CME
Physicians: Maximum of 1.00 AMA PRA Category 1 Credit™
Released: March 16, 2023
Expiration: March 15, 2024
Amer Zeidan, MBBS, MPH:
We will start by looking at the newly published classification systems for MDS. The WHO‑based classification system, which has long been available for myeloid neoplasms including MDS, recently was updated to the fifth edition.1 In 2022, a new classification system for MDS was published called the International Consensus Classification (ICC) of Myeloid Neoplasms and Acute Leukemias.2 Both are starting to be used in the clinic, but they differ in how they define certain aspects of MDS. For example, the ICC system does not include MDS with ring sideroblasts as a morphologic category, and the nomenclature for MDS with excess blasts differs between them.
Amer Zeidan, MBBS, MPH:
In a retrospective study performed at Moffitt Cancer Center under the aegis of the International Consortium for MDS, these 2 systems were validated and compared against each other using an institutional database of 2231 patients with MDS (based on WHO 2016 criteria).3
Updates to both classification systems mainly focused on the introduction of genetically defined groups based on the presence of TP53, deletion 5q (del[5q]), or SF3B1 abnormalities. Of importance, there also was an effort to define a new category within the ICC system called “MDS/AML” for patients who have 10% to 19% blasts. In the WHO classification, this group is defined as “MDS with increased blasts 2,” or MDS-IB2.
It was apparent that both classification systems function well in terms of correlation between the specific subgroup and median leukemia-free survival (LFS) and median OS. Patients with MDS who had SF3B1 mutations had the best median LFS (105.7-109.4 months) and OS (111.6 months), whereas those with multihit TP53 had the worst survival (median LFS: 10.0-11.5 months; median OS: 13.2-14.5 months). Patients with del(5q) had moderate median LFS (62.9 months) and relatively prolonged median OS (75.6 months).3
Amer Zeidan, MBBS, MPH:
The subgroup comparisons revealed some aspects of each classification system that seem to be better than the other.3 For example, the presence of multilineage dysplasia (which was shown to worsen outcomes) and single-lineage dysplasia was maintained in the ICC system but not the WHO system. In this population, using the ICC criteria, the median LFS for those with single-lineage dysplasia vs multilineage dysplasia was 74.2 months vs 41.5 months, respectively, and the median OS was 79.4 months vs 49.6 months.
Another important area was the introduction of new subgroups in the WHO classification system that are absent from the ICC system—for example, the introduction of hypoplastic MDS. The tendency of patients with hypoplastic MDS to better respond to immunosuppressive therapies was long known, but it was not a formal group in the WHO classification system. Also of importance is the formal introduction of MDS with fibrosis, as fibrosis worsens outcomes in patients with MDS. Here, the median OS for MDS with fibrosis was 18.9 months vs 24.3 months for MDS with increased blasts.
Amer Zeidan, MBBS, MPH:
P53-mutated MDS was consistently shown to be the most significant predictor for inferior survival in patients with MDS.3 The HR was 3.09 (95% CI: 2.06-4.61) for LFS and 3.39 (95% CI: 2.25-5.12) for OS.
Amer Zeidan, MBBS, MPH:
One of the important outcomes of this study on behalf of the International Consortium for MDS was the recommendation for a unification of these 2 classification systems. A big concern is that having 2 divergent systems could lead to confusion in terms of diagnosis, clinical trial enrollment, communication with patients, and—eventually—treatment recommendations. There is a growing sense in the community that having 1 system is certainly the way to go, and I would personally say that having a single imperfect system is better than having 2 theoretically perfect but different systems.
Amer Zeidan, MBBS, MPH:
Another important presentation on prognostication for MDS looked at validation of the Molecular IPSS (IPSS‑M) for MDS.
The IPSS and revised IPSS (IPSS-R) long have been used to recommend treatment for patients with lower‑risk or high‑risk MDS. Recently, mutations with important implications for patient prognosis have been identified, but these were not formally included in either original IPSS risk score. Thus, the International Working Group for Prognosis of MDS developed the IPSS-M, incorporating a very large sample database (N = 2957) from many centers. The IPSS‑M was shown to improve the accuracy of prediction compared with IPSS-R.4
Amer Zeidan, MBBS, MPH:
At ASH 2022, an external validation of IPSS-M was presented using a large European database of patients with MDS (N = 2876) called GenoMed4All.5
The C-index for IPSS-M was better than IPSS-R in the overall cohort (survival C-index: 0.81 vs 0.74; P <.0001; LFS C-index: 0.89 vs 0.76; P <.0001) and even among patients without detectable mutations (survival C-index: 0.89 vs 0.73; P <.0001; LFS C-index: 0.97 vs 0.81; P <.0001).
The IPSS‑M improved prognostic accuracy not only for the all-patient cohort, but also for those receiving HMAs. It was able to risk stratify patients according to their chance of posttransplant survival (C-index: 0.76 vs 0.60; P <.0001)5
Of importance, 46% of patients were restratified, with 23.6% being upstaged and 22.4% being downstaged with IPSS-M compared with IPSS-R.
IPSS-M uses 31 gene variables to calculate risk scores, in addition to clinical criteria such as hemoglobin level, platelet count, and cytogenetic abnormalities.4 Because many centers do not detect all of these genes, the study authors explored the minimum number of genes needed to accurately predict prognosis. It was found that a minimum set of 15 genes was sufficient to ensure risk prediction accuracy >70%.5 There still will be centers that do not have the capability to test for all of these genes, but the IPSS-M allows entry of missing data.
Amer Zeidan, MBBS, MPH:
The IPSS‑M is easily calculated using the web calculator and is starting to be used in the clinic, and I think over time it probably will replace the IPSS‑R as the primary method of prognostication for patients with MDS.
Although this work validated the use of IPSS-M in terms of clinical decisions, the use of IPSS-M risk groups still must be validated in the real-world setting, as historically these have been based on the IPSS-R. The next step will be to look at different databases to see how patients from different groups within the IPSS-M fare when receiving different therapeutic interventions and incorporate the IPSS-M prospectively into clinical trials design.
Amer Zeidan, MBBS, MPH:
Next, I am going to talk about developments in therapy for lower‑risk MDS.
Most clinical trials in lower‑risk MDS have included patients who are already transfusion dependent. Lenalidomide is approved by the FDA for the treatment of adult patients with transfusion-dependent anemia due to low- or intermediate-1‒risk MDS associated with a del(5q) cytogenetic abnormality, with or without additional abnormalities.6
The multicenter phase III Sintra‑REV trial is special because it is looking at lenalidomide in patients with MDS who are not yet transfusion dependent. In total, 61 patients with IPSS low- or intermediate-1‒risk MDS were randomized 2:1 to receive lenalidomide or placebo. All enrolled patients had symptomatic anemia (hemoglobin <12 g/dL) and a del(5q) abnormality but did not need a transfusion. Lenalidomide treatment was limited to 2 years, and the patients were monitored thereafter, with the primary endpoint being the time to transfusion dependence.7
Amer Zeidan, MBBS, MPH:
The use of lenalidomide improved the time to transfusion dependency, which was not reached in the intervention arm compared with 11.6 months with placebo (P = .003).7
Amer Zeidan, MBBS, MPH:
Several secondary endpoints were tracked, all of which were in favor of lenalidomide, including erythroid response (P <.001) and cytogenetic response (P <.001). The median OS was similar (8.4 years with lenalidomide vs 7.4 years with placebo), but this is not unusual in the context of lower‑risk MDS.7 Of importance, lenalidomide did not increase the risk of conversion to AML. Also, there was no observed increase in clonal expansion in patients with TP53 mutations, which has been seen in other analyses and still should be monitored closely.
Amer Zeidan, MBBS, MPH:
As expected, there was an increased incidence of adverse events (AEs) within the lenalidomide group, mostly gastrointestinal (GI) events and rash. However, most were grade 1/2 and reversible.7
Amer Zeidan, MBBS, MPH:
The clinical implications of this very important study are still being discussed among experts in the field. This study provides a rationale for consideration of lenalidomide in patients with del(5q) who are anemic but who have not yet started regular transfusion. Ideally, we need a longer follow-up to ensure that patients are not having clonal expansion of mutant TP53, although the preliminary data suggest that this does not happen.
Amer Zeidan, MBBS, MPH:
Another important study on the lower‑risk MDS front is the phase II/III IMerge study, which is looking at treatment with imetelstat, a first-in-class telomerase inhibitor, for patients with lower‑risk MDS who were refractory and relapsed to erythropoiesis-stimulating agents (ESAs).8,9 Imetelstat has previously shown preclinical activity in patients with MDS.
IMerge enrolled patients who were heavily transfusion dependent (needing ≥4 units of red blood cells per 8 weeks at baseline), could not have del(5q), and were lenalidomide and HMA naive. The primary endpoint was ≥8-week TI, with key secondary endpoints including safety, OS, PFS, and time to progression to AML.
Patients in the initial phase II cohort (N = 38) received IV infusion with imetelstat every 4 weeks, achieving a good 8-week TI rate (n = 16; 42%), and 11 patients (29%) had TI longer than 1 year.8
Amer Zeidan, MBBS, MPH:
An update focusing on those 11 patients who had sustained responses (TI ≥1 year) was presented at ASH 2022.9 This patient subset had a median duration of TI of 92.4 weeks (95% CI: 69.57-140.86) and substantial increases in hemoglobin. At median follow-up (57 months), median PFS was 34.2 months (95% CI: 25.1-39.2) and median OS was 56.1 months (95% CI: 29.4-NE), and no progression to AML was seen in any of the 11 patients with sustained responses.
Top-line results of the ongoing phase III part of the IMerge trial recently were described.10 The study met its primary endpoint, with a significant improvement in 8-week TI for imetelstat (39.8%; 95% CI: 30.9%-49.3%) compared with placebo (15%; 95% CI: 7.1%-26.6%; P <.001). Imetelstat importantly showed activity in patients with MDS independent of ring sideroblast status. Of importance, patients achieving 8-week TI with imetelstat also showed a substantial increase in hemoglobin levels compared with placebo (3.6 g/dL vs 0.8 g/dL; P <.001), and there was a reduction in red blood cell transfusions and hematologic improvement‒erythroid (42.4% vs 13.3%; P <.001).
Amer Zeidan, MBBS, MPH:
There is a lot of excitement about the positive results of this phase III study and we are eagerly awaiting the full presentation of these results. From the AE point of view, we know from the phase II part that imetelstat can cause cytopenia and liver enzyme abnormalities, but these generally are reversible with treatment interruptions.8 The preliminary safety signals for the phase III part seem to be in the same line,10 but we will have to wait and see the full data once the phase III study is fully presented.
Amer Zeidan, MBBS, MPH:
Now, moving on to high-risk MDS, I want to highlight 2 studies.
Sabatolimab is a monoclonal antibody that targets TIM-3, an inhibitory receptor that is expressed not only on T‑cells, but also on native immune cells, as well as leukemia blasts and leukemia stem cells. Sabatolimab is postulated to have a dual mechanism of action, both as an immune checkpoint inhibitor and as potentially causing direct toxicity to MDS blasts.11
A previous phase Ib study suggested durable clinical benefit for sabatolimab in combination with an HMA in patients with high-risk or very high‒risk MDS (overall response rate [ORR]: 56.9%; median duration of response: 17.1 months).12 Based on these promising results, further clinical trials are now underway.
Amer Zeidan, MBBS, MPH:
The randomized phase II STIMULUS‑MDS1 trial is studying sabatolimab vs placebo in combination with an HMA in adult patients (N = 127) with intermediate- to very high‒risk MDS who were unsuitable for intensive chemotherapy.13 All enrolled patients received an HMA, which could have been either decitabine or azacitidine at the choice of the treating physician.
The coprimary endpoints of the study were CR and PFS, with secondary endpoints including OS, duration of CR, ORR, event-free survival, LFS, TI, and safety, among others.
Amer Zeidan, MBBS, MPH:
The co-primary endpoints of the study were not met, as there was no statistically significant improvement in either CR rate (21.5% vs 17.7%) or median PFS (11.1 months vs 8.5 months) after sabatolimab treatment compared with placebo (both combined with HMAs).13
Of importance, there appeared to be a late separation in the PFS curves in favor of the intervention arm, suggesting a possible lag in therapeutic effect, which is commonly seen in immune checkpoint inhibitor treatments.
Amer Zeidan, MBBS, MPH:
An exploratory subgroup analysis of patients with lower IPSS-R risk categories (sabatolimab plus HMA: n = 35; placebo plus HMA: n = 34) showed improved median PFS (17.6 months vs 8.5 months; HR: 0.465; 95% CI: 0.243-0.891) and CR (28.6% vs 14.7%).13 Similarly, subgroup analysis of patients with <10% blasts at baseline (sabatolimab plus HMA: n = 32; placebo plus HMA: n = 29) revealed longer median PFS (11.3 months vs 8.3 months) and improved CR (28.1% vs 17.2%).
Thus, a longer follow-up is going to be important for this study, especially because patients who had lower disease burden seemed to respond best.
Amer Zeidan, MBBS, MPH:
Numerically, there was a lower incidence of grade ≥3 AEs in the sabatolimab arm vs the placebo arm.13 The most common were neutropenia (53.2% vs 63.5%) and thrombocytopenia (37.1% vs 42.9%). There was a higher rate of febrile neutropenia with sabatolimab (35.5%) compared with placebo (23.8%). The incidence of immune-mediated AEs with sabatolimab plus HMA vs placebo plus HMA was 9.7% vs 17.5% for all grades and 4.8% vs 7.9% for grades ≥3.
Amer Zeidan, MBBS, MPH:
The clinical implications of sabatolimab in MDS remain to be seen. STIMULUS-MDS1 did not meet its primary endpoint, but trials with sabatolimab are ongoing. The follow-up phase III STIMULUS‑MDS2 trial (NCT04266301) already has accrued more than 500 patients with intermediate- to very high‒risk MDS, who will be randomized to receive sabatolimab plus azacitidine or placebo plus azacitidine. Results are expected in 2024, and we hope they will provide the definitive answer on the role of sabatolimab in patients with higher-risk MDS.
Amer Zeidan, MBBS, MPH:
Another intervention I wanted to highlight is the use of ruxolitinib for patients with chronic myelomonocytic leukemia (CMML). CMML is a difficult disease to treat, and not many agents are approved for high‑risk disease beyond the use of HMAs. In a phase III trial in patients with advanced proliferative CMML (N = 170), decitabine was shown to improve response rate compared with hydroxyurea (63% vs 35%; P = .0004), but it did not improve median OS (18.4 vs 21.9; P = .67).14 Clearly, this is an area of unmet need.
Ruxolitinib is a JAK1/2 inhibitor that is approved by the FDA for patients with MF, as well as patients with polycythemia vera (PV) who are intolerant or resistant to hydroxyurea.15 This drug was previously tested in a phase I/II study for patients with CMML, showing 38% clinical ORR, including spleen and erythroid response, and acceptable safety.16
At ASH 2022, an open-label, multicenter phase II expansion trial was presented that looked at ruxolitinib in patients with CMML.17 Enrolled patients (N = 29) were HMA naive or nonresponsive and had a significant splenomegaly or MPN TSS of ≥17 and good counts (absolute neutrophil count: ≥250/µL; platelets: ≥35,000/µL).
The primary endpoint was ORR according to the International Working Group criteria for MDS/MPN. Secondary endpoints included time to AML transformation, OS, and duration of response.
Amer Zeidan, MBBS, MPH:
The study authors observed that two thirds of patients achieved clinical benefit, and the study ORR was 17%; 54% of patients had improvement of >50% in their TSS, and 33% had ≥35% SVR (SVR35).17
It does seem like ruxolitinib could have some benefit in CMML despite it being off label. However, performing a randomized study is very challenging because of the rarity of CMML.
Amer Zeidan, MBBS, MPH:
The full clinical implications of this study remain to be determined, but I think there could be consideration on a case-by-case basis for the use of ruxolitinib in patients with CMML. This is particularly true if they are not transplant candidates or they have not responded or stopped responding to HMAs, and after informed discussion with the patient about the risks and benefits. Getting this approved could be a struggle with insurance companies because it is not an FDA-labeled indication for ruxolitinib, but I think it can be considered for patients who have constitutional symptoms or splenomegaly as the predominant manifestation of their CMML.
Amer Zeidan, MBBS, MPH:
We are making significant progress on MDS, and it is important to incorporate our understanding of molecular alterations into routine disease management not only at tertiary care centers, but also in the community setting. Molecular data are increasingly included in the new classification systems (ICC, WHO fifth edition), as well as prognostic tools such as IPSS‑M. IPSS‑M probably will replace the nonmolecular risk stratification tools over the next few years. This implies that there should be wider use—and hopefully more insurance coverage—of next-generation sequencing for all patients with MDS and in countries with limited resources. However, this will be important not only for patient counseling and prognosis, but also for identification of new targets for intervention. For example, if a patient has an IDH mutation, they could be considered for off‑label use of IDH inhibitors. I also think unifying the ICC and WHO classification systems is going to be very important, as is further validation of IPSS‑M for clinical decision-making.
Amer Zeidan, MBBS, MPH:
In terms of lower‑risk MDS, this year actually has seen significant progress. As discussed previously, positive results were announced from the phase III IMerge study in the ESA‑refractory setting. The open-label phase III COMMANDS trial (NCT03682536) was not presented at ASH 2022, but preliminary results recently were announced. This trial was designed to compare luspatercept with epoetin alfa for the treatment of IPSS-R very low‒, low-, or intermediate-risk MDS in ESA-naive patients who required transfusion. The study met its primary endpoint of improved TI and hemoglobin levels compared with epoetin alfa, and we await presentation of the data. These 2 studies could potentially change the landscape of MDS management, pending full presentation of data and regulatory review and assuming the drugs receive approval for these indications. Luspatercept could become a novel option in the frontline setting in addition to ESAs, and imetelstat could be a second‑line agent, regardless of the patient’s ring sideroblast status.
There were other good presentations on the use of lower doses of oral decitabine, which also showed activity, as well as Sintra-REV, which—as discussed previously—is testing lenalidomide in patients with del(5q) who are not yet transfusion dependent.
We hope all of this is pointing to earlier intervention in terms of the disease timeline. In lower‑risk MDS, we historically have not treated patients until they have symptoms or reach transfusion dependence, and we generally have not been able to improve their OS. Now, as we have more drugs that are active, we can potentially move toward patients who have milder symptoms and who do not yet require transfusion. We hope to also start seeing improvements in OS, which has been very difficult to improve for lower-risk MDS, so those studies will need to be done.
Amer Zeidan, MBBS, MPH:
Management of high‑risk MDS continues to be quite challenging, as there really have not been any new agents since the approval of azacitidine (2004) and decitabine (2016), despite testing of many drugs and combinations with HMAs. Although we have seen lots of good early phase II data, many of the randomized phase III trials have been negative.
As discussed previously, although the primary endpoint of STIMULUS‑MDS1 (sabatolimab plus HMA vs HMA alone) was not met, the potential for delayed benefit remains, especially in patients with lower disease burden. The results of the randomized phase III STIMULUS-MDS2 trial (NCT04266301) are anticipated in 2024, which we hope will provide a definitive answer about the clinical benefit of sabatolimab in combination with HMA.
In addition to sabatolimab, numerous other agents are being tested in phase III studies including magrolimab (anti‑CD47 antibody), venetoclax (BCL-2 inhibitor), and tamibarotene (RARα agonist). All of these are ongoing phase III trials, and if at least 1 of them turns out positive, it would potentially change the standard of care (SoC) for patients with higher‑risk MDS, which today continues to be single‑agent HMAs for most patients.
Amer Zeidan, MBBS, MPH:
The refractory/relapsed setting for MDS continues to be a very difficult situation, with a median survival of approximately 6 months. Generally, patients do not do well because most are not able to go to transplant, and there are no approved therapies in the relapsed/refractory setting. There was a presentation at ASH 2022 about an inhibitor of CXCR1/2 (SX-682) that was well tolerated and showed some activity. In the interim analysis of a phase I dose-escalation study in patients after HMA failure (N = 17), a dose-dependent increase in ORR (up to 50%) was seen with SX-682 treatment; 47% of patients across all doses had reduction in marrow blasts, and marrow CR was achieved in 33% of patients receiving the selected optimal dose.18
Regardless, this is a very small cohort of patients with limited follow-up, so we need longer follow up and a bigger sample size. In general, management of MDS after HMA failure continues to be a very difficult area, and more drug development is needed urgently to help improve the outcomes of our patients.
Back
