Luspatercept Significantly Reduces Red Blood Cell (RBC) Transfusion Burden, Regardless of Gene Mutation Frequency, Spectrum, and Prognostic Significance, Among Patients (Pts) with LR-MDS Enrolled in the MEDALIST Trial

Background: Luspatercept is a first-in-class erythroid maturation agent that binds TGF-β superfamily ligands to reduce aberrant Smad2/3 signaling and enhance late-stage erythropoiesis. The phase 3 MEDALIST trial evaluated luspatercept in pts with RBC transfusion-dependent, IPSS-R-defined very low-, low-, and intermediate-risk MDS with ring sideroblasts (RS+) who were refractory, intolerant, or ineligible to receive erythropoiesis-stimulating agents. This study explored associations of gene mutations, as analyzed by next-generation sequencing (NGS), with response to luspatercept, as well as dynamics of gene mutations on therapy in MEDALIST pts.

Methods: DNA was isolated from bone marrow (BM) mononuclear cells from 222 of 229 pts enrolled in the study (148 luspatercept, 74 placebo) at screening and, when available, following treatment. NGS of 23 MDS-relevant genes was performed at screening and every 24 weeks; mean coverage was 1,000-fold and the variant allele frequency (VAF) cutoff was ≥ 1%. BM cell populations were analyzed by cytomorphology. Response criterion of RBC transfusion independence (RBC-TI) of ≥ 8 weeks within the first 24 weeks of treatment was used for correlative analyses.

Results: Mutations in SF3B1 were found in 91.0% of pts analyzed at screening (median VAF 42%, range 6-71%), consistent with the study population being RS+. Overall, a median of 2 (range 0-5) of the 23 MDS-relevant genes analyzed were mutated per pt. In addition to SF3B1, the most frequently mutated genes were TET2 (41.9%), DNMT3A (18.9%), ASXL1 (13.1%), and SRSF2 (8.1%). Mutation profiles were similar to those found in previous studies of refractory anemia with RS (RARS; Malcovati L, et al. Blood. 2015;126:233-41) and balanced between luspatercept and placebo arms.

Numbers of mutated genes at baseline were distributed similarly in luspatercept responders (R) and non-responders (NR) (Figure A), and comparable response rates were achieved irrespective of number of mutations, with response rates of 36.4%, 34.9%, 42.4%, and 33.3% for pts with 1 mutation, 2 mutations, 3 mutations, and 4 or 5 mutations in the 23 MDS-relevant genes analyzed, respectively. Response to luspatercept was independent of the presence of mutations in any of the genes analyzed individually (Figure B) or when grouped by functional categories (e.g. spliceosome, epigenetic regulation, transcription factor, etc.) (Figure C). Circos plots of co-occurring mutations showed similar mutation profiles in R and NR (Figure D). Response rates were also similar regardless of baseline SF3B1 allelic burden (R: 43%, NR: 42%; P = 0.11). At baseline, BM erythroid precursors were higher in R (R: 32.8%, NR: 26%; P = 0.008; while R and NR had similar levels of RS+ cells [R: 80%, NR: 84%; P = 0.25], Figure E), consistent with the postulated activity of luspatercept on the erythroid lineage.

When comparing the frequency of mutation changes in luspatercept- vs placebo-treated pts at week 24 of the study, no statistically significant differences were observed in the frequency of newly acquired mutations (13/126 [10.3%] pts in luspatercept vs 8/64 [12.5%] pts in placebo, P = 0.63) or mutation losses (4/126 [3.2%] in luspatercept vs 5/64 [7.8%] in placebo, P = 0.17). Evaluation of changes in allele burden (median VAF at week 24 vs baseline) for mutations in genes associated with adverse prognosis (ASXL1, SRSF2, U2AF1, NRAS, IDH2, GATA2, TP53, RUNX1, and EZH2; Bejar R. Curr Opin Hematol. 2017;24:73-8) showed no change between luspatercept- or placebo-treated pts (1.01-fold, n = 58 and 0.95-fold, n = 19, respectively, P = 0.69).

Conclusions: Pts enrolled in the MEDALIST study had mutations consistent with RS+, lower-risk MDS with a preponderance of SF3B1 mutations; genes associated with poor prognosis (and other genes) were balanced between study arms. RBC-TI responses with luspatercept were achieved regardless of SF3B1 allelic burden, number of baseline mutations, and presence of individual mutations, including adverse mutations, or co-mutations.

View largeDownload slide

View largeDownload slide

Close modal