Background
Phase 3, randomized, double-blind, placebo-controlled trial. 229 patients with RBC transfusion-dependent lower-risk MDS with ring sideroblasts (RS; ≥15% of erythroid precursors, or ≥5% RS with SF3B1 mutation), refractory, intolerant, or ineligible for erythropoiesis-stimulating agents (ESA). Randomized 2:1 (luspatercept 153, placebo 76). IPSS Low or Int-1 risk.
Interventions and follow up
Arm A: Luspatercept 1.0 mg/kg subcutaneously every 3 weeks, titrated up to 1.75 mg/kg based on response (n=153)
Arm B: Placebo subcutaneously every 3 weeks (n=76)
Primary endpoint: RBC transfusion independence (TI) ≥8 consecutive weeks during weeks 1–24
mFollow up: Median 41.0 months (extended follow-up)
Arm B: Placebo subcutaneously every 3 weeks (n=76)
Primary endpoint: RBC transfusion independence (TI) ≥8 consecutive weeks during weeks 1–24
mFollow up: Median 41.0 months (extended follow-up)
Results
RBC-TI ≥8 weeks (weeks 1–24): 38.2% vs 13.2%, P<.001
RBC-TI ≥12 weeks: 28.1% vs 7.9%, P<.001
Erythroid hematologic improvement: 52.9% vs 11.8%
Median Hgb increase in RBC-TI responders: +1.5 g/dL
RBC-TI ≥12 weeks: 28.1% vs 7.9%, P<.001
Erythroid hematologic improvement: 52.9% vs 11.8%
Median Hgb increase in RBC-TI responders: +1.5 g/dL
Adverse events
Most common any-grade (luspatercept): fatigue 27%, diarrhea 22%, asthenia 20%, nausea 20%
Overall: Grade ≥3 treatment-emergent AEs 38% (luspatercept) vs 26% (placebo); discontinuation 16% vs 8%
Other: No clinically meaningful increase in AML transformation rate observed during the trial period
Overall: Grade ≥3 treatment-emergent AEs 38% (luspatercept) vs 26% (placebo); discontinuation 16% vs 8%
Other: No clinically meaningful increase in AML transformation rate observed during the trial period
Conclusions
Luspatercept significantly increased transfusion independence rates compared with placebo in ESA-refractory/ineligible LR-MDS with ring sideroblasts, establishing a new treatment option for this molecular subset. The 8-week TI rate nearly tripled relative to placebo.
Key Limitations
Enrichment for ring sideroblast status limits generalizability to non-RS LR-MDS (~20% of LR-MDS). Results in SF3B1-wild-type RS patients are less robust. The trial did not compare luspatercept to ESA (head-to-head addressed in COMMANDS). No OS advantage shown during the trial period. IPSS-R risk stratification was not used; IPSS-M may reclassify some patients.
Clinical Context
MEDALIST supported FDA approval of luspatercept for LR-MDS with ring sideroblasts in April 2020. Luspatercept is guideline-recommended for RS-positive or SF3B1-mutated LR-MDS after ESA failure. The subsequent COMMANDS trial demonstrated superiority of luspatercept over ESA in the first-line setting. ESMO-MCBS: not formally scored for MDS.