Management of Acute Myeloid Leukemia (AML) in Older Patients

Abstract

Purpose of review:

The acute myeloid leukemia (AML) treatment landscape has rapidly evolved over the past few years. These changes have several implications for the care of older adults (≥ 60 years), who have inferior clinical outcomes. We review decision-making in older adults, focusing on patient- and disease-related factors. We then summarize current treatment options, including multiple recently approved therapies, based on hypothetical clinical scenarios.

Recent findings:

In lieu of using chronological age to determine fitness, we highlight the importance of standardized fitness assessments using geriatric assessments. Next, we review intensive and lower-intensity treatment options in the upfront setting. We focus on multiple newly approved medications, including venetoclax, midostaurin, CPX-351, gemtuzumab, glasdegib, enasidenib, and ivosidenib, and their specific indications. Lastly, we briefly discuss supportive care of older adults with AML.

Summary:

Outcomes of older adults with AML remain poor; fortunately, there are many new promising treatment options. Personalized treatment plans based on patient- and disease-specific factors are essential to the care of older adults with AML.

Introduction

Acute myeloid leukemia (AML) is the second most common leukemia in adults. In 2019, an estimated 21,450 new cases and 10,920 deaths were reported in the United States.¹ Older adults account for most cases, with a median age of 68 years at diagnosis. Incidence rates of AML increase with age from 1.3 per 100,000 in adults < 50 years, to 5.1 per 100,000 in adults 50–64 years, to 20.1 per 100,000 in adults ≥ 65 years. Five-year overall survival (OS) is poor and estimated to be 28.3% for all patients with AML and as low as 3–8% in patients ≥ 60 years.^1,2

The National Comprehensive Cancer Care Network (NCCN) defines older patients as age ≥ 60 years,³ while the European Leukemia Network (ELN) defines older age less rigidly as ≥ 60–65 years.⁴ This aged-based distinction is relevant since AML in older adults is a distinct clinical entity compared to AML in younger patients.⁵ Both disease- and patient-specific differences exist.⁵ Unfavorable cytogenetics and expression of multidrug resistance are more common in older adults.⁵ Also, the incidence of secondary AML (sAML), due to underlying myelodysplastic syndrome (MDS) or therapy for other malignancies, increases with age.⁶ Each of these factors is an independent risk factor for worse prognosis and is associated with poorer response to chemotherapy and lower complete remission (CR) rates compared to younger patients. Comorbidities, as well as cognitive and functional impairments, are more common in older adults and contribute to poorer physiologic reserves and greater treatment-related toxicity.⁷

Between 2000 and 2010, only 40% of patients ≥ 65 years received AML-directed therapy.¹ This treatment disparity was more pronounced in the oldest group of patients (> 80 years); in this group, only 20% received antileukemic treatment.^1,8 Recent studies have shown a trend towards more frequent use of leukemia-directed therapy in adults 65–80 years, which has been associated with improved survival and quality of life.^9,10 With the decreasing treatment disparity in older adults, and the emergence of multiple new therapies, the question of how best to treat AML in older or less fit adults becomes increasingly relevant.

In this article, we highlight various factors clinicians should consider when managing older patients with AML and review the current therapeutic landscape with a focus on novel therapies.

Initial Evaluation of an Older Adult with Newly Diagnosed AML

Treatment Decision-Making & Prognosis

Treatment decision-making is a complex yet critical component of care of all patients with AML and it is arguably more challenging in older patients. Treatment tolerability and benefits are important factors that patients and physicians consider during treatment decision-making.^11,12 Predicting toxicity and treatment success is difficult in older adults. Not only are older adults underrepresented in cancer-related clinical trials, but those enrolled on trials may not be representative of patients seen in practice, thereby limiting generalizability of findings.¹³ Often clinicians use expertise or experience alongside evidence to guide decision-making.¹⁴ However, many currently approved and NCCN-recommended therapies for adults ≥ 60 years are novel, so providers may have limited experience with them.³ Older adults may also value maintenance of independence, time spent at home, cognition, emotional well-being, and health-related quality of life (HRQoL) over survival.⁸ Fortunately, clinical trials have begun to include patient-reported outcomes such as HRQoL measures in addition to survival outcomes.¹⁵

Patient-related factors:

Increasing age is associated with poor prognosis.⁵ This is likely at least partially due to the presence of unmeasured factors such as poor performance status and high comorbidity indexes.^8,16 Performance status is frequently measured using the Eastern Cooperative Oncology Group Performance Status (ECOG PS) scale, which is based on clinicians’ general impression. In patients with poor performance status (ECOG ≥ 3), it may predict functional status and prognosis well; however, this is not consistent across scores.^8,16,17 As such, reliance on clinical gestalt and ECOG performance status to evaluate an older patient is often insufficient. Older patients diagnosed with AML often develop rapid disability due to the acuity of the disease, which further complicates assessments of their health.

A geriatric assessment (GA) is a comprehensive method that can be used to identify vulnerabilities in older adults that may impact their cancer treatment. ^17,18 Components evaluated in a GA include functional status, physical performance and falls, comorbid medical conditions, psychological health, social activity/support, medications, nutritional status, and cognition.¹⁹ Each of these domains is assessed with standardized questionnaires and/or objective measures. For example, functional status can be evaluated by assessing activities of daily living (ADLs) and instrumental activities of daily living (IADLS) and by using performance-based measures such as grip strength or the Short Physical Performance Battery.²⁰ Multiple options for assessments by non-geriatrician have been developed, including screening tools and online geriatric assessment–based calculators, which have been reviewed previously.²¹

Studies have demonstrated that GA can predict prognosis in older patients with AML.¹⁷ The NCCN, the American Society of Clinical Oncology, and the International Society of Geriatric Oncology recommend geriatric assessment in patients ≥ 65 years, especially those receiving chemotherapy, to identify impairments that may be missed by routine oncologic assessment tools.^19,22,23

Disease-related factors:

With advancing age, leukemia cell biology changes which contributes to poor prognosis. Unfavorable risk cytogenetics, such as monosomal karyotypes and complex cytogenetics (≥ 3 acquired chromosome aberrations), are more common with increasing age.^5,24 Secondary AML is also more frequent with increasing age, as described above.^24,25

Molecular genetic testing and conventional cytogenetic analysis are important in the diagnostic work-up for AML.²⁶ Next generation sequencing (NGS) and gene expression analysis have restructured our understanding of AML.²⁷ This is reflected by the incorporation of mutations in the 2016 updated World Health Organization (WHO) AML subtype classifications and the NCCN treatment guidelines.²⁸

Important actionable mutations that may be identified include fms-like tyrosine kinase 3–internal tandem duplication (FLT3-ITD) and isocitrate dehydrogenase (IDH) 1 and 2.²⁶ Additional mutations that provide prognostic information include FLT3 tyrosine kinase domain (FLT3-TKD), CCAAT/enhancer-binding protein a (CEBPA), DNA methyltransferase 3 alpha (DNMT3A), tumor protein p53 (TP53), tet methylcytosine dioxygenase 2 (TET2), Wilms tumor 1 (WT1), DNA methyltransferase 3 alpha (DNMT3A), Nucleophosmin (NPM1), and runt-related transcription factor 1 (RUNX1).²⁶ Early molecular genetic studies included few, if any, older adults²⁹ but more recent data has shown that molecular profiles differ by age.^30,31 For example, in one report of patients with AML and normal karyotype, the frequencies of both NPM1 and FLT3-ITD mutations were shown to decrease with age (NPM1: 50–59 years, 60%; 60–69 years, 43%; ≥ 70 years 41%; FLT-ITD: 50–59 years, 31%; 60–69 years, 25%; ≥ 70 years 20%).³² On the other hand, the frequencies of different mutations including FLT3-TKD and CEBPA did not change with increasing age.³² IDH 1/2 mutations are also more frequent in older patients and often co-occur with NPM1 and FLT3-ITD.³³ Understanding the implications of gene expression profiling is an area of active research. We summarize important genetic mutations in older adults with AML in Table 1.

Table 1:

Common genetic mutations in older adults with AML

Genetic mutations	Prognostic implications and genetic risk stratification group*	Comments	Frequencies
NPM1	• Prognosis is complex and depends on gene-gene interactions4  • Genetic risk stratification group:    ○ Favorable: Mutated NPM1 without FLT3-ITD (or with FLT3-ITDlow †)    ○ Intermediate: Mutated with FLT3-ITDhigh †    ○ Adverse: Wild-type NPM1 with FLT3-ITDhigh †4	• Screen in all patients at diagnosis, mutation defines risk category4  • Specific for de novo AML and most common in cytogenetically normal AML4  • Frequency of mutations decrease with age32	• Mutated in 25–30%92  • Frequency increases with age: 50–59 years, 60%; 60–69 years, 43%; ≥ 70 years 41%32
FLT3-ITD	• Depends on NPM1, see above	• Screen in all patients at diagnosis, mutations are prognostic, and may guide treatment decision-making related to targeted novel therapies (i.e., FLT3 inhibitors)4  • Frequency of FLT3-ITD mutations decrease with age (frequency of FLT3-TKD does not change with age)32	• Mutated in 20–25% of AML in adults32,49,92  • Frequency increases with age: 50–59 years, 31%; 60–69 years, 25%; ≥ 70 years 20%32
CEBPA††	• Genetic risk stratification group: favorable (Biallelic mutated CEBPA)4	• Screen in all patients at diagnosis, mutation defines risk category4	• Biallelic mutations present in 6–10%92  • No change with age32
IDH 1/2	• Genetic risk stratification group: not assigned due to insufficient evidence4	• Screen in all patients at diagnosis, may guide treatment decision-making related to targeted novel therapies (i.e., IDH 1/2 inhibitors)4  • More frequent in older patients32  • Often co-occur with NPM1 and FLT3-ITD4,32,92	• IDH1: 6–14%72,92  • IDH2: 8–19%92  • Frequencies of IDH1/2 increase with age32,33,92
RUNX1	• Poor prognosis; genetic risk stratification group: adverse 4  • Presence in MDS predicts increased risk of progression to AML4  • Associated with secondary AML and resistance to induction92	• Screen in all patients at diagnosis, mutation defines risk category4	• 10% of patients with de novo AML93  • Frequency increases with age92
TP53	• Poor prognosis; genetic risk stratification group: adverse 4  • Presence in MDS predicts increased risk of progression to AML4  • Associated with monosomal and complex karyotypes (combinations have the worst outcomes)4	• Screen in all patients at diagnosis, mutation defines risk category4  • Associated with chemoresistance and high relapse risk94	• TP53 is a tumor suppressor gene that is mutated in many cancers but is rarely mutated in de novo AML: 8% of de novo AML; 27% of therapy-related AML; in older patients with complex karyotypes incidence is up to 70%94

Abbreviations: AML, acute myeloid leukemia; NPM1, Nucleophosmin; FLT3-ITD, fms-like tyrosine kinase 3-internal tandem duplication; IDH, isocitrate dehydrogenase; FLT3-TKD, fms-like tyrosine kinase domain; CEBPA, CCAAT/enhancer-binding protein a; DNMT3A, DNA methyltransferase 3 alpha; TP53, tumor protein p53; WT1, Wilms tumor 1; RUNX1, runt-related transcription factor 1.

^*Genetic risk stratification groups are based on the 2017 European Leukemia Network Classification.⁴

^†Low, allelic ratio <0.5; High, allelic ratio ≥0.5

^††Only AML with biallelic mutated CEBPA is included in the clinicopathologic entity

Several targeted treatment options have recently received Food and Drug Administration (FDA) approval for use in AML with specific mutations, as discussed next.

Treatment

AML treatment is broadly divided into intensive and lower-intensity options. Select trials are shown in Table 2 and 3. Since 2017, eight new targeted therapies have received FDA approval for AML treatment.³⁴ In the setting of these recent approvals, hematologists are faced with additional challenges when deciding on appropriate treatments. Early identification of actionable mutations and the ability to anticipate and manage drug complications are essential.^3,35 As discussed above, information from fitness assessments rather than chronological age alone may assist in predicting outcomes and with treatment decision-making.^13,36

Table 2:

Select clinical trials of intensive therapy in acute myeloid leukemia that are applicable to older adults

Intensive induction regimen	Indication in study	Median age (years)	N	CR (%)	OS	Toxicity	Study
Daunorubicin 90 mg/m2/day with cytarabine	Newly diagnosed AML or high-risk MDS	67	402	64	31% at 2 years	• Longer interval between 1st and 2nd cycle in the higher-dose group (43 vs. 38 days, P=0.001) • No significant difference in the rate of moderate, severe, or life-threatening side effects between the two groups	Lowenberg et al., 200940
vs. Daunorubicin 45 mg/m2/day with cytarabine			411	54 (P=0.002)	26% at 2 years (P=0.16)
Daunorubicin 90 mg/m2/day with cytarabine		53	604	73	59% at 2 years	• More grade 3–4 GI toxicity in the higher-dose group	Burnett et al., 201539
vs. Daunorubicin 60 mg/m2/day with cytarabine			602	75 (P=0.60)	60% at 2 years (P=0.15)
Midostaurin added to cytarabine and daunorubicin (50 mg twice daily for 14 days starting induction day 8)	Newly diagnosed AML with FLT3 mutation	48	360	59	Median: 74.7 months	• Anemia, grade ≥3 higher in the midostaurin group (92.7% vs. 87.8%, P=0.03) • Rash higher in the midostaurin group (14.1% vs. 7.6%, P=0.008)	Stone et al., 201751
vs. Cytarabine and daunorubicin			357	54 (P=0.15)	Median: 25.6 months (P=0.009)	• Nausea (9.6% vs. 5.6%, P=0.05) higher in the midostaurin group
CPX-351 (100 mg/m2 cytarabine + 44 mg/m2 daunorubicin on induction days 1, 3, 5)	Newly diagnosed secondary AML	68	153	48	Median: 9.6 months	• Febrile neutropenia higher in the CPX-351 group (68.0% v 70.9%) • Pneumonia higher in the CPX-351 group (19.6% v 14.6%) • Hypoxia lower in the CPX-351 group (13.1% v 15.2%) • Prolonged time to neutrophil (median 35 vs. 29 days) and platelet count recovery (median 37 vs. 29 days) in the CPX-351 group	Lancet et al., 201855
vs. Cytarabine and daunorubicin			156	33 (P=0.016)	Median: 6.0 months (P=0.003)
Gemtuzumab ozogomycin added to cytarabine and daunorubicin (3 mg/m2 on days 1, 4, 7)	Newly diagnosed AML	62	139	75	41.9% at 2 years	• Persistent thrombocytopenia more common in the GO group (16% vs 3%, P<0.0001) • Veno-occlusive disease (occurred in 3/139 patients, 2 of the 3 patients died) in the GO group	Castaigne et al., 201242
vs. Cytarabine and daunorubicin			139	81 (P=0.25)	53.2% at 2 years (P=0.03)
Gemtuzumab ozogomycin added to cytarabine and daunorubicin (3 mg/m2 on day 1)	Newly diagnosed AML or high risk MDS	67	556	70	25% at 3 years	• Nausea and vomiting, grade 3–4 higher in the GO group (9% vs. 4%, P=0.002) • Hyperbilirubinemia, grade 3–4 higher in the GO group (7% vs. 6%, P=0.001) • Increased number of platelet unit transfusions in the GO group (13.7 vs. 9.6, P<0.001) • Days of intravenous antibiotics longer in the group without GO (19.2 vs. 18.1, P=0.04)	Burnett et al., 201244
vs. Cytarabine and daunorubicin			559	68 (P=0.3)	20% at 3 years (P=0.05)

Abbreviations: AML, acute myeloid leukemia; N, number of patients enrolled; CR, complete remission; OS, overall survival; ‘7+3’, 7 days of cytarabine and 3 days of anthracycline; BSC best supportive care; LDAC, low dose cytarabine; HMA, hypomethylating agent; MDS, myelodysplastic syndrome; FLT3-ITD, fms-like tyrosine kinase 3-internal tandem duplication; HR, hazard ratio; GI, gastrointestinal

Table 3:

Select clinical trials of lower-intensity treatment options in acute myeloid leukemia that are applicable to older adults

Lower-intensity regimens	Indication in trial	Median age (years)	N	CR (%)	Median OS (months)	Toxicity	Study
Low dose cytarabine ± ATRA (20 mg twice daily, for 10 days, every 4 weeks until progression)	Patients not considered candidates for intensive chemotherapy	74	102	18	Not available, improved with LDAC	• No difference compared to BSC	Burnett et al., 200786
vs. Hydroxyurea ± ATRA			99	1 (P<0.05)
Decitabine (20 mg/m2, for 5 days, every 4 weeks until progression)		73	242	17.8	7.7	• Rate of death attributed to AE did not differ between groups	Kantarjian et al., 201256
vs. LDAC or supportive care			243	7.8 (P<0.05)	5.0 (P=0.20)
Azacitidine (75 mg/m2/day for 7 days, every 4 weeks for at least 6 cycles)		70		18	24.5	• No increased toxicity compared to CCR • Fewer hospital admissions compared to CCR (3.4 vs. 4.3 per patient-year) • Higher rate of fever requiring intravenous antibiotics in the CCR group (1.1 vs. 0.6 instances per patient-year)	Fenaux et al., 201058
vs. CCR (IC, LDAC, BSC)				16 (P>0.05)	16 (P<0.05)
Venetoclax + HMA (Venetoclax ramp-up* with decitabine 20 mg/m2/day for 5 days or azacitidine 75 mg/m2/day for 7 days)		74	145	67	17.5	• Common AEs (>30%): GI symptoms, febrile neutropenia and fatigue • AE leading to venetoclax dose interruption (47%) • No tumor lysis syndrome	DiNardo et al., 201960
Venetoclax + LDAC (Venetoclax ramp-up* with LDAC 20 mg/m2/day)		74	82	54	10.1	• Febrile neutropenia (42%) • Thrombocytopenia (38%) • WBC count decreased (34%)	Wei et al., 201961
Glasdegib with LDAC (Glasdegib 100 mg/day on days 1 to 28)		76	88	17	8.8	• Higher pneumonia, grade 3–4 in the combination group (16.7% vs. 14.6%) • Fatigue, grade 3–4 (14.3%)	Cortes et al., 201967
vs. LDAC			44	2 (P<0.05)	4.9 (P<0.05)
Gemtuzumab ozogomycin (6 mg/m2 on day 1, 3 mg/m2 on day 8)	Newly diagnosed CD33 mutated AML	77	118	27	4.9	• No difference in adverse events compared to BSC	Amadori et al., 201669
vs. BSC			119	30 (P>0.05)	3.6 (P<0.05)
Ivosidenib (Dose range, 200–1200 mg daily for 28 days)	Newly diagnosed IDH1 mutated AML	77	34	30	Not available yet	• Diarrhea (53%), nausea (38%) • Peripheral edema (26%) • Differentiation syndrome: leukocytosis and nonspecific symptoms (18%)	Roboz et al., 201987
Enasidenib (Total doses of 50–650 mg daily for 28days)	Newly diagnosed IDH2 mutated AML	77	39	18	11.3 (data for responders not yet available)	• Grade 3–4 cytopenias (21%) • Indirect hyperbilirubinemia (31%) • GI complaints (23%) • Fatigue (18%)	Pollyea et al., 201973

^*Venetoclax ramp-up, daily dose: day 1, 100 mg; day 2, 200 mg; day 3, 300 mg; day 4, 400 mg when combined with HMA or 600 mg when combined with LDAC. Dose adjustments are recommended in specific scenarios.

Abbreviations: AML, acute myeloid leukemia; N, number of patients enrolled; CR, complete remission; OS, overall survival; BSC, best supportive care; CCR, conventional care regimen; IC, intensive chemotherapy; LDAC, low dose cytarabine; HMA, hypomethylating agent; IDH, isocitrate dehydrogenase; FLT3-ITD, fms-like tyrosine kinase 3-internal tandem duplication; AE, adverse events; GI, gastrointestinal; WBC, white blood cell

Any patient diagnosed with AML who is interested in pursuing treatment should be considered for a clinical trial. We will review several potential clinical scenarios of older patients with newly diagnosed AML, summarizing current treatment options and relevant key clinical trials. Figure 1 shows our proposed treatment algorithm for these clinical scenarios.

Figure 1: Treatment algorithm.

A proposed approach for older adults with newly diagnosed acute myeloid leukemia⁹¹

The dotted line represents a controversial treatment option: in fit older adults with poor-risk disease, options include intensive induction therapy with ‘7+3’, CPX-351, or lower-intensity options.

Abbreviations: AML, acute myeloid leukemia; GA, geriatric assessment; “7+3”, 7 days of cytarabine and 3 days of anthracycline; HMA, hypomethylating agent; LDAC, low-dose cytarabine; IDH, isocitrate dehydrogenase; FLT3-ITD, fms-like tyrosine kinase 3-internal tandem duplication

Clinical Scenario 1A: De novo AML in an adult ≥60 years considered a good candidate for induction, without unfavorable cytogenetics, and without CD33-positivity, FLT-3 mutation

Intensive therapy

A traditional induction chemotherapy regimen consists of 7 days of cytarabine and 3 days of anthracycline, often referred to as “7+3”. The goal is to achieve CR, which is successful in 40–60% of adults ≥ 60 years compared to 60–80% in younger adults.⁴

Anthracyclines and cytarabine

Daunorubicin is the most commonly used anthracycline for induction therapy in AML (60–90 mg/m²/day).^3,37 In 2009, daunorubicin 45 mg/m²/day was compared to 90 mg/m²/day for induction in older adults with AML (median age 67 years).³⁸ Higher CR was seen in adults treated with 90 mg/m²/day compared to 45 mg/m²/day (64% vs. 54%, P=0.002), without higher toxicity.³⁸ However, due to concerns about toxicity with the higher dose, daunorubicin 60 mg/m²/day was frequently used thereafter.³⁷ In 2015, daunorubicin 60 mg/m²/day was compared to 90 mg/m²/day (median age 53 years) for induction. The higher dose was associated with increased 60-day mortality (10% vs. 5%, P=0.001), without improved OS at 2 years (60% vs. 59%, P=0.15).³⁷ Idarubicin is another commonly used anthracycline in AML induction. In older adults, idarubicin 12 mg/m² for 3 days has been shown to have similar efficacy and tolerability to daunorubicin 80 mg/m² for 3 days.³⁹ Standard dose cytarabine (Ara-c) is 100–200 mg/m²/day for 7 days.³

Based on these studies, “7+3” induction with cytarabine 100–200 mg/m²/day for 7 days and daunorubicin 60–90 mg/m²/day or idarubicin 12 mg/m²/day for 3 days can be considered in patients ≥ 60 years who are candidates for intensive induction and do not have unfavorable-risk cytogenetics.

Clinical Scenario 1B: De novo AML in an adult ≥60 years considered a good candidate for induction, without unfavorable cytogenetics and with CD33 positivity

Gemtuzumab ozogamicin (GO)

Gemtuzumab ozogamicin (GO), an anti-CD33 conjugate, has been studied as an addition to frontline intensive therapy in patients with de novo AML. A randomized phase 3 trial evaluated the addition of GO (3 mg/m²/day on days 1, 4, and 7 during induction) to intensive chemotherapy in patients with AML aged 50–70 years (median age 62 years).⁴⁰ Addition of GO was associated with improved event-free survival (EFS) [Hazard Ratio (HR) 0.58, P=0.0003], OS (HR 0.69, P=0.036), and relapse-free survival (RFS) (HR 0.52, P=0.0003). GO was associated with higher hematologic toxicity (mainly persistent thrombocytopenia) but not increased mortality.^40,41 Another randomized trial compared adding GO (a single 3 mg/m² dose) to induction chemotherapy in older adults with AML or high-risk MDS (median age 67 years). Compared to induction chemotherapy without GO, 3-year OS was improved in the group who received GO (20% vs. 25%, P=0.05), without additional toxicity.⁴² A third randomized controlled trial (RCT) evaluated the addition of GO (6 mg/m²/day on days 1 and 15) to standard induction therapy in older patients with AML (median age 67 years).⁴³ This trial failed to demonstrate OS benefit. In a subgroup of patients 70–75 years, increased early mortality risk was noted. The aforementioned trials, along with a few other phase III trials evaluating the addition of GO to induction therapy^44–47 showed that GO’s efficacy was observed mainly in patients with favorable risk disease and sometimes those with intermediate risk disease.⁴⁸ Based on these data, in patients with favorable-risk AML and CD33-positivity, the addition of GO to induction can be considered.

Clinical Scenario 1C: De novo AML in an adult ≥60 years considered a good candidate for induction, without unfavorable cytogenetics, with a FLT3-mutation

FLT3 inhibitor

Midostaurin, an oral multitargeted kinase inhibitor, was evaluated in patients with AML (median age 48 years) and a FLT3 mutation.⁴⁹ Midostaurin was added to induction (daunorubin 60 mg/m²/day for 3 days and cytarabine 200 mg/m²/day for 7 days) and consolidation chemotherapy. Although the midostaurin group had similar CR (58.9% vs. 53.5%, P=0.15), OS was longer (74.7 months vs. 25.6 months, P=0.009).⁴⁹ The average patient age at trial entry was < 50 years, limiting the generalizability of these results to older patients. Nonetheless, midostaurin received FDA approval without an upper age-limit.⁵⁰ The NCCN recommends the addition of midostaurin to induction with 7+3 in fit older adults with FLT3 mutations.³

Clinical Scenario 2: Secondary AML or AML with MDS-related changes in an adult ≥60 years considered a good candidate for induction

CPX-351

CPX-351, a liposomal encapsulation of cytarabine 100 mg/m² and daunorubicin 44 mg/m² at a synergistic 5:1 ratio, received FDA approval for the treatment of newly diagnosed sAML in 2017.⁵¹ CPX-351 was studied in the frontline setting in a randomized phase II trial of older patients (median age 68 years) with newly diagnosed de novo and secondary AML.⁵² CPX-351 showed a non-significant trend towards higher remission rates (defined as complete + incomplete remission) when compared to conventional induction therapy (66.7% vs 51.2%, P=0.07). In a subgroup analysis of patients with sAML, CPX-351 was associated with higher remission rates (57.6% vs 31.6%, P=0.06).⁵² A subsequent phase III randomized trial in older patients (median age 68 years) with newly diagnosed high-risk/secondary AML compared induction with 7+3 to CPX-351.⁵³ CPX-351 led to an improved CR rate (47.7% vs. 33.3%, P=0.016) and OS (HR 0.69, P=0.003) compared to conventional 7+3; it was, however, associated with a longer treatment phase and prolonged cytopenia.⁵³

In summary, CPX-351 is recommended for patients ≥ 60 years with sAML, including those with therapy-related disease, an antecedent hematologic disorder, or AML with myelodysplasia-related changes.

Lower-intensity treatment

Clinical Scenario 3: Adult ≥60 years with good functional status but unfavorable-risk cytogenetic or who is not a candidate for (or declines) intensive treatment

In older adults considered fit for intensive therapy with poor-risk cytogenetics or unfit for intensive therapy, lower-intensity treatment options are considered. It is important to note that assigning treatment based on cytogenetics is controversial (i.e., among older adults fit for intensive therapy with poor-risk cytogenetics, intensive therapy is often considered). Lower-intensity treatments are generally given as outpatient therapies with palliative intent and patients are generally not considered candidates for HSCT. Lower-intensity treatments may be administered as monotherapy or combinations, as discussed next.

Hypomethylating agents: Decitabine & Azacitidine

Decitabine (20 mg/m²/day for 5 days every 28 days) was compared with provider’s treatment choice [low-dose cytarabine (LDAC) or best supportive care] in a randomized phase III trial of older patients with AML (median age 73 years).⁵⁴ In initial analyses, decitabine showed a nonsignificant but favorable trend towards increased median OS. Mature survival data showed that the trend eventually reached statistical significance (P=0.037).⁵⁴

Azacitidine (AZA) was studied in older patients (median age 75 years) with AML with > 30% blasts in a randomized phase III study.⁵⁵ Azacitidine was associated with a trend towards better OS compared to conventional care regimens (CCR, standard induction chemotherapy, LDAC, or supportive care only) (10.4 months vs. 6.5 months, P=0.10).⁵⁵ AZA was compared to CCR in another phase III trial, in patients with low bone marrow blast counts of 20–30% (median age 70 years).⁵⁶ In this study, azacitidine was associated with improved OS (24.5 months vs. 16 months, P=0.05), as well as shorter fewer hospitalizations and number of days spent in the hospital.

Both decitabine and azacitidine modestly improve OS compared to BSC. Hypomethylating agents (HMAs) are also more tolerable than intensive induction.

Venetoclax-containing regimens

Based on two key trials, venetoclax, an oral B-cell lymphoma-2 (BCL-2) protein inhibitor, in combination with HMA or LDAC, received FDA approval in 2018 for treatment of older adults with de novo AML considered unfit for intensive chemotherapy.⁵⁷ The trials that led to venetoclax’s FDA approval^58,59 included patients 60–74 years with ECOG PS 0–3, ≥ 75 years with ECOG PS 0–2, or ≥ 60 years with predefined comorbidities (or comorbidities judged to be incompatible with intensive chemotherapy by the treating physician).

Venetoclax + HMA:

A phase Ib escalation trial for older patients (median age 74 years) ineligible for intensive chemotherapy evaluated venetoclax with a HMA (decitabine 20 mg/m²/day for 5 days or azacitidine 75 mg/m²/day for 7 days).⁵⁸ Venetoclax combined with HMA was associated with 73% CR + CRi. CR + CRi rates were 60% in patients with poor-risk cytogenetics and 65% for those ≥ 75 years.⁵⁸ An ongoing confirmatory phase III trial evaluating venetoclax with azacitidine compared to azacitidine alone in patients with AML who were ineligible for IC has recently reported having met the primary endpoints of improvement in OS and composite CR (complete and incomplete). ⁶⁰

Venetoclax + LDAC:

A phase Ib/II study of venetoclax (600 mg daily) with LDAC in older adults (median age 74 years) demonstrated 54% CR + CRi, with a manageable safety profile.⁵⁹ Venetoclax (600 mg daily) with LDAC was compared to LDAC alone in a phase III randomized, multicenter trial of patients with AML who were ineligible for IC (median age 76 years).⁶¹ At the pre-planned median follow-up time of 12 months, median OS was 7.2 months in the venetoclax/LDAC group and 4.1 months in LDAC alone, with a non-statistically significant difference (HR 0.75, P=0.11). The lack of significant difference in OS may be in part due to the higher rates of secondary AML and poorer cytogenetic risks noted in the group treated with venetoclax.⁶¹ Secondary endpoints included CR/CRi which was higher in the venetoclax arm compared to LDAC alone (48% and 13%, P<0.01).⁶¹

Prior to studies on venetoclax, HMA or LDAC alone had been the standard of care of patients not considered fit for intensive therapy.⁶² Venetoclax combination therapies have shown promising results as described above. ^58,59 We will await the final results of the phase III venetoclax/azacitidine trial as well as the final analysis of the venetoclax/LDAC trial with longer follow-up. ^63,64 There are several challenges when using venetoclax-based regimens include mitigation of tumor lysis syndrome, management of severe bone marrow suppression, and optimization of venetoclax dose when other CYP34-inhibiting drugs are utilized.⁶⁵ Data from real-world experience are emerging⁶⁶ and practical guidance are available.³⁵

In conclusion, venetoclax with an HMA or LDAC can be considered for patients with AML who are not candidates for or decline intensive treatment.

Clinical Scenario 3 - Alternative treatment options:

Glasdegib + LDAC

Glasdegib, a hedgehog pathway inhibitor, was studied in a phase II randomized trial for older patients with AML or high-risk MDS (median age 76 years) unfit for intensive chemotherapy.⁶⁷ Glasdegib combined with LDAC was associated with higher CR (17% vs. 2.3%, P<0.05) and improved OS (8.8 months vs. 4.9 months, P=0.0004) compared to LDAC alone, without significant toxicities.⁶⁷ Based on these results, glasdegib received FDA approval for treatment of AML in older or unfit patients.⁶⁸

GO

In patients with CD33-positive AML, GO monotherapy can be considered.³ A randomized phase III study in patients with AML not considered candidates for intensive therapy compared GO to BSC (median age 77 years).⁶⁹ Patients who received GO (6 mg/m² on day 1 and 3 mg/m² on day 8) had an improved median OS of 4.9 months, compared to 3.6 months in the BSC group (HR 0.69, P=0.005) without additional toxicity.⁶⁹

Ivosidenib and Enasidenib

In patients with IDH1/2 mutant AML, ivosidenib and enasidenib are being evaluated.³³ Ivosidenib and enasidenib target IDH1 and IDH2 mutations respectively and were initially approved in the relapsed/refractory (R/R) setting.^70,71

An ongoing phase 1 dose escalation study of ivosidenib has enrolled 34 older patients who were not candidates for intensive treatment (median age 77 years).⁷² ORR was defined as CR, CRi, partial response or morphologic remission. Patients who received ivosidenib had a CR of 30% and ORR 55%. Treatment was well-tolerated: grade 1–2 toxicities were most common (~25% of patients) and were predominantly gastrointestinal (GI) symptoms, fatigue and cytopenias. IDH differentiation syndrome occurred in 16% of patients.⁷² IDH differentiation syndrome is a rare syndrome theorized to occur after rapid neutrophil recovery; it presents with significant neutrophil-predominant leukocytosis and nonspecific symptoms such as fever, hypotension, and intra-thoracic effusions.⁷²

Enasidenib was evaluated in a phase I/II trial of older patients (median age 77 years) with AML unfit for intensive treatment.⁷³ ORR was 30.8% and CR was 18%. Toxicity included indirect hyperbilirubinemia, GI complaints and fatigue. Grade 3–4 cytopenias occurred in 21% of participants. Overall, enasidenib was found to be safe and well-tolerated.⁷³

The NCCN recommends considering ivosidenib and enasidenib, respectively, for newly diagnosed patients with IDH1 and IDH2 mutations.³ Ivosidenib has received FDA approval in the upfront setting for patients with IDH1 mutant AML who are not suitable candidates for intensive therapy.⁷⁴

In summary, in the absence of actionable mutations, options include LDAC (monotherapy, with glasdegib, or with venetoclax), HMA (monotherapy or with venetoclax), or best supportive care. In the presence of IDH1 mutation, IDH2 mutation, and CD33-positivitity, alternative options include ivosidenib, enasidenib and GO, respectively.

Post-remission therapy

Relapse rates for older adults with AML who achieve CR are high which contributes to poor overall survival.⁷⁵ More aggressive disease biology contribute to higher relapse rates in older compared to younger patients.^5,76 Post-remission therapy options include allogeneic hematopoietic stem cell transplant (HSCT) and/or conventional chemotherapy (e.g., high dose cytarabine); both were more commonly utilized after intensive induction therapies. Patient- and disease-related factors must be re-evaluated at the time of decision-making for consolidation regimens. Some patients who receive intensive induction may experience significant treatment toxicity precluding further intensive treatment.

Allogeneic HSCT is recommended for older adults with poor-risk AML in first remission who are deemed appropriate candidates for HSCT.³ Historically, treatment toxicity has been considered prohibitive to the use of allogeneic HSCT in older patients.⁷⁷ However, advances in HSCT such as the use of reduced-intensity conditioning regimens have been developed to decrease toxicities. Devine et al conducted a study of 114 patients (median age 65 years) who received transplantation following reduced-intensity conditioning. Survival was compared to historical cohorts treated without HSCT. OS was improved (48%, P<0.05) and complications including non-relapse mortality (15% at 2 years) were lower than expected.⁷⁷

For patients who are not candidates for allogeneic HSCT, chemotherapy should be considered. Various regimens are used in clinical practice and are recommended by the NCCN which may include intermediate dose-cytarabine (1.0–1.5 g/m²/day for 4–6 days for 1–2 cycles). Addition of targeted therapies such as GO and midostaurin is recommended for CD-33 positivity and FLT-3 mutations, respectively. For those with secondary and therapy-related AML, consolidation with CPX-351 should be considered if used for induction.³

Health-Related Quality of Life and Hospitalization

Historically, overall survival has been the most measured outcome in AML trials. More recently, interest in patient-centered outcomes such as HRQoL, has increased markedly. HRQoL incorporates physical, mental, emotional, and social functioning domains; however, it is often difficult to quantify since there are no universally accepted measures of HRQoL. We recently provided a summary of different HRQoL tools that have been incorporated in trials of patients with AML.¹⁵ HRQoL assessments provide critical information that can influence decision-making. In the context of the multiple newly approved treatments, many with otherwise comparable efficacy and tolerability, HRQoL may be the differentiating factor.

Patients with AML have significant morbidity, often requiring inpatient admissions.⁷⁸ Common causes for admission include urgent evaluation following a new diagnosis, managing acute illness, induction chemotherapy, complications from outpatient therapies, and supportive care. Based on a national inpatient sample database, patients ≥ 60 years with AML often have shorter lengths of stay (LOS) than those < 60 years (5.4 vs. 6.8 days). Factors associated with shorter LOS in patients ≥ 60 years include older age, higher income, and admission to nonteaching hospitals.⁷⁸ The lower length of stay in older adults may be due to lower utilization of AML-directed therapy and higher in-hospital mortality.⁷⁸ Understanding factors associated with hospitalization and length of stay can help with treatment decision-making as the ability to stay at home and spend time with family may be important to older adults.⁷⁹

Supportive care: Growth Factors

Neutropenia is a common cause of significant morbidity in patients with AML. It is multifactorial, due to the disease itself and the myelosuppressive treatments, particularly traditional “7+3” induction and venetoclax.⁶⁵ Myeloid growth factors such as granulocyte-colony stimulating factor (G-CSF) have been extensively studied in patients with hematologic malignancies.⁷ Their benefits include accelerated neutrophil count recovery, decreased duration of fever, and decreased length of hospital stay. However, when studied in older patients with AML, G-CSF did not portend survival benefits.^7,80,81 There are also theoretical concerns about stimulation of leukemic cell growth and cost-effectiveness.⁷ The routine use of G-CSF is not recommended; however, G-CSF is sometimes used during consolidation therapy when remission has been achieved.

Supportive care: Exercise, Cognitive Behavioral Therapy, and Palliative Care

Patients with cancer are at high risk of functional and cognitive declines due to the disease itself and its treatment.⁸² Older patients are most at risk for these consequences, and lower levels of physical activity have been associated with more deconditioning in this patient population. Exercise helps with symptom control, fatigue, and quality of life during and after treatment.^83,84 In a systematic review of exercise in adults with hematologic malignancies, aerobic exercise was associated with positive effects on mood and fatigue.⁸⁵ A recent trial of a supervised inpatient exercise program was conducted in patients with AML admitted for induction (median age 53 years).⁸⁴ This 4-week mixed-modality program resulted in improved patient-reported fatigue and improved physical function compared to placebo. Despite the small sample size (n=17), results of this study are consistent with previously reported studies in patients with leukemia.^86,87 Increased awareness about and promotion of the benefits of exercise among providers are important as patients are more likely to exercise if recommended by their oncologist.⁸³ Randomized controlled trials are needed to determine the appropriate type, timing and “dose” of exercise.

Symptoms of cancer-related cognitive decline have been shown to improve with cognitive behavioral therapy (CBT).⁸² An RCT is currently being conducted in adults with AML and lymphoma to evaluate the role of CBT in mitigating cancer-related fatigue, an almost universally reported symptom.⁸⁸

Given the poor outcomes associated with AML in the older adult population, incorporation of palliative care services should be considered.⁸⁹ The integration of palliative care with typical oncology care has recently been shown to be beneficial in a multi-site randomized control trial of patients > 60 years with AML undergoing IC.⁹⁰ The palliative care intervention group included at least twice weekly visits by a palliative care provider during each hospitalization. Patients were found to have improved QOL as well as psychological distress. Intervention effects were significant at 2 weeks and sustained up to week 24.⁹⁰ The integration of palliative care may help patients mitigate the many challenges they face related to symptoms and psychological distress throughout their disease.

Conclusion

In an aging population, the prevalence of AML in older adults is rising. Although treatment disparities are decreasing in patients 60–80 years, studies are still lacking in patients > 80 years. Treatment of AML in older adults is complex. Whenever possible, patients should be included in clinical trials,⁹¹ which will improve our understanding of AML therapies. The advent of multiple novel agents is promising, yet oncologists/hematologists now face new challenges. These include how best to assess fitness and anticipate risks and benefits of treatment. Comprehensive assessments such as GA predict functional reserve and assess overall health more accurately than chronological age alone. Providers and patients are also discussing and considering broader definitions of “treatment benefit” with a focus on measures beyond longevity, such as HRQoL and length of hospitalization. Personalizing both therapeutic and non-therapeutic treatments is critical given the heterogeneity of AML and older patients.

Key References	Brief Description
** Ref 8: Klepin et al.8	Discusses the changing treatment landscape of AML in older adults and associated issues
* Ref 16: Klepin et al16	Predictive value of GA in older adults with AML
** Ref 18: Loh and Klepin18	Role of GA in personalized care of older adults with AML and discussion of future directions
* Ref 35: DiNardo et al35	Review of treating AML in the era of novel agents, not specific to older adults
** Ref 53: Lancet et al53	Trial demonstrating efficacy of CPX in sAML
** Ref 58 and 59: DiNardo et al58, Wei et al59	Trials that led to Venetoclax-based therapy becoming approved for older patients and those who are not candidates for intensive induction
*Ref 72 and 73: Roboz et al72, Pollyea et al73	Key trials of ivosedinib and enasidenib for use in newly diagnosed IDH1/2 mutant AML