CME
Physicians: Maximum of 1.00 AMA PRA Category 1 Credit™
Released: February 16, 2021
Expiration: February 15, 2022
Amy E. DeZern, MD, MHS:
In the treatment of MDS, to date, allogeneic hematopoietic stem cell transplant (HSCT) has been the sole curative therapy.24 However, although HSCT is more widely used in younger patients, recent data show no survival disadvantage for older patients.25 Through the use of reduced‑intensity conditioning (RIC) regimens, which have become more mild but still efficacious, older patients have become candidates for HSCT. However, in the community setting, patients with MDS who are older are not commonly referred for transplant consultation. This means older patients often do not have the option of undergoing HSCT. In addition, due to the uncertainty of relative benefit, HSCT is not currently covered by Medicare in the United States.
To evaluate the benefit of HSCT in older patients with higher-risk MDS, the BMT CTN 1102 clinical trial is comparing RIC allogeneic HSCT with non‑HSCT therapy in patients 50‑75 years of age.26
Amy E. DeZern, MD, MHS:
At ASH 2020, Nakamura and colleagues26 presented results from BMT CTN 1102, an ongoing multicenter, open-label phase II trial by the Clinical Trials Network for Transplant in the United States (N = 384). Enrolled patients had IPSS intermediate‑2–risk/high‑risk MDS and were candidates for RIC allogeneic HSCT—this usually suggests they have a certain level of fitness.
Patients were randomized to donor or no-donor arms based on high‑resolution HLA typing and whether or not the patient had an 8/8 HLA match. Patients were initially assigned to the no-donor arm and only moved to the donor arm when a suitable donor was identified. Participants in the donor arm underwent RIC HSCT within 6 months of enrollment.
The primary endpoint of this trial was 3‑year OS, adjusted to account for any potential bias from biologic donor/no donor assignment. The secondary endpoint was leukemia-free survival (LFS).
Amy E. DeZern, MD, MHS:
The baseline characteristics of the donor arm (n = 260) and the no-donor arm (n = 124) were quite comparable. Approximately 60% in each arm were older than 65 years of age, which is important because that is the age when patients become Medicare eligible. As in MDS generally, approximately two thirds of the patients in BMT CTN 1102 are male, in both the donor and no-donor arms. Patients’ fitness for transplantation is measured by Karnofsky performance status, which showed only a subtle difference in lower performance status between the donor arm and no-donor arm. The median time since MDS diagnosis was longer than 2 months in both arms. Remember that these are high‑risk patients and their IPSS scores were approximately two thirds intermediate‑2–risk and approximately one third high‑risk disease.
The patients were treated to reduce their blasts before considering RIC allogeneic transplant; at baseline, in the donor arm, the patients who had achieved CR to hypomethylating agents was 4%, and in the no-donor arm, it was 6%. However, a CR is not a requirement to go to transplant, and approximately 20% of patients in this study had achieved a PR. The study also included patients (~ 30%) who did not respond to a hypomethylating agent such as azacitidine or decitabine.
The Hematopoietic Cell Transplantation–Specific Comorbidity Index (HCT-CI) is associated with response and outcome, and this was tabulated at baseline for those in the donor arm. In this study, most patients had a fairly high HCT-CI score ≥ 3, likely due to age.
Amy E. DeZern, MD, MHS:
In terms of 3‑year OS, in this analysis, there was a statistically significant benefit in the donor arm, with 47.9% alive compared with 26.6% of the no-donor arm (P = .0001). This was an absolute OS improvement of 21.3%. The absolute improvement was similar for the sensitivity analysis at 19.9% (P = .0004). Among the patients who ultimately went to transplant (the “as‑treated” analysis), there was a 31.4% absolute improvement (P < .0001). Of note, age did not affect the difference in OS between arms.
Amy E. DeZern, MD, MHS:
The secondary endpoint of LFS, which is clinically quite important to patients with MDS, also showed a benefit in the donor arm, with a nearly 36% 3‑year LFS compared with 21% for the no-donor arm (P = .003). This split was even wider for patients who went to transplant, at 39.3% vs 10.9% (P < .0001).
It should be noted that, as with OS, a subgroup analysis revealed no difference in LFS for the donor vs no-donor arm in both patients who were younger or older than 65 years of age.
Amy E. DeZern, MD, MHS:
Something that has always been a concern for healthcare professionals referring patients with MDS to transplant is QoL: We are taking these patients to a potentially curative transplant, but does this diminish their QoL such that they are not able to feel the benefit? To help address this question, the BMT CTN 1102 trial incorporated QoL standardized and validated scoring systems, including FACT‑G, SF‑36, and EQ-5D, to compare patients who had a donor and went on to transplant vs patients with no donor.
It is a very important take-home message that patients in the donor arm did not have a significant decrease in QoL compared with the no-donor arm, across all time points and tools. Of interest, the no-donor arm showed a statistically significant poorer QoL score compared with the donor arm at 8 months and 36 months on 2 separate scoring systems (P < .05). However, the authors of the trial thought this was not clinically significant given these time points.
Amy E. DeZern, MD, MHS:
Results from the BMT CTN 1102 trial show that allogeneic HSCT with a suitable donor vs no donor leads to improved outcomes for patients 50-75 years of age with higher-risk MDS. Patients with a donor had a very significant OS benefit (21% improvement), improved LFS (by 15%), and no decrease in QoL. Outcomes were similar for those of Medicare age (> 65 years) compared with patients 65 years of age or younger.
I think this is an incredibly meaningful study and one of the most practice‑changing abstracts presented at ASH 2020. It demonstrates that transplantation should be considered in patients with higher‑risk MDS, even those 65 years of age and older. Clearly, obtaining a referral to a transplant center early in a patient’s MDS course should be considered, and I hope to see increased referral patterns fairly soon after the publication of this abstract and the subsequent paper. These results will hopefully be a wake-up call for Medicare to allow patients with MDS much greater access to allogeneic HSCT.
Amy E. DeZern, MD, MHS:
Next I will discuss a post hoc analysis of the open-label, randomized phase II P-2001 trial of pevonedistat plus azacitidine compared with azacitidine monotherapy in patients with higher‑risk MDS, chronic myelomonocytic leukemia (CMML), or low-blast acute myeloid leukemia (AML).27
Pevonedistat is a unique drug that inhibits the NEDD8‑activating enzyme.28-30 This particular enzyme prevents protein ubiquitination upstream of the proteasome, disrupting the cell cycle and causing apoptotic death in tumor cells. In preclinical murine AML models, synergism was observed between pevonedistat and azacitidine, and this was the basis for the current study.31
Results from in P-2001 presented at ASCO in June 2020 showed that, in the intention-to-treat population, the median event‑free survival (EFS) times for combination therapy and monotherapy were 21.0 vs 16.6 months, respectively (HR: 0.67; P = .076).32 The median OS was similar between arms (21.8 vs 19.0 months; HR: 0.80; P = .334).
Amy E. DeZern, MD, MHS:
At ASH in December 2020, Sekeres and colleagues27 presented a post hoc analysis of the cohort of patients with higher-risk MDS in P-2001 (N = 67). Enrolled patients had not received hypomethylating agents and were considered ineligible for allogeneic HSCT at the time of entrance into the study. Patients were randomized 1:1 to receive pevonedistat plus azacitidine (n = 32) or azacitidine alone (n = 35).
The primary endpoint of the study was changed to OS due to regulatory input. Secondary endpoints were EFS (the original primary endpoint), and this analysis focused on clinical/cytogenetic risk by the revised IPSS (IPSS‑R) and genetic factors affecting efficacy.
Amy E. DeZern, MD, MHS:
As mentioned, patients had to have higher-risk MDS at the time of enrollment. By IPSS-R, patients could have intermediate‑risk disease (if ≥ 5% marrow myeloblasts), high‑risk disease, or very high–risk disease.33 Risk was also assessed using the Cleveland Clinic model formula, which assigns weighted coefficients to factors associated with OS and incorporates those with IPSS-R.34
Amy E. DeZern, MD, MHS:
Baseline characteristics in both arms were very similar. Approximately 70% of the patients were male, with a median age of approximately 73 years. Almost all had de novo MDS, and according to the WHO tumor classification, > 50% of the enrolled patients had MDS with excess blasts 2 (refractory anemia with excess blasts 2). The breakdown between intermediate-risk, high-risk, and very high–risk IPSS-R categories was similar in both arms and fairly equally distributed. The median time from initial diagnosis was approximately 2 months in both arms, which may speak to the fact that it takes some time to enroll these patients on clinical trials.
Amy E. DeZern, MD, MHS:
Although patients are frequently more concerned with OS, EFS is incredibly important in terms of avoiding AML transformation and extending time to death. The EFS curves show an advantage for combination therapy vs azacitidine alone. The separation of the EFS curves happened fairly early, between 6 and 9 months. The median EFS was 20.2 months for combination pevonedistat and azacitidine compared with 14.8 months for azacitidine (HR: 0.539; 95% CI: 0.292-0.995; P = .045), and this is clinically meaningful. Although the curves come closer together at 27 months, that is quite far out and, again, is clinically meaningful.
Regarding OS, the curves also separated fairly rapidly. The median OS was 23.9 months for combination therapy compared with 19.1 months for azacitidine alone (HR: 0.701; 95% CI: 0.386-1.273; P = .240). As with EFS, the OS curves start to come together at approximately 27 months.
Amy E. DeZern, MD, MHS:
The EFS and OS in patients with high-risk MDS as classified by the Cleveland Clinic formula are similar to EFS and OS as classified by IPSS-R. However, by looking at specific factors encountered with the Cleveland Clinic model, there is a greater split between the 2 arms, showing an EFS of 20.2 months in the combination arm vs 11.7 months for monotherapy (HR: 0.388; 95% CI: 0.166-0.902; P = .023). A similar benefit was observed for OS, with a difference of approximately 10 months (HR: 0.447; 95% CI: 0.190-1.05; P = .056).
Amy E. DeZern, MD, MHS:
The ORR in the combination arm was quite impressive with a CR of 52% and an ORR of 79%, whereas the ORR for the monotherapy arm was 57% with only a 27% CR rate. Although these are not statistically significant differences, they are clinically meaningful to patients. Hematologic improvement occurred in 24% of patients on the combination arm, which is important for patients who are ineligible for transplantation—this might mean less time in the clinic receiving transfusions, which improves QoL.
The duration of response is one of the things that I find most interesting about this study. The median duration of response was much longer in the combination arm compared with the monotherapy arm, at 34.6 months vs 13.1 months, respectively.
Amy E. DeZern, MD, MHS:
The rate of TI (including both RBC and platelets) trended higher in the combination arm, at 69.2% for the combination of pevonedistat and azacitidine vs 47.4% for azacitidine alone (Relative risk: 1.46; 95% CI: 0.81-2.65; P = .228). Likewise, the duration of TI was significantly longer with the combination vs azacitidine alone, at 23.3 months vs 11.6 months, respectively (HR: 0.11; 95% CI: 0.01-0.94; P = .016). Again, this is important because it shows that this combination therapy is not more myelosuppressive than azacitidine alone and may be more blast reducing, allowing more space for normal hematopoiesis.
Amy E. DeZern, MD, MHS:
AML transformation is a risk for any patient with higher‑risk MDS. Although there was not a statistically significant difference in AML transformation in this study, there is some separation in the curves suggesting that there may ultimately be an advantage for the combination arm.
The median time to AML transformation in the combination arm was 12.2 months compared with 5.9 months for monotherapy. There were only 14 patients in this cohort, but I still view these data as favorable despite the lack of statistical significance.
Amy E. DeZern, MD, MHS:
The adverse event rate may be one of the most clinically meaningful things we can glean from this study. Historically, combination therapy has not been successful for MDS due to increased toxicity without a marked increase in efficacy. As clinicians and patients, we are definitely willing to tolerate a little bit more pain, if you will, for more gain, but combination therapies in MDS have always been markedly more toxic than a hypomethylating agent monotherapy. P-2001 is one of the first studies where this was not the case, as the rates of adverse events are very similar in both arms, and I think this is one reason that lack of statistical significance in OS benefit is acceptable because the combination is not markedly more toxic and there is a separation in the curves.
One historical problem in clinical trials of combination therapy was azacitidine dose adjustment to mitigate toxicity. In this study, the median azacitidine dose intensity was similar in both arms (98%), meaning that dose adjustment was not necessary for patients to be able to tolerate azacitidine in combination with pevonedistat.
Amy E. DeZern, MD, MHS:
Patients with MDS increasingly undergo next-generation sequencing. The authors of this trial chose to incorporate data on molecular mutations to attempt to identify a prognostic signal for combination therapy. They looked at myeloid genes known to be commonly mutated and associated with poor prognosis. For example, mutation of SRSF2 was more prevalent in the combination arm (29% vs 12% with monotherapy) whereas IDH1 mutation was more prevalent in the azacitidine arm (16.0% vs 8.3% with combination therapy).
Amy E. DeZern, MD, MHS:
Looking at the ORR data by specific mutation, there is a substantial difference between the arms. This is something that needs to be explored further in randomized studies to find molecular signatures predicting which patients are most likely to respond to combination therapy or who may only need monotherapy.
Amy E. DeZern, MD, MHS:
In patients with higher-risk MDS, the combination of pevonedistat plus azacitidine significantly prolonged EFS, with a trend for extended OS, compared with azacitidine alone. Median EFS was 20.2 months vs 14.8 months, respectively (HR: 0.54; P = .045), and median OS was 23.9 months vs 19.1 months, respectively (HR: 0.70; P = .240). Moreover, adding pevonedistat to azacitidine also increased CR rate, median duration of response, TI, ORR, and delayed time to AML transformation. Toxicity was not significantly increased in the combination arm.
The phase III PANTHER trial will further compare combination pevonedistat and azacitidine to azacitidine monotherapy in patients with higher-risk MDS, and we look forward to seeing the results.35
Amy E. DeZern, MD, MHS:
Moving on, I will now discuss the DACOTA phase III study comparing decitabine with hydroxyurea for advanced proliferative CMML.36 CMML has features of both MDS and MPN, and the prognosis for patients with CMML is poor.37,38 Patients with myeloproliferative CMML (MP-CMML) tend to do worse than patients with myelodysplastic CMML due to higher white blood cell counts and proliferation.
Hydroxyurea was previously shown to improve ORR and OS in patients with MP-CMML.39 The efficacy of hypomethylating agents like azacitidine and decitabine in patients with advanced/high‑risk MP-CMML has also been evaluated.40 In phase II trials, decitabine monotherapy achieved an ORR of 38% to 48% and a median OS of 17.0-18.3 months.41,42
Amy E. DeZern, MD, MHS:
CMML overlaps MDS and MPNs and the prognosis for patients who carry this diagnosis is poor. Specifically, MP-CMML is defined by a white blood cell count of ≥ 13,000/μL, and those patients tend to do worse than patients who have a shift toward the MDS component of their CMML, who tend to have lower counts and less proliferation.
DACOTA was a multicenter, open-label, randomized phase III trial comparing decitabine with hydroxyurea therapy in patients with previously untreated MP-CMML (N = 170).36 Enrolled patients had white blood cell counts ≥ 13 x 109/L and were ineligible for HSCT at the time of enrollment.
Patients were randomized 1:1 to decitabine (n = 84) or hydroxyurea (n = 86). Decitabine was administered at the standard dose of 20 mg/m2 for 5 days and hydroxyurea was administered at a dose of 1‑4 g/day for 28 days in 28‑day cycles. It should be noted that patients in the decitabine arm could receive hydroxyurea in the first 3 cycles.
The primary endpoint was EFS, and events were defined as AML transformation, progressive disease, or death. Secondary endpoints included ORR, duration of response, OS, and safety.
Amy E. DeZern, MD, MHS:
Baseline characteristics were comparable across the arms. Overall, patients were fairly fit, with an Eastern Cooperative Oncology Group performance status 0/1 in the majority of patients. The majority of patients had CMML-1 according to the WHO classification. More than 50% of patients in both arms had intermediate‑2–risk disease, and approximately 40% were intermediate‑1 risk. This patient population needs to be treated but is not as high risk as those who meet criteria for true high CMML-specific prognostic scoring system criteria. Patients who received previous hydroxyurea treatment accounted for 40% to 45% of the patients in each arm. The median previous duration of hydroxyurea treatment was 26-27 days in both arms.
Amy E. DeZern, MD, MHS:
As shown here, the EFS curves overlap after a median follow-up of 13.9 months. The median EFS for the decitabine arm was 12.6 months compared with 10.3 months for the hydroxyurea arm (HR: 0.88; 95% CI: 0.61-1.25; P = .46). There is a small clinical difference, but it is not statistically significant.
Amy E. DeZern, MD, MHS:
Looking at the secondary endpoint of OS, patients receiving decitabine had a shorter median OS at 18.4 months compared with patients receiving hydroxyurea at 23.1 months, but this too was not statistically significant (HR: 1.07; 95% CI: 0.73-1.58; P = .73).
Amy E. DeZern, MD, MHS:
The cumulative incidence in both those who had AML and those who died was relatively similar between the 2 arms.
Amy E. DeZern, MD, MHS:
The ORR data for DACOTA are interesting, given that there was no EFS or OS benefit. Of the patients treated with decitabine, 63% had an ORR vs 34% of the patients treated with hydroxyurea (P = .0002). The median duration of response was 16.3 months across arms.
Of interest, although perhaps not unexpected, 8% of patients receiving decitabine achieved a CR after 3 cycles. Unsurprisingly, there were no CRs for patients receiving hydroxyurea, as it is only intended for cytoreduction. But when you look at stable disease with hematologic improvement, which is probably the most clinically meaningful of these response metrics for this patient population, it is very similar: 18% of those treated with decitabine and 21% of those treated with hydroxyurea.
The same response metrics at 6 cycles were similar although the gap between the arms is narrowing with an ORR for decitabine of 32% compared with 17% for hydroxyurea. Why this difference in responses did not translate into an OS benefit is confounded by a number of things, one of which may be safety.
Amy E. DeZern, MD, MHS:
Patients in the decitabine arm had a greater frequency of hospitalization compared with patients in the hydroxyurea arm (P = .05).36 This may be explained by patients in the decitabine arm having a slightly higher frequency of infections, which may have precipitated the hospitalizations.
Rates of grade ≥ 2 adverse event were similar across both arms. The types of adverse events are what we see in these patients in general: hemorrhage, cardiac, and pulmonary events. There were a fairly high number of musculoskeletal events in both arms, at 17% to 20%. This is a reasonable safety profile for a disease that is high risk and proliferative.
Amy E. DeZern, MD, MHS:
Although DACOTA was deemed a negative trial, I think it provides important clinical information for treating MP-CMML.36 Although there was not a difference in the EFS or OS between decitabine and hydroxyurea, of importance for the clinical trial community and for future therapeutic options for these patients, the proliferative nature of this disease did not prevent them from being randomized in a clinical trial. Both patients and their physicians were willing to embark upon a clinical trial paradigm and patients were randomized in a timely fashion.
Practically speaking, the data suggest that hydroxyurea is an appropriate treatment option for patients with MP-CMML. This is important since there has been some concern that hydroxyurea alone might have higher rates of transformation to AML or cause increased toxicity and was not sufficiently disease modifying, but perhaps that is not true.
In my opinion, continued investigation of the safety of hypomethylating agents in this patient population is important, partly because of the higher hospitalization rate. These data demonstrate that, compared with intravenous therapy, an oral pill option with hydroxyurea is not more toxic and does not diminish survival. This is an important discussion to have with patients.
Healthcare professionals have to determine which patients in this challenging group would benefit from treatment with hypomethylating agents. To date, no molecular signature has been identified that might predict which patients with CMML would benefit from a hypomethylating agent and which patients can be treated with hydroxyurea alone. Of importance to the CMML clinical trial community, results from DACOTA showed that patients with MP-CMML could be randomized in a clinical trial in a timely fashion, allowing more clinical trials for this unique patient population in coming years.
Amy E. DeZern, MD, MHS:
Symptomatic anemia is a major problem in the lower‑risk MDS community: More than 90% of these patients have this issue at the time of diagnosis.43 Current treatment options include RBC transfusions, erythropoiesis‑stimulating agents, and luspatercept20 (approved in April 2020). However, all of these treatments have imperfect response rates and limited duration of response, so we are in dire need of additional agents to manage the anemia seen in lower-risk MDS patients.
Roxadustat is an oral drug, unlike the other treatments mentioned, and it is a hypoxia‑inducible factor prolyl hydroxylase inhibitor shown to increase RBC production.44 At ASH 2020, Henry and colleagues45 presented results from a 52‑week update of a roxadustat trial in patients with primary, lower‑risk MDS who presented with transfusion‑dependent anemia.
Amy E. DeZern, MD, MHS:
This is an ongoing, 2‑part, randomized, placebo‑controlled phase III study of roxadustat in patients with lower‑risk MDS.46 At ASH 2020, results were presented from the open‑label portion of the trial with 3 sequential dose cohorts: 1.5, 2.0, and 2.5 mg/kg three times weekly.45 Doses are titrated every 8 weeks depending on hemoglobin response and transfusion metrics.
Patients in this study had low-risk to intermediate‑risk MDS by IPSS‑R, < 5% blasts in their bone marrow at the time of enrollment, and hemoglobin ≤ 10 g/dL. Eligible patients had a low transfusion burden (1-4 RBC units within 8 weeks of randomization), baseline endogenous erythropoietin levels ≤ 400 mIU/mL, and no del(5q). The criterion of low transfusion burden is increasingly seen in MDS study designs. Baseline erythropoietin levels are quite predictive of response.
The primary endpoint was TI for ≥ 8 weeks during the first 28 treatment weeks. There was also a follow-up at 52 weeks. Secondary endpoints included a ≥ 50% reduction in RBC transfusion burden over any 8 weeks vs baseline, and TI for ≥ 20 weeks.
Amy E. DeZern, MD, MHS:
Two thirds (66.7%) of the 24 patients completed 52 weeks of therapy, which is good considering they could choose transfusions and less time in the clinic. The median age in this study was 73 years, and the median MDS duration was 3.5 years. Four of the 8 patients who discontinued early did so due to a lack of efficacy. These patients were iron overloaded as measured by a median ferritin level > 1200 ng/mL at presentation. Of importance, given the available drugs in the field, 54.2% of the cohort had ring sideroblasts at presentation and 45.8% did not.
Amy E. DeZern, MD, MHS:
TI for ≥ 8 weeks (the primary endpoint of this study) was achieved in 38% of the patients. Remember that this was a dose-titration study, and 7 of the 9 patients who achieved TI (78%) received 2.5 mg/kg of roxadustat. Patients who received either 1.5 mg/kg or 2.0 mg/kg had lower response rates.
TI in patients with and without ring sideroblasts was interesting. Of the patients who did not have ring sideroblasts, 55% achieved TI vs only 23% of the patients with ring sideroblasts. By contrast, baseline erythropoietin level did not seem to matter regarding TI.
Amy E. DeZern, MD, MHS:
Reductions in RBC transfusion dependence by ≥ 50% during the first 28 weeks were very similar to data for the entire 52 weeks, at 58% vs 63%, respectively. This means patients spend less time on transfusions and, as an oral therapy, will minimize clinic visits. This suggests that the efficacy of roxadustat will be visible early in treatment and is not improved with longer duration of treatment. The frequency of TI for ≥ 20 weeks was low—only 13% of patients in the first 28 weeks and 17% over 52 weeks.
Amy E. DeZern, MD, MHS:
Treatment‑emergent adverse events occurred in 87.5% of the patients. However, investigators deemed that < 30% of these adverse events were attributable to roxadustat. The adverse events observed were similar to what has reported for other drugs for this type of disease. Diarrhea, dyspnea, and nausea were prominent, occurring in approximately one quarter of patients. Bronchitis and dizziness were also notable, occurring in approximately 20% of patients, but nothing else was particularly of clinical concern, and I think it is a reasonable safety profile. There were no deaths attributed to therapy and only 1 patient discontinued study treatment specifically due to adverse events.
Amy E. DeZern, MD, MHS:
The investigators concluded that roxadustat reduces RBC transfusions and enables TI in patients with lower risk MDS and TD anemia. Roxadustat was well tolerated at all doses tested, but 2.5 mg/kg is the optimal starting dose. In fact, that is the starting dose used in the ongoing double blind phase of this study. I think we will see more commentary on the ring sideroblast data following the completion of the double-blind phase, as roxadustat could be an answer for anemic patients without ring sideroblasts going forward in contrast to luspatercept, which is indicated for patients with ring sideroblasts.
Back
