Before discussing the role of maintenance therapy in the management of AML, it is important to discuss how we identify patients who have a high risk of relapse. One of the long-standing ways that we do this is through risk categorization.

The most commonly used system for risk categorization, I would argue, is the European LeukemiaNet (ELN) system, which was most recently updated in 2022.¹ The 2022 ELN risk stratification system for AML is based on patients’ genetics or molecular mutations and the cytogenetic features of their disease. This information is usually obtained at the time of diagnosis and often checked throughout the time of their treatment course.

According to the ELN system, patients with AML are categorized as having favorable, intermediate, or adverse risk. For example, mutated NPM1 on its own is considered favorable risk, whereas an NPM1 mutation with an accompanying FLT3-ITD mutation is considered intermediate risk. Examples of genetic abnormalities associated with adverse risk include RUNX1, ASXL1, and EZH2.

Patients who are favorable risk typically have a lower risk of relapse and often are treated with approaches that do not require bone marrow or stem cell transplant. Patients who are intermediate or adverse risk often will receive alloHSCT as part of their treatment course because of their increased risk of relapse.

This is a recent study that validated ELN 2022 by looking at relapse-free survival (RFS) and OS according to risk group.² This shows that the patients with AML stratify nicely into the 3 groups of favorable, intermediate, and adverse risk in both the RFS curves and the OS curves.

 When you look at RFS on the left side, you will see that after 5 years, approximately 85% of patients who have an adverse risk in cytogenetics (orange curve) are going to have either relapsed or died. So, this is definitely a group with an increased risk of relapse.

Here are the results of a meta-analysis of many studies looking at MRD status at induction or during or after consolidation and the impact of survival and relapse.⁴ You can see on the OS curves on the left and the disease-free survival (DFS) curves on the right that patients who are MRD negative fare much better than patients who are MRD positive (5-year OS: 68% vs 34%; 5-year DFS: 64% vs 25%). There is a big difference in these curves favoring the MRD-negative population.

So, back to the original concept of this section, identifying patients who are more likely to relapse: These would be patients who have higher-risk disease status at baseline and patients who are MRD positive.

The goal of initial induction is to bring AML into remission, and treatment can vary greatly depending on patient characteristics. Fit patients typically will be treated with intensive chemotherapy with the addition of targeted therapy (eg, FLT3) as clinically appropriate.⁵ Patients ineligible for intensive chemotherapy usually are prescribed prescribed a hypomethylating agent (HMA) plus venetoclax, targeted therapy with or without an HMA, or low-dose cytarabine with glasdegib. After remission, patients may go on to receive consolidation therapy consisting of limited-duration chemotherapy or alloHSCT. In some situations, maintenance therapy may be used after consolidation during a longer period of time to reduce the risk of relapse.

Before getting into the data, I’d like to begin with a summary of the most recent National Comprehensive Cancer Network (NCCN) guidelines on postchemotherapy maintenance for patients with AML. Patients with intermediate- or adverse-risk disease who received intensive chemotherapy and who have their AML in remission with no transplant planned are recommended to receive maintenance therapy until progressive disease or unacceptable toxicity. Patients could complete no consolidation, some consolidation, or a recommended consolidation course. Currently, the category 1 recommendation for maintenance therapy in this setting is oral azacitidine. There are also data for the use of azacitidine or decitabine, the more conventional HMAs, as postchemotherapy maintenance. It is important to note that oral azacitidine is not intended to replace consolidation chemotherapy or transplant. There are no data to suggest that it can be used in lieu of those approaches. Those approaches should still be done if the patient is a candidate.

For patients with a history of FLT3-ITD mutation who received quizartinib as frontline therapy and who do not have alloHSCT planned, there is now a recommendation for the use of quizartinib maintenance therapy.

First, let’s look at the data from the randomized phase III HOVON97 study, which looked at azacitidine maintenance vs observation after intense chemotherapy in 116 patients.⁶ Median DFS favored azacitidine maintenance vs observation, with a median DFS of 15.9 months vs 10.3 months (P = .04). OS, on the other hand, was not statistically significantly different between the 2 arms (P = .21), although there was a numerical advantage in the azacitidine maintenance arm vs observation (12-month estimated OS: 82% vs 63%).

The HOVON97 data suggested a beneficial role for HMA maintenance in some patients after chemotherapy. Subsequently, an oral azacitidine formulation was developed and tested in the QUAZAR AML-001 trial.

QUAZAR AML-001 was a randomized, double-blind, placebo-controlled phase III trial in 472 patients 55 years of age and older with AML and intermediate or poor cytogenetics who had achieved a CR/CRi after intensive chemotherapy with or without consolidation and were not eligible for transplant.^7,8

Patients were randomized to receive either oral azacitidine 300 mg daily for 14 days every 4 weeks or placebo for 14 days every 4 weeks, and they had to begin therapy within 4 months of achieving a CR.

Response was assessed every 3 cycles, and if patients achieved or maintained a CR/CRi, they would continue treatment. If there was a residual blast population of 5% to 15%, they could have an optional treatment escalation, including both azacitidine and placebo, and if they had 15% or more blasts, they would have to discontinue treatment. Patients otherwise would continue receiving treatment until death, withdrawal of consent, and/or loss to follow-up.

The primary endpoint for QUAZAR AML-001 was OS, and the key secondary endpoint was RFS.

QUAZAR AML-001 met its primary endpoint with a superior median OS for oral azacitidine vs placebo of 24.7 months vs 14.8 months (HR: 0.7; P = .0009).⁷ At 3 years, the estimated OS was 37.4% vs 27.9%, which was an absolute difference of approximately 10%. These results led to the FDA approval of oral azacitidine for continued treatment of adult patients with AML who achieved first CR/CRi following intensive induction chemotherapy and are not able to complete intensive curative therapy.⁹

Looking at subgroups, patients with NPM1-mutated AML did quite well on the oral azacitidine maintenance, with a significantly longer median OS vs placebo (47.2 months vs 15.9 months).¹⁰ NPM1 mutation status is overall prognostically favorable and independently predictive of increased OS with oral azacitidine.

Patients who were MRD positive at baseline had improved survival with oral azacitidine vs placebo (14.6 months vs 10.4 months).¹¹ Something that was interesting about this data set is that there were patients who actually responded—in the sense that they were able to clear their MRD while they were receiving oral azacitidine. In fact, some cleared their MRD while receiving placebo. The MRD clearance rate was 37% with oral azacitidine and 19% with placebo. Many of these responses were early (≤6 months), although there were responses seen after 6 months of treatment

In terms of safety, there were more adverse events (AEs) in the oral azacitidine arm compared with the placebo arm.⁸ The AEs that stand out included gastrointestinal (GI) toxicities such as nausea, vomiting, and diarrhea, as well as hematologic AEs such as neutropenia, thrombocytopenia, anemia, febrile neutropenia, and leukopenia. Hematologic malignancies were associated with more grade 3/4 AEs. Of note, although the GI toxicities were more numerous, there were not many grade 3/4 toxicities, but the grade 3/4 events generally were higher than what was seen in the placebo arm.

In QUAZAR AML-001, AEs led to discontinuation in 13% of patients receiving oral azacitidine vs 4% receiving placebo. Similarly, there were more treatment-emergent AEs (TEAEs) that led to dose reduction (16% vs 3%) or dose interruption (43% vs 17%) on the oral azacitidine arm vs the placebo arm.

Here are the recommendations for managing neutropenia and thrombocytopenia with oral azacitidine treatment.⁹ When the neutrophil count is <500 cells/mL, the recommendation is to hold oral azacitidine until improved to >500 cells/mL, and then one can resume at the same dose. For febrile neutropenia with neutrophil count <1000 cells/mL, treatment should be held and then resume after resolution. If febrile neutropenia occurs in 2 consecutive cycles, then this prompts a dose reduction of the oral azacitidine to 200 mg. If the febrile neutropenia persists after dose reduction, the recommendation is to reduce the dosing cycle by 7 days. After adjustment in dose and cycle length, persistent febrile neutropenia would prompt a discontinuation of the oral azacitidine. There are similar recommendations for thrombocytopenia with platelets <50 cells/mL with bleeding. On the first occurrence, oral azacitidine should be held and then resumed after resolution. If the thrombocytopenia occurs in 2 consecutive cycles, the treatment needs to be held and then resumed at 200 mg after resolution, and the treatment cycle should be shortened by 7 days. If thrombocytopenia happens again, then treatment should be discontinued.

In practice, GI toxicity can be a bothersome AE. The recommended management approaches are similar to what I described for cytopenias. If one develops grade 3/4 GI toxicity, then you would need to hold the drug and resume at the same dose once it resolves to grade 1. If it recurs, then reduce the dose to 200 mg, and if it recurs again, drop to a 7-day cycle. If the severe GI toxicity persists despite these measures, oral azacitidine will need to be discontinued. In my own practice, I give patients an antiemetic before each dose of oral azacitidine to reduce the chance that patients will experience nausea or vomiting. Loperamide can be useful for patients who have diarrhea.

I mentioned earlier that the NCCN guidelines now recommend the use of quizartinib as maintenance therapy, and that is based on the recently reported QuANTUM-First study.

QuANTUM-First was a randomized, double-blind, placebo-controlled phase III trial that looked at quizartinib plus chemotherapy in 539 patients 18-75 years of age with newly diagnosed FLT3-ITD‒positive AML.^12,13 Patients received up to 2 cycles of induction chemotherapy with cytarabine and daunorubicin or idarubicin—the classic 7+3 regimen—with quizartinib or placebo given on Days 8-21 of each cycle. The patients with a CR could get up to 4 cycles of high-dose cytarabine consolidation with quizartinib (or placebo) and/or transplant per investigator discretion and institution policies. Then, patients who either finished consolidation or went to transplant could continue dosing of quizartinib monotherapy for up to 3 years or 36 cycles.

The primary endpoint of the study was OS. Secondary endpoints included event-free survival, response rates, and MRD negativity. Other exploratory endpoints included RFS and duration of CR.

QuANTUM-First met its primary endpoint of improved OS. The blue line represents quizartinib, with a median OS of 31.9 months vs 15.9 months with placebo (HR: 0.78; P = .032). The absolute difference was 16.8 months.

These data led to FDA approval of quizartinib for patients with newly diagnosed AML with FLT3-ITD mutation, in combination with standard cytarabine and anthracycline induction and cytarabine consolidation, and as maintenance monotherapy after consolidation chemotherapy.¹⁴

This is a post hoc analysis of OS in patients with CR with or without transplant. This is a descriptive analysis, but it is interesting.¹³

On the left is the OS curve for patients who have achieved a CR and received a transplant on first remission. On the right are patients who had a CR who did not receive a transplant on first remission. Both figures are numerically in favor of the quizartinib arm, with HRs of approximately 0.6 in both analyses. This suggests that 36 months of quizartinib is providing benefit with or without transplant. We will discuss additional data on posttransplant quizartinib maintenance later in this module.

Quizartinib, like all drugs, has some toxicity. On the QuANTUM-First trial, there was more neutropenia in the quizartinib arm vs the placebo arm (20% vs 10% all grades; 18% vs 9% grade ≥3) and more febrile neutropenia (44% vs 42% all grades; 43% vs 41% grade ≥3).^12,13

These are some potential cardiac AEs specific to quizartinib. There was more QTc prolongation with quizartinib compared with placebo. In total, 13.6% of patients receiving quizartinib had any form of treatment-emergent QTc prolongation compared with 4.1% of patients receiving placebo. Approximately 2.3% of patients receiving quizartinib had QTc that was >500 ms, for example, and 10% had a QTc increase from baseline of >60 ms. Other cardiac TEAEs reported with quizartinib include cardiac arrest/ventricular fibrillation and ventricular tachycardia, which occurred in <1% of patients.

In this trial, 2 patients had cardiac arrest with recorded ventricular fibrillation in the setting of severe hypokalemia while receiving quizartinib, and 1 patient who received quizartinib died in their sleep.

QTc prolongation led to treatment discontinuation for 2 patients receiving quizartinib.

Quizartinib has a new approval, and the QT prolongation and other cardiac AEs are potentially fatal, so it is important for healthcare professionals to be aware of the monitoring and management recommendations.

ECGs should be performed prior to initiation of induction or consolidation chemotherapy, once weekly, and then as clinically indicated. ECGs also are recommended prior to initiation of maintenance therapy, then once weekly for at least the first month and following dose initiation and escalation. To manage QT prolongation, dose modification is not needed unless QTcF is >480 ms. For QTcF 481-500 ms, we would reduce but not interrupt the quizartinib dose. If QTcF is >500 ms, we would interrupt the quizartinib and then resume at a lower dose until QTcF is <450 ms. If QTcF >500 occurs more than once, we would permanently discontinue the drug.

It is also important to note that quizartinib has a Risk Evaluation and Mitigation Strategy program that healthcare professionals need to enroll in because of the risk of QT prolongation.

I would like to wrap up this section with a review of a few ongoing postinduction chemotherapy maintenance trials. This is by no means every trial that is out there, but it is an interesting group.

Several trials are evaluating targeted therapies as maintenance following induction chemotherapy. The FLT3 inhibitor gilteritinib is being compared with midostaurin in the phase III HOVON 156 AML trial (NCT04027309) and as monotherapy maintenance in the phase III GOSSAMER trial (NCT02927262). Quizartinib maintenance is being evaluated in patients with untreated non–FLT3-ITD AML in a phase II trial. Crenolanib is being compared with midostaurin in a phase III trial of patients with newly diagnosed FLT3-mutated AML (NCT03258931). For newly diagnosed IDH-mutated AML, the HOVON 150 AML trial (NCT03839771) will be evaluating the efficacy of the IDH inhibitors ivosidenib or enasidenib in combination with induction and consolidation chemotherapy and then as maintenance therapy. I am interested to see how these trials will turn out.

Another interesting trial is a phase I trial (NCT05010772) that is aiming to personalize maintenance based on the patient’s specific genotype. This trial will combine decitabine with venetoclax if there are no particular mutations. Patients will receive decitabine with gilteritinib if they have a FLT3 mutation or decitabine plus ivosidenib or enasidenib if they have an IDH1 or IDH2 mutation, respectively. So, this is kind of a personalized medicine approach to maintenance, and it will be interesting to see how this trial pans out.

There is also a question about whether we can improve on oral azacitidine outcomes. The VIALE-M trial (NCT04102020)—where the M stands for maintenance—is ongoing. VIALE-M is comparing venetoclax plus oral azacitidine vs oral azacitidine alone in patients with newly diagnosed AML in CR/CRi after intensive induction and consolidation chemotherapy. The primary endpoints are safety and RFS. VIALE-M is active and no longer recruiting, so data are eagerly anticipated.

Next, I want to discuss maintenance in the posttransplant setting. The current NCCN guidelines recommend the use of FLT3 inhibitors as maintenance therapy for patients who underwent alloHSCT, are in remission, and have a history of a FLT3 mutation.⁵ The panel recommends consideration of sorafenib, midostaurin, gilteritinib, and quizartinib. Of note, quizartinib is only for patients with FLT3-ITD‒positive disease.

There is literature on using HMAs as posttransplant maintenance, but the data have not translated into standard practice.

Oral decitabine and IV azacitidine have been investigated in posttransplant maintenance. An early study evaluating IV azacitidine compared with supportive care has shown improvements in RFS and OS, but the results were not statistically significant.¹⁵ Oral azacitidine also has shown promise in a single-arm study. Other favorable studies have shown reduction in the rate of relapse with decitabine,¹⁶ but again it has not necessarily led to widespread use of these drugs in the posttransplant setting.

One interesting study is the RELAZA2 trial, which looked at patients who developed kind of an MRD relapse state, either post induction or post transplant. This study showed that the initiation of azacitidine could delay or prevent frank hematologic relapse.¹⁷

There are several studies in the literature showing the benefit of the FLT3 inhibitor sorafenib in the posttransplant setting for patients with AML. One of these trials was the SORMAIN trial. SORMAIN was a randomized, double-blind phase II trial that evaluated sorafenib maintenance in 83 patients with FLT3-ITD‒positive AML who underwent alloHSCT.¹⁸ Patients 60-100 days post transplant and in remission with normal peripheral blood counts were randomized to receive sorafenib 2 x 400 mg daily or placebo for up to 24 months. We also should note that this study was closed early because of slow accrual.

The primary endpoint for this trial was RFS, which was defined as death from any cause or relapse of AML. Secondary endpoints included OS, RFS and OS in patients with NPM1-mutated vs wild-type disease, RFS and OS based on FLT3-ITD ratio, safety, and biomarker analysis.

The primary endpoint of RFS improvement was met despite the study being stopped early because of slow accrual. The median RFS was not reached vs 30.9 months for sorafenib vs placebo. The HR was 0.39 (P = .013), and this was statistically significant in favor of sorafenib vs placebo. The 24-month RFS estimates were 85% for sorafenib vs 53.3% for placebo.

Median OS was not reached in both arms of the trial, with an HR of 0.516 in favor of sorafenib (P = .0855). The 24-month estimated OS with sorafenib was 90.5% vs 66.2% in the placebo arm (HR: 0.241; P = .007).

So, this really is a positive trial showing a benefit to sorafenib maintenance in patients with FLT3-ITD‒positive AML. Sorafenib can be difficult to use, and dose reductions due to AEs are common. On the SORMAIN trial, 48% of patients receiving sorafenib vs 40% receiving placebo had dose reductions, and 22% vs 5% of patients had treatment discontinuation due to AEs. Common AEs on the sorafenib arm were graft-vs-host disease (GvHD), infections, electrolyte alterations, and GI or skin toxicity. The most common drug-related AEs were electrolyte alterations (7.1%), GI toxicities (4.8%), skin toxicities (4.8%), and infections (2.4%).

A study similar to SORMAIN was performed in China. This was an open-label, randomized phase III trial comparing sorafenib 400 mg twice daily with placebo from 30-60 days post transplant.^19,20

The 5-year follow-up of this trial was published recently, and it was also positive.²⁰ The 5-year OS HR of 0.55 was statistically significant (P = .011) in favor of sorafenib vs placebo. The cumulative incidence of relapse was lower with sorafenib (5-year HR: 0.33; P = .0003), and the leukemia-free survival was also better for sorafenib vs placebo (5-year HR: 0.47; P = .0007).

As I mentioned with the SORMAIN study, AEs can be problematic with sorafenib. In this trial, there were more AEs with sorafenib vs placebo. Common AEs include infections, GvHD, and hematologic toxicities. Overall, however, this is a second study showing that sorafenib maintenance improves outcomes for patients with FLT3-ITD‒positive AML after alloHSCT.

RADIUS was another randomized phase II trial that enrolled patients with FLT3-ITD‒positive AML in CR1 after alloHSCT. This trial compared midostaurin plus standard-of-care maintenance with standard of care alone after alloHSCT. Patients received maintenance for up to 12 cycles, and the primary endpoint was RFS 18 months post alloHSCT. Secondary endpoints of importance were OS, RFS at 24 months post alloHSCT, and safety.

This was a negative trial. The midostaurin arm potentially had a numerical benefit, with 18-month RFS 89% vs 76% (HR: 0.46), but this was not a statistically significant difference (P = .2655). The story was similar for OS. The estimated 24-month OS was 85% for the midostaurin arm vs 76% for standard of care (HR: 0.58; P = .3418).

As I mentioned earlier, there are data for the use of quizartinib in the posttransplant setting. This again comes from the QuANTUM-First trial, where patients continued to receive quizartinib for up to 36 cycles as maintenance.¹² Here are the results of a post hoc analysis showing OS based on MRD status in patients who received quizartinib and alloHSCT.²¹ In patients who were MRD negative, the median OS was not reached for both quizartinib and placebo, but the results were numerically in favor of quizartinib (HR: 0.7). In patients who were MRD positive, the OS again favored quizartinib vs placebo (HR: 0.47), but again—similar to patients who were MRD negative—the median survival was not reached for either arm.

MORPHO was an important study presented at the European Hematology Association 2023 Congress. This multicenter, randomized, double-blind, placebo-controlled phase III trial evaluated gilteritinib vs placebo as posttransplant maintenance in FLT3-ITD‒positive AML.²² Patients in CR who were >30 but <90 days after transplant were randomized after transplant to receive either gilteritinib 120 mg/day or placebo for 24 months.

The primary endpoint was RFS, and key secondary endpoints included OS, event-free survival, incidence of GvHD, safety, and nonrelapse mortality. Marrow samples were collected for MRD analysis at various time points during the study.

The primary endpoint for the study (RFS) was not met. There was a benefit numerically to gilteritinib, as seen in the figure, but this was not statistically significant (HR: 0.679; P = .0518). The secondary endpoint of OS for the entire population also was not met (HR: 0.846; P = .4394).

When one looks at RFS by MRD status going into the transplant or going into treatment, the data are in favor of gilteritinib vs placebo (HR: 0.515; P = .0065) for patients who were MRD positive. For patients who were MRD negative, on the other hand, there was no benefit to the gilteritinib maintenance (P = .5750).

In terms of safety, there were more TEAEs with gilteritinib. Neutrophils were decreased more frequently (42.1% vs 15.8%), and there was more chronic GvHD (52.2% vs 42.1%). There were also more TEAEs leading to discontinuation of gilteritinib.

Let’s finish with a review of a few interesting studies in the posttransplant maintenance setting for AML. Again, a lot of these trials are testing targeted therapies, as this is an exciting area being investigated. Crenolanib, another FLT3 inhibitor, is currently being investigated in a phase II trial (NCT02400255). IDH inhibitors also are being studied in the posttransplant maintenance setting, and some potentially interesting data are emerging from these studies. A phase I trial of enasidenib posttransplant maintenance in IDH2-mutated AML reported a 2-year progression-free survival of 69%,²³ and a phase I trial of ivosidenib posttransplant maintenance in IDH1-mutated AML reported a 2-year progression-free survival of 81%.²⁴ These data are early but show promise, and we look forward to later-phase trials investigating these inhibitors.

Studies are still ongoing with HMAs in the posttransplant setting. The phase III AMADEUS study (NCT04173533) is investigating oral azacitidine vs placebo in patients in first or second CR. A phase II trial (NCT04128501) is evaluating the efficacy of azacitidine plus venetoclax in various leukemias (NCT04128501).

The phase III VIALE-T trial (NCT04161885) is also of interest. This trial, which is still recruiting, is evaluating venetoclax plus azacitidine posttransplant maintenance. Enrolled patients will plan to receive or have received alloHSCT within the past 60 days and have <10% blasts before pretransplant conditioning and <5% blasts post transplant. This trial will have 2 parts. Part 1 will be used to establish the recommended phase III dose for the combination of venetoclax plus azacitidine and best supportive care. The primary endpoint for part 1 is dose-limiting toxicity. In part 2, enrolled patients will be randomized to receive venetoclax plus azacitidine with best supportive care vs best supportive care alone for up to 24 cycles. The primary endpoint for part 2 is OS, and secondary endpoints are morphologic and composite RFS, GvHD rate, GvHD-free RFS, MRD-negativity rate, and quality of life. It will be interesting to see whether combination therapy will have a role in the posttransplant setting.

These are some interesting studies, and I think their results will be eagerly anticipated in the event that they might expand the patient populations who are offered posttransplant maintenance.