`longer disease-free interval.
`
`Case 4: Treating an older patient with FLT3-ITD-
`mutated AML
`
`A 75-year-old obese male, known to have diabetes and
`restrictive lung disease with hypoventilation syndrome, pre-
`sents to the emergency room with dyspnoea. On examina-
`tion, his O2 saturation is 86% in ambient air, and chest
`radiography demonstrates bilateral pulmonary infiltrates. His
`CBC shows a WBC count of 140 000/ll with 85% circulating
`monoblasts, 60 000/ll platelets and haemoglobin 8 4 g/dl.
`He is
`treated with supplement oxygen, broad-spectrum
`antibiotics and hydroxyurea 6 g/day. Two days later, his
`WBC is 25 000/ll and his lungs are clear. The molecular
`AML work-up comes back with the diagnosis of normal
`karyotype, NPM1 wt/FLT3-ITD-mutated (allelic ratio 1 5)
`AML.
`FLT3 is a transmembrane tyrosine kinase receptor, which,
`upon binding to its ligand, promotes proliferation and sur-
`vival of normal haematopoietic progenitors through various
`intracellular pathways. Mutations in FLT3, resulting in con-
`stitutive activation of FLT3, are among the most common in
`AML in younger patients and are less frequent in advanced
`age. These mutations can be sub-divided into mutations in
`the juxtamembrane domain (FLT3-ITD mutation) and the
`tyrosine kinase domain (typically D835 point mutation).52
`FLT3-ITD is a well-known adverse prognostic marker, and is
`associated with high proliferation index and reduced RFS
`and OS.53 The poor prognosis is more pronounced if high
`allelic ratio (AR) FL3-ITD mutation (defined as >0 5) is
`noted.54 Guidelines of the National Comprehensive Cancer
`Network and the European LeukemiaNet categorise AML
`with FLT3-ITD mutation as poor-risk AML, based on the
`high relapse risk after intensive chemotherapy, and recom-
`mend allo-SCT as the preferred post-remission treatment. As
`for FLT3-TKD mutations, their prognostic impact is more
`uncertain due to conflicting evidence,55–57 and therefore they
`are not considered currently as poor prognostic markers.
`FLT3-ITD-mutated AML patients treated with low-inten-
`sity regimens with no FLT3 inhibitors have poor responses.
`CR rates of 44% for venetoclax + LDAC and 33% for glas-
`degib + LDAC were reported, which were lower than results
`in patients with no FLT3 mutations. The OS was also infe-
`rior, with only one-third of patients treated with veneto-
`clax + LDAC being alive after a year. Similar results were
`
`686
`
`recently published, demonstrating 53 3% CR/CRi rates and
`median OS of 12 4 months for FLT3-ITD-mutated AML
`receiving venetoclax-based combinations.58 Thus,
`patients
`FLT3-ITD-mutated AML in older patients should be consid-
`ered separately with novel agents. FLT3 inhibitors have a
`substantial role in treatment of AML patients presenting with
`FLT3 mutations. Midosturin, a first-generation FLT3 inhibi-
`tor, was the first to be approved as first-line therapy, in com-
`bination with 7 + 3 intensive chemotherapy, based on CR
`and survival benefit demonstrated in a prospective phase III
`study.59,60 Sorafenib, another first-generation FLT3 inhibitor,
`which is biologically active only against FLT3-ITD and not
`against TKD mutations, was also tested prospectively in com-
`bination with either intensive chemotherapy or azacytidine.
`Addition of sorafenib to chemotherapy during induction and
`consolidation, with use as a single agent in 1-year mainte-
`nance, was found to improve the CR and survival rates in
`FLT3-ITD-mutated young AML patients.61 The same survival
`benefit was demonstrated for FLT3-ITD-mutated AML
`patients aged 60–69, but not in a more elderly group, pre-
`sumably because of adverse effects.62 This is in line with a
`prior report regarding increased early mortality for patients
`>60 years treated with chemotherapy combined with sorafe-
`nib.63
`The combination of sorafenib and azacytidine was tested
`in a phase II trial for newly-diagnosed AML and relapsed/re-
`fractory patients and resulted in considerable response rates
`of 32% for the relapsed patients and 67% for newly-diag-
`nosed ones. Notably, about half of responses to sorafenib
`combined with azacytidine were CRi and not CR.64,65
`Next-generation FLT3 inhibitors are more specific and
`more potent than first-generation inhibitors. Quizartinib was
`tested against chemotherapy for relapsed/refractory FLT3-
`ITD + AML patients in a phase III
`trial, demonstrating
`improved OS with manageable
`toxicity. Over 20% of
`patients
`receiving quizartinib achieved only PR as best
`response considered an event, which might have contributed
`to a less-pronounced EFS advantage.66 Due to the very short
`EFS with this drug, the FDA has not approved it for use yet.
`Gilteritinib is another next-generation FLT3 inhibitor, which
`was tested as monotherapy against intensive and non-inten-
`sive salvage chemotherapy in a phase III trial. Based on
`improved OS in this setting, it was approved by the FDA for
`relapsed/refractory AML. Interestingly, only 21% of patients
`achieved CR/CRh, but nevertheless many others became
`transfusion-independent and/or could be referred to allo-
`SCT. It remains to be seen whether adding chemotherapy or
`a targeted agent to gilteritinib will improve the outcome in
`advanced FLT3-positive AML. These two next-generation
`inhibitors are currently being explored in various combina-
`tions in other scenarios,
`including newly-diagnosed AML
`patients.
`This patient has serious comorbidities that preclude him
`from standard chemotherapy. He also has a highly prolifera-
`tive AML with high allelic-ratio FLT3-ITD mutation, which
`
`ª 2020 British Society for Haematology and John Wiley & Sons Ltd
`British Journal of Haematology, 2020, 191, 682–691
`
`

`

`would argue for the addition of an FLT3 inhibitor. There are
`no good options here, and the therapeutic intent is limited.
`In the absence of a clinical trial and the lack of mature
`data using second-generation FLT3 inhibitors in newly-diag-
`nosed AML, our preferred choice would be to treat him
`with the combination of sorafenib and azacytidine. The
`importance of FLT3 inhibition should be appreciated.
`Recently, DiNardo et al. reported treatment failure in AML
`patients receiving venetoclax combinations, frequently attrib-
`uted to expansion of kinase bearing clones, with domination
`of FLT3 mutation.12 Interestingly, a recent report67 described
`the feasibility of adding an FLT3 inhibitor to venetoclax-
`based regimens, without additive toxicity and with improved
`outcomes, but the results are not mature enough to allow
`clear recommendation. In any event, the monocytic pheno-
`type described in this case may predict
`for venetoclax
`failure.
`
`Case 5: An older patient with treatment-related
`AML
`
`A 73-year-old active female (ECOG PS-1) is diagnosed with
`AML. Eight years ago, she had triple-negative breast cancer,
`for which she received 4 cycles of adriamycin and cyclophos-
`phamide followed by taxol
`for 12 successive weeks. After
`completion of neo-adjuvant therapy, she underwent mastec-
`tomy with negative sentinel lymph node biopsy. She has had
`routine annual oncologic surveillance with no evidence of
`disease. In the past year, she developed mild leukopenia and
`deteriorating anaemia, for which she did not undergo any
`investigation. She now presents with WBC of 5600/ll, 30%
`of which are circulating blasts, 23 000/ll platelets and hae-
`moglobin 6 1 g/dl. Molecular analysis is negative for NPM1,
`FLT3 and IDH mutations. Cytogenetic analysis reveals 47XX
`with trisomy 8. On geriatric evaluation, she is considered fit
`for intensive chemotherapy.
`For this fit patient, is 3 + 7 the best option?
`This case illustrates a fit patient with a secondary AML
`based on prior exposure to anthracyclines and alkylators.
`Her cumulative anthracycline dose is 240 mg/m2, which is
`low enough to allow further anthracycline exposure. The
`probability to achieve CR with intensive chemotherapy is
`estimated to be 50% at most in treatment-related AML4,68
`but with a 2-year OS of only 20%.4
`
`CPX351
`
`formulation of cytarabine and
`CPX351 is a liposomal
`daunorubicin at a fixed molar ratio of 1:5. A phase II trial
`comparing CPX351 to 7 + 3 demonstrated higher response
`rates, particularly more CRi without increased toxicity. In
`secondary AML patients,
`it also demonstrated a survival
`advantage.68 A successive phase III trial conducted in patients
`aged 60–75 years
`reported that
`therapy with CPX351
`achieved better outcomes in treatment-related AML patients,
`
`ª 2020 British Society for Haematology and John Wiley & Sons Ltd
`British Journal of Haematology, 2020, 191, 682–691
`
`Treatment of Older Patients with AML
`
`as well as in AML patients with antecedent MDS/chronic
`myelomonocytic leukaemia and AML with complex kary-
`otype. Toxicity profile of CPX351 is similar to that of inten-
`sive chemotherapy, but early mortality rate is reduced. Time
`will tell if CPX351 is indeed superior to 3 + 7, or if this
`advantage merely reflects optimised anthracycline/cytarabine
`ratio.69 In a prespecified subgroup analysis, two subgroups,
`including patients previously treated with HMAs and patients
`with FLT3-mutated AML, seemed to have no benefit from
`CPX351, the latter despite an acceptable CR rate.70 However,
`the standard deviation in these groups was wide and no
`definitive conclusions could be made. As FLT3 inhibitors
`were prohibited in that study, it is possible that, if used, the
`combination of CPX351 and FLT3 inhibitor would have the
`same advantage in FLT3-mutated AML patients. CPX351 is
`currently the only drug found to be better than standard IC
`for secondary AML patients in a phase III trial.
`Following cardiac evaluation with echocardiography, con-
`firming normal heart function, the patient can be offered
`treatment with CPX351. On