Acute Myeloid Leukemia Treatment in the Elderly: A Comprehensive Review of the Present and Future

Acute myeloid leukemia (AML) is a heterogeneous cluster of myeloid neoplasm that remains a challenging disease to treat. It is a disease of the elderly with a median age of around 67 years, which limits the use of intensive chemotherapy and allogeneic transplant (allo-HSCT) due to poorer tolerance in older patients. As such, Improvements in complete remission (CR) and overall survival (OS) rates have paled in comparison with younger adults [1, 2]. Much of this difference is attributed to factors specific to older adults and features unique to leukemia in this population, such as higher prevalence of adverse cytogenetics and molecular markers and antecedent hematological disorders [3, 4]. While it has been generally accepted that intensive chemotherapy with cytarabine and an anthracycline is beneficial in elderly patients, many adults do not tolerate such regimens and succumb to morbidity from disease progression, drug resistance, or toxicity. Hence, a new model for therapy moving forward must account for more precise metrics of tolerability and benefit in elderly adults with AML. Indeed, the most important consideration prior to therapy is to estimate the benefits and risks associated with any regimen for the elder individual. With this in mind, we review the use of established therapy and new agents to pave a roadmap for the future of AML treatment in elderly patients. In learning about the multitude of agents and clinical trials, we posit the view that a single “gold-standard” (i.e., 7 + 3 induction) for AML is not the canonical paradigm. This once-slated backbone of AML therapy continues to diverge into several parallel backbones of treatment.

Two main risk factors contribute to poor prognosis in the elderly with AML: treatment-related mortality and resistance to therapy [5, 6]. Due to a higher burden of comorbidities, poor performance status, and impaired ability to clear chemotherapy, older individuals are less tolerant of aggressive treatments and more susceptible to therapy-related adverse events. The standard 7 + 3 induction regimen typically results in inferior survival in patients older than 75 years of age [1, 7]. Performance score is of particular importance as elderly patients with ECOG ≥3 have significantly elevated risk of death with intensive induction chemotherapy (IC) [1]. Accordingly, the NCCN guidelines now recommend against using intensive regimens in elderly patients with performance score >2 [8]. AML has distinct cytogenetic and molecular features that cluster in the elderly populations: they are more likely to have unfavorable cytogenetic abnormalities involving changes in chromosomes 5 and 7 or complex chromosome abnormalities and less likely to carry favorable cytogenetics such as t(8;21) and inv(16) compared to younger cohorts [1, 9]. Older adults are also more likely to present with secondary AML (sAML) arising in the setting of myelodysplastic syndrome (MDS) or AML related to prior chemotherapy, additional independent risk factors for decreased responsiveness to standard therapies. Higher proportions of sAML also carry deleterious TP53 mutations, which portend particularly worse prognosis [10, 11]. Furthermore, overexpression of the multidrug resistance 1 (MDR1) protein is more prevalent in the elderly at up to 87% of patients, resulting in a higher rate of chemotherapy efflux by P-glycoprotein and subsequent chemotherapy resistance [12]. Mutation rates in genes of prognostic significance such as nucleophosmin 1 (NPM1) and the Fms-like tyrosine kinase 3 (FLT3) also differ between the young and older populations. FLT3-internal tendum duplication (ITD), for example, may not be associated with inferior prognosis in patients aged ≥60 years [4, 13, 14]. Use of these modalities should be brought into standard practice when choosing therapies that can confer the most benefit to patients.

Novel treatments for AML include agents that can be used complementary to intensive chemotherapy, alternatives to induction regimens, and post-induction therapy. A brief review of the data and response rates of conventional chemotherapeutic regimens helps argue for the development and use of more selective agents that are not toxic to a wide variety of cells. The major conventional therapies reviewed below include IC with the antimetabolite/anthracycline combination (cytarabine/daunorubicin), clofarabine, and more recently hypomethylating agents (HMA) in combination with venetoclax.

It is well established that IC can be a viable therapeutic option for carefully selected elderly patients. In fit patients without significant comorbidities, IC can result in increased OS and response rate in those aged 70 years or older [15, 16]. Nonetheless, the regimen is not a designated “standard of care” and is often not indicated in many elderly patients, especially in those unfit for chemotherapy due to comorbidities. In the past trials, the standard 7 + 3 induction regimen resulted in an inferior median survival of <1 year in patients older than 75 years of age, and these patients had a very high likelihood of dying within 30 days of induction [1, 7]. Furthermore, older patients with poor performance score of ECOG ≥3 had a mortality rate of 82% compared to 29% in younger patients with the same performance score [1]. The rates of CR with 7 + 3 induction vary widely in elderly adults (35–60%) and this range drops significantly over time, with no more than 10–20% of patients in their first CR living beyond 3 years from initial diagnosis [10, 11, 17].

Another high-intensity regimen using high-dose cytarabine (HDAC) plus mitoxantrone has been attempted in patients who are older than 60 years of age and/or have at least two adverse prognostic factors [18]. 27% of patients with Charlson comorbidity index (CCI) >2 had HDAC dose reduced to 2 g/m². The CR/CR with incomplete hematological recovery (CRi) was 45% overall and 35% in patients >70 years of age. The induction death rate was 9% in all participants but was significantly higher in patients >70 years of age at 29% [18]. Recently, the same group showed in a phase I study that adding azacytidine (AZA) to HDAC/mitoxantrone combination for epigenetic priming can improve the response rate in the similar patient population, but again, the question of tolerability in elderly population remained [19]. Hence, aggressive therapeutic strategy may be considered for selected fit elderly adults with good-risk AML to prolong survival, or even in intermediate-risk AML as curative intent with eventual hematopoietic stem cell transplant, but careful consideration of performance status, comorbidities, cognitive functions, and social situations is critical. Several prognostic models are available to risk stratify and predict outcomes of patients undergoing induction therapy, which can help select patients who will benefit from the intensive regimens [20‒23], as discussed later in this article. Use of additional novel agents together with the standard therapies guided by molecular analysis of the individuals may enhance outcomes in the elderly who can tolerate them. In older patients unable to proceed with allo-HSCT, alternative induction therapy should be explored.

A new intensive chemotherapy added recently to the arsenal of AML induction therapy is CPX-351 (Vyxeos). It is a liposomal formulation of cytarabine and daunorubicin in a fixed 5:1 molar ratio that has shown promising results among elderly patients with sAML or treatment-related AML. The formulation optimizes the synergistic drug ratio and prolongs the intracellular drug exposure time. The liposome carrier also allows for bypassing of drug efflux pumps to help the chemotherapy overcome resistance mechanism developed by leukemic cells [24, 25]. In a phase II study by Lancet and colleagues, CPX-351 produced higher response rates compared to traditional 7 + 3 treatment in 126 patients aged 60–75 years with newly diagnosed AML, in both the entire study population (57.6% vs. 31.6%, p = 0.06) and in the sAML subgroup (57.6% vs. 31.6%, p = 0.06) [26]. Although differences in OS and event-free survival (EFS) were similar between two groups, a further planned analysis showed a statistically significant improvement in OS (12.1 vs. 6.1 months, p = 0.01) and EFS (4.5 vs. 1.3 months, p = 0.08) in favor of the CPX-351 cohort in sAML patients [26]. This finding led to opening of a phase III randomized trial to investigate the efficacy of CPX-351 in elderly patients with untreated AML with a history of antecedent cytotoxic therapy, MDS, chronic myelomonocytic leukemia, or AML with MDS-related cytogenetic abnormalities (AML-MRC) [27]. Compared to the 7 + 3 regimen, CPX-351 treatment resulted in a superior mOS (9.56 vs. 5.95 months; HR = 0.69; p = 0.005) and CR rate (47.7% vs. 33.3%; p = 0.016). 34% and 25% of patients underwent allo-HCT following the CPX-351 and 7 + 3 treatments, respectively (p = 0.098), and the exploratory landmark survival analysis from the time of allo-HCT favored CPX-351 (HR 0.45, p = 0.009) [27].

A recently published 5-year follow-up result of the phase III trial confirmed improved mOS in the CPX-351 group at 9.3 months versus 5.95 months in the 7 + 3 group (HR 0.8) with 5-year OS of 18% versus 8%, respectively [28]. It is important to recognize that even though the approval is for all adult population, the study only included patients 60–75 years of age. A retrospective analysis demonstrated lower response rates (CR/CRi 43.5%) and shorter OS (5.2 months) following CPX-351 therapy in younger patients compared to results seen in the phase 3 trial involving older patients 60–75 years of age [29]. A clinical trial to evaluate the efficacy of CPX-351 in patients younger than 60 is ongoing (NCT04269213), as well as a phase III trial comparing CPX-351 to standard intensive chemotherapy in all adult patients with newly diagnosed AML with intermediate or adverse genetics (NCT03897127). A phase I trial combining CPX-351 with GO in AML patients 55 years or older is also active and recruiting patients, despite initial concern for significant dose-limiting toxicities (NCT03878927).

Clofarabine is a second-generation purine nucleoside analog that has been investigated in elderly patients not fit for IC. Upon cellular entry clofarabine is phosphorylated to its active triphosphate form by cellular kinases and inhibits DNA polymerase and ribonucleotide reductase [30]. Burnett and colleagues treated 106 treatment-naive AML patients who are either older than 70 years or 60–69 years of age unfit for IC with single agent clofarabine. The CR/CRi rate was 48%, but 18% of patients died within 30 days. Interestingly, CR and OS rates were similar in adverse and intermediate cytogenetic risk groups [31]. In the subsequent study with a similar patient population, clofarabine nearly doubled ORR and CR rates compared to low-dose cytarabine (LDAC) (22% vs. 12% and 38% vs. 19%, respectively), but it did not result in a significant improvement in the OS [32]. In a phase II study combining clofarabine and cytarabine in 60 patients aged 50 years and older with previously untreated AML or high-risk MDS, the ORR was 60%, demonstrating its efficacy comparable to conventional induction regimens [33]. However, again, the OS did not significantly improve compared with other regimens. Another phase II study involving patients over the age of 60 demonstrated that a lower intensity regimen consisting of clofarabine and LDAC alternating with decitabine (DEC) was well tolerated and highly effective. The CR was 60% and median OS was 11.1 months, with main adverse events of nausea, rash, and elevated hepatic enzymes [34]. A more recent UK NCRI AML16 trial of 806 patients with AML or high-risk MDS (median age 67 years) randomized treatment to clofarabine plus daunorubicin or cytarabine and daunorubicin. The ORR was 69% with no differences between two groups (66% vs. 71%), and there were no differences in 30 and 60 day mortality rates as well. Hence, the results presented by this study provide no supporting evidence to suggest that cytarabine should be replaced by clofarabine [35].

HMAs are commonly used to treat older patients who are not fit for intensive induction therapy. AZA and DEC are cytosine analogs that are phosphorylated inside a cell by deoxycytidine kinase. Their incorporation into DNA and RNA results in DNA methyltransferase inhibition, disassembly of polyribosomes, and inhibition of peptide synthesis, leading to reversal of gene silencing secondary to aberrant hypomethylation [36]. AZA has shown clinical activity in older patients with AML and high-risk MDS. In a large phase III clinical trial comparing AZA versus conventional care regimens, which included intensive IC, LDAC, or supportive care, in patients >65 years with AML, the AZA group had higher OS of 10.4 months versus 6.5 months (p = 0.0829) and improved 1-year survival rates of 46.5% versus 34.2% [37]. The ORR of AZA typically ranged from 30 to 60% across multiple studies [37‒39]. DEC has also showed similar efficacy in older patients. When administered on a 5-day schedule in 28-day cycles in AML patients >65 years, DEC achieved significantly higher CR rates of 18% compared to 7.8% with conventional therapy (p = 0.037), as well as a longer mOS of 7.7 months versus 5.0 months (p = 0.108) [40]. HMAs appear to generate higher response rates among patients with high-risk cytogenetics and molecular profiles such as TP53 mutation. DEC administered at a dose of 20 mg/m² for 10 days per cycle resulted in an ORR of 67% in patients with unfavorable-risk cytogenetics and 100% in those with tp53 mutations [41]. However, 10-day regimen did not prove to be more efficacious compared to the 5-day regimen in older AML patients overall [42]. The result from an international open-label randomized phase III trial comparing the efficacy of 10-day DEC versus intensive IC regimen followed by allo-HCT has recently been presented at the 27th Congress of the European Hematology Association (EHA). The trial enrolled 303 patients in each arm aged 60 years or older with median age of 68, and 34% of patients were ≥70 years old. The CR/CRi rate was 48% with DEC and 61% with IC, and 40% and 39% of patients underwent allo-HSCT, respectively. The OS was not significantly different between the two arms: 15 months in the DEC and 18 months in the IC groups (HR = 1.04, p = 0.68), while the incidence of grade 3–5 adverse events, such as febrile neutropenia, oral mucositis, and thrombocytopenia, was higher in the IC arm [43] (NCT02172872). As of today, AZA is approved by the European Medicines Agency Committee for AML with 20–30% bone marrow blasts arising from MDS, and DEC is approved for patients older than 65 years with AML who are not considered candidates for standard therapy.

The addition of venetoclax to HMAs strikingly improved outcomes in the battle to treat AML. Venetoclax (ABT-199) is a second-generation BH3 mimetic that selectively inhibits B-cell lymphoma-2 (BCL-2), a family of proteins that regulate apoptosis [44]. It has proven to be one of the most successful novel therapies that have been developed in the past decade particularly for older and less fit patients with AML. An initial venetoclax monotherapy in patients with R/R AML or newly diagnosed AML who are unfit for standard therapy showed a limited response rate of 19% [45]. A follow-up phase Ib clinical trial was conducted to evaluate the safety and efficacy of the combination of venetoclax with DEC or AZA in patients with newly diagnosed AML aged >65 years with intermediate or poor prognostic cytogenetics who were unfit to receive IC [46, 47]. The combination therapy achieved a remarkable CR/CRi rate of 67%, with a median duration of response (mDOR) and mOS of 11.3 and 17.5 months, respectively. These results were significant improvements from historical outcomes seen with HMA monotherapy, where CR rates remain around 15–25% and mOS has been limited to less than 12 months [37, 38].

Venetoclax is also effective when combined with LDAC as demonstrated in a phase I/II dose-escalation and dose-expansion trial involving treatment-naive patients aged ≥65 years [48, 49]. CR and CRi rates, mDOR, and mOS in this venetoclax-LDAC study were 64%, 8.1 months, and 10.1 months, respectively [49]. DEC, AZA, and LDAC showed similar response rates when administered with venetoclax, although the LDAC combination resulted in a lower response rate of 35% compared with 57% and 79% with DEC and AZA, respectively. mOS was also longer for DEC and AZA than for LDAC [47, 49]. Inferior outcomes in the LDAC trial likely reflect its distinct study population that included patients with sAML for whom HMAs had previously failed for MDS [46]. In both studies, venetoclax was well tolerated, with the most common adverse events (AEs) being nausea, diarrhea, thrombocytopenia, and febrile neutropenia. A phase III trial comparing the efficacy of AZA versus AZA-venetoclax has been completed, and its results reaffirm significantly higher CR rates and longer mOS for the venetoclax group [50]. The median patient age was 76 years in both groups. The AZA-venetoclax group demonstrated significantly longer at 14.7 months versus 9.6 months in the control AZA only group (p < 0.001), as well as higher composite CR rate at 66.4% versus 28.3% (p < 0.001). AEs of grade 3 or higher were more frequent in the AZA-venetoclax group notably thrombocytopenia (45% vs. 38%), neutropenia (42% vs. 28%), febrile neutropenia (42% vs. 19%), and leukopenia (21% vs. 12%), but the 30-day mortality rate was similar in the two groups (7% vs. 6%) [50]. Based on these data, the U.S. Food and Drug Administration (FDA) granted approval for the use of venetoclax in combination with AZA, DEC, or LDAC to treat patients with newly diagnosed AML who are aged ≥75 years or unfit for IC. Kadia et al. [51] have conducted a trial with a modified version of venetoclax combination therapy. Patients over the age of 60 or those unfit for IC were treated with alternating regimen of venetoclax plus cladribine (CLAD) and LDAC for initial two 28-day courses, followed by consolidation with venetolax plus AZA for 2 additional courses, for the total of 18 courses. A total of 60 patients with the median age of 68 years were recruited. The rate of CRc was 93% with 84% MRD negativity rate, and the mOS and DFS have not been reached after a follow-up of 22.1 months. The regimen was well tolerated with 1 death in 4 weeks.

Outcomes of HMA with venetoclax have been compared to IC in older/unfit patients with AML. A retrospective analysis of 10-day DEC with venetoclax versus IC showed a significantly higher CR/CRi rate in the (81% vs. 52%, p < 0.001), a lower relapse rate (34% vs. 56%, p = 0.01), lower 30-day mortality (1% vs. 24%, p < 0.01), and longer OS (12.4 months vs. 4.5 months, p < 0.01) [52]. A retrospective study compared HMA/venetoclax with CPX-351 in elderly or unfit patients with AML. The CR/CRi was 56.4% in the HMA/venetoclax cohort and 47.2% in the CPX-351 arm, and the mOS was similar in both cohorts at 13.8 months versus 11.1 months (p = 0.82), as well as early mortality rates at day 60 (4.9% vs. 3.8%). The CPX-351 included more patients with AML with MRC and adverse risk molecular changes [53]. Another larger retrospective analysis comparing the outcomes of 217 patients receiving CPX-351 to 439 patients receiving AZA/venetoclax produced similar results. The mOS did not differ based on therapy (13 months for CPX-351 vs. 11 months for AZA/venetoclax; p = 0.22), and early mortality was similar (10% vs. 13% at 60 days). However, infection and febrile neutropenia rates were higher with CPX-351 as well as longer hospital stay [54]. Overall, these trials suggest that HMA/venetoclax represent a reasonable frontline therapeutic choice and prospective randomized trials are justified.

These novel regimens provide potential intermediate-level intensity approach to treat AML in elderly populations with impressive outcomes and warrant follow-up in a larger trial. Building on these encouraging findings, venetoclax is being studied in combination with other small molecule inhibitors such as IDH1/2 inhibitors and FLT3 inhibitors, and we will further discuss this aspect with novel therapies.

Various studies have demonstrated the prognostic value of minimal residual disease (MRD) in AML patients treated with intensive chemotherapy. The detection of MRD has been associated with significantly higher relapse rate and shorter OS [55‒57]. The role of MRD in lower intensity therapy is a fairly new concept and has been recently investigated. Maiti et al. [58] assessed MRD in older, unfit patients with AML treated with the regimen of 10-day DEC and venetoclax. MRD was evaluated in bone marrow specimens using multicolor flow cytometry (sensitivity 0.1%). The CR/Cri rate was 85% with the MRD negative rate of 54%, and the median time to MRD negativity was 2 months. MRD negative status resulted longer RFS (NR vs. 5.2 months, p = 0.004), EFS (NR vs. 5.8 months, p < 0.001), and mOS (25.1 vs. 7.1 months, p < 0.001) compared to MRD positive patients. The MRD status and its prognostic relevance were also explored in patients treated with venetoclax and AZA in the VIALE-A trial. The MRD negative response of <10⁻³ was achieved by 41% of patients. In these patients, the median DOR, EFS, and OS were not reached, and the 12 months DOR, EFS, and OS were 81.2%, 83.2%, and 94%, respectively. The multivariate analysis showed that CR with MRD less than 10⁻³ was a strong predictor of OS (adjusted HR, 0.285, p < 0.001). It was also noted that approximately the same proportion of patients became MRD negative after cycles 1, 4, 7, and beyond 7, recognizing the delay in best outcome that can be achieved with this regimen [59]. These retrospective analyses underscore the importance of measuring MRD status, and achieving CR with MRD-negativity should be the goal of therapy in future trials. It will be particularly important for patients who are older but fit enough to be considered for hematopoietic cell transplant, as the MRD status at the time of transplant was shown to be an independent risk factor for relapse post-transplant [60].

An allo-HSCT is considered a part of consolidation therapy in AML, especially for intermediate and high-risk AML. Given the frailty of elderly patients due to comorbidities, a decision to transplant must be made carefully by assessing treatment-related morbidity from allo-HCT. Historically, due to intensity and risk of allo-HSCT, only a fraction of patients older than 70 years old utilized allo-HSCT, but the number has increased markedly in the past decades, rising from 0.1% of transplants in 2000 to 3.85% by 2013 [61] and the number keeps increasing every year since. Better understanding of transplant complications, increasing usage of unrelated donors, and development of non-myeloablative conditioning strategies helped to improve transplant outcome and survival over time, leading to an approximate 2 year survival rate of 40% in the recent decade compared to around 25% in the 2000s [61]. Low-intensive chemotherapy such as HMA/venetoclax can serve as bridging therapy to allo-HSCT. Approximately 10–40% of patients who achieve a response proceed to have allo-HSCT and around 40% of patients can achieve durable remissions, with OS significantly greater among transplanted patients compared to those who were potentially eligible but did not undergo transplant [62, 63].

Nevertheless, challenges remain for older and frailer populations as treatment-related toxicity remains significant. In patients who underwent allo-HSCT for AML between 2005 and 2014, non-relapse mortality (NRM) at 2 years was higher for patients aged ≥70 years at 34% compared to 24% in those aged 50–69 (p < 0.001). The OS at 2 years was also lower for older patients at 38% compared to 50% in the younger population (p < 0.001) [64]. However, age should not be a sole factor determining the eligibility for transplant. Additional variables such as comorbidities, psychosocial support, functional status, and disease status should be considered to calculate expected treatment tolerance and outcomes [65, 66]. Several tools have been developed, such as the hematopoietic cell transplantation comorbidity index (HCT-CI), to risk stratify older patients with AML based on a number of medical comorbidities and predict 2 year NRM and OS [67]. Other systems combine the HCT-CI with cytogenetics, molecular profiles, and clinical characteristics to create AML composite models for transplant eligibility in patients older than 60 years of age [68]. Geriatric assessment tools including cognitive function evaluation are shown to be highly informative in predicting outcomes in older patients when planning allo-HSCT [69]. The MRD status should also be a part of decision making, as discussed in the previous section. The outcomes were significantly better for those who underwent allo-HSCT with MRD-negative CR than MRD-positive CR irrespective of the number of remissions at HSCT, although this should be further tested in lower intensity treatment settings [60]. Clearly, an increasing number of patients are able to receive allo-HSCT as a consolidation therapy with better outcomes but given ever-increasing number of medical therapies that are patient-specific and better tolerated, non-transplant options are rapidly emerging.

As our molecular and cytogenetic understanding of AML biology continues to expand, the use of newer and more specific targeted agents may show unseen benefits in patients with AML, especially in elderly adults >60 years old. Below is a review of the novel agents (Table 1) that have surfaced in the past few years, which will be of increasing value for elderly patients with AML.

Alterations in FLT3 gene causing internal tandem duplications (ITD) or tyrosine kinase domain (TKD) point mutations result in constitutively active tyrosine kinase, leading to neoplastic proliferation [112, 113]. FLT3 mutations are observed in approximately 30% of patients with AML and are generally associated with poor prognosis, especially with high allele burden [114]. Consequently, the development of FLT3 tyrosine kinase inhibitors (TKI) has heralded a new avenue for induction and maintenance therapeutics in AML. Some of the early inhibitors that were tested include sorafenib, midostaurin, sunitinib, and lestaurtinib. The SORAML study was a randomized placebo-controlled trial that investigated the benefit of adding sorafenib (400 mg twice daily) to the standard induction therapy in adults with AML aged 60 years or younger [115]. Statistically significant improvements were seen in the 3-year EFS (22% vs. 40%, p = 0.013), albeit with an increased rate of toxicity in the experimental group, notably grade 3–4 fever, diarrhea, bleeding, cardiac events, and hand-foot-skin reactions [115]. In elderly patients, however, adding sorafenib to 7 + 3 regimen did not result in clinically significant benefit, as demonstrated in another randomized, placebo-controlled trial involving AML patients 60 years of age or older. EFS and OS were similar between the sorafenib and placebo groups (7 vs. 5 months and 15 vs. 13 months, respectively), while the early death rate was higher in the sorafenib group than in the control group with borderline significance (17% vs. 7%, p = 0.052) [116].

Midostaurin is a first-generation multi-kinase inhibitor that showed strong responses in AML. It was initially tested as a monotherapy in R/R AML and MDS (median age 62), which resulted in a decrease in peripheral blast count by 50% in 70% of the patients [113]. In a phase Ib trial testing midostaurin in combination with the 7 + 3 induction regimen in newly diagnosed younger AML patients, similar CR and 2-year OS rates were observed in both FLT3 mutant and wild-type patients (92% vs. 74% and 62% vs. 52%, respectively), establishing a seminal role for kinase inhibitor therapy in AML [117]. The subsequent phase III CALGB RATIFY trial has demonstrated the standard-changing efficacy of midostaurin as a part of induction therapy in newly diagnosed FLT3-mutated AML patients younger than 60 years old. Administering midostaurin 50 mg twice daily on days 8–21 in combination with 7 + 3 followed by continuous midostaurin during maintenance resulted in significantly improved EFS (HR 0.78, p = 0.002) and OS (HR 0.78, p = 0.009) compared to placebo [118]. Midostaurin is currently approved for treatment of adult patients newly diagnosed with AML with FLT3-positive mutation in combination with standard 7 + 3 induction and cytarabine consolidation. Its efficacy and safety in older population are still being investigated, although it is typically utilized with close monitoring for side effects. In a phase II hypothesis-generating trial, 284 patients with newly diagnosed FLT3 mutation-positive AML, including 86 patients age 61–70 years, were treated with 7 + 3 plus midostaurin induction therapy followed by allogenic stem cell transplant and single agent midostaurin maintenance therapy of 12 months. The rate of CR/CRi was similar between younger and older patients (75.8% vs. 77.9%, p = 0.76), but the mortality rate was higher in older patients (10.5% vs. 3.5%, p = 0.03). Two-year EFS and OS were also similar between younger and older patients in the multivariable analysis (p = 0.51), but a trend toward better OS in younger patients was evident (p = 0.07). An analysis that compared this study with historical cohorts of five previous AML trials suggested that the improvement of EFS was even better in older patients with a HR of 0.42, suggesting strong applicability of the drug in older population [70].

A recent phase I/II study evaluated safety and efficacy of midostaurin at a dose of 75 mg twice daily in combination with AZA for unfit or elderly patients with untreated FLT3 wild-type AML, considering midostaurin has broad target kinases [71, 119]. The mOS and duration of response were 244 days and 255 days, respectively. However, the regimen was not well tolerated, resulting in 8 out of 24 patients discontinuing treatment before cycle 2 due to adverse events, suggesting increased toxicity with 75 mg twice daily dosage. Additional trials enrolling elderly patients with AML with or without FLT3 mutation were opened to investigate the tolerability and efficacy of midostaurin in combination with intensive induction therapy (the UNIFY trial NCT03512197) or with 10-day DEC (NCT04097470). The UNIFY trial published its interim data at 2021 American Society of Hematology (ASH) annual meeting and concluded that the clinical efficacy of midostaurin is primarily in the FLT3-mutated setting only [72].

Gilteritinib is a potent second-generation TKI with high selectivity that has shown to have strong clinical activity. It is a dual inhibitor of FLT3 and AXL, an important trait as overexpression of Axl-1 is implicated in resistance to FLT3 inhibitors [120]. In a first-in-human phase I/II trial of 252 patients with R/R AML, gilteritinib produced an ORR of 40% (30% CR), with a higher response rate shown in FLT3-mutated patients at 49% compared with 12% in FLT3-unmutated patients. The response rate was higher for patients who received >80 mg daily without increased adverse events. The mOS was 31 weeks in this patient cohort. Notably, 26% of patients who have failed prior FLT3 inhibitors achieved a CR with gilteritinib, suggesting its role in overcoming resistance mechanism [121]. In ADMIRAL study, adult patients with R/R AML were randomized to receive gilteritinib monotherapy versus salvage chemotherapy in a phase III trial comparing gilteritinib monotherapy with salvage chemotherapy in patients with R/R FLT3-mutated AML. The average age was 62 for both groups, ranging up to 85 years of age. Results favored the gilteritinib group, which demonstrated a significantly higher CR/CRp (34% vs. 15.3%) and a longer mOS (9.3 months vs. 5.6 months, p < 0.001) than the chemotherapy group. AEs of grade 3 or higher and serious AEs were expectedly much less frequent in the gilteritinib group, with the incidence of exposure-adjusted ≥ grade 3 AEs of 19.34% versus 42.44% in the chemotherapy group [73]. Gilteritinib is now FDA approved to be used for R/R AML patients with FLT3 mutations, allowing more elderly and less fit patients to receive a salvage therapy.

For those with newly diagnosed FLT3-mutated AML who are ineligible for intensive induction therapy, results from phase III LACEWING trial were published [122]. Gilteritinib plus AZA showed an increased composite CR rate compared to the AZA only treatment (58.1% vs. 26.5%, p < 0.001), but it failed to demonstrate an OS benefit (9.82 vs. 8.87 months, p = 0.753). Patient subgroups that showed improved OS with the combination therapy included those with better functional status (ECOG PS 0–1) and high FLT3-ITD allelic ratio of ≥ 0.5 [122]. Gilteritinib was also tested in combination with venetoclax in a phase Ib open-label trial [75]. In this study, a total of 61 patients with FLT3-mutated R/R AML and the median age of 63 were treated with venetoclax plus gilteritinib 80 mg or 120 mg for dose escalation. The composite CR rate was 39% with a median DOR of 4.9 months and a mOS of 10 months [75]. A triplet therapy combining gilteritinib, AZA, and venetoclax appears more promising. In a phase I/II study treating FLT3-mutated AML patients who are unsuitable for IC, the regimen demonstrated a 100% response rate in treatment-naive patients with a CR/CRi rate of 90%, while R/R patients achieved an ORR of 74% with a CR/CRi rate of 24%. The 1-year OS was 80% for treatment-naive and 27% for R/R patients. Outcomes were superior for those who had not received prior HMA plus venetoclax or gilteritinib [76]. Benefits of this “triplet therapy” were also apparent with other FLT3 inhibitors as well, as demonstrated in a recent retrospective study that compared the outcomes of HMA, venetoclax, and various FLT3 inhibitor combination therapy to those of HMA/venetoclax double therapy in older and/or unfit adult patients with newly diagnosed FLT3-mutated AML. Triplet regimens with both first- and second-generation FLT3 inhibitor were associated with significantly improved CR/CRi rate, FLT-PCR, and multiparametric flow cytometry (MFC)-MRD rates, as well as longer OS, without increasing early mortality [123]. In summary, while benefit with FLT3 inhibitor monotherapy has been clearly established, more studies assessing combination therapy could shed light on the additive efficacy in both FLT3 inhibitor-exposed and inhibitor-naive adults with AML. In the future, we may see an evolution for the role of FLT3 inhibitors as part of induction regimens and maintenance therapy, or as a bridge to hematopoietic transplantation in R/R patients. Additional advanced-phase clinical trials will be vital to explore these possibilities.

Quizartinib is a potent second-generation FLT3 kinase inhibitor with higher selectivity and longer half-life, designed to reduce off-target toxicity and prolong drug exposure. It was first tested in a phase I study as a monotherapy in 76 patients with R/R AML (median age 60). 30% of treated patients showed responses, including 13% CR/CRi and 17% partial response (PR). The response rate was higher at 53% for FLT3-ITD positive patients compared to 14% for FLT3-ITD negative patients. mDOR and mOS were 13.3 weeks and 14.0 weeks, respectively [124]. The subsequent phase II trial involved patients aged 60 years or older with R/R AML in cohort 1 (n = 157) as well as those aged 18 years or older with R/R disease in cohort 2 (n = 176). In cohort 1, the composite CR rate was 56% in FLT3-ITD mutated patients and 36% in patients without the mutation while cohort 2 showed 46% and 30% composite CR rates in FLT-ITD mutated and unmutated patients, respectively. Grade 3 or worse QT prolongation was more frequently seen in the older age group of cohort 1, who were older than 70 years of age compared to younger patients aged 60–69 years (17% vs. 5%). Treatment-related death occurred in 6% and 5% of patients in cohort 1 and 2, respectively [77]. The follow-up phase IIb study investigated the optimal dosing of quizartinib and determined that 60 mg daily dosing results in longer duration of remission and mOS while minimizing the frequency of QTc prolongation, a commonly observed side effect [125].

The recent phase III QuANTUM-R study has further sought to prove the beneficial role of quizartinib monotherapy in R/R AML patients. A total of 367 patients were randomized in a 2:1 ratio to receive either 60 mg of quizartinib daily or investigator’s choice chemotherapy. The mOS was longer for quizartinib compared to chemotherapy at 6.2 months and 4.7 months, respectively (p = 0.02), while both groups showed similar rates of AEs [78]. Despite these promising data, FDA declined to approve quizartinib as a monotherapy for R/R FLT3-mutated AML due to concern for lack of improvement in EFS and high incidence of cardiac-related AEs. Quizartinib has also been tested in combination with AZA or LDAC in adults with AML. In a phase II study, patients aged >60 years with untreated MDS/chronic myelomonocytic leukemia/AML, or any age receiving first salvage treatment for FLT3-ITD mutated AML, were treated with a combination of quizartinib and either AZA or LDAC. Fifty-nine patients had a median age of 67 years and they demonstrated an ORR of 75%, and the mOS of 18.6 and 11.5 months in treatment-naive and previously treated patients, respectively. Notably, 11 of 12 older patients >60 years of age in the frontline arm achieved a response (CR 83%), and 4 of 5 patients who previously received FLT3 inhibitors showed a response [79]. A triple combination of DEC, quizartinib, and venetoclax also produced clinically meaningful response especially in heavily pretreated and prior FLT3 inhibitor-exposed populations [80]. These data suggest quizartinib in combination with chemotherapy is a potential therapeutic agent that can improve clinical outcome in R/R or treatment-naive FLT3-mutated AML patients who are elderly and have limited treatment options. However, its ongoing concern about cardiac-related toxicity should be addressed before clinical implementation. For newly diagnosed patients with FLT3-ITD positive AML, QuANTUM-First trial data were recently published. Patients received intensive chemotherapy induction plus quizartinib or placebo, which was maintained throughout consolidation and post-allo-HSCT for up to 3 years. The median age was 56 years. The study met its primary end point of longer mOS of 31.9 months for quizartinib versus 15.1 months for placebo (p = 0.032). However, in the post-hoc subgroup analysis by age, the HR for OS was not significantly different for patients aged 60 years or older (HR 0.91 [0.66–1.26]) [126]. We are expecting soon the FDA approval again for quizartinib for FLT3-mutated AML patients.

Crenolanib besylate is a second-generation TKI with potent activity against both FLT3-ITD and FLT3-TKD mutant kinases. Possessing extensive “pan-kinase” inhibition of secondary TKD mutation, it can potentially suppress kinase domain mutation-mediated clinical resistance [127]. In a single center phase II trial, patients with median age of 61 who had undergone a median of 3.5 prior therapies (including prior-generation FLT3 kinase inhibitors) were given crenolanib administered at 200 mg/m²/day three times a day continuously in 28-day cycles. Response rates were higher for patients naive to FLT3 inhibition compared to those with prior exposure to FLT3 inhibitor therapy (CRi 23% vs. 5% and mOS 55 weeks vs. 13 weeks, p = 0.027). The superior results in FLT3 inhibitor-naive patients suggest that potent FLT3 inhibition up front enhances efficacy by repressing resistant FLT3-ITD leukemic clones [81]. In another larger phase II study using crenolanib monotherapy in 69 R/R FLT3-mutated AML patients, a response rate of 50% was achieved in patients’ naive to FLT3 inhibitors and 28% in those who had received prior FLT3 inhibitors [82]. The encouraging results allowed the development of trials evaluating crenolanib in combination with other cytotoxic chemotherapy in treating both untreated and R/R FLT3-mutated AML. A phase II trial investigated the efficacy of adding crenolanib 100 mg three times a day after standard 7 + 3 induction in patients with a median age of 55 years who have untreated, FLT3-mutated AML. Nine of 26 patients (35%) were >60 years old. CR/CRi was achieved in 24 out of 25 evaluable patients (96%) and 21 patients were alive with a median follow-up of 6 months [128]. The combination regimen was well tolerated including in the elderly, suggesting that crenolanib remains a possible option for elderly patients with newly diagnosed AML, who cannot tolerate IC.

Isocitrate dehydrogenase 1 and 2 (IDH1/IDH2) encode enzymes that play a critical role in regulating cell cycle and metabolism by catalyzing the oxidative decarboxylation of isocitrate to α-ketoglutarate (α-KG) and reducing NADP to NADPH. Mutations in IDH confer neomorphic enzymatic activity by reducing α-KG to 2-hydroxyglutarate (2-HG), accumulation of which leads to epigenetic dysregulation and a block of myeloid differentiation [129, 130]. Ivosidenib is an IDH1 inhibitor. In a phase I dose-escalation and dose-expansion study of ivosidenib monotherapy in IDH1-mutated AML population, 179 patients with R/R disease were enrolled, with a median age of 67. The study produced the CR/CR with partial hematologic recovery (CRh) of 30.4%, the ORR of 41.6%, and a mOS of 8.8 months, with 50% 18-month survival for those in CR/CRh [83]. The phase I/II multicenter dose-escalation and dose-expansion clinical trial evaluated the pharmacokinetic profiles of enasidenib, the first-in-class, selective oral mutant IDH2 inhibitor, as monotherapy. The study enrolled 239 patients with R/R AML with IDH2 mutation with the median age of 70. The CR/CRi rate was 28.6% with an ORR of 40.3%, and the mOS among all patients was 8.8 months. The mOS was 22.9 months for patients attaining a CR [131]. Both ivosidenib and enasidenib were well tolerated at 500 mg and 100 mg daily doses, respectively, and produced low early death rates. IDH-differentiation syndrome was reported in 11% of patients in each of the ivosidenib and enasidenib trial, of which 7% and 4% were grade 3 or higher, respectively [83, 131]. Based on these two single-arm studies, both drugs received FDA approval for treating R/R IDH-mutated AML with their respective IDH mutations. In addition, ivosidenib received FDA approval for newly diagnosed IDH1-mutated AML in patients unfit for intensive chemotherapy based on a recent follow-up analysis of the subset of patients enrolled in the prior phase I study. In this population, the median age was 76, and it included 76% of patients with sAML. The single agent ivosidenib given at 500 mg daily demonstrated a CR/CRh rate of 42.4% and a mOS of 12.6 months. Even with PR, patients benefited from other alleviating factors such as transfusion independence in 42.9% of patients. Notably, IDH1 molecular clearance was observed in 9 out of 14 patients with CR/CRh [84].

Studies of IDH inhibitors in combination with other chemotherapy are ongoing. A phase I/II open-label multicenter trial of ivosidenib or enasidenib in combination with AZA in newly diagnosed AML patients who are too old or unfit for intensive chemotherapy is currently enrolling patients. Results for the ivosidenib portion that enrolled 23 patients with a median age of 76 were recently published, demonstrating a CR/CRh rate of 70% and a 12-month OS of 82%. The mDOR was not reached at the 16-month follow-up [132]. The subsequent confirmatory phase III study of ivosidenib versus placebo in combination with AZA in newly diagnosed AML patients has recently published its results in April 2022 (NCT03173248). The trial randomized 146 patients with ND AML into ivosidenib plus AZA arm and placebo + AZA arm with a median age of 76. The estimated 12-month EFS was significantly longer in the combination group than in the placebo group (37% vs. 12%) as well as mOS (24 months vs. 2.9 months, HR 0.44; p = 0.001). A CR/CRi was achieved in 53% of patients who received ivosidenib versus 13% with placebo [85]. Based on these data, FDA approved the use of ivosidenib in combination with AZA for treating newly diagnosed AML with IDH1 mutation in adult patients 75 years or older, or who are ineligible for IC [133]. Both phases I and II results of enasidenib plus AZA portion were also published recently. After the initial dose-finding portion, phase II enrolled 101 patients into either enasidenib plus AZA or AZA only groups in a 2:1 fashion, with the median age of 75 years. Outcomes were significantly improved with enasidenib plus AZA, which produced an ORR of 74% and a CR/CRi/CRp rate of 63% compared to 36% and 30%, respectively, in the AZA only group (p = 0.003). However, no significant difference was noted in EFS or OS between treatment groups as the study was not powered to detect survival outcomes [87].

Another notable combination is IDH inhibitors with venetoclax. Venetoclax is a rational partner of IDH inhibition as IDH-mutated malignancies demonstrate increased reliance on BCL-2 anti-apoptotic proteins, rendering them susceptible to BCL-2 inhibition [134]. A small study reported interim findings of safety and efficacy analysis of ivosidenib combined with venetoclax ± AZA in 25 patients with newly diagnosed AML, MDS, sAML, or R/R AML, with the median age of 67. The composite CR was remarkable at 92%, 63%, 80%, and 100% in those with newly diagnosed AML, R/R AML, sAML, and MDS, respectively. The 1-year OS was 68% for the entire study population and 71% for the newly diagnosed patients [86]. Enasidenib was also tested in combination with venetoclax in patients with IDH2-mutated R/R AML or high-risk MDS/MPN. As of data cutoff date of July 28, 2021, 11 patients were enrolled, and the median age was 72 years. A CR was achieved in 2 and a CRi in 3 patients for an ORR of 55%, with mOS not reached [88]. Patient enrollment in dose-escalation cohorts is ongoing (NCT04092179).

The newest IDH1 inhibitor on the market is olutasidenib (FT-2102). The phase I/II trial assessing its safety, pharmacokinetics, and clinical activity as monotherapy or in combination with AZA with IDH1-mutated AML patients is ongoing and interim data were recently published. Outcomes from total of 78 patients were reported, 31 in the monotherapy group and 46 in the combination group, with the median age of 72 and 67, respectively. In treatment-naive patients, 25% receiving monotherapy and 77% receiving combination therapy had an overall response, while in R/R patients, 41% receiving monotherapy and 46% receiving combination therapy had an overall response. Olutasidenib was well tolerated both as monotherapy and as a combination therapy, with thrombocytopenia and febrile neutropenia being most associated AEs. Differentiation syndrome was observed in 11% of patients in both groups [135]. On December 1, 2022, the FDA approved olutasidenib for adult patients with R/R AML with IDH1 mutation.

TP53 mutation is more commonly seen in AML arising from antecedent myeloid disease or secondary to prior chemotherapy exposure. Naturally, these patients tend to be older than those with de novo AML. Treatment outcomes with currently available therapies for TP53-mutated AML are poor, with CR rated of 20–30% either with HMAs or with intensive chemotherapy [136‒139]. Eprenetapopt (APR-246) is a novel, first-in-class drug that targets TP53 mutant cells. It is a small molecule that stabilizes mutant p53 protein by covalently binding to cysteine residues, restoring its wild-type conformation and function as an apoptotic regulator [140]. Its antitumor activity was demonstrated in a variety of solid tumors and AML in vitroand in vivo studies, as well as in a phase I study that treated TP53 mutated AML patients with eprenetapopt monotherapy [141‒143]. Subsequently, a phase Ib/II open-label multicenter study was opened to assess the safety and efficacy of eprenetapopt combined with AZA in treatment-naive TP53-mutated high-risk MDS and oligoblastic AML (20–30% blasts) patients [91]. Among 55 patients recruited, 11 had AML and a median age of 68. The ORR and CR rates were 73% and 50% in MDS and 64% and 35% in AML, respectively, with 38% of responding patients achieving complete molecular remission defined by variable allele frequency less than 5%. The OS was 10.8 months and was similar between the MDS and AML disease cohorts, with a significant improvement in responding versus nonresponding patients by landmark analysis (14.6 vs. 7.5 months; p = 0.0005). 35% of patients underwent allogeneic HSCT and demonstrated a mOS of 14.7 months. AE rates were similar to those reported in AZA or eprenetapopt monotherapy, and the most common grade >3 events were febrile neutropenia (33%), leukopenia (29%), neutropenia (29%), thrombocytopenia (25%), and lung infections (25%) [91].

A similar phase II European counterpart trial was conducted in France, except this study included AML with more than 30% blasts and allowed maintenance therapy of eprenetapopt and AZA for up to 1 year after allo-HSCT [92]. Fifty-two patients were enrolled with a median age of 74, out of which 18 had AML and 39% had >30% marrow blasts. CR/CRi rates of 53%, 45%, and 14% and durations of response of 10.4, 14, and 11.5 months were observed in MDS, AML with 20–30% blasts, and AML with >30% blasts, respectively. At a median follow-up time of 9.7 months, mOS was statistically similar at 12.1 months for MDS and 10.4 months for AML, although a trend for shorter OS was observed in AML with >30% blasts at 3 months (p = 0.34). Again, AEs were comparable to those seen in AZA monotherapy [92].

These two trials demonstrated, albeit in a limited patient number, clinical activity and safety of adding eprenetapopt to treat patients with TP53 mutated MDS or AML, especially for those with low blast count. An international phase III, multicenter, randomized study of eprenetapopt in combination with AZA versus AZA alone in patients with TP53-mutant MDS is active but currently not enrolling patients (NCT03745716). In AML patients, HMA/venetoclax combination has become a standard therapy, as described above. The combination therapy has increased the CR/CRi to 55% in those with TP53 mutations compared to 0% in AZA monotherapy, but it failed to improve OS [50]. As such, a multicenter, dose-finding, and expansion phase I study of eprenetapopt in combination with venetoclax and AZA in TP53 mutant AML was conducted, and interim data have been published in the 2021 Annual ASH meeting. Forty-seven patients were enrolled as of the data cutoff date. The median age was 67 years, and the median marrow blasts was 28%. The triple therapy had higher grade ≥3 AE rates than the doublet therapy of eprenetapopt/AZA, including 52% febrile neutropenia and 9% treatment discontinuation due to AEs, with a 30-day mortality rate of 9%, which were largely comparable to those seen with AZA-venetoclax combination therapy. Pre-planned efficacy analysis of first 30 patients demonstrated a CR rate of 37% and a CR/CRi rate of 53%, warranting further study in this difficult to treat population [93].

CD47 is a transmembrane protein that is overexpressed in malignant cells. It provides a potent “don’t eat me” signal that allows for tumor cell evasion of immune destruction by phagocytic cells and functions as a macrophage checkpoint [144]. Magrolimab (5F9) is a first-in-class antibody targeting CD47 that has been shown to induce potent macrophage-mediated phagocytosis of AML cells in vitro and eradicate human AML in vivo, leading to a survival benefit in xenograft models [144‒146]. A phase Ib study was opened to evaluate the safety and efficacy of magrolimab in combination with AZA in patients with newly diagnosed AML patients unfit for IC. Interim results of analysis of 52 patients with median age of 73, of whom 65% had TP53 mutation, have been published [89]. The combination was well tolerated with safety profile similar to AZA alone. The ORR was 65% with a CR/CRi rate of 56%. Remarkably, 67% of TP53mutant AML patients achieved a CR/CRi, an enviable result considering historically poor response rates with traditional chemotherapy in patients with TP53 mutations, which tend to be present more commonly in elderly patients. mOS were 12.9 months and 18.9 months for TP53-mutated and T53-wild-type patients, respectively.

Given encouraging early clinical data, magrolimab was further tested in a triple therapy along with venetoclax and AZA to treat patients with newly diagnosed and R/R AML. After the RP2D was established in the phase Ib trial involving R/R AML patients, the phase II trial enrolled patients in 3 arms: frontline, venetoclax-naive R/R, and venetoclax-exposed R/R AML. The frontline cohort only enrolled patients >75 years of age or those who were unfit for intensive chemotherapy. The interim analysis of 38 patients (median age 70) was presented at the 2021 Annual ASH meeting. The CR/CRi was achieved in 81%, 63%, and 27% of frontline, venetoclax-naive, and venetoclax-exposed patients, respectively, with the MRD negativity shown in 55% of CR/CRi patients in the frontline cohort. Patients with newly diagnosed AML with TP53 mutation also fared well with a CR/CRi in all 7 evaluable patients, with MRD negativity in 57% of them. The triple therapy had similar rate of severe AEs as venetoclax-AZA combination, with 8-week mortality of 9.7%, all of which occurred in R/R patients [147]. Follow-up data were presented during the 2022 ASH Annual Meeting. In the frontline population, the CR/CRi rate was 72%, and the MRD negative rate was 67%. In the R/R cohort, the CR/CRi rate was 11% in venetoclax-exposed patients and 44% in venetoclax-naive patients [90]. Randomized studies to assess whether AZA, venetoclax, and magrolimab can improve on AZA and venetoclax in frontline patients are active in the phase 3 ENHANCE-2 (NCT04778397) and ENHANCE-3 (NCT05079230) studies.

Menin inhibitors are a novel class of inhibitors of protein-protein interactions that target the interaction between menin and the histone-lysine-N-methyltransferase 2A (KMT2A), formerly known as MLL1 [148]. Rearrangements of KMT2A occur in 5–10% of acute leukemias and account for 10–15% of treatment-related AML [149]. The presence of KMT2A rearrangements is associated with a dismal outcome with long-term OS rates of <10%, and allo-HSCT is generally offered in their first CR [150]. KMT2A fusion proteins most notably result in aberrant upregulation of homeobox (HOX) genes, which are key mediators of cell proliferation and self-renewal, and menin is a necessary cofactor that allows KMT2A to bind HOX gene promotors and maintain leukemogenesis [151, 152]. NPM1 mutant AML has a gene expression profile similar to KMT2A-rearranged leukemia and is dependent on the menin-KMT2A interaction [153]. Several different menin inhibitors have been developed for clinical trials.

AUGMENT-101 trial (NCT04065399) is a prospective, multicenter phase I/II dose escalation and expansion study investigating a menin inhibitor SNDX-5613, or revumenib, in patients with R/R KMT2A-rearranged or NPM1-mutated acute leukemia. As of data cutoff in March 2023, there were 68 patients enrolled, of which 56 patients were AML. The median age of adult patients was 42.5 years, but patients up to 79 years of age were included. Most patients were heavily pretreated with a median of four previous lines of therapy. Most frequent grade 3 or higher AEs were febrile neutropenia (31%), thrombocytopenia (19%), and sepsis (18%). Differentiation syndrome occurred in 16%, all grade 2. The CR/CRh rate was 30%, and 78% of these patients achieved undetectable MRD. The duration of response was 9.1 months, and the mOS was 7 months overall but 14.3 months in the efficacy population [94]. KO-539, or ziftomenib, is another menin inhibitor currently in early phase I/II study, KO-MEN-001 (NCT04067336). Phase Ib data showed a 30% CR rate among 20 patients with NPM1-mutant AML at 600 mg dose. Differentiation syndrome occurred in 7 patients including 1 death. Overall safety profile was manageable and phase II trial with 600 mg dose is underway [95]. Menin inhibitors represent a promising novel agents that are better tolerated for heavily pretreated AML patients and will benefit many older patients.

ADCs offer an avenue for directed immunologic cell targeting. Various strategies have been developed to target cell surface proteins that are ubiquitous in leukemic cells, most notably CD33 and CD123.

CD33A is a cell surface transmembrane protein expressed on myeloid precursor and leukemic cells. Owing to its endocytic functions, it is an attractive marker for targeted drug delivery [154].

Gemtuzumab ozogamicin (GO) is a humanized recombinant antibody against CD33 conjugated to calicheamicin, a DNA-cleaving cytotoxic antibiotic [155]. GO was initially approved by the FDA in 2000 for treatment patients 60 years or older with CD33-positive AML at first relapse based on an open-label phase II study that produced a CR/CRp rate of 30%. GO was given at 9 mg dosing every 14 days for 3 times. The notable AEs with this dosing were hepatotoxicity and veno-occlusive disease (VOD), leading to a black box warning [156]. SWOG S0106 was the first phase III trial assessing GO with IC, using a lower dose of 6 mg/m² on day 4. It has failed to demonstrate improved clinical activity compared to the standard arm and has resulted in a higher early mortality rate in the experimental arm, leading to withdrawal of GO by the FDA in 2009 [157]. Nonetheless, more recent meta-analysis data from 5 clinical studies utilizing GO suggested a modest benefit among elderly patients with favorable or intermediate risk AML, producing significantly reduced risk of relapse (OR 0.81; p = 0.0001) and improved OS at 5 years (OR 0.9; p = 0.01) [158]. As a monotherapy, GO demonstrated superiority in mOS compared to the best supportive care (4.9 months vs. 3.6 months) in adults aged ≥75 years with newly diagnosed AML or at least 61 years old and unfit to receive standard chemotherapy in a phase II/III trial [96, 159].

In relapsed AML, GO monotherapy was tested in 57 older patients with median age of 64 years with a newly devised fractionated dosing of 3 mg/m² on days 1, 4, and 7 for one course. CR/CRp was achieved in 34% of patients, and the median RFS was 11 months. No VOD was observed including in 7 patients who underwent allo-HSCT [97]. Combining GO with IC has also demonstrated clinical benefits. In the phase III ALFA-0701 study, 280 patients 50–70 years of age were randomized 1:1 to receive standard 7 + 3 induction with or without GO at fractionated lower doses of 3 mg/m² on days 1, 4, and 7 of induction cycle 1 and on day 1 of two consolidation cycles. The CR rates were similar in two groups, but the 2-year EFS, OS, and RFS were significantly improved in the GO group at 40.8% versus 17.1%, 53.2% versus 41.9%, and 50.3% versus 22.7%, respectively. In the updated analysis, however, the statistical advantage in OS was lost (47.6 vs. 41 months), although the GO arm maintained significantly longer EFS than control arm (17.3 vs. 9.5 months, p = 0.0002) [99]. Benefits were more pronounced in patients with FLT3-ITD mutations and in those with favorable and intermediate-risk cytogenetics. More thrombocytopenia and hepatotoxicity were observed in the GO group, including the VOD incidence of 4.5% overall, highlighting importance of monitoring liver function if allo-HSCT is considered [99, 160]. Another randomized phase III study investigated the benefits of adding GO to the induction therapy in untreated AML patients of at least 60 years of age using the similar lower dose of 3 g/m² on day 1 of the induction cycle. The 3-year incidence of relapse was 68% versus 76%, and the 3-year survival was 25% versus 20% in favor of GO group. No significant difference in early mortality was observed [98]. Based on these data that showed improved benefit when used at lower fractionated doses, GO was reapproved by the FDA for newly diagnosed AML in combination with intensive chemotherapy, as well as monotherapy for R/R CD33-positive AML.

Hence, GO is a suitable option for older patients with favorable or intermediate risk profile, especially when post-remission transplant is not strongly considered. In those with adverse cytogenetic profile, however, benefit of GO has not been clearly demonstrated and thus is not strongly recommended. The added toxicity is a major concern when GO is combined with the traditional intensive chemotherapy regimen. Therefore, substituting anthracycline with GO in older patients has been tested in a randomized ALFA1401-Mylofrance 4 study [99]. The trial results were negative, as GO was associated with more toxicities without significantly improving response rates or survival outcomes. Therefore, replacing GO with anthracycline as a frontline therapy is not feasible in treating older patient populations with de novo AML [99].

Vadastuximab talirine (SGN-CD33A or 33A) is a CD33-directed antibody conjugated to two molecules of a pyrrolobenzodiazepine (PBD) dimer [100]. Upon target binding, 33A is internalized and transported to lysosomes, where the drug dimer is released via proteolytic cleavage, resulting in DNA cross-links and cell death [161]. It has been shown to be more potent than GO across a panel of AML cell lines and was able to overcome multidrug resistance in pre-clinical studies [161, 162]. A phase I dose-escalation safety study using vadastuximab talirine as a single agent was conducted in patients with R/R AML, followed by additional expansion study in a cohort of older, treatment-naive patients with median age of 73 years. At the recommended dose of 40 μg/kg, the CR/CRi was 28%, with 50% of patients who responded achieving MRD negativity [100]. Unfortunately, a phase III trial comparing HMA with or without vadastuximab talirine was terminated due to a higher rate of deaths in the experimental arm (NCT02758900). Clinical hold has been placed on another phase I/II trial due to the deaths of 4 patients with new-onset VOD who were treated with 33A either prior to or following allo-HSCT (NCT02614560), emphasizing the need of more studies before therapeutic implementation.

The interleukin-3 receptor (IL-3R, also known as CD123) transmits signals via the soluble cytokine IL-3 to support proliferation and differentiation of myeloid and lymphoid progenitors. After IL-3 binding to its receptor, the complex undergoes receptor-mediated endocytosis and activates the cell signaling pathways that are essential for cell survival [163]. Notably, myeloid leukemic progenitors overexpress IL-3R compared to negligible expression on normal hematopoietic stem cells, making this receptor an attractive target, especially in the context of R/R disease, where persistence in leukemic stem cells may be a contributor to disease progression [164, 165]. Early stage clinical trials targeting CD123 include fusion antibodies combined with recombinant toxins, cytotoxic drugs, and recombinant T cells.

SGN-CD123A is a recombinant antibody-drug conjugate consisting of a humanized anti-CD123 antibody conjugated to a potent DNA-binding pyrrolobenzodiazepine (PBD) dimer drug. The compound is rapidly internalized upon IL-3R binding, thereby inducing DNA damage and cell death. It has been shown to induce potent cytotoxicity of CD123+ AML cells in vitro, as well as AML eradication in xenograft models [166, 167]. Tagraxofusp (SL-401) is a recombinant fusion protein composed of the catalytic and translocation domains of diphtheria toxin fused to IL-3, which targets leukemic stem cells that express CD123 [168]. In phase I/II trials, tagraxofusp demonstrated tolerability and therapeutic activity as a single agent in context of AML and blastic plasmacytoid dendritic cell neoplasm (BPDCN), an uncommon disease with high expression levels of IL-3R. Among 59 patients with R/R AML, two CRs with 8-month and >25-month durations were observed [169].

Safety and efficacy of combining tagraxofusp with AZA or AZA/venetoclax are being assessed in an ongoing phase Ib study involving CD123 positive AML, MDS, or BPDCN patients. The trial included treatment-naive patients unfit for IC as well as those with R/R diseases. The interim data presented at the 2021 Annual ASH meeting showed that 8 of 9 previously untreated AML patients (median age 74) were able to achieve CR/CRi after receiving tagraxofusp-AZA-venetoclax triplet therapy, including 2 patients with TP53 mutations and 3 with sAML. At the time of data cutoff, 50% of patients achieving a response went on to receive allo-HSCT [102]. An important drug-specific adverse effect to consider is capillary leak syndrome that occurred in 18–33% of patients who received tagraxofusp, although most cases were mild [102, 170]. Further longer term data from the trial is eagerly awaited (NCT03113643).

Another novel agent that has gained FDA approval to in recent years is glasdegib. It is an oral small molecule inhibitor of the Hedgehog pathway signaling, which has been implicated in the development of hematologic malignancies and is shown to be critical for leukemia stem cell survival and expansion [171, 172]. BRIGHT AML 1003 was an open-label, randomized multicenter phase II study that enrolled patients aged ≥75 years with newly diagnosed, previously untreated AML or high-risk MDS, or those who are ≥55 years and ineligible for intensive chemotherapy (NCT01546038) [103]. Total of 132 participants with median age of 77 and 75 were randomized 2:1 to receive LDAC 20 mg BID for 10 days with or without glasdegib 10 mg daily per 28 days, respectively. The mOS was significantly longer for patients receiving glasdegib/LDAC compared to LDAC alone at 8.8 months and 4.9 months, respectively (HR 0.51, p = 0.0004). The CR rate was 17% in the glasdegib/LDAC arm and 2.3% in the LDAC arm (p < 0.05). The safety profile was manageable and similar between the two groups, with similar discontinuation rates (35.7% vs. 46.3%). A longer follow-up data up to 36 months after randomization was published, which continued to demonstrate clinical benefit with glasdegib plus LDAC in mOS at 9.1 versus 4.1 months. A randomized, double-blind, multicenter phase III trial of glasdegib in combination with IC (NCT03416179) or AZA (NCT04842604) in patients with untreated AML is ongoing.

Bispecific T-cell engagers (BiTE) are recombinant antibody conjugates that harness endogenous T-cell responses against tumor cells and are becoming an exciting avenue for targeted immunotherapy of malignant cells. Blinatumomab is an example of a BiTE construct with dual affinity for CD3-positive T cells and CD19-positive B cells and is demonstrating promise in context of R/R ALL and as a potential bridge to hematopoietic stem cell transplantation [173]. Though a similar approach has not gained enough momentum in AML, promising studies are underway. AMG 330 is a CD33/CD3 BiTE antibody construct under investigation that couples CD3-positive T cells and CD33-positive myeloid leukemic cells. Harrington etal. [174] showed that AMG 330 potently kills primary human-derived CD33+ leukemic cells in vitro. Interestingly, though the engager showed strong cytotoxic effects even in low CD33-expressing cells, there appears to be yet undefined CD33-independent resistance mechanisms in specific patient subsets. Safety and tolerability profile and preliminary efficacy results from a phase I dose escalation study have recently been published. The study enrolled 55 patients up to the date of assessment, who had AML R/R to multiple lines of therapy and a median age of 58 (18–80). Grade ≥3 cytokine release syndrome (CRS) was observed in 67%, which was reversible and occurred in a dose-dependent manner within the first 24 h. The clinical effect was limited with a CR/CRi rate of 16% [104]. High rates of CRS and low remission rates were rather disappointing, although it remains a viable strategy to investigate further in heavily treated patients who do not have alternative options, or as a means to eradicate MRD as seen with blinatumomab in ALL.

Another similar bispecific antibody that targets CD123 and CD3 but with a longer half-life is APVO436 that has recently been tested in R/R AML and MDS patients in a phase Ib study. In mostly heavily treated patient population, APVO436 exhibited a favorable safety profile with manageable side effects, with grade 3–4 CRS occurring in 4 out of 46 evaluable patients (8.7%). The efficacy data is too premature to draw a conclusion, but 8 out of 34 patients experienced clinically meaningful stabilization of their leukemia [105]. A new dose expansion phase is currently open (NCT03647800). As demonstrated with tagraxofusp, combining APVO436 with AZA, venetoclax, or both is a rational next step, and this strategy is being tested in a newly diagnosed AML in a clinical trial, which is currently recruiting patients (NCT04973618).

Dual affinity retargeting (DART) proteins are a newer class of recombinant molecules related to BiTE constructs that may show enhanced ability to cross-link T cells and B cells for more efficient antitumor activity [175]. In laboratory studies, a CD3-CD123 DART induced dose-dependent killing in AML cell lines and primary AML blasts, providing the basis for testing in early phase clinical trials [176]. Flotetuzumab, a CD123-CD3 bispecific DART molecule, has been tested in a phase I/II study in patients with R/R AML. Among 30 patients with median age of 59, the ORR was 30% and the CR/CRh rate was 26.7%. Interestingly, all CR/CRi occurred in primary refractory patients. A further follow-up analysis demonstrated a CR rate of 32.1% in primary refractory patients, which is a notable improvement from historical response rates of <10% in similar patient population. Infusion-related reaction/CRS were the most common AEs, occurring in majority of patients including grade 3 events in 8% of patients [106, 177, 178]. Further results from a longer follow-up are expected from the trial (NCT02152956).

Chimeric antigen receptor T (CAR-T) cells are genetically modified with the ability to utilize antigen specificity of a monoclonal antibody with potent effector functions of T cells [179]. In AML xenograft models, CAR-T cells targeting CD33 resulted in eradication of leukemia and prolonged survival. Unlike CD19-targeting in ALL, the CAR construct against CD33 resulted in cytopenia and reduction of myeloid progenitors in xenograft models, suggesting that permanently expressed CD33-specific CAR-T cells would have unacceptable toxicities in AML patients [101, 180]. As a means of avoiding long term CD33-mediated myelosuppression, investigators have tested transient mRNA expression of CARs into T cells, demonstrating potent and self-limiting activity [180]. Other avenues for more specific molecular targeting of AML cells via CARs are highly sought after. More promising targets under investigations include CD7, CD123, FLT3, and CLL1 [181]. Whether CAR-T cells will be a viable modality in elderly patients is less clear, and its role may be limited by comorbidities and intrinsic marrow-related anomalies in AML.

Immunologic checkpoint inhibition has demonstrated durable responses in a variety of tumors, notably for melanoma, lung, and genitourinary cancers even at advanced stages. Programmed death-1 and programmed death-1 ligand (PD-1 and PD-L1) constitute a signaling pathway that suppresses proliferation of immune mediators, particularly CD8-positive T cells. Pembrolizumab/nivolumab and ipilimumab are monoclonal antibodies targeting PD-1 and CTLA4, respectively, approved for treating various solid tumors, and are being actively investigated for the treatment of hematological malignancies. Aside from potential class specific immunotherapy related adverse events or infusion reactions, checkpoint inhibitors are generally less toxic than traditional chemotherapy and are often better tolerated in older and less fit patients [182]. Several murine models demonstrated upregulation of PD-L1 and PD-1 in leukemia cells, and PD-1 knockout suppressed leukemic cell proliferation and improved survival in mice with AML [183].

A phase Ib multicenter trial, KEYNOTE-013 (NCT01953692), demonstrated tolerability of pembrolizumab in patients with primary or secondary MDS after failing to respond to a HMA agent. Among the 28 patients enrolled in the MDS cohort, median age was 73 years. Twenty-seven patients were evaluated for efficacy, among whom there were no CRs, suggesting that checkpoint inhibitor alone may not be sufficient for durable clinical responses in elderly individuals with MDS [109]. Another phase I study examined the safety and efficacy of ipilimumab in 28 patients with relapsed hematologic malignancies after allogenic HSCT, which included 12 AML patients. At a dose of 10 mg per kilogram, 4 patients with extramedullary AML and 1 patient with MDS achieved durable CRs, lasting 12–16 months [184]. The use of HMAs in AML was shown to upregulate the expression of PD-L1, PD-L2, PD-1, and CTLA4, suggesting utility of selective checkpoint inhibition in combination with a HMA agent [110]. A phase II study investigated a clinical role of combining AZA with nivolumab in older R/R AML patients. Seventy patients were treated, with a median age of 70 years and the median number of prior therapies received of 2. As a whole group, the ORR was 33%, with 22% CR/CRi, but a higher ORR was achieved in HMA-naive patients at 58%. The mOS of 10.6 months in the salvage-1 patients was significantly better than the mOS of 5.2 months in the historical cohort of similar patient populations treated with HMA only [111]. Given the proven efficacy of the HMA-venetoclax therapy, combining it with a checkpoint inhibitor would be an interesting next step.

AML is a disease of the elderly and inherently may coexist with other comorbid diseases and age-related debilitation. Therefore, decisions regarding patient fitness have guided whether patients received intensive or less-intensive induction regimens [185]. Presently, hematologist/oncologists have more treatment options in their arsenal than in the past, and decisions about how to treat AML have become more complex. Thus, a framework for determination of patient fitness for treatment, particularly in elderly patients, is needed. A number of factors that contribute to inferior outcomes in elderly include comorbid medical conditions, adverse cytogenetic risk profiles, and poor performance status [1]. Historically, studies assessing therapies for older patients with AML have included measures such as World Health Organization (WHO) performance status, Karnofsky performance status, and SWOG criteria [186‒188]. Additionally, a geriatric assessment with evaluation of cognitive function and physical function has been proposed [189]. Furthermore, the HCT-CI was developed to predict the effect of pre-transplant comorbidities on non-relapse mortality and survival [67].

An analysis of 446 patients 70 years or older with AML treated with intensive chemotherapy found that age 80 years or older, complex karyotypes, ECOG 2–4, and elevated creatinine were negative prognostic characteristics. Notably, in patients with 3 or more of these factors the 8-week mortality rate was 71%, though this only consisted of 9% of patients [190]. Ferrara et al. [191] published conceptual criteria for assessing unfitness for intensive and non-intensive chemotherapy in patients with AML. To define a patient unfit for intensive therapy, authors suggested that one or more of the following should be met: age over 75, severe cardiac, pulmonary, renal, or hepatic comorbidity, active infection resistant to treatment, cognitive impairment, poor performance status, any co-morbidity that is deemed incompatible for chemotherapy. For determining unfitness for non-intensive chemotherapy, patients must have one of the following criteria: very severe cardiac comorbidity, severe pulmonary comorbidity, severe hepatic comorbidity, active infection resistant to treatment, cognitive impairment, uncontrolled neoplasia [191]. Figure 1 incorporates elements of the decision schema of both Kantarjian et al. [190] and Ferrara et al. [191] in assessing fitness for chemotherapy in the elderly patient with newly diagnosed AML. Overall, deciding whether a patient is “fit” for intensive or non-intensive chemotherapy in the context of AML remains complex, and such a decision should ultimately be made by experienced clinicians after discussing benefits and risks with the patient, taking into account patient’s unique clinical characteristics and goals of care. Above-mentioned tools can assist in determining the right therapy for elderly patients. Figure 2 illustrates a schema to provide a guideline for choosing the best novel agent for elderly/unfit patients who have R/R AML.

Treatment of AML in elderly patients continues to pose significant challenges. Poorer outcomes in this population are related to patient and leukemia-specific factors. Namely, the evolution of comorbidities and aggressive cytogenetics in leukemia hinders benefit from most conventional therapies. The need for more targeted treatments is thus ever more necessary to better eradicate leukemic burden and prolong survival and quality of life in elderly patients with AML. Examples of newer treatments reviewed in this paper showed potential agents with promising features including improving CR rates, OS and even fitness for hematopoietic transplantation. Because older patients are more susceptible to toxicities, the implementation of novel therapies must be accompanied by close investigation for treatment-related toxicities in this population. Novel agents are particularly useful in R/R settings where therapeutic options are limited especially in those unfit for intensive therapies. Figure 2 provides guidance on determining therapy for R/R AML patients. It should be noted that most of these options are in clinical trial phase with exception of FLT3 inhibitors and IDH inhibitors.

An international expert panel from ELN recently published an updated guideline on diagnosis and management of AML patients, including therapeutic options for patients unfit for intensive chemotherapy, which should provide additional summary of how AML can be treated [192]. As we move into a new era in the treatment of AML in elderly adults, we abandon the “one-size fits all” paradigm of conventional 7 + 3 induction. Instead, we relegate therapeutic regimens based upon more precise metrics that incorporate functional status, comorbidities, and molecular features of the AML clones. It is worth pointing out is that as patient cohorts are “sliced” into specific subsets (based upon molecular, cytogenetic analyses), it may become more challenging to garner statistical significance in coming trials. Keeping this in mind, it will be important to maintain openness in the regulatory drug approval policies and patient inclusion criteria. Interaction between providers, patients, and pharmaceutical developers at all phases should be fortified in order to hasten progress in AML. Eventually, with more targeted, less toxic therapies, our hope is that the landscape for elderly patients with AML will feature longer survival rates and abrogation of treatment-related morbidity.

Jun Choi and Mihir Shukla: no conflict of interest or anything to declare. Maher Abdul-Hay: Rigel and Daiichi advisory boards; Jazz, Takeda, and Servier advisory boards and speaker bureau.

No funds, grants, or other support was received to assist with the preparation of the manuscript.

Jun Choi and Maher Abdul-Hay both substantially contributed to the conception and design of the article, interpreting the relevant literature, drafted the article and revised it critically for important intellectual content, and published or can otherwise be considered experts on the topic. Mihir Shukla contributed to interpreting the relevant literature, drafted the article, revised it for important intellectual content, and designed the flow charts.