CME
Physicians: Maximum of 1.25 AMA PRA Category 1 Credits™
Released: February 12, 2021
Expiration: February 11, 2022
Eunice S. Wang, MD:
FLT3 mutations are observed in approximately 30% of patients with AML and are associated with a poor prognosis, particularly in those with FLT3-ITD mutations.1 Gilteritinib is a potent second-generation FLT3 inhibitor approved by the FDA for treatment of adults with relapsed/refractory FLT3‑mutated AML.2 This approval was based on the phase III ADMIRAL trial, which demonstrated that gilteritinib significantly prolonged survival vs low-intensity or high‑intensity salvage chemotherapy in patients with relapsed/refractory FLT3-mutated AML.3 These findings prompted interest in the potential use of gilteritinib as a frontline alternative for the FLT3 inhibitor midostaurin, which is FDA approved for patients with newly diagnosed FLT3-mutated AML in combination with frontline standard cytarabine and daunorubicin induction and cytarabine consolidation.4
At ASH 2020, Pratz and colleagues5 presented the final results from a phase I trial assessing the safety and efficacy of gilteritinib in combination with cytarabine and an anthracycline in patients with newly diagnosed AML, with or without FLT3 mutations.
Eunice S. Wang, MD:
This was a multicenter, open‑label phase I trial evaluating frontline gilteritinib in combination with cytarabine and either idarubicin or daunorubicin.5 Following induction and consolidation, single-agent gilteritinib was administered as maintenance therapy for up to 26 cycles. The primary endpoints were dose‑limiting toxicities, adverse events (AEs), and the maximum tolerated dose of gilteritinib, with the pharmacokinetics of gilteritinib as a secondary endpoint.
This trial enrolled 79 patients with newly diagnosed AML. The median age was 59.0 years, with ages ranging from 23-77 years. This contrasts with the younger population enrolled in the phase III RATIFY trial establishing frontline midostaurin plus chemotherapy as standard of care for FLT3-mutated AML. The RATIFY trial only enrolled patients between 18 and 59 years of age.6
Eunice S. Wang, MD:
No dose‑limiting toxicities were observed with gilteritinib doses up to 120 mg/day, but 2 dose‑limiting toxicities (neutropenia and enterocolitis) were observed in patients receiving 200 mg/day.5 Therefore, investigators identified gilteritinib at 120 mg/day as the maximum tolerated dose and proceeded with this dose in combination with intensive chemotherapy for the dose-expansion part of the trial. Recall that 120 mg/day is the FDA-approved dose for gilteritinib monotherapy in the relapsed/refractory setting.2
Eunice S. Wang, MD:
In this trial, the serious treatment emergent AEs of note included febrile neutropenia in 12 patients and sepsis in 4 patients.5
Eunice S. Wang, MD:
Treatment with gilteritinib at 120 mg/day plus chemotherapy was associated with high efficacy, as indicated by a composite CR rate of 81.6% (CR rate: 39.5%) in all patients with FLT3-mutated AML.5 The median duration of response was 8.0 months, with a median OS that was not reached after a follow-up of 35.8 months.
The table here presents the responses broken out by phase and dose cohort. Efficacy does not appear to vary among different anthracyclines, although there were very small numbers in each group.
Eunice S. Wang, MD:
Mutational clearance, which was defined as FLT3-ITD:total FLT3 signal ratio ≤ 10-4 by next-generation sequencing, was assessed sequentially at different points in the treatment course in a subset of patients with FLT3-ITD–positive AML.5 As you can see in this table, mutational clearance was achieved in 4 of 12 patients by the end of induction and 3 of 8 patients by the beginning of consolidation. By the beginning of maintenance, mutational clearance was achieved in almost all (11 of 13) patients.
Eunice S. Wang, MD:
Dr. Cortes, I know that you and your colleagues at the MD Anderson Cancer Center have used multiple FLT3 inhibitors. What are your thoughts on these results?
Jorge Cortes, MD:
I would first note that this is a phase I trial with multiple dosing cohorts and small numbers in each cohort, making it difficult to compare outcomes among subsets. However, we can still come away with several important interpretations.
First, we would like to use FLT3 inhibition in the frontline setting, meaning we would have to use the FLT3 inhibitor in combination with chemotherapy. Thus, this study was quite important and relevant, despite being a phase I trial. As we discussed, single-agent gilteritinib is already approved as salvage therapy. I use gilteritinib frequently in that setting.
Second, I was not surprised by the safety profile observed in this trial. In general, we have found that adding an FLT3 inhibitor at its full dose does not greatly increase the toxicity of the chemotherapy partner, whether that is a hypomethylating agent; cladribine, idarubicin, and cytarabine; fludarabine, cytarabine, and filgrastim; or others. It was reassuring to see that there was not much additional toxicity with gilteritinib in the frontline setting.
Third, I consider the response rates here to be very good. As you mentioned, the cohorts are quite small, and it is difficult to draw conclusions and make comparisons. Nonetheless, the composite CR rate was 81.6% in patients with FLT3‑mutated AML, which is the population where we would use gilteritinib plus chemotherapy. Perhaps more important is that almost all patients with FLT3-ITD–positive AML who achieved a CR or an incomplete CR (CRi) also achieved mutational clearance by next-generation sequencing. These are good, strong remissions.
I would think that a randomized trial might find that this regimen is better than cytarabine plus anthracycline (7 + 3) plus midostaurin. How do you think gilteritinib will compare with midostaurin?
Eunice S. Wang, MD:
I consider these phase I results to be highly encouraging, although I was slightly disappointed in a composite CR rate of only 81.6%. I would have preferred that rate to be closer to 100%, but this was a phase I trial, and the analysis included patients who received < 120 mg/day of gilteritinib. Of note, the CR rate in this trial was 39.5%, possibly due to increased CR without count recovery due to myelosuppression. In contrast, the RATIFY trial reported a CR rate of 59% with midostaurin plus 7 + 3 and 55% with placebo plus 7 + 3,6 with no difference between the 2 arms. (P = .15).
In this phase I trial, gilteritinib appeared to be associated with a high composite CR rate in the frontline setting, but of course we need randomized comparative data. The ongoing randomized phase III trials PrE0905 and HOVON 156 AML are currently comparing induction and consolidation chemotherapy plus gilteritinib vs midostaurin in newly diagnosed FLT3-mutated AML.7,8
Many clinicians are looking forward to using newer technologies, particularly next-generation sequencing and molecular endpoints, to predict response in our patients. It was promising that the investigators were able to observe high levels of mutational clearance of FLT3-ITD using this new next-generation sequencing assay.
Many clinicians are also looking forward to moving second-generation FLT3 inhibitors into the frontline setting. This early-phase trial is a promising first step. I should note that 2 other newer generation FLT3 inhibitors have been evaluated or are being evaluated in the frontline setting: quizartinib and crenolanib. In the next few years, we will likely see these newer FLT3 inhibitors begin to supplant midostaurin.
Jorge Cortes, MD:
I agree. Regarding safety, I consider gilteritinib to be a very safe drug overall, although there is slightly more liver toxicity with this agent. This is particularly important in patients who are receiving azoles and/or other hepatotoxic drugs.
At this point, gilteritinib plus chemotherapy appears to be a generally safe combination. I would monitor for liver toxicity because this trial reported that 5 of the 12 drug-related treatment-emergent AEs leading to discontinuation were due to elevated alanine aminotransferase or aspartate aminotransferase.
Eunice S. Wang, MD:
I agree. One unanswered question posed by this trial is which anthracycline would be best to combine with gilteritinib. I suspect that many participating trial centers used idarubicin in combination with cytarabine and gilteritinib because that chemotherapy regimen is likely the institutional standard.
Regarding the patients who received daunorubicin, it is somewhat difficult to compare the results with this trial’s patients, who received daunorubicin at 90 mg/m2, to those reported in RATIFY, who received daunorubicin at 60 mg/m2.6 Many ongoing and planned randomized trials are using daunorubicin. In our center, we use daunorubicin at 90 mg/m2 based on data suggesting a potential advantage to using dose‑intensified daunorubicin in patients with FLT3‑mutated disease.9,10 Until we know more about which is the best combinatorial regimen, I consider it important for randomized trials to consistently use the same anthracycline in the induction regimen.
To summarize, I anticipate that these results validate the feasibility and potential efficacy of moving gilteritinib into the frontline setting. Of note, it is important to monitor for toxicity (such as liver abnormalities and cytopenias) and bear in mind that second-generation FLT3 inhibitors (gilteritinib, quizartinib, and crenolanib) have different safety profiles than the first-generation inhibitors.11
Eunice S. Wang, MD:
Moving on, the next report generated a lot of enthusiasm at ASH 2020. Sallman and colleagues12 presented preliminary results from an ongoing phase Ib trial of magrolimab plus azacitidine in a cohort of individuals with newly diagnosed AML considered ineligible for intensive chemotherapy. Magrolimab is a first-in-class macrophage immune checkpoint inhibitor that targets CD47.13 CD47 is a “don’t eat me” signal expressed at high levels in many cancers, including AML, as a means by which malignant cells evade macrophages. By blocking CD47, magrolimab restores macrophage phagocytosis of malignant cells and may act synergistically with azacitidine, which can also induce prophagocytic signals, in preclinical studies.
Eunice S. Wang, MD:
In this phase Ib trial, patients received a priming dose of magrolimab at 1 mg/kg followed by 30 mg/kg weekly in combination with a 7‑day regimen of azacitidine.12 The primary endpoints were safety and efficacy with secondary endpoints including the pharmacokinetics, pharmacodynamics, and immunogenicity of magrolimab.
This analysis included 64 patients with untreated AML ineligible for induction chemotherapy. As would be expected, the median age was advanced at 71 years. The trial population overwhelmingly had poor cytogenetic risk (70%), with more than half of patients with AML exhibiting myelodysplasia‑related changes and 14% having therapy‑related AML.
The trial’s protocol was amended early to preferentially enroll patients with TP53‑mutated disease. In this cohort analysis, 73% of patients had the TP53 mutation.
Eunice S. Wang, MD:
Shown here are the safety data for magrolimab plus azacitidine in untreated AML patients.12 No maximum tolerated dose was reached, and the overall tolerability was excellent, with no significant increase in rates of cytopenias, infections, or immune‑related AEs from baseline. Fewer than 5% of patients required dose modification or discontinuation due to drug-related AEs. The all‑cause mortality rate was just below 5% at 30 days and < 10% at 60 days.
Eunice S. Wang, MD:
This regimen was associated with the unique AE of on-target anemia.12 Aged red blood cells express surface CD47, leading to on‑target hemolytic anemia occurring mostly with the first magrolimab infusion. As shown here, the first dose was associated with a mean decrease in hemoglobin from baseline of 0.9 g/dL (range: +2.5 to -3.7).
Despite this initial decrease in hemoglobin levels due to on-target hemolytic anemia, most patients had improvement in their transfusion frequency and hemoglobin levels over several months of magrolimab-based therapy.
Eunice S. Wang, MD:
The graphs here highlight some of the longer term effects associated with staying on therapy.12 After 10-15 weeks, neutrophil and platelet levels began to improve. This occurred in parallel with improvements in hemoglobin levels and decreased transfusion frequency as we just discussed.
Eunice S. Wang, MD:
Regarding efficacy, 63% of all patients in this cohort achieved a response, with a CR rate of 42% and a CRi rate of 12%.12 Although these are impressive data and comparable to the efficacy of venetoclax/azacitidine in a similar patient population, a key difference is that this regimen was associated with efficacy regardless of TP53 mutation status. Thus, this novel immunotherapeutic approach appears be effective against malignant AML cells independent of mutational profile.
The median time to response was fairly short (< 2 cycles) and the median duration of response was almost 10 months in all patients, with almost half of patients achieving complete cytogenetic response and approximately one third achieving measurable residual disease (MRD) negativity.
There was a high rate of transfusion independence in patients with and without TP53 mutation, with 68% of all patients achieving red blood cell transfusion independence over the course of therapy. Consistent with what we saw for the neutrophil and platelet counts, multiple patients experienced improved responses over time.
Eunice S. Wang, MD:
These graphs depict the preliminary OS in 16 patients with wild-type TP53 and in 47 patients with TP53‑mutated AML.12 The median OS was particularly impressive at 12.9 months in the TP53-mutated cohort. This compares favorably to a median OS of 5.2-7.2 months with frontline venetoclax plus hypomethylating agents in similar patients with TP53-mutated AML.14,15
The median OS was 18.9 months in patients with wild-type TP53, although this cohort had only 16 patients enrolled in this preliminary OS analysis.12 This compares favorably to the results of the VIALE‑A trial, which reported a median OS of 14.7 months following upfront venetoclax/azacitidine in older unfit individuals with AML, of whom 23% had TP53 mutation.16
Eunice S. Wang, MD:
Dr. Cortes, what are your thoughts on these results and on the potential use of magrolimab for both TP53‑mutated and wild‑type TP53 AML?
Jorge Cortes, MD:
I was impressed by these results, particularly for the patients with TP53-mutated AML, because we currently lack effective therapies for that population. Even venetoclax/azacitidine has poor efficacy in TP53-mutated disease, as you discussed.
I was struck by the ORR being as high as 69% in those with TP53-mutated AML, with 45% achieving CR. As you mentioned, it is important to emphasize that this regimen is not specific to those with TP53 mutation. Magrolimab simply is not affected by TP53 status, unlike chemotherapy and other agents.
These results will change our care for those with TP53-mutated disease. The question is, what should we do for patients with wild-type TP53? Our standard of care is venetoclax plus either azacitidine or another hypomethylating agent, which is quite effective. That being said, the safety profile seems to be very favorable for magrolimab plus azacitidine, particularly given its lower rates of myelosuppression.12,17 We could also try combining the backbone of venetoclax/azacitidine with magrolimab.
I am not that concerned about the anemia observed in this trial. We are accustomed to observing anemia in patients with AML when starting chemotherapy, with transfusion frequency increasing during the initial cycles. We saw here that over time, as a result of the responses, approximately two thirds of patients became transfusion independent. We would simply need to support patients through that first to second month of therapy with magrolimab plus azacitidine.
What is your perspective, Dr. Wang?
Eunice S. Wang, MD:
I completely agree. I believe that magrolimab will likely be a transformative agent for AML. Moreover, this macrophage immune checkpoint inhibitor is the first immunotherapeutic treatment of AML or myelodysplastic syndrome to demonstrate very high response rates > 50%.
Like you, I was also struck by the high response rates in those with TP53‑mutated AML. Recent data published by Sallman and colleagues18 at Moffitt Cancer Center indicated that TP53‑mutated AML is associated with an immunosuppressive microenvironment, which may explain why there are such high response rates in that patient population.
Regarding the preliminary OS benefit, these data compare very favorably with the VIALE-A data on venetoclax/azacitidine, as we discussed.12,16 Another important point of contrast is the difficulty of administration and toxicity management requirements. I have treated patients with this combination on this clinical trial, and our center was able to administer magrolimab plus azacitidine in the ambulatory setting. Admissions were necessary only for patients who might have, for example, experienced an acute drop in hemoglobin levels because of hemolytic anemia. In contrast, the profound myelosuppression with venetoclax/azacitidine requires us to hospitalize many of our older patients for at least the first week, and potentially much longer, for daily transfusion support and monitoring for tumor lysis syndrome.
I am looking forward to some of the phase III trials that are ongoing or being planned. The ENHANCE trial is recruiting patients with newly diagnosed high‑risk myelodysplastic syndrome regardless of molecular subtype to evaluate magrolimab plus azacitidine19 vs azacitidine alone. Another trial being planned will compare magrolimab plus azacitidine vs venetoclax/azacitidine in patients with newly diagnosed TP53-mutated AML.12 Those phase III results are highly anticipated given that this agent is generating so much enthusiasm.
Jorge Cortes, MD:
We will next discuss emerging data for another interesting agent, flotetuzumab. This is a bispecific construct that targets CD123 and CD3 to bring a CD123-expressing leukemic cell close to a CD3-positive T-cell.20 CD123 is an antigen that has been very attractive for the development of monoclonal antibodies because it is expressed in most patients with AML. We used to think that CD123 was more selective for leukemic stem cells, although now we know that is not the case.
This approach of bringing T‑cells close to leukemic cells with bispecific antibodies is similar to the approach used by CAR T‑cells. The difference is that with CAR T‑cells, the T‑cells are collected from the patient and engineered to express the means to recognize the leukemic cell, whereas bispecific antibodies must recognize both the T‑cell and the leukemic cell to bring them together.
Results from a phase I/II trial indicating that flotetuzumab is associated with a tolerable safety profile were recently published in Blood. This early trial also established the recommended phase II dose.20 Other interesting findings from this trial were, first, that they observed responses in patients who had experienced primary induction failure (PIF) or early relapses. Second, their data suggested that patients who are resistant to cytarabine‑based therapy have a tumor microenvironment infiltrated by immune cells, which may explain why these patients could particularly benefit from flotetuzumab. Third, they identified a 10-gene signature predicting response to flotetuzumab.
At ASH 2020, Aldoss and colleagues20 presented updated results of the expansion cohort of that phase I/II trial, with a focus on patients who had experienced PIF and/or early relapses.
Jorge Cortes, MD:
This phase I/II trial recruited patients with AML refractory to up to 2 cycles of cytarabine‑based chemotherapy or 2-4 venetoclax‑based combinations (eg, a hypomethylating agent plus venetoclax or low‑dose cytarabine plus venetoclax) or—if they had experienced PIF—gemtuzumab ozogamicin. Those who had early relapse < 6 months after achieving a CR (ER6) were also eligible.20 These patients historically do poorly, with a response rate of approximately 10% to salvage therapy.
The trial had earlier identified the recommended phase II dose as 500 ng/kg/day.20 Bispecific constructs are generally administered as a continuous infusion with a step-up lead-in dose, which is how flotetuzumab was administered in 28‑day cycles. The primary endpoint was the CR rate, and secondary endpoints included the duration of response, rate of transfusion independence, and time in hospital.
This analysis included 44 patients. This was a somewhat older patient population with a median age of 63.5 years; 61.4% had experienced PIF, and 38.6% had experienced ER6. Most of the patients with PIF had received a cytarabine‑based chemotherapy (74.1% of the PIF population), and approximately one quarter had received venetoclax‑based combinations. The European LeukemiaNet risk stratification was adverse for 72.7% of patients, and 36.4% had secondary AML. By design, this study enrolled a challenging, high-risk population.
Jorge Cortes, MD:
This updated analysis found a tolerable safety profile consistent with what was previously published in Blood.20 All patients experienced infusion-related reactions/cytokine-release syndrome (CRS), although the overwhelming majority of events were low grade. Common AEs included rashes (38.6%) and arthralgias (25.0%), which were all low grade. The most common high-grade AE was febrile neutropenia (13.6%), and that was probably the leukemia itself.
A handful of patients also experienced liver toxicity and diarrhea, but none of these AEs is particularly concerning.
Jorge Cortes, MD:
The greatest safety concerns with both bispecific constructs and CAR T-cells are CRS and neurologic events. In this analysis, the investigators observed neurologic events ranging from headaches in almost 30% of patients to other less common manifestations.20
The neurologic event that stands out is confusional state, which occurred in 3 patients and was grade 3 in each. The investigators reported that all of these confusion events were transient and completely reversible with treatment interruption. Confusion needs attention, but these neurologic events were certainly manageable and not permanent. Again, these neurologic events are consistent with the safety profiles of bispecific constructs, and as long as they can be managed, they are not a cause for concern.
Jorge Cortes, MD:
Shown here are the responses, and we will focus on the table first. In the overall population, the composite CR rate was 31.8%.20 That may sound low, but remember that these are patients who have experienced PIF or ER6. The composite CR rate was perhaps driven a little more by the patients with PIF (37.0%) than by those with ER6 (23.5%), consistent with the initial data indicating an immunosuppressive microenvironment in PIF AML.20
The analysis also examined responses in the 10 patients with TP53 mutation. The composite CR rate was 50% in this small group, a good response for patients with TP53 mutation who had failed prior therapy. That is encouraging and worth following further.
Turning to the waterfall graph, which depicts the best bone marrow change in each patient, we see that most patients had a reduction in blast count (median blast decrease: 81%). The bar graph suggests that flotetuzumab has a little more activity and benefit than was measured by strict criteria including CR and CR with partial hematologic recovery. Combining flotetuzumab with other agents may enhance that benefit.
Moving on to the graphs depicting duration of response and OS, we again see that the patients with PIF appear to have markedly better response durations and survival than do the patients with ER6. The overall duration of remission is 8.1 months, with a dramatic split of 15.2 months in the PIF subgroups vs 2.4 months in the ER6 subgroups. We also see a marked split in the median OS (PIF: 15.9 months; ER6: 5.0 months). Recall that these are patients with a median age of 63.5 years who had received a median of 2 prior therapies, with approximately one third having secondary AML.
Dr. Wang, what are your thoughts on these response data with flotetuzumab?
Eunice S. Wang, MD:
I have a somewhat different take on this agent. I do consider this bispecific construct to have a lot of potential. These results—particularly the blast reduction rate—are impressive given that this is a very high-risk patient population of interest. As you mentioned, these patients were refractory to a median of 2 prior therapies, and some were primary refractory. I also share your interest in the transient responses observed in the patients with TP53‑mutated disease.
However, I am concerned about the durability of response, which was only a few months. Furthermore, while I agree that this is a very refractory population, we should note that only one third of the patients experienced a response. We must consider that this rather unimpressive response rate in context with the 100% incidence of CRS, with 1 patient (2.3%) experiencing grade ≥ 3 CRS. In the first part of this trial, 8% of patients experienced grade 3 CRS.20 At ASH 2020, the investigators shared that the patients were hospitalized for a minimum of 8 days at the start of therapy, without specifying how much longer they needed to be hospitalized on average.
I agree that the blast reduction is encouraging, but only about one third of patients are achieving CR, CRi, or CR with partial hematologic recovery—with all patients needing to be hospitalized for some time due to the risk of CRS and neurologic events. Given this risk:benefit ratio, I have reservations about the widespread use and feasibility of treatment with flotetuzumab.
That being said, given the blast clearance and the responses observed with flotetuzumab, this agent appears to be a good approach to treat patients with primary refractory AML to bridge to a potentially curative allogeneic stem cell transplant.
Jorge Cortes, MD:
I certainly see your point. We always want to see high response rates, but I do find these results very encouraging given the patient population—particularly in those with PIF. This agent could have a positive clinical impact.
Jorge Cortes, MD:
Regarding how to move this agent forward, it would be valuable to confirm the genetic signature associated with responses that the Blood paper reported. This would help us identify—even at diagnosis—patients who are more likely to benefit from this agent and to potentially add flotetuzumab earlier to backbones (eg, standard chemotherapy or a venetoclax‑based combination).
Bispecific antibodies are going to be very useful in AML. However, it is difficult to determine right now which will make it. I do like these results with flotetuzumab, and I have a more optimistic take on CRS—a toxicity present with all bispecific constructs and CAR T‑cells—given that once the investigators for this trial figured out the ramping up dose, the CRS cases were largely low grade and manageable.
I would like to hear more of your perspective, Dr. Wang.
Eunice S. Wang, MD:
Flotetuzumab is associated with intriguing activity and responses in the patients with PIF, although the patients with ER6 clearly did not respond in this setting, and I consider that to be a population with great unmet need. As we discussed, I have more reservations because of the 100% rate of CRS and the long hospitalization stays for some of these patients, especially when paired with a response rate of only 31.8%. Thus, I am enthusiastic about flotetuzumab’s potential only for this specific, small patient population with PIF, particularly if they can be bridged to an allogeneic stem cell transplantation. I am curious to see where this agent goes in development.
I am enthusiastic overall about immunotherapy. Another setting where we may see the most clinical benefit of bispecific constructs for AML patients is in the MRD setting, similar to how we currently use blinatumomab for treatment of patients with MRD-positive ALL in CR or second CR.21 Given the lower burden of cancer cells in the MRD setting, we may see less CRS and potentially greater efficacy with flotetuzumab in this situation.
Eunice S. Wang, MD:
CC‑486 is an investigational oral formulation of the hypomethylating agent azacitidine.22 This is not a 1:1 substitution for azacitidine, but rather a novel formulation with a distinct pharmacokinetic/pharmacodynamic profile, which allows the hypomethylating activity of the agent to be durable over 14 days out of a 28-day cycle.
At ASH 2020, we saw updated data from the phase III QUAZAR AML-001 trial comparing CC-486 vs placebo as maintenance treatment in patients with AML who were aged ≥ 55 years and in first CR/CRi after induction therapy with or without consolidation.22,23 Patients were enrolled within 4 months of achieving their first CR and had to have evidence of morphologic residual disease but not MRD. Initial data from QUAZAR AML-001 were presented in a late-breaking abstract at ASH 2019.24 These initial results demonstrated that CC-486 significantly prolonged both median OS (24.7 vs 14.8 months with placebo; HR: 0.69; 95% CI: 0.55-0.86; P = .0009) and median relapse-free survival (RFS; 10.2 vs 4.8 months with placebo; HR: 0.65; 95% CI: 0.52-0.81; P = .0001). Based on these positive data, the FDA recently approved CC‑486 as maintenance treatment for patients with AML who achieved CR/CRi after intensive induction chemotherapy and who are unable to complete intensive curative therapy.25
The data presented at ASH 2020 provided nuances to these results, notably by reporting findings evaluated per other response criteria—thereby allowing us to fully appreciate the benefit of this agent. Use of CC-486 as maintenance treatment is a relatively novel concept in AML; historically, the only other drug has been FLT3 inhibitors (eg, sorafenib) in the post-transplant setting.26
Eunice S. Wang, MD:
As we discussed, the double-blind phase III QUAZAR AML‑001 trial enrolled patients with AML who were aged ≥ 55 years with intermediate-risk or poor‑risk cytogenetics who achieved CR/CRi after intensive chemotherapy with or without consolidation.22,24,27 Of importance, these patients had to be considered transplant ineligible and were therefore not proceeding to a potentially curative allogeneic stem-cell transplantation. The trial randomized 472 patients to CC‑486 for 1-14 days out of every 28 days vs placebo.
The primary endpoint was OS. The ad hoc analyses presented at ASH 2020 assessed outcomes in patients who had progressive disease (defined as relapse with 5% to 15% blasts on study) who then received escalated dosing for 1-21 days out of every 28 days, patient‑reported health-related quality of life (HRQoL), and outcomes in patients with detectable vs undetectable MRD-positive AML.22,24,28
Eunice S. Wang, MD:
In QUAZAR AML-001, patients who experienced early AML relapse on study (defined as increased blasts from 5% to 15%) were eligible to receive dose escalation of CC‑486 from every 14 days per 28‑day cycle to every 21 days per 28-day cycle22 Those who were randomized to CC‑486 remained on CC‑486, and those randomized to placebo remained on placebo.
In the CC‑486 arm, 23% of those who experienced early relapse and received CC-486 achieved a second CR with escalated dosing vs 11% in the placebo arm. Among those who achieved second CR, 6 patients in the CC-486 arm also achieved MRD negativity vs none in the placebo arm. In this population with early relapse, unsurprisingly, the median OS was nonsignificantly prolonged with CC-486 vs placebo (22.8 vs 14.6 months, respectively; HR: 0.66; 95% CI: 0.42-1.04; P = .07).
These intriguing data show that we can successfully re-achieve response in patients with early relapse by using oral chemotherapy.
Eunice S. Wang, MD:
Regarding safety in this patient population, dose escalation of CC-486 was well tolerated overall.22 Only 8% of patients in the CC 486 arm discontinued due to AEs with this dose-escalation strategy vs 3% in the placebo arm.
Eunice S. Wang, MD:
The second ad hoc analysis of QUAZAR AML-001 examined how maintenance treatment with CC-486 vs placebo was tolerated and its impact on HRQoL. This is important because negative effects on HRQoL are a major deterrent in patients to the use of long‑term maintenance strategies in other malignancies.
Encouragingly, Roboz and colleagues27 observed no statistically significant differences between the arms in terms of change from baseline HRQoL, as assessed by questionnaires given to patients over the course of therapy. The metrics assessed were the FACIT-Fatigue, EQ-5D VAS, and EQ-5D-3L health utility index scores. Longitudinal analyses showed no clinically meaningful differences with CC-486 vs placebo.
When the investigators compared rates of deterioration in HRQoL, they again found no significant differences with CC-486 vs placebo.
These are positive findings because we would generally expect that patients receiving an oral chemotherapy agent with an active component to experience AEs that would diminish their HRQoL. These results are given further importance because the patients continued on study until they developed intolerable AEs, experienced overt relapse (> 15% blasts), or proceeded to hematopoietic stem cell transplant. Thus, many patients would be anticipated, in the absence of being able to proceed to a transplant, to remain on prolonged maintenance therapy with oral azacitidine for maximal benefit.
Eunice S. Wang, MD:
Dr. Cortes, I am interested in hearing what you thought of these first 2 analyses before I present the MRD results. Are you using oral azacitidine in your practice?
Jorge Cortes, MD:
I find CC-486 particularly useful in the maintenance setting because, as you mentioned earlier, we have generally lacked beneficial options once this patient population achieves remission and stops induction/consolidation therapy. The QUAZAR AML-001 trial provides extra opportunity to patients who may be losing response. If the loss of response is caught early, we can now offer a maintenance therapy that could improve outcomes.
In the data on patients with early relapse, I was struck by the high background rate of patients regaining CR/CRi in the placebo arm (11% with placebo vs 23% with CC-486). This suggests that what we sometimes characterize as early relapse may just be transient elevations of blasts and/or may be because of how the blasts were counted. We need to be careful to assess and confirm these early relapses to avoid overreacting to a transient elevation, because even with just the placebo, some patients will regain response.
The data on patients with early relapse also illustrated that you can safely increase the dose and regain response. This is a reasonably well-tolerated drug, and the safety data show that the dose escalation is associated with some more grade 3/4 cytopenias and gastrointestinal toxicity. The second analysis that you presented showed that these AEs do not adversely affect HRQoL.
I consider these findings important while bearing in mind that this dose-escalation approach will benefit only a subset of patients. Given that it is well tolerated, I consider it worth trying.
Eunice S. Wang, MD:
I agree. The data on the placebo arm do suggest that some patients characterized as experiencing early relapse may have been misclassified because of sampling error. Nonetheless, it is encouraging that we have the option to dose-escalate CC-486, rather than initiate an aggressive salvage regimen, if a patient begins to experience an increase in blasts. Simply being able to maintain the response a little longer could give you some time to look for other options.
I also consider the HRQoL data to be very important, because when patients learn about using a drug potentially for the rest of their lives, they are relieved to learn that the agent is unlikely to cause significant toxicities, affect their ability to travel, and to simply live their life.
Eunice S. Wang, MD:
We will now move on to a third important analysis of QUAZAR AML-001 presented at ASH 2020. Roboz and colleagues28 compared outcomes in enrolled patients by MRD status. It has already been established that patients with MRD-positive AML, as measured by multiparameter flow cytometry and/or next-generation sequencing, are more likely to experience relapse and have significantly shortened OS and disease-free survival vs those who achieve MRD negativity.29,30 MRD negativity also identifies the patients most likely to achieve durable cure with subsequent allogeneic stem cell transplantation.31
Eunice S. Wang, MD:
As you can see in these graphs of OS by MRD status, the patients who were MRD negative had better median OS regardless of treatment.28 That being said, in both the MRD-negative and MRD-positive patients, CC-486 improved OS (HR for MRD negative: 0.81; 95% CI: 0.59-1.12; HR for MRD positive: 0.69; 95% CI: 0.51-0.93).
Eunice S. Wang, MD:
RFS was similarly improved with CC-486 vs placebo, independent of MRD status.28 Again, we saw that although MRD negativity conferred a worse prognosis, CC-486 was still more beneficial than placebo in both groups (HR for MRD negative: 0.71; 95% CI: 0.52-0.98; HR for MRD positive: 0.58; 95% CI: 0.43-0.78).
Eunice S. Wang, MD:
A multivariate analysis further confirmed that patients who were MRD positive had significantly worse outcomes for OS (HR: 1.85; P < .0001) and RFS (HR: 2.04; P < .0001). Significant OS and RFS benefits were observed with CC-486 vs placebo regardless of MRD status (HR for OS: 0.74; P = .067; HR for RFS: 0.63; P < .0001).28
Eunice S. Wang, MD:
When we look at outcomes in the MRD-positive cohort, we see that 37% of the patients who received CC-486 subsequently converted to MRD‑negative disease.28 In total, 19% of the patients receiving placebo also converted to MRD-negative disease. This high background rate may again be due to issues with the sensitivity and specificity of the assay being used.
Nevertheless, the duration of MRD negativity was significantly prolonged with CC-486 vs placebo (median: 11.0 vs 5.0 months, respectively; HR: 0.62; 95% CI: 0.48-0.78).
Eunice S. Wang, MD:
Dr. Cortes, I believe these MRD data are consistent with other studies, but what do you think of CC-486’s capacity to alter MRD status? Are you concerned that these data were confounded by technological issues, as suggested by the high background rate in the placebo group?
Jorge Cortes, MD:
There is no question that we need to further refine our methodologies for defining MRD status to make them more predictive, better standardized, and more available. Despite those limitations, MRD status is well established as being predictive for relapse, worse OS, and worse RFS, as you mentioned.
Overall, the results you just presented have several implications. First, these data confirm the value of MRD, because in both arms the MRD‑negative patients had improved OS and RFS outcomes vs the MRD‑positive patients. However, we saw the greatest benefit in the MRD-positive patients—the population that needs the most help. In the MRD‑negative patients, the OS benefit was relatively modest and not statistically significant. It still improved some, and I welcome that benefit, but it was a small benefit. The RFS benefit was also more pronounced in the MRD-positive patients. The MRD-positive population represents those with the greatest unmet need. The high rate of conversion to MRD negativity further supports the benefit of CC-486 in the MRD-positive population.
This analysis of MRD strengthens the value of this agent. We already saw that the results were significantly positive in the overall population. This analysis provides more granularity as to where the benefit comes from and the value of assessing MRD in routine practice.
Eunice S. Wang, MD:
These data confirmed what we suspected, and I whole-heartedly agree that the MRD‑positive patients are most likely to benefit from this agent. Currently, we feel a bit nervous when faced with an MRD‑positive patient, because the question now is what do you do with those patients? Do you send them to transplant, even though the transplant team is concerned because the patient is MRD positive?
As we discussed, a certain percentage of these patients will not be able to proceed to transplant. The baseline trial data were quite interesting because they showed that, at the time of morphological remission, approximately half of these patients were MRD positive. That is a high rate of MRD positivity at the end of potentially intensive chemotherapy and 1‑2 cycles of consolidation. Clinicians have generally believed that in patients achieving morphological CR—perhaps 10% to 20%—would be MRD positive. Thus, I was a little surprised to see how many were MRD positive.
We do not know yet whether CC-486 is the most efficacious agent for achieving MRD conversion. The comparison of MRD conversion rates in this analysis is plagued by the high background rate in the placebo arm. Nonetheless, I consider this to be an encouraging step in the right direction.
Jorge Cortes, MD:
I have a brief note about the MRD-positive patients who achieved CRi vs CR. The MRD-positive population achieving CRi is more heterogeneous, with some patients being true CRs who are just experiencing some myelosuppression from their therapy, whereas others are not experiencing as good a response as a true CR. This is another important reason we should perform MRD assessments, because MRD status can differentiate a true CR from a CRi, with the latter being less clinically valuable.
Eunice S. Wang, MD:
Yes. This brings up the question of why we still use morphological definitions—they were more valuable when we used cytotoxic chemotherapy. Achieving count recovery is good, but now that we are using biological therapy, does it make a difference whether the patient is a CR vs CRi vs CR with partial hematologic recovery? MRD status appears to have more bearing on long‑term outcomes.
Back
