Luspatercept and imetelstat for myelodysplastic syndrome: updates from ASH 2025 annual meeting Authors: Yagya Ahlawat, Heng Jiang, Delong Liu Published: 2026 ABSTRACT AbstractLuspatercept, an erythroid maturation agent, and imetelstat, a first-in-class telomerase inhibitor, have expanded the therapeutic landscape for patients with myelodysplastic neoplasms (MDS). Recent data presented at the 2025 American Society of Hematology (ASH) Annual Meeting further clarify their evolving clinical roles. New analyses evaluated the real-world use of luspatercept in patients with lower-risk MDS (LR-MDS), while additional studies characterized the efficacy and outcomes associated with imetelstat in transfusion-dependent (TD) LR-MDS and in advanced MDS or acute myeloid leukemia following hypomethylating agent failure. We summarized the latest updates on luspatercept and imetelstat for MDS from the 2025 ASH Annual Meeting.Keywords: Luspatercept, Imetelstat, Myelodysplastic neoplasms, Telomerase inhibitor To the editor Recent presentations at the 2025 American Society of Hematology (ASH) Annual Meeting provided new insights into therapies for anemia in myelodysplastic neoplasms (MDS), particularly luspatercept in lower-risk MDS (LR-MDS) and imetelstat in transfusion-dependent (TD) LR-MDS and advanced disease following hypomethylating agent (HMA) failure. Luspatercept in LR-MDS Luspatercept is an erythroid-maturation agent that promotes late-stage red cell development by binding to the TGF- superfamily members [1]. The ASH 2025 luspatercept data can be grouped on three clinically relevant themes: activity compared with erythropoiesis-stimulating agents (ESAs), earlier use, and dosing strategy. A post hoc analysis of the COMMANDS trial evaluated the efficacy of luspatercept in ESA-naïve, TD, LR-MDS. Although patients with earlier-stage features, including hemoglobin (Hgb) ≥ 8 g/dL, transfusion burden (TB) of 2 units, and serum erythropoietin levels (sEPO) ≤ 100 U/L, had higher response rates in both treatment arms, luspatercept achieved higher rates and longer duration of RBC transfusion independence (RBC-TI) than epoetin alfa across these subgroups. Among luspatercept-treated patients, those with Hgb ≥ 8 g/dL achieved higher rates of ≥ 12 weeks RBC-TI than those with Hgb < 8 g/dL (87.5% vs. 69.1%), with longer median durations of RBC-TI with luspatercept (150.0 vs. 108.3 weeks); similar trends were observed when stratifying by TB and sEPO levels [2]. These findings support consideration of luspatercept as a preferred first-line option in selected ESA-naïve TD LR-MDS patients, particularly those with lower TB and other favorable baseline features. Real-world studies further support earlier use of luspatercept in LR-MDS (Table 1). In a U.S. retrospective cohort, 182 patients receiving luspatercept and 181 patients receiving ESAs were evaluated; 69.6% patients receiving luspatercept achieved Hgb ≥ 10 g/dL within 6 months compared with 58% with ESA, with a shorter median time to response (2.3 vs. 5.3 months) [3]. Additional data reported RBC-TI ≥ 12 weeks in 64.2% patients within 6 months [4]. Another real-world analysis reported dose escalation in 27.5% and discontinuation in 28.1%, mainly due to lack of response [5]. These findings are limited by retrospective design, heterogeneous baseline characteristics, and variable prior ESA exposure. Table 1Luspatercept studies in low risk MDS from 2025 ASH annual meetingStudy (reference)Patient populationNumber of patients (n)Major endpointsKey resultsCOMMANDS trial post hoc analysis [2]ESA-naïve TD LR-MDSluspatercept vs. ESA363luspatercept, n=182ESA, n=181≥12‑week RBC‑TIDuration of ≥12‑week RBC‑TIErythroid hematologic improvement (HI-E)>2‑fold higher odds of achieving ≥12‑week RBC‑TI with luspaterceptHigher response rates with favorable baseline (Hb ≥8 g/dL, TB=2U, sEPO ≤100 U/L; rate of ≥12‑week RBC‑TI ~86%-88%) and longer duration of ≥12‑week RBC-TIU.S. real-world retrospective cohort [3]LR-MDSluspatercept vs. ESA418luspatercept, n=213ESA, n=205Time to achieve HgB ≥10 g/dLHigher rate and faster response with luspatercept (69.9% vs. 58.0%; median response time 2.3 vs. 5.3 months)Baseline Hbg≥8 g/dL predicted better response (aHR 2.0) and TD status predicted worse response (aHR 0.5)France and Germany real-world retrospective cohort [4]LR-MDS [with known ring sideroblast (RS) status in Germany]luspatercept only110France, n=53Germany, n=57Hgb response within 6 months treatment initiationTime to achieving TITime to discontinuation64.2% TD LR-MDS achieved ≥12‑week RBC‑TI within 6 months of treatment initiation68.8% RS positive patients achieved ≥12‑week RBC‑TI within 6 months of treatment initiationMedian time to TI in TD LR-MDS of 2.4 monthsHgb increased by ≥1.5 g/dL in 64.7% patientsU.S. real-world retrospective cohort [5]LR-MDS (with known RS status)luspatercept only320≥8‑week and ≥12‑week RBC‑TIModified HI-E (mHI-E)Dose escalation and discontinuation61.5% TD LR-MDS achieved ≥12‑week RBC‑TI at a median 8.8-month follow up (RS negative: 53.3%; sEPO< 200 U/L: 69.2%)50.9% of overall cohort achieved mHI-E27.5% needed at least one dose escalation and 28.1% discontinued luspaterceptMAXILUS trial interim analysis [6]TD LR-MDS (ESA-naïve and ESA-R/R/I with known RS status)luspatercept only105ESA-naïve, n=52ESA-R/R/I, n=53≥8‑week RBC‑TI with concurrent mean Hgb increase≥1.0 g/dL≥12‑week RBC‑TIHigher rate of achieving both outcomes in ESA-naïve group (73.3% vs. 65.3%; 70% vs. 49%)ESA-naïve, RS positive disease, and sEPO≤ 200 U/L are favorable response baseline features; no new safety signals and no new thromboembolic events or AML reportedQOL-ONE Phoenix trial interim analysis [7]TD LR-MDS with del5q and ESA- or lenalidomide R/R/Iluspatercept only22≥8‑week RBC‑TIDuration of RBC-TIQOL measurement changes≥8‑week RBC-TI achieved in 38% (n=8) of heavily pretreated del(5q) patients with concurrent Hgb >10 g/dLMedian time to response ~13 weeks; median response duration ~41 weeks with sustained responses and QOL improvement (fatigue/physical function) Luspatercept studies in low risk MDS from 2025 ASH annual meeting ESA-naïve TD LR-MDS luspatercept vs. ESA 363 luspatercept, n=182 ESA, n=181 ≥12‑week RBC‑TI Duration of ≥12‑week RBC‑TI Erythroid hematologic improvement (HI-E) >2‑fold higher odds of achieving ≥12‑week RBC‑TI with luspatercept Higher response rates with favorable baseline (Hb ≥8 g/dL, TB=2U, sEPO ≤100 U/L; rate of ≥12‑week RBC‑TI ~86%-88%) and longer duration of ≥12‑week RBC-TI LR-MDS luspatercept vs. ESA 418 luspatercept, n=213 ESA, n=205 Higher rate and faster response with luspatercept (69.9% vs. 58.0%; median response time 2.3 vs. 5.3 months) Baseline Hbg≥8 g/dL predicted better response (aHR 2.0) and TD status predicted worse response (aHR 0.5) LR-MDS [with known ring sideroblast (RS) status in Germany] luspatercept only 110 France, n=53 Germany, n=57 Hgb response within 6 months treatment initiation Time to achieving TI Time to discontinuation 64.2% TD LR-MDS achieved ≥12‑week RBC‑TI within 6 months of treatment initiation 68.8% RS positive patients achieved ≥12‑week RBC‑TI within 6 months of treatment initiation Median time to TI in TD LR-MDS of 2.4 months Hgb increased by ≥1.5 g/dL in 64.7% patients LR-MDS (with known RS status) luspatercept only ≥8‑week and ≥12‑week RBC‑TI Modified HI-E (mHI-E) Dose escalation and discontinuation 61.5% TD LR-MDS achieved ≥12‑week RBC‑TI at a median 8.8-month follow up (RS negative: 53.3%; sEPO< 200 U/L: 69.2%) 50.9% of overall cohort achieved mHI-E 27.5% needed at least one dose escalation and 28.1% discontinued luspatercept TD LR-MDS (ESA-naïve and ESA-R/R/I with known RS status) luspatercept only 105 ESA-naïve, n=52 ESA-R/R/I, n=53 ≥8‑week RBC‑TI with concurrent mean Hgb increase≥1.0 g/dL ≥12‑week RBC‑TI Higher rate of achieving both outcomes in ESA-naïve group (73.3% vs. 65.3%; 70% vs. 49%) ESA-naïve, RS positive disease, and sEPO≤ 200 U/L are favorable response baseline features; no new safety signals and no new thromboembolic events or AML reported TD LR-MDS with del5q and ESA- or lenalidomide R/R/I luspatercept only ≥8‑week RBC‑TI Duration of RBC-TI QOL measurement changes ≥8‑week RBC-TI achieved in 38% (n=8) of heavily pretreated del(5q) patients with concurrent Hgb >10 g/dL Median time to response ~13 weeks; median response duration ~41 weeks with sustained responses and QOL improvement (fatigue/physical function) The open-label phase 3b MAXILUS trial evaluated luspatercept initiated at the maximum-approved dose of 1.75 mg/kg in TD LR-MDS regardless of prior ESA exposure [6]. Interim analysis (n = 105) showed good tolerability, with no new safety signals or thromboembolic events. Among ESA-naïve patients, 73.3% achieved ≥ 8-week RBC-TI and 70.0% achieved ≥ 12-week RBC-TI during a 24-week course; rates were lower in ESA-relapsed/refractory or intolerant (R/R/I) patients (65.3% and 49.0%). Dose reductions occurred in 13.2% of ESA-naïve patients and delays in 36.5% [6]. QOL-ONE Phoenix, a phase 2 single-arm multicenter study in TD LR-MDS patients with deletion 5q, reported Hgb improvement in some evaluable patients and improvement in fatigue-related quality-of-life measures, although longer follow-up and larger patient numbers are needed to clarify the durability and clinical impact of these findings [7]. Imetelstat in MDS Imetelstat is a competitive telomerase inhibitor that induces apoptosis in malignant haematopoietic progenitor cells and was approved for TD LR-MDS based on results from the IMerge trial, particularly in ESA-R/R/I patients [8, 9]. Secondary analyses from the IMerge trial further identified features associated with improved response and suggested a potential survival benefit [10, 11]. With median follow-up time of 45 months, median overall survival (OS) in the intent-to-treat population was 47.8 months with imetelstat group versus 44.8 months with placebo [10]. Favorable trends were also observed in progression-free survival (PFS) and time to AML progression (Table 2). Another post hoc analysis demonstrated that early treatment-emergent cytopenias correlated with greater hemoglobin increases and higher erythroid response and transfusion independence rates [11]. Although this may represent an on-target pharmacodynamic effect, imetelstat-associated myelosuppression remains clinically important and may require treatment delays, dose modifications, close laboratory monitoring, and proactive management of cytopenia-related adverse events. In the phase 2 IMpress study, imetelstat showed no marrow responses in patients with relapsed or refractory AML/high-risk MDS after HMA-based therapy, suggesting minimal single-agent activity in this higher-risk setting,, while increased exposure was not associated with increased toxicity [12]. Table 2Imetelstat studies in MDS from 2025 ASH Annual meetingStudy (reference)Patient populationNumber of patients (n)Major endpointsKey resultsIMerge trial [8]ESA-R/R/I TD LR-MDS without del5q and lenalidomide- and hypomethylating agent-naïveimetelstat vs. placebo178imetelstat, n=118placebo, n=60≥8‑week or longer RBC‑TIRate of 24-week RBC-TIDuration of RBC-TIHI-EHigher rate of achieving ≥8‑week or longer RBC‑TI in imetelstat group (40% vs. 15%)Longer median response duration of ≥8‑week or longer RBC‑TI in imetelstat group (51.6 vs. 13.3 weeks)Higher rate of mHI-E in imetelstat group (30% vs. 10%)Key secondary endpoint---Higher rate of ≥24‑week RBC‑TI in imetelstat group (28% vs. 3%)IMerge trial post hoc analysis [10]ESA-R/R/I TD LR-MDS without del5q and lenalidomide- and HMA-naïveimetelstat vs. placebo178imetelstat, n=118placebo, n=60OSPFSprogression to AMLImproved survival with imetelstat vs. placebo (median OS 47.8 vs. 44.8 months; median PFS 47.6 vs. 42.2 months)Lower risk of progression to AML in imetelstat cohort (1.7% vs. 3.3%)IMerge trial and IMerge QTc substudy post hoc analysis [11]ESA-R/R/I TD LR-MDS without del5q and lenalidomide- and HMA-naïveimetelstat only226Association of imetlstat-related cytopenias and Hgb responseHI-EEarly thrombocytopenia and neutropenia correlated with greater response (≥50% platelet reduction: Hgb increase 2.07 vs. 1.17 g/dL; ≥75% neutrophil reduction: 2.53 vs. 1.58 g/dL)Higher HI-E rate with ≥50% platelet reduction (34.2% vs. 11.9%; ≥24-week RBC-TI 32.6% vs. 7.1%)IMpress trial interim analysis [12]HR-MDS or AML patients R/R/I azacitidine or decitabine or venetoclax/azacitidineimetelstat only23MDS, n=6AML, n=17Overall response (CR, CRi, PR, or HI)Limited single-agent activity in HR-MDS/AML after HMA failure (no CR/PR observed; one HI)No increase in toxicity despite increased exposure Imetelstat studies in MDS from 2025 ASH Annual meeting ESA-R/R/I TD LR-MDS without del5q and lenalidomide- and hypomethylating agent-naïve imetelstat vs. placebo 178 imetelstat, n=118 placebo, n=60 ≥8‑week or longer RBC‑TI Rate of 24-week RBC-TI Duration of RBC-TI HI-E Higher rate of achieving ≥8‑week or longer RBC‑TI in imetelstat group (40% vs. 15%) Longer median response duration of ≥8‑week or longer RBC‑TI in imetelstat group (51.6 vs. 13.3 weeks) Higher rate of mHI-E in imetelstat group (30% vs. 10%) Key secondary endpoint---Higher rate of ≥24‑week RBC‑TI in imetelstat group (28% vs. 3%) ESA-R/R/I TD LR-MDS without del5q and lenalidomide- and HMA-naïve imetelstat vs. placebo 178 imetelstat, n=118 placebo, n=60 OS PFS progression to AML Improved survival with imetelstat vs. placebo (median OS 47.8 vs. 44.8 months; median PFS 47.6 vs. 42.2 months) Lower risk of progression to AML in imetelstat cohort (1.7% vs. 3.3%) ESA-R/R/I TD LR-MDS without del5q and lenalidomide- and HMA-naïve imetelstat only Association of imetlstat-related cytopenias and Hgb response HI-E Early thrombocytopenia and neutropenia correlated with greater response (≥50% platelet reduction: Hgb increase 2.07 vs. 1.17 g/dL; ≥75% neutrophil reduction: 2.53 vs. 1.58 g/dL) Higher HI-E rate with ≥50% platelet reduction (34.2% vs. 11.9%; ≥24-week RBC-TI 32.6% vs. 7.1%) HR-MDS or AML patients R/R/I azacitidine or decitabine or venetoclax/azacitidine imetelstat only 23 MDS, n=6 AML, n=17 Limited single-agent activity in HR-MDS/AML after HMA failure (no CR/PR observed; one HI) No increase in toxicity despite increased exposure In summary, the updates from ASH 2025 may help refine treatment selections for LR-MDS anemia. Although many analyses were post hoc, retrospective, or interim, the consistency of findings across data sources supports earlier luspatercept use, particularly in patients with lower TB, higher baseline Hgb, and lower sEPO. Imetelstat remains an important option for ESA-R/R/I TD LR-MDS, but cytopenia-related toxicity requires careful patient selection and monitoring. Acknowledgements DL is the editor-in-chief of this journal. Abbreviations American society of hematology Myelodysplastic syndromes Lower-risk MDS Transfusion-dependent Acute myeloid leukemia Hypomethylating agent Epoetin alfa Erythropoiesis-stimulating agent Hemoglobin Serum erythropoietin levels Red blood cell Transfusion independence Food and drug administration Relapsed/Refractory/Intolerant Overall survival Progression-free survival Deletion 5q Quality of life High-risk MDS Complete remission Complete remission with incomplete hematologic recovery Partial remission Hematologic improvement Author contributions YA and DL conceptualized and designed the study. YA and HJ collected the data, and drafted the manuscript. All authors revised and approved the final version of the manuscript. Funding none. Data availability No datasets were generated or analysed during the current study. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement No datasets were generated or analysed during the current study.