Precision Medicine for Older AML Patients

Simple Summary

Acute Myelogenous Leukemia (AML) is a hematological malignancy that predominantly affects individuals of >60 years of age. However, older AML patients have poor prognosis due to individual and disease-related factors. Individual factors are mainly dependent on age and are represented by the accompanying comorbidity burden and by an individual, age-related personal vulnerability. Disease-related factors are mainly linked to the peculiar AML biology observed in older patients, with an enrichment of adverse risk factors. These factors considerably limit the feasibility of intensive induction treatment in older AML patients, thus requiring the development of low-intensity treatments. The discovery of a considerable number of drug-targetable genetic abnormalities in a significant proportion of AML, including older patients, now offers an opportunity for less toxic treatments, compatible with older AML patients.

Abstract

The development of molecular profiling approaches for AML patients such as whole genome sequencing, whole exome sequencing and transcriptomic sequencing have greatly contributed to better understanding of leukemia development, progression and treatment responsiveness/resistance. These studies have generated a new knowledge about driver events operating in AML that can be translated into clinics, thus favoring the mutations; using this approach, more than 50% of older AML patients display molecular alterations, such as IDH1, IDH2, FLT3 (FLT3-TKD and FLT3-ITD), NPM1 and KMT2A rearrangements that can be targeted by specific drugs. Preclinical and clinical studies have supported the use of drugs targeting these molecular alterations as first-line therapy in association with induction chemotherapy in chemotherapy-fit patients or with a hypomethylating agent in association with a Bcl-2 inhibitor (Venetoclax) in chemotherapy-unfit patients. These studies have shown promising results that need to be confirmed through randomized clinical studies specifically involving the enrollment of older AML patients.

1. Introduction

Acute Myeloid Leukemia (AML) is an aggressive and genetically heterogeneous hematologic malignancy that originated through the clonal expansion of neoplastic myeloid progenitor cells in the bone marrow, in which treatment response and individual patient survival are strongly influenced by leukemia biology and the patient’s age. Tremendous progress achieved during the last three decades has consistently contributed to the understanding of the complex genetic alterations underlying the different subtypes of AML. The definitions of molecular alterations and driver gene mutations have shown the consistent heterogeneity of this disease and have led to new AML classifications reflecting the heterogeneity of leukemia biology and of genetic pathways underlying leukemia development. Risk stratification criteria have considerably progressed through the introduction of novel molecular markers and of the most recurrent chromosomal abnormalities, thus providing an improved prediction of individual patient outcomes following the current treatments, including both intensive and non-intensive treatments.

Landscape of Genetic Alterations in Older AML

Three main types of AML can be defined for their ontogenesis: de novo AML, not related to a prior malignant or pre-malignant condition; secondary AML (sAML), derived from the transformation of a prior hematologic malignancy (myelodysplastic syndromes (MDSs) or myeloproliferative disorders or clonal hematopoiesis); therapy-related AML (t-AML), triggered by oncogenic mechanisms induced by chemotherapeutic agents or radiation.

Remarkable differences in the genetic landscape of older AML compared to younger AMLs have been observed; these differences mainly involve (i) a higher incidence of DNMT3A, TP53, TET2, ASXL1, SRSF2, RUNX1, STAG2 and SETBP1 mutations; (ii) a higher incidence of adverse cytogenetic abnormalities (5q/5 and 7 chromosome monosomies and complex karyotype); (iii) a lower frequency of FLT3-ITD, FLT3-TKD, KIT and NRAS mutations and KMT2A rearrangements; (iv) a higher frequency of AMLs bearing myelodysplasia-related mutations; (v) a higher frequency of AMLs with adverse-risk profile, compared to younger AMLs [1,2,3,4,5,6,7].

The ASXL1 gene, located on the chromosome 20q11.21 encodes a 1541 amino acid protein acting as an epigenetic regulator: the ASXH and PHD domains of this protein are putative DNA and histone binding domains and the ASXH domain interacts with BAP1, conferring deubiquitinase activity and leading to gene repression [8]. Mutations in ASXL1 are observed in hematopoiesis of undetermined potential (CHIP) and are significantly associated with aging and smoking [9]. Longitudinal studies suggest that the presence of ASXL1 mutation in a hematopoietic stem cell (HSC) clone confers a fitness advantage, particularly when co-mutated with SRSF2 or EZH2, and markedly increases the risk of developing a myeloid neoplasia [10]. ASXL1 gene is frequently mutated in AML and the incidence of these mutations markedly increases with the age of patients: about 15% of older AML patients compared to 6% in younger AML patients [2]. ASXL1 loss in mice determines an MDS that could transform to AML with age, suggesting cooperation with additional mutations to induce myeloid leukemia development. In line with this finding, ASXL1 mutations frequently coexist with other mutations, such as TET2, RUNX1, SETBP1 and NRAS. A recent study showed that AMLs bearing ASXL1 mutations display a peculiar epigenetic profile, associated with a peculiar sensitivity to specific epigenetic-targeted agents, such as BET inhibitors [11]. Jahn et al., using targeted DNA sequencing have explored a group of 604 older AML patients treated in the context of a phase III clinical trial; through oncogenetic tree modeling and hierarchical clustering, main leukemic trajectories have been delineated [12]. The oncogenetic modeling algorithm defined in this study identified a tree with five branches with ASXL1, DNMT3A, TET2, TP53 and DDX41 emanating from the root as leukemia-initiating events [12]. The tree originating from the ASXL1 node generates various individual clones with EZH2, NRAS, RUNX1, SRSF2 and U2AF1 mutations [12].

A total of 5.1% of younger AML patients compared to 15.9% of older AML patients display RUNX1 mutations; these mutations are preferentially associated with age and with intermediate-risk cytogenetics, including normal karyotype, in the elderly AML patients [13]. RUNX1 mutations were mainly associated with ASXL1 in elderly AML patients (37.55 of RUNX1-mutant AMLs display ASXL1 mutations) [13]. In older AML patients treated with a hypomethylating agent plus Venetoclax (HMA + VEN) the response and survival were unaffected by RUNX1 mutational status; older RUNX1-mutant AML patients with prior MDS displayed a particularly dismal outcome [13].

Spliceosomes are complexes composed of small nuclear RNA that remove introns in protein-encoding genes (SRSF2, SF3B1, U2AF1 and ZRSR2) mutations are frequently encountered in secondary AML (sAML) patients and usually are associated with inferior outcomes to standard induction therapy. SRSF2 and U2AF1 mutations are significantly more frequent in older than in younger AML patients (for SRSF2, 17% vs. 7%, respectively, and for U2AF1, 8% vs. 3%, respectively) [2]. In line with these findings, myelodysplasia-related mutations (MDS-R) were observed more frequently in older than in younger AML patients (44.9% vs. 23.4%, respectively) [7]. Both in younger and older AML patients, the presence of MDS-R mutations was associated with reduced outcomes following treatment with standard IC [7,14]. Meclenbrauck et al. have characterized a group of 560 AML patients of different ages and observed that older patients displayed a higher variant allele frequency (VAF) of at least one MDS-related gene and a higher frequency of ASXL1 mutations [7]. The appearance of myelodysplasia-related change-like features in AML, such as ASXL1, U2AF1, SRSF2 and SETBP1 mutations is associated with older age [15].

TET2 is one of the ten-eleven translocation (TET) family genes encoding DNA dioxygenases and is involved in the regulation of the process of genome demethylation and in histone modification. TET2 mutations are frequent in CHIP and in older patients with myeloid neoplasia. TET2 mutations are frequent in AML patients, particularly in older patients: 20% in older AML compared to 6% in younger AML [2]. The epigenetic dysregulation caused by TET2 mutations affects the differentiation and proliferation of HSCs, thus contributing to CHIP and leukemia development [16]. In the oncogenetic tree model of AML development, the TET2 AML node generated one branch containing CEBPA [12]. The prognostic impact of TET2 mutations in older AML patients is related to the type of TET2 mutations and to the type of co-mutations: patients with “significant” TET2 mutations (nucleotide alterations in TET2 consisting in truncation of TET2 protein) have a shorter OS than “non-significant” TET2-mutated group (missense TET2 mutations); the presence of ASXL1 co-mutations was associated with poor prognosis and of NPM1 co-mutations with a better prognosis [17]. Other studies distinguished TET2-mutant AMLs in two groups: one group with TET2 mutations and a second group with both TET2 and DNMT3A mutations: the OS of patients with TET2 mutations was similar to that of those with TET2-WT; patients with concomitant TET2 and DNMT3A mutations have a shorter OS than those with TET2 mutations alone or DNMT3A and TET2-WT [18].

TP53 mutations are significantly more frequent among older than younger AML patients (27% vs. 12%, respectively) [2]. The presence of this markedly higher frequency of TP53 mutations in older AML represents one of the main mechanisms responsible for lower response of these patients to current treatments. TP53-mutated AMLs are now classified as a separate disease entity from other AML subtypes for the European Leukemia Net (ELN) and the international Consensus Classification (ICC). The AML TP53-mutant node did not generate any further branch, suggesting that this gene does not depend on or constitute preconditions to further alterations [12]. TP53 mutations are associated with poor prognosis with limited response to standard conventional therapies, including 7 + 3 induction chemotherapy or HMA + VEN, and allo-HSCT may ensure long-term survival in only 10–15% of eligible patients [19]. Owing to poor outcomes, many hematologists believe it is futile to treat TP53-mutated high-risk myeloid neoplasms in older patients [20].

Molecular markers are present in most AML patients, and their identification is important (i) for the definition of AML molecular subtype and outcome prediction; (ii) for the possible identification of biomarkers suitable for targeted therapy. Finally, the identification of genetic alterations specifically linked to some AML subtypes has greatly contributed to the development of targeted therapies increasingly used in the treatment of AML patients. The standard treatment of adult AML patients with <65 years implies therapy with intensive chemotherapy, followed, when possible, by bone marrow transplantation. Most older AML patients are unfit for intensive induction chemotherapy and must be treated with low-intensity treatments, the most efficacious being the Bcl-2 inhibitor Venetoclax (VEN), associated with a hypomethylating agent (Azacitidine or Decitabine). In fact, VEN in combination with an HMA resulted in a synergistic antileukemic effect with a higher rate of responses and an improved survival compared to HMA alone [21,22,23]. Some AML subtypes, such as those with IDH1/IDH2 or NPM1 mutations had a clear benefit from this treatment, while AMLs bearing TP53, KRAS, FLT3-ITD mutations display resistance. A recent study showed that HMA + VEN treatment approach can be extended to octogenarian and nonagenarian patients with AML, with an acceptable safety profile, particularly introducing some modifications of the schedule of VEN administration (<14 days). A total of 25% of these patients displayed prolonged survival [24].

Shimony et al. have reported the response to treatment with HMA + VEN compared to HMA in a group of 314 older AML patients (median age 74 years, 11 patients with TP53 mutations and 115 with sAML); they observed that the response to treatment was different for various subgroups of patients [25]. Thus, HMA + VEN, compared to HMA alone, did not improve the outcome of TP53-mutant patients (5.7 months vs. 6.1 months, respectively), as well as AML de novo patients (13.2 months vs. 10.3 months, respectively); however, HMA + VEN treatment improved the survival of sAML and displayed a longer mOS compared to HMA (14.1 months vs. 6.9 months, respectively) [25]. These findings clearly show that in older AML patients treated with HMA + VEN, the response is different for various AML subgroups and the prognostic impact of many genetic alterations is different from that observed in younger AML patients treated with intensive chemotherapy [25]. Furthermore, the analysis of a large cohort of 395 sAML patients showed that (i) mOS was comparable in older sAML patients treated either with 7 + 3 induction chemotherapy or liposomal daunorubicin, and cytarabine (CPX-551) or HMA + VEN [26] the presence in these patients of KRAS/NRAS co-mutations was associated with a decreased response and the presence of SF3B1 mutations with an increased response to HMA + VEN [26].

2. Development of Precision Medicine for AML Patients

The development of precision medicine for AML patients, as well as for other types of cancer patients, required not only an improvement in techniques of genomic profiling but also the development of techniques for monitoring the response of patients with high sensitivity and specificity, the expansion of artificial intelligence and machine learning approaches to manage large data sets and to derive algorithms improving diagnostic and prognostic criteria, and the use of ex vivo drug sensitivity assays to predict drug sensitivity of AML blasts.

2.1. Genomic Profiling

Genomic profiling represents a fundamental tool for characterization of the heterogeneity of AMLs at all ages. The genetic heterogeneity of AMLs influences its clinical features, treatment, responses and outcomes. The advent of next generation sequencing (NGS) has mediated tremendous progress in genomic research, offering the opportunity of a highly efficient analysis of DNA and RNA in a high-throughput approach. NGS technique allows the sequencing of large numbers of DNA fragments simultaneously, thus providing fundamental information concerning structural abnormalities of the genome, mutational variability, gene expression profiles and epigenetic changes [27]. The analysis of the genome by NGS may involve different techniques such as whole genome sequencing, whole exome sequencing, targeted sequencing, transcriptomic and epigenomics.

Recent advances in genomic techniques have enabled a deeper understanding of the molecular mechanisms underlying AML, leading to more personalized approaches. Key findings of recent studies of genomic profiling have contributed to (i) the identification of genetic markers and risk stratification; (ii) the development of multi-gene prognostic models; (iii) the discovery of targetable gene alterations with consequent therapeutic implications; (iv) the analysis of immune profiling.

2.2. Diagnostic Tools for AML

The diagnostic tools used for diagnosis and follow-up of AML are numerous. A combination of histologic, immunophenotyping, cytogenetcis and molecular analyses are required for AML diagnosis, including NGS using different gene extension analyses to screen all genetic alterations with genetic, prognostic and/or therapeutic value [28].

Particularly, genomic studies based on NGS have made a great contribution to the dissecting of molecular abnormalities observed in AML, with the identification of new mutations, copy number alterations and recurrent gene structural abnormalities, with gene fusions [29]. These molecular studies have greatly contributed to new updated classifications of AML following WHO, ICC and ELN; these classifications were largely based on genetic data and were integrated into diagnostic and prognostic algorithms. Recently, these classifications were adapted to the study of older AML patients, providing a better risk stratification for these patients [30].

Improvement of diagnostic tools for monitoring the response of AML patients to treatments have been developed in these last years. Classical criteria for the assessment of drug response were based on morphological blasts count; however, recent studies have shown that multiparametric flow cytometry (MFC) and molecular techniques of higher sensitivity and specificity to detect the presence of a minority of residual leukemic cells after therapy. The MFC and, particularly, the molecular techniques (quantitative polymerase chain reaction, qPCR, droppel digital PCR, ddPCR, and NGS) allow the detection of measurable residual disease (MRD), whose assessment offers a unique parameter to predict relapse and may provide criteria for early intervention (MRD-directed therapy) [28].

2.3. Artificial Intelligence and Machine Learning in AML Diagnosis, Prognosis and Treatment

Artificial intelligence (AI) is a broad definition comprising all devices that imitate the mechanisms and the achievement of human intellect. Machine learning (ML) is a subset of AI, concerning the development of computer algorithms to generate predictions based on experience [31].

AI and ML have given relevant contributions to cell classification and detection of hematological malignancies through the introduction of AI models [32], the development of models predicting response to treatment, even in the context of real-life data [33], and classification of AML through flow cytometry analysis [34].

A study by Awada and coworkers, based on the analysis of 6788 AML patients for whom genetic data were available, showed the power of ML-derived algorithms to detect subgroups of AML, not previously recognized and associated with different survival after induction chemotherapy: a genomic cluster 1 comprising AMLs with normal karyotype, enriched for NPM1, DNMT3A, FLT3-ITD and IDH2140 mutations, but without TP53, ASXL1, EZH2 and RUNX1 mutations and associated with an mOS of 34 months; a genomic cluster 2, comprising biallelic CEBPA, IDH2R172K, but absent TP53, ASXL1, RUNX1 and NPM1 mutations and associated with an mOS of 26 months; a genomic cluster 3, including SF3B1, SRSF2 and EZH2 mutations, associated with an mOS of 15.8 months; a genomic cluster 4 with abnormal karyotype and TP53 mutations, associated with an mOS of 9 months [35].

Other recent studies have shown that algorithms derived using ML-based approaches improve epigenomic diagnosis and prognosis of AML [36], immunophenotypic evaluation of MRD [37] and the evaluation of synergistic drug combinations for AML patients [38].

Importantly, a machine learning approach contributed to the performance of an improved age stratification of molecular alterations in a cohort of 3062 AML patients of different ages with an extensive molecular characterization [39]. Machine learning models helped to identify an algorithm that quantified prognostic contribution of each genetic alteration across six age groups, including infants, children, adolescent/young adults, adults, seniors and elderly. Some genetic alterations, such as NPM1, CEBPA, inv(16) and t(8;21) (favorable risk), and TP53, RUNX1, ASXL1, del(5q), −7 and −17 (adverse risk) and FLT3-ITD (variable risk) provided a strong support to model decisions [39]. Importantly, age modified the prognostic impact of genetic alterations, with some specific alterations associated with aging, such as TP53, ASXL1, del(5q), −7 and −17, markedly increased the risk of elderly AML patients [40]. Furthermore, the presence of these age-related genetic alterations in younger AML patients disproportionately conferred a higher risk of these patients [39].

2.4. Drug Sensitivity Assays for AML

Predicting the response to chemotherapy or to other drugs for treatment of AML patients is a major challenge. All the molecular studies used for AML diagnosis and characterization, including analysis of mutational profiles and standard cytogenetic studies provide only an approximate prediction of treatment outcome. To improve the capacity to predict the response to treatment, ex vivo drug assay provides a tool to better evaluate the risk of individual AML patients and to support the decision of optimal treatment.

Several studies have performed ex vivo drug screening for AML patient-derived samples. These studies have explored the genetic heterogeneity of AML cells and have shown in many instances the capacity to predict drug sensitivity for patients with given genetic alterations. Particularly, some studies have shown the existence of an association between mutational AML profile and ex vivo drug sensitivity to some drug classes, even in AML samples showing the existence of distinct leukemic subclones [40]. Other studies showed the existence of a relationship between blast cell differentiation and ex vivo drug sensitivity [41].

Other studies have explored the feasibility of ex vivo drug sensitivity assays in the context of clinical studies. Andersen et al. in a retrospective study, used ex vivo drug sensitivity screening to predict the survival of AML patients undergoing standard chemotherapy [42]. These authors developed a computational method of evaluating the data of ex vivo drug assays that reduced the noises, the bias and the technical confounders of these tests; using this approach, they showed that ex vivo drug sensitivity profiles robustly predict AML patients’ outcomes and integrate well with genome profiles [42].

The SMART was a prospective non-interventional trial investigating the feasibility of ex vivo drug response profiling for the treatment of guidance in hematologic malignancies, including a validation cohort of 95 AML patients [43]. Drug profiling reports were given within 7 days in 91% of patients; ex vivo drug profiles improved ELN-2022 brisk stratification of high-risk AML patients [43].

Preclinical studies performed in the context of the BEAT AML Program have shown that the combination of whole exome sequencing, RNA sequencing and analysis of ex vivo drug sensitivity allowed a consistent characterization of molecular mutational profile and a definition of drug sensitivity profile that resulted to be associated with mutational status [44]. These studies were extended to a total of 805 AML patients and showed a strong association between ex vivo drug sensitivity and is governed broadly by AML cell differentiation stage [45].

A second program of development of precision medicine in AML was based on ex vivo drug response multiomics profiling data [46]. A functional precision medicine tumor board (FPMTB) integrated clinical, molecular and functional data for application in clinical medicine [46]. In this study, the ex vivo drug sensitivity and resistance test involved the evaluation of 347 investigational and 168 approved cancer drugs. Importantly, actionable drugs were found for 97% of AML patients; in 37 relapsed/refractory AML patients, the recommendations derived from drug assay were clinically implemented, reporting an objective response rate of 59%, with 35% of CRs and with five patients bridged to allo-HSCT [46].

As mentioned above, several studies have shown a remarkable success of treatments based on VEN combinations in the upfront setting in adult AML patients. However, about one third of patients treated with these drug combinations do not achieve remission. Thus, some studies have evaluated the capacity of ex vivo assays to predict VEN sensitivity in AML patients. The VenEX trial evaluated a drug sensitivity score using a high-throughput drug sensitivity test for VEN sensitivity in a group of 39 AML patients undergoing treatment with VEN-based regimens [47]. The ex vivo assays predicted response to VEN with an accuracy of 79% in relapsed/refractory patients and of 100% in upfront patients [47]. Importantly, for patients predicted to be VEN resistant, the response rate was 13% [47]. Finally, the ex vivo tests predicted benefit for VEN-based treatment with an mOS of 14.6 months for patients predicted to be sensitive and of 3.5 months for those predicted to be resistant [47]. The same authors more recently published the extended results of this clinical trial further showing that ex vivo VEN sensitivity predicts clinical response in AML: in untreated AML patients, ex vivo sensitivity corresponded to an 85% CR rate, with an mOS of 28.7 months compared with 5.5 months for ex vivo-resistant patients; for R/R AML and sAML, ex vivo sensitivity corresponded to 62% CR rate and mOS 9.7 months vs. 3.3 months for ex vivo-resistant AMLs [48]. The ex vivo test can be integrated into clinical studies to predict patients who respond to VEN [48].

Another study evaluated ex vivo drug assays with VEN-inclusive combinations that could be more efficient than HMA + VEN association. In this study 25 VEN-based combinations were evaluated using VEN + Azacitidine (VEN + AZA) as a reference: a total of 25 VEN-based combinations were evaluated, identifying several combinations with enhanced efficacy relative to VEN + AZA; some of these combinations were associated with specific patient characteristics [49]. According to the results obtained, this study subdivided three types of VEN-based regimens according to the AML phenotype: drug combinations more active in early progenitor leukemic phenotypes; drug combinations more active in differentiated leukemic phenotypes; drug combinations with broad efficacy, independently of differentiated state [49]. The authors then proposed a model for the VEN-based drug combinations for the different subtypes of AMLs, according to the most recurrent mutations: primitive AML phenotype, associated with IDH1-IDH2 or NPM1 mutations, is associated with sensitivity to VEN + AZA, cytarabine, BTK, MEK inhibitors; primitive or differentiated phenotypes with NRAS/KRAS, TP53 or SF3B1 mutations is associated with sensitivity to VEN with TKI, PARP, CDK4/6 or FLT3 inhibitors; differentiated phenotypes with NRAS/KRAS, TP53 or SF3B1 mutations or with all other types of mutations are associated with sensitivity to VEN with p38/MAPK, PI3K, JAK1/2 or multikinase inhibitors [49]. One of these drug combinations is represented by VEN+ with CDK4/CDK6 inhibitor Palbociclib; this drug combination was active in AML cells with loss of BAX or PMAIP, resistant to VEN [50].

2.5. Development of Precision Medicine for Older AML Patients

More recently, in some patients molecularly targeted treatments have been introduced, predominantly for refractory/relapsing patients. However, two factors have precluded a precision medicine approach in AML patients: (i) only a subset of AML patients have in their leukemic cells a molecular abnormality suitable for targeting therapy; (ii) the expedited initiation of standard treatments and the inability to rapidly obtain mutational data [51]. However, in spite of these difficulties, the BEAT AML Master Trial provided evidence that precision medicine treatment in elderly AML patients using prospective genomic profiling is feasible and compatible with an improvement in clinical efficacy [51]. Thus, the BEAT AML trial enrolled 395 older AML patients, cytogenetically and mutationally profiled within 7 days from sample receipt and then selected either for standard care or assigned to a sub-study based on the dominant clone [51]. Importantly, day 30 mortality was less frequent, and overall survival was significantly longer for patients enrolled on BEAT AML sub-studies compared to patients selected for standard of care [51]. Thus, this study showed that a precision medicine therapy is feasible with 7 days from initial hospital admission, allowing the incorporation of data on genomic alterations into decisions about the optimal treatment, without negatively affecting patients’ outcomes [51]. After this initial presentation of the BEAT AML trial, results obtained in the various sub-studies for specific AML subtypes have strongly supported the efficacy of the precision medicine approach. AML in older patients is a molecularly heterogenous disease with a genomic profile that is distinct from that observed in younger AML patients, characterized by higher prevalence of TP53 mutations and MDS-related gene mutations and by cytogenetic abnormalities, negatively impacting their response to treatments [52]. However, it is important to consider that almost 50% of AMLs observed in older AML patients display a mutation that is potentially targetable at molecular level for treatment, including FLT3-ITD, FLT3-TKD, IDH1, IDH2, NPM1 mutations and KMT2A rearrangements, thus supporting the evaluation of targeted therapy approaches in these patients [2].

It is important to note that other sub-studies of the BEAT umbrella master trial involve the enrollment of AML patients bearing mutations associated with unfavorable prognosis and treatable with drugs under experimental evaluation. An example of this type of study is given by a phase II sub-study of the BEAT AML master trial involving the treatment of newly diagnosed TP53-mutated AML patients with the spleen tyrosine kinase (SYK) inhibitor Entospletinib in association with decitabine; the study, however, showed only a low rate of CR and a short OS [53].

Similar to the BEAT study, the National Cancer Institute’s Myeloid Malignancies Molecular Analysis for Therapy Choice (MyeloMATCH) study provided an evaluation of patients with myeloid malignancies within a Master Screening and Reassessment Protocol (MSRP), providing a comprehensive genomic evaluation with 72 h and providing the information required for the patients to undergo a targeted treatment (either at induction or MRD evaluation or post-HSCT) [54]. In the study, patients with newly diagnosed AML or MDS are stratified into three different groups: MDS, older adult AML and younger adult AML, and the personalized treatments for these patients are proposed [54]. The NCI myeloid assay version 2 (NMAv2) uses the Genexus System, an automated platform enabling with <48 h turnaround from receipt to report for any leukemic specimen with 95% of sensitivity for 291 known AML-related mutations and 100% specificity [55].

3. FLT3-Mutant AMLs

FLT3 mutations represent one of the genetic alterations most frequently observed in adult AML. Two types of FLT3 mutations are observed in adult AML patients: more frequently, FLT3-ITD mutations are observed in 20–25% of cases, and less frequently, FLT3-TKD (tyrosine kinase domain) point mutations are observed in 5–10% of cases [56]. FLT3 mutations are more frequent in younger than in older AML patients [56].

For older patients who are unsuitable for intensive chemotherapy, there is not an established role of FLT3 inhibitors for the treatment of newly diagnosed patients. The standard of care for older AML patients is represented by the treatment based on the combination of a hypomethylating agent with VEN; however, this treatment does not seem to be very favorable in these patients in that FLT3-ITD AMLs showed a lower response to this treatment compared to FLT3-WT response, with an mOS of 9.9 months compared to a survival of 14.7 months, respectively [57]; furthermore, relapses are frequent in FLT3-mutated patients treated with HMA + VEN [58].

Gilteritinib is one of the most potent FLT3 inhibitors active against both FLT3-ITD and FLT3-TKD mutations. In the randomized phase III clinical study ADMIRAL, the administration of Gilteritinib in association with intensive induction chemotherapy resulted in improved OS in AML patients with R/R AML (mOS 9.3 months vs. 5.6 months, respectively) [59]. Importantly, in this study clinical benefit was observed both in patients aged <65 years and >65 years (HR 0.61 vs. 0.64, respectively) [59].

Given the inability of many older AML patients to fit the minimal requirements for intensive chemotherapy treatment, other studies have explored Gilteritinib in association with VEN. Thus, preclinical studies have shown a synergistic interaction between Gilteritinib and VEN [60]. A phase Ib clinical study carried out in AML patients with R/R FLT3-mutated AML and treated with Gilteritinib plus VEN showed a CRR of 75% and an mOS of 10 months, findings which compare favorably with those observed for Gilteritinib alone in the ADMIRAL trial [61]. However, a randomized phase III clinical trial failed to show any significant improvement in mOS of R/R FLT3-mutated AML patients treated with Gilteritinib plus Azacitidine with respect to that observed in patients treated with Azacitidine alone [62] (Table 1).

A recent phase I/II study evaluated AZA, VEN and Gilteritinib in two cohorts of FLT3-mutated AML patients: (i) newly diagnosed patients unfit for intensive chemotherapy; (ii) relapsed/refractory patients [63]. In the frontline cohort, the median age was 71 years and 96% of patients achieved a CR; with a median follow-up of 19.3 months, the mRFS and mOS were not reached; the 18-month RFS and OS rates were 71% and 72%, respectively; 65% of FLT3-ITD patients achieved an MRD negativity at <5 × 10⁻⁵ by NGS; 44% of NPM1-mutant patients achieved MRD negativity by NGS at 5 × 10⁻⁵; 43% of patients underwent allogeneic HSCT; five patients relapsed (17% of total), one after HSCT and four in the absence of HSCT [63]. Dose reductions were required in most patients because of myelosuppression. Since this triplet regimen showed a high rate of CR and of MRD negativity, it could be a good regimen for older AML patients, eligible or not for transplant, as it can spare these patients the toxicities of an intensive chemotherapy regimen (Table 1).

Table 1. Main clinical trials involving FLT3 inhibitors in older FLT3m AML patients.

Drug Name	Molecular Target	Clinical Trial (Phase)	Patient Number and Disease Status	Therapeutic Regimen	Trial Outcomes	Toxicity and Adverse Events
GILTERITINIB	FLT3	LACEWING (III)	123 (median age 77 yrs) ND FLT3m AML GIL + AZA 74 pts AZA 49 pts	AZA (75 mg/m2) Gilteritinib (80 or 120 mg/QD)	GIL + AZA mOS 9.8 mp AZA mOS 8.9 mo	Common AE Pyrexia 47% Diarrhea 38% Febrile Neutropenia 35% Constipation 34% Nausea 33%
GILTERITINIB	FLT3	NCT 041440487 (I/II)	52 (median age 71 yrs) 22, R/R FLT3m AML 30, ND FLT3m AML 73% FLT3-ITD (ND) 45% FLT3-ITD (R/R)	AZA (75 mg/m2) VEN (200–400 mg/QD)	ND AMLL CRR 96% 18-mo RFS 71% 18-mo OS 72% mRFS ans mOS no reached R/R AML CRR 27% mRFS 4.3 mo mOS 5.8 mo	AE, grade 3 or 4 Febrile Neutropenia 33% (ND). 45% (R/R) Infection 50% (ND) 59% (R/R)
QUIZARTINIB	FLT3	Yilmaz et al. 2025 [64]	73 (median age 70 yrs) 26, ND FLT3-ITD AML 47, R/R FLT3-ITD AML	DECITABINR (20 mng/m2) VEN (400 mg/QD) Quizartinib (30 mg or 40 mg/QD)	ND AML CR + CRi 92% MRD-neg (PCR) 71% mOS not reached R/R AML CR + CRi 60% MRD-neg (PCR) 28% mOS 6.3 mo	ND AML Pneumonia 38% Neutropenic fever 57% Infections 22% R/R AML Pneumonia 72% Neutropenic fever 62% Infections 47%

Table 1. Main clinical trials involving FLT3 inhibitors in older FLT3m AML patients.

Drug Name	Molecular Target	Clinical Trial (Phase)	Patient Number and Disease Status	Therapeutic Regimen	Trial Outcomes	Toxicity and Adverse Events
GILTERITINIB	FLT3	LACEWING (III)	123 (median age 77 yrs) ND FLT3m AML GIL + AZA 74 pts AZA 49 pts	AZA (75 mg/m2) Gilteritinib (80 or 120 mg/QD)	GIL + AZA mOS 9.8 mp AZA mOS 8.9 mo	Common AE Pyrexia 47% Diarrhea 38% Febrile Neutropenia 35% Constipation 34% Nausea 33%
GILTERITINIB	FLT3	NCT 041440487 (I/II)	52 (median age 71 yrs) 22, R/R FLT3m AML 30, ND FLT3m AML 73% FLT3-ITD (ND) 45% FLT3-ITD (R/R)	AZA (75 mg/m2) VEN (200–400 mg/QD)	ND AMLL CRR 96% 18-mo RFS 71% 18-mo OS 72% mRFS ans mOS no reached R/R AML CRR 27% mRFS 4.3 mo mOS 5.8 mo	AE, grade 3 or 4 Febrile Neutropenia 33% (ND). 45% (R/R) Infection 50% (ND) 59% (R/R)
QUIZARTINIB	FLT3	Yilmaz et al. 2025 [64]	73 (median age 70 yrs) 26, ND FLT3-ITD AML 47, R/R FLT3-ITD AML	DECITABINR (20 mng/m2) VEN (400 mg/QD) Quizartinib (30 mg or 40 mg/QD)	ND AML CR + CRi 92% MRD-neg (PCR) 71% mOS not reached R/R AML CR + CRi 60% MRD-neg (PCR) 28% mOS 6.3 mo	ND AML Pneumonia 38% Neutropenic fever 57% Infections 22% R/R AML Pneumonia 72% Neutropenic fever 62% Infections 47%

A randomized trial was proposed to confirm the results observed in this study [65]. Thus, the MyeloMATCH treatment trial will enroll FLT3-ITD- and FLT3-TKD-mutated AML patients older than 60 years and will randomize them to the treatment with AZA + VEN ± Gilteritinib; the patients will be stratified according to the age of patients (<70 years and >70 years) and to the burden of FLT3-ITD mutations (VAF < 33% vs. VAF > 33%) [65].

The S8 group 2 sub-study, part of the multicenter BEAT AML Trial, evaluated the safety and the efficacy of Gilteritinib in combination with Decitabine and VEN in untreated FLT3-mutated older AML patients with high and low variant allele frequency [66]. Triplet therapy induced in these patients a high response rate (CRc 61%), with an mOS not reached after a median follow-up of 19.7 months [66]. The standard dose of Gilteritinib (120 mg/day) in association with DEC and VEN induced significant hematological toxicities and required a dose reduction (80 mg/day), allowing a better safety profile [66].

A recent study by Yilmaz and coworkers confirmed the findings of this study, showing that the triplet Decitabine, VEN and quizartinib induced a high rate of responses among both ND and R/R FLT3-ITD patients. Thus, these authors have enrolled 73 FLT3-ITD patients (26 ND and 47 R/R) with a median age of 70 years, in a phase I/II study involving treatment with a triplet drug combination based on Decitabine, VEN and quizartinib [64]. Among ND patients, 92% CR + CRi with an MRD negativity of 71% and an mOS not reached were observed; among R/R patients, 60% of CR + CRi, with 28% of MRD negativity and an mOS of 6.3 months were observed [64]. The most frequent non-hematological adverse events were pneumonia, neutropenic fever and infections, more frequent among R/R than ND AML patients [64].

Two other clinical trials associating an FLT3 inhibitor with intensive chemotherapy enrolled both younger and older FLT3-mutant AML patients: the AML SG 16-10 trial explored induction chemotherapy (7 + 3) in association with Midostaurin and reported in patients older than 60 years a CR + CRi rate of 72.4% and 1-year OS rate of 59% [67]; the QUANTUM-first study involving treatment with quizartinib in association with induction chemotherapy reported an mOS of 17.5 months [68].

Particularly, the phase III QUANTUM-First study (NCT 02668653) demonstrated that in ND FLT3-ITD AML patients, adding quizartinib to standard chemotherapy ± allo-HSCT, followed by quizartinib or placebo for up to 3 years, decreased the relative risk of death of 22% vs. placebo [68,69]. A recent study analyzed the effect of continued therapy post-chemotherapy or post-allo-HSCT, showing that a clinical benefit for continued therapy with quizartinib versus placebo in ND FLT3-ITD AML patients, particularly for those without allo-HSCT, suggesting that quizartinib delayed or prevented relapse or death; the continued therapy with quizartinib increased the chances for patients who are MRD-positive at the end of the initial therapies to become MRD-negative; quizartinib administration in continued therapy provided benefit to both MRD-negative and MRD-positive patients [69].

The phase II INTERVENE trial involved the enrollment of 120 newly diagnosed AML patients who were at least 60 years old, ineligible for intensive chemotherapy and negative for adverse cytogenetics; these patients were randomly assigned to receive Midostaurin plus low-dose Cytarabine and VEN or low-dose Cytarabine and Venetoclax only [70]. The addition of Midostaurin improved responses and survival among patients with FLT3-ITD but did not improved outcomes in the other molecular subgroups: among patients with FLT3-ITD AML, the rate of CR + CRi was 82% in the Midostaurin arm and 57% in the control arm; among patients with FLT3-ITD AML, the mOS was 16.6 months in the Midostaurin arm and 8.8 months in the control arm [70].

More recently, the safety and the efficacy of Crenolanib, a second-generation tyrosine kinase inhibitor with activity against FLT3-ITD and FLT3-TKD-mutant AML, was evaluated. A recent study evaluated Crenolanib in association with intensive chemotherapy in both younger and older AML patients [71]. The ORR was 86%, of which 96% was in patients < 60 years and 80% in older patients; MRD negativity (as assessed by flow cytometry analysis) was observed in 89% of younger and 45% of older AML patients achieving a CR; in patients 60 years and younger, mOS was not reached and 3-year OS was 71.4%, while in older AML patients the mOS was 19.8 months, with an estimated 3-year OS of 33.3%; the mEFS was not reached for younger AML and was 7.9 months for older patients [71]. Older AML patients undergoing HSCT displayed a significantly lower cumulative rate of relapse and longer OS than those treated with chemotherapy alone [71].

A recent retrospective study on 619 newly diagnosed FLT3-mutated AML patients, all treated at the University of Texas, MD Anderson Cancer Center, was particularly informative about the current status of therapy of these patients [72]. A total of 337 patients received IC treatment (median age 53 years) and 282 patients received low-intensity therapy (median age 70 years). This analysis confirmed that in FLT3-ITD-mutated patients, the addition of an FLT3 inhibitor to IC improved both RFS (32.3 vs. 14.3 months) and OS (35.5 vs. 18.9 months), compared to treatment with IC alone [72]. In FLT3-TKD-mutant patients, FLT3 inhibitors elicited a statistically significant prolongation of OS and RFS [72]. In patients with FLT3-ITD mutations, those treated with triplet LIT regimens (LIT plus VEN and FLT3 inhibitor) displayed a longer OS compared to those treated with other treatment combinations (LIT alone, LIT plus VEN, LIT plus an FLT3 inhibitor): 19.1 months vs. 9.1–11.2 months, respectively [72]. Patients with FLT3-ITD plus NPM1 co-mutation showed a trend toward improved response to various treatments [60]. Patients with FLT3-TKD mutations favorably respond to triplet regimens, with a 2-year OS rate of 61% [72]. Allo-HSCT significantly improved and markedly improved the survival of patients undergoing IC and LITs, respectively [72]. In conclusion, these studies have supported the important role of FLT3 inhibitor inclusion into IC in chemotherapy-fit patients and into triplets including LIT and VEN in chemotherapy-unfit patients.

Mechanisms of Resistance of FLT3 Mutations to Venetoclax

Molecular signatures based on mutations in TP53, FLT3-ITD, NRAS and KRAS allowed a classification of patients treated with HMA + VEN combination into three groups with higher, intermediate and lower benefit: the higher benefit group including patients without TP53 mutations, FLT3-ITD and KRAS/NRAS mutations, associated with VEN resistance in AML (26.1 months OS); the intermediate-benefit group included patients without TP53 mutations and with FLT3-ITD or KRAS/NRAS mutations (12.1 months); the lower-risk benefit included patients with TP53 mutations (5.5 months) [73].

Although the combination of VEN and Gilteritinib doubles the response of AML patients, most of patients relapse in the absence of allo-HSCT. An understanding of the molecular mechanisms mediating resistance is required for the identification of therapeutic strategies that could bypass these resistances in AML cells resistant to VEN/Gilteritinib treatment. Although the mechanisms of resistance to the VEN/Gilteritinib combination are still to be determined, RAS pathway mutations and/or activation have been implicated in the resistance either to Gileritinib [73] or to VEN [74].

FLT3 mutations promote leukemic cell survival via the activation of RAS-MAPK, PI3KB and STAT5 pathways [75]. These molecular pathways downstream of FLT3 play a key role in modulating antiapoptotic proteins, such as Bcl-XL_L and MCL-1 that promote resistance to both VEN and HMAs [76].

Multiomic single-cell DNA/protein and RNA/protein analysis of patients treated with VEN and Gilteritinib showed that while VEN–Gilteritinib effectively eliminated FLT3-mutant clones, it promoted the selection of RAS pathway activation, RAS mutations and RAS-monocytic differentiation [77]. These observations indicate that RAS signaling pathway plays a key role in mediating FLT3 and BCL-2 inhibitor resistance [77].

4. IDH-Mutant AMLs

IDH mutations are frequent among older AML patients. An analysis of 3141 AML patients of different ages, from infants to very-old patients, showed that both IDH1 and IDH2 mutations are strongly correlated with increased age (3–4% in pediatric and 21% in older patients) [78] (Figure 1).

A recent study carried out in the context of the BEAT AML trial in 1023 newly diagnosed AMLs with a mean age of 72 years, reported 9.7% of IDH1 mutant, 18.9% of IDH2 mutant and 1% of double mutant [79]. IDH1 mutations were frequently associated with DNMT3A (42.4%) and with NPM1 (44.4%) mutations; IDH2 mutations were recurrently associated with DNMT3A (35.8%), NPM1 (31.1%) and SRSF2 (38.3%) mutations [79]. The co-mutation frequency of several genes is affected by patients’ age, with NPM1, FLT3 and NRAS mutations being less frequent in older than in younger AML patients and RUNX1 and ASXL1 being more frequent in older than in younger AML patients [78] (Figure 1). Both IDH1 and IDH2 mutant AMLs display consistent heterogeneity related to the presence of different co-mutations, affecting the sensitivity/resistance to IDH and menin inhibitors [80,81].

Newly diagnosed IDH1-mutant AML in older patients respond better to low-intensity treatments based on IDH1 inhibitors or to intensive chemotherapy monotherapy regimen in patients suitable for this treatment [79]. IDH2-mutant AML patients display benefit in OS deriving from treatment based on low-intensity treatments involving VEN and HMA [79].

Several clinical studies support a benefit deriving from the administration of IDH inhibitors to newly diagnosed older AML patients. Thus, the phase III clinical study AGILE showed that the administration of Ivosidenib (IVO) in combination with Azacitidine (AZA) resulted in a significant clinical benefit compared to AZA plus placebo, with a complete remission rate (CRR) of 54% vs. 16%, respectively, and a median overall survival (mOS) of 24 months vs. 7.9 months, respectively [82] (

Table 2. Main clinical trials using IDH inhibitors in older AML patients.

Drug Name	Molecular Target	Clinical Trial (Phase)	Patient Number and Disease Status	Therapeutic Regimen	Trial Outcomes	Toxicity and Adverse Events
Ivosidenib (IVO)	IDH1	AGILE (III)	146, ND IDH1m-AML	Azacitidine (75 mg/m2) + Ivosidenib (500 mg/QD) vs. Azacitidine + placebo	ORR 62% vs. 19% CR + CRi 58% vs. 19% mEFS at 12 mo 22.9 vs. 4.1 mo mOS 24 vs. 7.9 mo	Differentiation syndrome 4% vs. 4% Febrile neutropenia 20% vs. 34% Thrombocytopenia 20% vs. 15% Infection 21% vs. 31%
Ivosidenib (IVO)	IDH1	NCT03471260	37, ND IDH1m-AML	Azacitidine (75 mg/m2) + Ivosidenib (500 mg/QD) Venetoclax (400 or 800 mg)	CR + CRi 86% MRD-neg 81% mOS not reached 2-yr OS 73% 2 yr CIR 24%	Grade 3 or 4 events Hyperbiliribinemia 5% Differentiation syn 3% Transmaninitis 2%
Olutasidenib	IDH1	NCT02577406 (phase I/II)	67 (mean age 66 yr) R/R IDH1m AML	Olutasidenib (150 mg BID) AZA (75 mg/m2)	ORR 51% CR + CRi 31% mOS 12.5 mo mOS in patients achieving CR 36 mo	AE Grade 3–4 Anemia 25% Thromobocytopenia 37% Neutropenia 19% Leukocytosis 6%
Enasidenib (ENA)	IDH2	AG221-AML-005 (phase I/II)	101, ND IDH2m AML	Phase Ib: AZA (75 mg/m2) + ENA (100 mg or 200 mg/day) Phase II: AZA + ENA vs. AZA	AZA + ENA vs. AZA ORR: 76% vs. 33% CRR: 54% vs. 12% mOS: 22 vs. 22.3 months	Grade 3 or 4 events AZA + ENA vs. AZA Thrombocytopenia 37% vs. 19% Neutropenis 37% vs. 25% Anemia 19% vs. 22%
Enasidenib (ENA)	IDH2	NCT092719524 (phase I)	319 (mean age 72 yr) R/R IDH2m AML	ENA (100 mg/QD) vs. Conventional therapy	ORR 40.5% vs. 9.9% CR + CRi 29.7% vs. 6.6% OS at 12 mo 38% vs. 26% mEFS 4.1 mo vs. 2.6 mo	AE grade 3–4 Diff Synd 5% vs. 0% Hyperbilirubinemia 10.8% vs. 0% Anemia 5% vs. 19% Leukopenia 3% vs. 33%
Enasidenib (ENA)	IDH2	NCT04092179 (phase I/II)	27 (median age 70 yr) R/R IDH2m AML	VEN (400 mg) AZA (100 mg/QD)	ORR 70% CR 57% mOS 9.4 mo IDH2R172 ORR 83% CR 67% IDH2R140 ORR 55% CR 45%	AE grade 3–4 Neutropenia 41% Lung infection 22% Thrombocytopenia 26%

). Long-term results of the AGILE study confirmed these results, showing an mOS of 29.3 months in the group treated with IVO + AZA, compared to 7.9 months in the group treated with AZA alone [61]. Furthermore, hematologic recovery was faster and more durable, and conversion to transfusion independence was more common with IVO + AZA than with AZA alone [83].

A good therapeutic activity was observed also in R/R IDH1m AML patients using the combination of AZA with Olutasidenib, another IDHI inhibitor. Thus, Cortes et al. reported the study of 67 IDH1-mutant AML patients with refractory/relapsed disease (43% refractory and 57% relapsed); 31% of the patients achieved a CR + CRi, with a median duration of response of 15 months; 27% of the patients achieved a CR with a median duration of response of 20 months; 51% of the patients showed a response to treatment; the median OS in the whole population of patients was 13 months, 24 months in the population of all responding patients, 30.6 months in the group of patients with CR + CRi, and not reached in the patients with CR [84] (Table 1). Hematological grade 3–4 adverse events such as thrombocytopenia and anemia are the most common (Table 2).

Given the results of this study and the sensitivity of IDH1-mutant AMLs to VEN and AZA, it seemed logical to assess the response of these AMLs to the triplet IVO, AZA and VEN, showing a remarkable rate of responses: 96% of ORR, 75% of MRD negativity by flow cytometry, 5-year OS of 71.4%, mEFS of 50.4 months [85,86] (Table 2).

The phase Ib/II clinical study AG221-AML-005 compared Enasidenib (ENA) plus AZA versus AZA alone in patients with newly diagnosed, IDH2-mutant AML [87]. Compared to the AZA-only arm, ENA plus AZA resulted in a significant increase in ORR (36% vs. 76%, respectively) and CR rates (12% vs. 54%, respectively), but no significant differences in EFS and OS (22 vs. 22.3 months, respectively) between the two arms were observed [87] (Table 2). Grade 3 or 4 events of hematological toxicities were similarly observed in the two arms of treatment, thus suggesting that they are mainly related to the myelosuppressive effects of AZA.

Enasidenib was evaluated vs. conventional care in a population of 319 older IDHm R/R AML patients (mean age 72 years) [88]. Enasidenib elicited a significant improvement in the EFS and of the hematological parameters but failed to determine a significant increase in mOS compared to standard treatment; however, the interpretation of the data of this study concerning OS could be confounded by early dropout of non-responding patients and the use of subsequent AML therapies [88] (Table 2).

The association of ENA with VEN was explored in a group of R/R patients with AML or other myeloid malignancies, showing that ENA + VEN induced an increase in the rate of OR, compared to ENA alone [89]. The level of ORR and CRR was higher in AML patients bearing the IDH2^R172- than in the IDH2^R140-mutations (83% vs. 55%, respectively; 67% vs. 45%, respectively) [89]. Importantly, in the whole population of treated patients, the rate of ORs was high (70%), with 57% of CRs, and the median duration of responses was >1 years (16.6 months); three patients proceeded to HSCT following CR [89] (Table 1).

A phase Ib/II clinical trial evaluated the safety and the efficacy of the stepwise approach to the treatment of older patients unable to tolerate major hematologic toxicities IDH2-mutant AML, based on the initial treatment with ENA, followed by subsequent addition of hypomethylating agent therapy in patients who do not achieve an adequate response [90]. The phase Ib of the study evaluated the safety of adding AZA to ENA and the phase II assessed the ORR to ENA monotherapy [65]. In this study, the enrolled patients were selected according to the presence of the IDH2 mutant in the dominant leukemic clone [90]. The phase II of the study showed a composite complete response rate (CR/CRi) of 46%; the phase Ib, limited to 17 patients for whom ENA monotherapy did not induce a CR/CRi, transitioned to combination ENA + AZA therapy, showing a CR + CRi of 41%, with one dose-limiting toxicity [90] (Table 2).

The ongoing I-DATA phase II randomized clinical trial compares the safety and the efficacy between two sequences of first-line and second-line therapies in patients with newly diagnosed IDH1/2 AML patients not available for intensive chemotherapy [91]. In this study, the patients were randomized 1:1 to receive treatment first with an IDH inhibitor (IVO or ENA) plus AZA, then followed by AZA plus VEN or, alternatively, AZA plus VEN followed by an IDH inhibitor plus AZA [91].

IVO or ENA were combined with intensive chemotherapy in patients with newly diagnosed IDH-mutant AML, resulting in IDH1 or IDH2 mutation clearance in 39% and 23% of cases, respectively [92].

Since many older AML patients are unfit for intensive induction chemotherapy, low-intensity treatments have been developed using HMAs. In 2018, the HMA AZA combined with VEN treatment was approved for the treatment of AML patients unfit for intensive chemotherapy, according to the results of the VIALE-A trial showing an mOS of 14.7 months and a CRR of 66.4% for patients treated with AZA plus VEN, compared to an mOS 9.6 months and a CRR of 28.3% for patients treated with AZA alone [21,93]. Analysis of AML subgroups identified according to their mutational profile provided evidence that patients bearing IDH1/IDH2 mutations are particularly responsive to VEN + AZA treatment [94]. Thus, among patients with IDH1 mutations, mOS was 10.2 months and 2.2 months for VEN + AZA and AZA + placebo, respectively; among patients with IDH2 mutations, mOS was 27.5 months and 13.0 months for VEN + AZA and AZA + placebo, respectively [94]. Interestingly, 35% and 7% of the responder patients who survived >2 years in the VEN + AZA and AZA + placebo, respectively, had IDH1/IDH2 mutations [94].

Real-world data have confirmed a good sensitivity of IDH-mutant AML to frontline treatment with VEN + AZA treatment, with an mOS of 13.1 months for IDH1-mutant AMLs and 42 months for IDH2-positive AMLs [95].

Importantly, a recent real-world study compared IVO + HMA (181 patients) to HMA (99 patients) in IDH1m ND AML showing that treatment with IVO + HMA compared to HMA alone was associated with higher percentage of patients achieving CR + CRi (63% vs. 48.5%) and displaying a 6-month EFS (55.8% vs. 38.4%) [96]. The results of this study, particularly concerning the increased efficacy of IVO + HMA treatment, associated with a favorable toxicity profile, support this drug combination as the preferred standard of care treatment for patients with IDH1m ND-AML [96].

The large majority (90%) of patients achieving a CR following treatment with AZA + VEN were MRD-negative by flow cytometry evaluation; however, evaluation of MRD status by NGS showed that only 33% were MRD-negative; the mOS was significantly longer in patients achieving MRD negativity by NGS, compared to those MRD-negative by flow cytometry [97]. Given the sensitivity of IDH-mutant AMLs to both IDH inhibitors and to VEN + AZA, the three drugs were evaluated in association. Thus, a phase Ib/II study evaluated IVO with VEN ± AZA in IDH1-mutated AMLs. A total of 31 patients were enrolled in this study, with a mean age of 67 years, including treatment-naïve (29%), sAML (16%) and R/R-AML (26%), 9.7% of patients with prior IDH inhibitor treatment and with 2022 ELN risk group favorable, intermediate and adverse in 30%, 14% and 50% of cases, respectively. Both treatments, either IVO + VEN or IVO + Venetoclax + AZA resulted in 83% and 90% of CRs, respectively [98]. A total of 63% of treated patients achieved MRD negativity, with clearance of IDH1 mutations in 64% of patients receiving >5 cycles of treatment [76]. mOS was 42 months [98]. An updated analysis of this study with inclusion of additional patients showed that the 3-year OS was 71.4%, with an mDOR of 43.4 months [83]. A recent study evaluated the outcomes of frontline triplet regimens with an HMA, VEN and an IDH inhibitor for intensive chemotherapy-ineligible patients with IDH-mutant AML [62]. The patients were treated in the context of two different studies: the NCT03471260 study involved treatment with AZA + VEN + IVO in IDH1-mutated patients; the NCT04774393 study involved treatment with Decitabine + VEN + IVO or ENA in IDH1-mutant or IDH2-mutant AMLs [84]. The two studies enrolled a total of 60 patients treated with triplet regimens; the safety profile of the triplet regimens was like that observed in HMA + VEN or doublet regimens based on IDH inhibitors [84]. The composite CRR was 92%, with an ORR of 95%; the 2-year OS was 69%; patients with treated sAML exhibited inferior outcomes [62]. mOS was not reached in patients with IDH1 mutations vs. 35.2 months among those with IDH2 mutations; the 2-year OS rates were 73% and 65%, respectively; the 2-year EFS was not reached and 26.5 months, respectively; the cumulative incidence of relapse (CIR) was 24% and 24%, respectively [84]. The results of these studies based on triplet regimens strongly support additional studies of comparison with doublet regimens based on IDH inhibitors. Furthermore, a longer follow-up of these patients will be required to assess the long-term survival of IDH-mutant patients treated with triplet regimens.

A retrospective meta-analysis carried out in studies involving the treatment of newly diagnosed IDH-mutant AMLs with IDH inhibitors or AZA alone or AZA + VEN has provided indirect evidence that for IDH1-mutant and IDH2-mutant, AMLs, AZA + VEN represents the best treatment in terms of CRR and OS [99].

5. NPM1-Mutant and KMT2A-Rearranged AMLs

Both in younger and older adult AML patients, NPM1-mutant AMLs represent one of the most frequent AML subtypes. The treatment of NPM1-mutant AML is a complex medical problem in that this group of leukemias is heterogeneous at molecular, clinical and therapeutic levels [100]. The biology of NPM1-mutant AMLs is largely determined by their heterogeneous co-mutational profile, with a subgroup characterized by concomitant DNMT3A and FLT3-ITD mutations, associated with a poor response to current treatments and high rate of relapse [100]. NPM1m AMLs are characterized by several genetic features distinct from NPM1-WT AMLs, including a significantly higher frequency of DNMT3A, FLT3-ITD and TET2 mutations and of normal diploid karyotype [101].

NPM1 mutations are less frequent in older than in younger AML patients; particularly, NPM1m/FLT3-ITD AMLs are clearly less frequent in older than in younger AML patients [102] (Figure 2).

Furthermore, some genes such as TET2 and SRFSF2 are more co-mutated in older than in younger AML, while other genes such as WT1 and FTL3-ITD are less frequently co-mutated in older compared to younger AMLs [77] (Figure 3).

NPM1 gene mutations in AML patients are associated with favorable responses to chemotherapy and improved survival. However, older AML with NPM1 mutations have a lower survival compared to younger AML patients with NPM1 mutations. An early study from Falini and coworkers showed that NPM1m AML patients display a higher rate of CR and a lower rate of refractory disease compared to NPM1-WT AML patients; this benefit was limited to patients with <60 years [104]. A subsequent study by Ostronoff and coworkers provided clear evidence that the survival benefit of NPM1m is age-dependent and is lost in older patients; thus, the study of 743 cytogenetically normal AML patients comprised in the 55–65 year range showed an improvement in two-year OS compared to NPM1-WT patients, while patients comprised in the >65 years group displayed no improvement in 2-year OS compared to NPM1-WT AML patients [105].

A retrospective analysis carried out in 178 NPM1m AML undergoing treatment with IC showed that (i) the presence of both DNMT3A and FLT3-ITD co-mutations showed poorer CRR and OS; (ii) NPM1/DNMT3A/FLT3-ITD triple mutants displayed an extremely poor prognosis; (iii) the presence of IDH1/IDH2 mutations or PTPN11 mutations localized in the phosphatase (PTP) domain was associated with improved OS [106].

The BEAT-AML study reported the analysis of 246 older NPM1-mutant AML patients (18.9% of total), including 11% of t-AML and 3% with complex karyotype [107]. Most recurrent mutations included DNMT3A, TET2, FLT3-ITD, KRAS/NRAS and SRSF2 [107]. At gene expression level, older NPM1-mutated AMLs are characterized by enrichment in immunosuppressive signatures, while younger NPM1-mutant AMLs are characterized by enrichment in inflammation signature [108]. These patients were treated with different therapeutic approaches: intensive chemotherapy with an mOS of 41.6 months; VEN plus an HMA with an mOS of 21 months; in a sub-study, a therapy involving a menin inhibitor, with an mOS of 22 months; non-intensive therapy with an mOS of 6.3 months [107].

Farhat reported the results on clinical outcomes in a group of 396 newly diagnosed NPM1-mutant AML patients, including both younger and older patients and undergoing different treatments: patients treated with high-intensity chemotherapy (HIC) had an mEFS and mOS of 87.8 and 87.8 months, respectively; patients treated with Cladribine and low-dose Cytarabine plus VEN had an mEFS and mOS of 59.6 and 59.6 months, respectively; patients treated with HMA + Venetoclax (all >60 years) had an mEFS and mOS of 15.4 and 23.6 months, respectively; patients with an age of >75 years and treated with HMA + VEN had an mEFS and mOS of 10.3 and 10.9 months, respectively [101]. Among patients with NPM1m treated with HIC, predictors of worse outcome were age and the presence of FLT3-ITD mutations or extramedullary disease; in patients treated with HMA + VEN the only predictor of worse outcome was age [101].

The CAVEAT study explored VEN (at various doses) in combination with chemotherapy (Cytarabine and Idarubicin) in older AML patients fit for chemotherapy induction; the treatment was well tolerated with 4% of mortality [109]. In the whole population, the mOS was 19.3 months; in newly diagnosed AML patients, mOS was 33.1 months; in NPM1-mutant AML patients, mOS was 43.9 months [109] (

Table 3. Main clinical trials in NPM1m and KMT2Ar AML patients, involving VEN or menin inhibitors.

Drug Name	Molecular Target	Clinical Trial (Phase)	Patient Number and Disease Status	Therapeutic Regimen	Trial Outcomes	Toxicity and Adverse Events
VENETOCLAX	BCL-2	CAVEAT (Ib)	85, mean age 71 yr De novo AML 61% sAML 39%	IC (Cytarabine, Idarubicin) VEN (50–600 mg)	All AMLs CRR 75% mOS 19.3 mo De novo AML CRR 89% mOS 33.1 mo	AE grade 3 or 4 Febrile Neutropenia 55% Spesis 35% Localized infection 10%
VENETOCLAX	BCL-2	Retrospective study Bewersdorf et al. [84]	221, median age 68.6 yr ND NPM1m AML IC 147 HMA + VEN 74	IC HMA + VEN	IC vs. HMA + VEN CRR 85% vs. 74% 24 mo OS (all)b39% vs. 38% 24 mo OS (60–75 yr) 60% vs. 44%	Not Reported
REVUMENIB	MENIN	AUGMENT-101 (I/II)	84, median age 63 yr R/R NPM1m AML	Revumenib (160 mg/m2) Q 12 h in 28-day continuous cycles	CR + CRi 26% ORR 48% MRD-neg 63& in CR Duration of CR 4.7 mo	AE grade 3 or greater QTc prolongation 23.4% Anemia 14.3% Neutropenia 13.1% Diff syndrome 13.1% Thrombocytopenia 10.7%
REVUMENIB	MENIN	NCT 03013998 (I, in the context of BEAT AML Master Trial)	43, >60 yr ND NPM1m AML ND KMT2Ar AML DL1: mean age 75 yr DL2: mean age 69.5 yr	AZA (75 mg/m2) VEN Revumenib at two doses: DL1 (113 mg); DL2 (163 mg)	DL1 ORR 90.5% DL2 ORR 86.4% KMT2Ar ORR 100%, CRR 88.9% NPM1m ORR 85.3%, CRR 79.4% KMT2Ar mOS 18 mo NPM1m mOS 15.5 mo	AE grade 3 or greater Neutroppenia 26% Acute kidney injury 14% QTc prolongation 12% Hypokalemia 12% Diff syndrome 4%
BLEXIMENIB	MENIN	NCT 05453903	120, median age 66.5 yr 68 NPM1m AML 52 KMT2Ar AML R/R 86 patients ND 52 patients	AZA VEN Bleximenib (BL, 50 mg, 100 mg, 150 mg)	R/R AML BL 50 mg ORR 76% CRR 32% R/R AML BL 100 mg ORR 79% CRR 54% ND AML BL 50 mg ORR 77%. CRR 62% ND AML BL 100 mg ORR 92% CRR 85%	AE, grade 3 or greater Thrombocytopenia 53% Anemia 48% Neutropenia 46% Diff syndrome 4%
ZIFTOMENIB	MENIN	KOMET-007 (NCT 05735184) (I-II)	82 49 ND NPM1m AML 33 ND KMT2Ar AML	IC (7 + 3) Ziftomenib 600 m QD	NPM1m ORR 93% MRD-neg 68% KMT2Ar ORR 89% MRD-neg 83% At 25 wk, NPM1m OS 96% At 16 wk KMT2Ar OS 88%	AE, grade 3 or greater Neutropenia 47% Thrombocytopenia 35% Anemia 22%

). Bewersdorf and coworkers have compared intensive chemotherapy induction (IC) to HMA + VEN in a group of newly diagnosed NPM1m AML patients [110]. In the whole population of patients explored, CR + CRi was comparable in IC and HMA + VEN groups, while 24-mo OS was higher in IC than in HMA + VEN groups (59% vs. 38%, respectively) [88]. However, the 24-mo OS of patients aged ≥60 years was similar with IC vs. HMA + VEN after adjustments for clinicopathologic characteristics (60% vs. 44%, respectively) [110]. In patients with normal cytogenetics and without FLT3-ITD mutations, a better mOS was observed in patients treated with IC compared to those treated with HMA + VEN [110] (Table 3).

A retrospective analysis of IC vs. HMA + VEN for older AML patients with a favorable risk profile without FLT3-ITD mutations showed an mOS of 6.2 years for patients treated with IC and of 4.9 years for those treated with HMA + VEN [111]. Although a lower proportion of patients treated with HMA + VEN, compared to those treated with IC, did undergo HSCT, the OS of patients treated with HMA + VEN was equally good, as well as for those treated with IC [111].

Recent studies have suggested a new potential treatment for NPM1-mutant AML using menin inhibitors. Menin is a nuclear protein that interacts with proteins involved in regulating gene transcription and cell signaling and plays a relevant functional role in many biological processes, and particularly in the regulation of hematopoiesis, controlling myeloid cell proliferation and differentiation [112].

Although mutations in NPM1 determine the translocation of the protein from the nucleus to the cytoplasm, a small fraction of this protein remains in the nucleolus where it interacts with KMT2A: the resulting NPM1m-KMT2A-menin complex induces a deregulated expression of homeobox (HOX) and MEIS expression that induces a block of stem cell differentiation [109]. The pharmacologic blockade of the interaction between KMT2A and menin induces a rescue of the aberrant expression of these leukemogenic targets and a consequent antileukemic effect [113]. Six different menin inhibitors (Revumenib, Ziftomenib, Bleximenib, BMF-219, DS-1594, Enzomenib) have been identified and one of these compounds, Revumenib, showed consistent antileukemic activity in KMT2A-rearranged (KMT2Ar) and NPM1m-AMLs, leading to its approval in November 2024 for the treatment of R/R KMT2Ar AMLs [113]. Revumenib was investigated both in R/R and newly diagnosed NPM1-mutated AMLs.

In the phase II portion of the AUGMENT-101 trial 84 R/R NPM1m AML patients were evaluated and 74 patients met efficacy evaluable criteria. In this study, R/R NPM1m AML patients were treated with REV at 160 mg/m² for 28-day continuous cycles. The median age of these patients was 63 years and 51% were >65 years old [114]. The CR + CRi rate was 26%, with an ORR of 48%; the median duration of CR was 4.7 months and the median time to response was 2.76 months [114]. Of patients who achieved a CR, 63% were MRD-negative either by flow cytometry or by PCR [88]. A total of 5 of the 30 responding patients proceeded to HSCT [114]. A total of 78.6% of the treated patients experienced an adverse event and 59.5% experienced an adverse event of grade 3 or greater. The most common grade 3 or greater adverse events were GT_C prolongation (21.4%), anemia (14.3%), febrile neutropenia (13.1%), differentiation syndrome (13.1%) and thrombocytopenia (10.7%) [92] (Table 3).

In the context of the BEAT AML Master Trial, a phase I dose-escalation and expansion study of VEN, AZA and REV (at two dose levels) in older newly diagnosed AML with NPM1m or KMT2Ar was carried out [89]. This study involved the enrollment and treatment of 43 patients. The ORR, CRc, CR rates were 85%, 79% and 65% in NPM1m and 100%, 89% and 78% in KMT2Ar AMLs, respectively [115]. At 24 months, mOS was 15.5 months for NPM1-mutated and 15.5 months for KMT2A-rearranged AMLs; at 24 months, mEFS was 13.3 months for NPM1-mutated and not yet evaluable for KMT2A-rearranged patients [115]. Among patients classified, the ORR, CRc and CR rates were 86%, 76% and 72%, respectively, in patients classified as intermediate or adverse risk according to the 2024 ELN classification [115]. MRD status after therapy was evaluated in the majority of patients. A total of 100% of patients achieved MRD negativity by flow cytometry; in NPM1-mutated patients evaluable for MRD NPM1 NGS, 31% received MRD-negative results [115] (Table 3). Given the promising results of this study, a randomized phase III study aiming to compare AZA, VEN, REV with AZA, VEN and placebo in older/unfit NPM1-mutated AML patients and to determine whether REV addition may improve OS in these patients (NCT06652438) is under development.

An ongoing phase Ib dose-finding study (NCT0543903) evaluated the safety profile and the response to treatment based on AZA/VEN + Bleximenib at 50 mg, 100 mg or 150 mg twice daily, in a population of R/R and ND AML patients with NPM1m or KMT2Ar [90]. A total of 120 patients were included in the first report on this study, including 86 R/R and 34 ND patients, and 52 patients with KMT2Ar and 68 with NPM1m [116]. The median age of these patients was 66.5 years. Pharmacodynamic studies supported a greater antileukemic activity at 100 mg of Bleximenib compared to 50 mg. In the R/R patients, the ORR and CR + CRi for the 50 mg cohort were 76% and 32%, respectively, while in the 100 mg cohort were 79% and 54%, respectively; no further improvement in efficacy was observed in the 150 mg cohort. In the ND cohort the ORR and CR + CRi rates were lower with Bleximenib at 50 mg (77% and 62%, respectively) compared to 100 mg (92% and 85%, respectively) [94]. The rates of response were similar among NPM1m and KMT2Ar patients [116]. The most common grade 3 or greater adverse events were thrombocytopenia (53%), anemia (48%) and neutropenia (42%); differentiation syndrome was observed in 4% of patients [116] (Table 3).

Two phase I studies have investigated the safety and the efficacy of a menin inhibitor in association with induction chemotherapy. Thus, a phase Ib study (NCT05453903) study explored the menin inhibitor Bleximenib in association with the intensive chemotherapy in 13 newly diagnosed NPM1m AML patients with a median age of 58 years, reporting an ORR of 100%, with 77% of CR-CRh; 3 patients remained in treatment and 6 continued therapies with Bleximenib monotherapy [117].

KOMET-007 (NCT05735184) is a phase Ia/Ib dose-escalation and dose-expansion study evaluating Ziftomenib in combination with VEN/AZA, VEN alone, a cytarabine and daunorubicin in patients with AML with either KMT2Ar or NPM1m. A part of the KOMET-007 study explored the safety and the efficacy of the treatment of 23 newly diagnosed NPM1m AML patients with the menin inhibitor Ziftomenib in association with intensive induction chemotherapy, these patients had a median age of 66 years; none of these patients had FLT3 mutations and 29% had IDH1/IDH2 mutations [118]. The ORR was 100%, with 100% of CR and 76% of patients achieving an MRD-negative status [96]. After a median follow-up of 31 weeks, mOS and mDOR were not reached; five of these patients received allo-HSCT; no discontinuations due to adverse events or relapse were observed [118]. An updated report on the KOMET-007 study showed the results on 49 NPM1m and 33 KMT2Ar AML patients: in NPM1m patients, CR rate was 94% and 83% for patients with KMTAr; MRD negativity was 68% for patients with NPM1m and 83% for patients with KMT2Ar; at 25 wks of follow-up, 96% of patients with NPM1m are alive and at 16 wks of follow-up, 88% of patients with KMT2Ar remained alive [119] (Table 3).

AML with KMT2A rearrangements (KMT2Ar) form a group of aggressive and prognostically unfavorable AML, characterized by frequent resistance to standard treatments and a high rate of relapse. The KMT2Ar AMLs form a clinical and biological distinct entity, characterized by the presence of gene rearrangements of the KMT2A gene with more than 100 different molecular partners, frequently represented by MLLT3, MLLT4 and ELL [98]. KMT2Ar is observed in 3–10% of adult AMLs and is less frequent in older adult than in younger adult AMLs. The pathophysiology of these leukemias is mainly related to the formation of oncogenic chimeric fusion proteins that activate oncogenic pathways, such as HOXA and MEIS1; the KMT2A fusion proteins recruit menin as a cofactor, driving leukemogenesis through abnormal histone methylation. Menin inhibitors block the interactions between KMT2A and menin and, through this mechanism, inhibit leukemogenic effects of KMT2Ar [120].

Clinical studies carried out in R/R KMT2Ar patients, mainly young AML patients, have consistently shown an efficacy of menin inhibitors, such as Revumenib, Bleximenib, Enzumenib and Ziftomenib in these patients, with an ORR ranging from 20% to 60% and with a CRR ranging from 20% to 30% [120,121].

In the context of the BEAT-AML trial, the characterization of 39 older AML patients with KMT2A rearrangements was reported; 23% of these patients had treatment-related AML (t-AML) and 31% displayed a complex karyotype [85]. These patients were treated using different therapeutic approaches, reporting results dramatically different to these treatments: mOS was 12.7 months for patients treated with IC and 5.0 months for those undergoing HMA + VEN therapy; mOS was 2.7 months for patients treated with non-intensive treatments; mOS was 31.6 months for patients enrolled in a sub-study involving targeted treatment with a menin inhibitor [107].

Two phase I clinical studies evaluated menin inhibitors in association with intensive induction chemotherapy in newly diagnosed KMT2Ar AML patients. A first study (NCT05453903) enrolled 13 KMT2Ar AML patients undergoing treatment with Bleximenib in association with intensive chemotherapy and reported 88% of ORR, with 88% of CR; three of these patients proceeded to allo-HSCT [113]. A second study, KOMET-007, involved the treatment of 23 newly diagnosed KMT2Ar AML patients with Ziftomenib plus intensive induction chemotherapy; 83% of these patients achieved a CR and 10 patients proceeded to allo-HSCT; after a median follow-up of 19 weeks, mOS and mDOR were not reached; 96% of patients remained alive [118].

6. The Development of Triplet Regimens for an Efficacious Targeting of AML Disease in Elderly Patients

VEN-based regimens have completely reshaped the treatment landscape of older AML patients, providing high rates of responses, associated with a lower toxicity compared to standard chemotherapy and offering access in suitable patients to allo-HSCT [122]. However, to ensure durable responses, VERN must be associated with other drugs, including an HMA, and in suitable patients a molecular targeting drug (such as IDH, FLT3 and menin inhibitors) [123].

DAC10-VEN was the first prospective trial to report outcomes of a triplet therapy based on VEN/HMA and an FLT3 inhibitor and provided evidence for future prospective triplet combinations [124].

FLT3 or IDH or menin inhibitors-based triplet regimens in combination with VEN and HMA have shown feasibility and efficacy in patients unfit for intensive chemotherapy.

In this context, some recent studies have shown very promising clinical results in the context of phase I/II clinical studies. In addition to the triplet studies above analyzed, two studies recently presented at EHA 2025 Meeting are worth mentioning. A first Ib/II study (NCT04774393) proposed a triplet regimen for IDH-mutated myeloid malignancies involving Decitabine–Cedazuridine in combination with VEN and Ivosidenib (IDH1m AML) or Enasidenib (IDH2m AML) [125]. The study involved the enrollment of 105 patients (62 ND and 43 R/R), with a median age of 71 years [125]. Among ND patients, 87% achieved a CR + CRi and 91% of these responding patients were MRD-negative by FC; among ND patients, CRRs were 91% in de novo AML and 73% in tsAML [125]. Among R/R patients, 60% achieved CR + CRi, with an MRD negativity rate of 38% [125]. The mOS was not reached in the ND cohort and the 2-year OS was 63%; the mOS in the R/R cohort was 12.3 months, with a 2-year OS 34% [125]. The response rates were similar among IDH1 and IDH2-mutant patients. Non-hematologic adverse events occurred in 52% of patients, with 30% experiencing grade ≥ 3 [125].

Tuspetinib (TUS) is an oral kinase inhibitor that targets several kinases including SUK, WT and mutant forms of FLT3, mutant and not WT forms of KIT, and indirectly suppresses the antiapoptotic protein MCL1 [126]. The ongoing phase I/II TUSCANY clinical study (NCT 03850574) provided evidence that the triplet therapy based on Tuspetinib, VEN and AZA induces a high rate of CR and MRD negativity in ND AML patients [127]. Early results from 10 patients treated across three cohorts (40 mg, 80 mg, and 120 mg of Tuspetinib) in combination with VEN and AZA showed 100% of CR + CRi among patients treated at 80 mg and 120 mg dose levels; furthermore, among responding patients, 78% achieved MRD negativity [127]. These results appeared promising in the prospective of future studies in that among responding patients, there were patients with TP53, RAS and FLT3-ITD mutations [127]. The safety data showed a good tolerability profile, with no limiting toxicities.

7. Ongoing Barriers to the Development of Precision Pargeted Therapies in Older AML Patients

Ongoing barriers to the development of precision-targeted therapies in older AML patients include the paucity of myeloid specific targets, the limited number of pharmacologically targetable genetic alterations, the risk of on-target off-tumor toxicities, and the consistent diversity of somatic mutations, both between patients with the same malignancy and within the tumor cell population of any given patient [128]. These limitations are particularly pronounced for older AML patients (>65 years compared to younger patients); in fact, Bataller et al. in a large cohort of AML patients molecularly characterized and treated in a single institution (875 patients, 34% <65 years and 66% >65 years) showed 44% of targetable patients among older AMLs compared to 60% among younger AML patients [2]. This difference is basically related to the enrichment of FLT3 mutations and KMT2A rearrangements in young compared to older AML patients (for FLT3, 37% vs. 24%, respectively; for KMT2Ar, 12% vs. 3%, respectively) [2]. Older AML patients presented with a higher incidence of adverse cytogenetic risk abnormalities, such as complex karyotype or chromosome 5, 7 and 17 abnormalities and of adverse-risk mutations, such as ASXL1, RUNX1 and TP53 mutations. These findings indicate that older patients with AML have poor outcomes due to general risks usually associated with aging, such as a higher incidence of comorbidities or concomitant diseases, but also because their AML diseases are enriched in unfavorable genomics that are currently untargetable.

The complexity of the hemopathology diagnoses needed for targeted treatment of AML patients requires the timely definition of molecular profiling and cytogenetic characterization. The studies carried out in the context of the BEAT AML master trial [51] and of the MyeloMATCH study [54] have shown that, using automated platforms for NGS analysis, 3–7 days are sufficient for a complete genomic analysis for each patient, a time compatible for any type of clinical study. However, these advanced automated genomic platforms are currently available only in few medical centers. Similarly, the personalized AML treatments require also a reduction in turnaround time of conventional analysis, whose data need to be integrated with those obtained through the analysis of mutational profiling [129].

Although precision medicine drugs in AML patients cause less toxic effects than standard chemotherapy, some important adverse events induced by these drugs must be considered. Thus, an important cause of toxicity elicited by IDH inhibitors, FLT3 inhibitors and menin inhibitors is represented by differentiation syndrome, a serious side effect related to the massive induction of differentiation of leukemic cells due to a large and rapid release of cytokines from differentiating leukemic cells; differentiation syndrome is characterized at clinical levels by various signs and symptoms, including fever, cough, trouble breathing, weight gain, swelling of the arm, legs, and neck, accumulation of excess fluid around the heart and lungs, lowering of blood pressure and kidney failure [130,131]. Differentiation syndrome can be life-threatening. In QuANTUM-First trial, it was reported that about 30% of patients treated with quizartinib displayed a QT interval of more than 450 msec compared to 17% in placebo arm; grade 3/4 QT prolongation were observed in 3% of quizartinib-treated patients [107].

Precision drug medicines are not curative, and no patients are cured without subsequent allo-HSCT. The lack of curative effects of these drugs are usually related to the development of resistance mechanisms due to on-target mechanisms such as mutations occurring at the level of the drug-target determining drug resistance or secondary to off-target emerging mutations (such as the development of RAS mutations in patients treated with FLT3 inhibitors or menin mutations in patients treated with menin inhibitors) [132,133,134].

Another important limitation to the development of targeted therapy in older AML patients could be related to the time required for the full molecular characterization of each patient before initiation of treatment. In ND AML patients, immediate initiation of treatment is standard of care. A recent study, retrospectively carried out on two large cohorts of older AML patients undergoing treatment with VEN-based therapy, provided evidence that deferral of antileukemic therapy did not impact outcome. Thus, the mOS was similar in patients starting treatment earlier or later than 10 days after initial diagnosis; importantly, the frequency and the extent of adverse events, such as severe infections, bleeding and thromboembolic events were similar in the group of patients starting earlier or later treatment [135]. Thus, delaying the start of VEN-based therapy in ND older AML patients, may represent a clinically acceptable option, under conditions of a close clinical monitoring [135].

Many older AML patients are unsuitable for standard treatments and for precision medicine-based treatments. The evaluation of individual patient fitness takes into account clinical parameters, assessment of the accompanying comorbidity burden, personal vulnerability and socioenvironmental frailties. The increasing availability of novel agents for the treatment of AML patients, expanding the therapeutic opportunities for less-intensive options, represents an element of additional complexity to the fitness assessment process [136]. Recently, an expert panel supported by the European Leukemia Net proposed new criteria for a more comprehensive evaluation of AML patient fitness, more carefully considering the complex problems raised by older patients for their consistent clinical, personal and social frailties [136]. The final aim of fitness assessment is to deliver the right therapy to the right patient and is a fundamental step in the process of the initial evaluation of an older AML patient.

8. Conclusions

The treatment of older AML patients remains particularly challenging. At the moment of patient hospital admission, it is fundamental to perform all the laboratory tests allowing appropriate risk stratification, including morpho-cytogenetic analysis and next generation sequencing to assess mutational profile. The evaluation of the fitness of older AML patients is a challenging process that requires a careful evaluation of the characteristics of each individual patients and of the AML disease. The development of HMA + VEN represents fundamental progress and has provided an important therapeutic option for patients unfit for intensive chemotherapy. Analysis of real-world studies have provided evidence that HMA + VEN regimens have an efficacy comparable to that observed in fit older patients treated with intensive chemotherapy and superior to that observed with other low-intensity treatments.

The development of studies of molecular characterization of the genetic abnormalities observed in AML have led to the identification of a significant proportion of these patients of driver genetic abnormalities that can be targeted using potent and specific pharmacologic inhibitors. Thus, in the last decade, the number of drugs available for the treatment of AML patients considerably increased through the regulatory approval of FLT3 inhibitors (Midostaurin, Gilteritinib and quizartinib), IDH inhibtors (Ivosidenib, Enasidenib and Olutasidenib), menin inhibitors (Revumenib). Now, these drugs can be used in the frontline treatment of older AML bearing FLT3 or IDH or NPM1 mutations or KMT2A rearrangements either in association with intensive chemotherapy or with hypomethylating agents and Venetoclax.

Future randomized clinical trial will carefully define the therapeutic impact of drug associations including FLT3 or IDH or menin inhibitors with the specific aim of improving their antileukemic efficacy and their safety through mitigation of adverse secondary/toxic effects.

The treatment of older NPM1-mutant AML patients remains challenging, both for the clinic-biological heterogeneity of these patients and for the limited response to current treatments; in this context, ongoing and future clinical studies involving the association with IC or the association of HMA + VEN+ a menin inhibitor could lead to an improvement in the survival of these patients.