Efficacy and safety of luspatercept with or without erythropoiesis‐stimulating agent in patients with myelodysplastic neoplasms with ring sideroblasts: A multicentre real‐life study by the <scp>GFM</scp>

Erythropoietin-stimulating agents (ESAs) were until recently the only first-line treatment option for anaemia of lower risk myelodysplastic syndromes (LR-MDS)1 with no deletion 5q. Luspatercept, a first-in-class erythroid maturation agent, was approved for patients with MDS with ring sideroblasts (MDS-RS) and MDS/myeloproliferative neoplasms with RS and thrombocytosis (MDS/MPN-RS-T) requiring red blood cell (RBC) transfusions after ESA failure, based on the phase 3 MEDALIST trial results.2, 3 Its reimbursement was granted in France in July 2022 and was rapidly prescribed by hospital physicians and dispensed through retail pharmacies. We conducted a study to assess efficacy and safety of luspatercept in routine clinical practice through centres of the Groupe Francophone des Myélodysplasies (GFM). Inclusion criteria were MDS-RS and MDS/MPN-RS-T (according to the WHO 2016 classification),4 age ≥ 18 years and refractoriness to a first-line high-dose ESA treatment. Based on a previous GFM clinical trial showing that, in anaemia of LR-MDS resistant to ESA, lenalidomide combined with ESA yielded better erythroid response than lenalidomide alone,5 and to the fact that ESA is mainly active on early erythropoiesis, whereas luspatercept is mostly active on late erythropoiesis,6 addition of ESA to luspatercept was suggested in case of luspatercept failure. This real-life multicentre study was approved by the GFM scientific committee. Informed consent was obtained before inclusion, and the study was conducted according to the Declaration of Helsinki. Participating centres should prospectively collect and record data in the French nationwide MDS registry at luspatercept onset. International Prognostic Scoring System (IPSS), IPSS-Revised (IPSS-R)7 and IPSS-Molecular (IPSS-M)8 were calculated at inclusion. Patients were categorized, according to the International Working Group (IWG) 2018 criteria9 in no transfusion dependency (NTD) and transfusion dependency (TD), including low transfusion burden (LTB) and high transfusion burden (HTB) groups. Response to luspatercept +/− ESA was defined according to the IWG 2018 criteria after 24 weeks of luspatercept monotherapy, and 16 weeks of luspatercept + ESA combination. The primary end-point was achievement of transfusion independence (TI) or maintenance of non-transfusion dependence (NTD) status for ≥8 weeks during the first 24 weeks (W24) of luspatercept alone or for ≥8 weeks during the first 16 weeks of luspatercept + ESA. Key secondary end-points were achievement of erythroid response (HI-E) and adverse events (AE). Treatment-related adverse events (TRAE) were defined according to the NCI-CTCAE 5.0. Factors influencing TI achievement were studied by univariate and multivariate analyses. Logistic regression was performed in patients who were TD at luspatercept onset, with all model assumptions thoroughly assessed. Multiple comparisons between categories were adjusted using the Tukey correction. Between July 2022 and January 2024, 187 patients (55% males, 45% females) were enrolled. Detailed patient characteristics and treatment exposure are summarized in Tables S1 and S2. Median age at inclusion was 77 (range, 36–92). MDS-RS subtype represented 87% and MDS-MPN-RS-T 11% of patients; SF3B1 mutation was detected in 103 of 139 tested patients (74%). According to the WHO 2016, five patients were not classified as MDS-RS at diagnosis but displayed SF3B1 mutation. All patients had previously received high-dose ESA (at least 60 000 units of epoetin or 300 μg of darbepoetin/week during at least 8 weeks) with an overall response rate of 58.8%; 47 had previously received lenalidomide and 17 a hypomethylating agent. The median follow-up was 522 days (47–882) for the entire study. Of the 171 RBC-TD patients (55 with LTB and 116 with HTB), 55 (32%) achieved TI by week 24 (64% of LTB and 17% of HTB patients). Of the 133 RBC-TD patients who received at least seven injections of luspatercept, TI was obtained in 55/133 (41%) by week 24 and 33 (25%) were still in TI through week 48. Of the 16 NTD patients, 15 maintained TI by week 24 (one premature discontinuation) and 12 maintained TI until week 48 (Figure S1A). In multivariate analysis, in TD patients, only baseline LTB was associated with better TI (OR 4.64, CI 95% (1.93, 11.2), p < 0.001) (Table S3). HI-E at week 24 was observed in 76 patients (47%): 10/16 NTD (62.5%), 35/55 LTB (66%) and 31/116 HTB (27%). HI-E at week 48 was maintained in 45 patients: 10/12 NTD (83.3%), 26/29 LTB (90%) and 9/19 HTB (47%) (Figure S1B). Thirty-nine patients (21%) prematurely discontinued treatment before week 24 evaluation due to TRAE in 21 patients, progression to higher risk MDS or acute myeloid leukaemia in four patients, treatment failure in eight, patient decision in one and one occurrence of lymphoma. Four patients died due to comorbidities (Figure S1C). TRAE were reported in 72 patients (39%), including fatigue (alone or with dizziness and/or arthralgia) reported in 31 (17%) patients. As previously described, the number of reported TRAE decreased over time: 72 reported by 6 weeks, 45 at 12 weeks and 16 at 21 weeks (Figure S1D). No dose reduction or ESA addition improved the symptoms. In 54 patients, ESA was added after failure of luspatercept monotherapy. At the time of ESA addition, four patients were TI, four had prematurely stopped luspatercept and 46 were TD (including 32 patients in primary failure and 14 patients having relapsed after response to luspatercept) (Figure S1A). Baseline characteristics of those patients are summarized in Table 1. Fifty (93%) of the 54 patients received luspatercept at 1.75 mg/kg/3 weeks, combined with epoetin (n = 38) or darbepoetin (n = 16). ESA was started at 30–40 000 UI (epoetin) or 150 μg (darbepoetin) weekly in 40 patients and dosage was subsequently doubled in 33 of them. Fourteen patients started ESA at full dose. Median duration of exposure to luspatercept and ESA combination was 5.5 months (range, 1–21) and 38 patients (70%) received at least 16 weeks of the combination. Of the 50 RBC-TD patients (10 with LTB, 40 with HTB), 7 (14%) achieved TI, including 8 (23.5%) of the 34 patients who received at least 16 weeks of combination (Figure 1A). In the entire cohort (54 patients), HI-E was obtained in 11 patients (20%: 0/4, 4/10 (40%) and 7/40 (18%) of NTD, LTB and HTB patients respectively), and 29% of the 38 patients who received at least 16 weeks of combination treatment achieved HI-E (Figure 1B). Three responders had relapsed after 2, 6 and 12 months, respectively, and eight patients were still in response after 12–19 months. Of the 11 responders, nine and five had previously achieved HI-E with ESA and luspatercept monotherapy respectively. Two patients without HI-E after ESA and single agent luspatercept reached HI-E with the combination. No predictive factor of response to luspatercept combination was found in univariate analysis. No patterns of somatic mutations (especially of mutations in addition to SF3B1) were observed in patients achieving or maintaining TI with luspatercept alone +/− ESA (Figure S2B). At last follow-up, treatment had been discontinued in 43 patients (78%) due to lack of response (n = 35), loss of response (n = 3), adverse events including fatigue (n = 2), systemic hypertension (n = 1) and hepatic cytolysis (n = 1). One patient discontinued treatment due to MDS progression. Although approximatively one-third of our patients had received another previous treatment (in addition to ESA), our results with luspatercept alone are consistent with MEDALIST results and previously published real-life experiences10, 11 (RBC-TI in 38% in MEDALIST, 30.8% and 31.9% in the real-life Italian and US studies, 32.2% in our study). We confirmed that the most important predictive factor for response was the transfusion burden. Those concordant data suggest that the benefit of luspatercept is higher if prescribed early after RBC-TD occurrence. Recently, RBC-TI with luspatercept was reported in more than 60% of TD patients treated in first line.12 Our population was probably frailer and with more comorbidities than patients usually included in phase 3 protocols but we reported a tolerance profile in concordance with the literature with a 12.8% incidence of grade 3–4 side effects and a decrease in TRAE reported over time of treatment. Our 29% HI-E (11/38 patients who received at least 16 weeks of the combination) with ESA plus luspatercept appears lower than previously reported responses13-15: 3/7 patients (43%) in the Italian13 and 10/28 (35.7%) in the US cohort.15 Of note, when ESA is combined simultaneously to luspatercept as a second line of treatment, the HI-E seems higher.14 Obviously, in our series of ESA association with luspatercept, patients were resistant to both drugs as single agents and were more heavily pretreated. Indeed, 27/54 (50%) patients had received at least two lines before luspatercept monotherapy, including 23 patients treated with lenalidomide + ESA. The retrospective nature of our study and the relatively small sample size represent limitations to our study. However, we provide new data on the luspatercept and ESA combination, which has limited side effects and requires further exploration in prospective studies. Maud D'Aveni and Thibault Comont wrote the manuscript including preparation of Tables. Laurence Schenonne, Juan Pardo and Fatiha Chermat collected patient data that were further quality checked by Maud D'Aveni and Thibault Comont. Maud D'Aveni and Jose Torregrosa performed analysis and interpretation of patient data. Marie Sébert, Lorea Aguinaga, Aspasia Stamatoullas, Sophie Dimicoli, Emmanuel Gyan, Sophie Park, Sylvain Thepot, Etienne Daguindau, Guillaume Beziat, Teresa Botin, Claire Calmettes, Lenaig Le Clech, Houria Debarri, Etienne Paubelle, Clemence Santana, Bohrane Slama, Annel-Laure Taksin, Lise Willems included patients and participated in patient care. Michaela Fontenay, Lionel Adès and Pierre Fenaux contributed to manuscript review and editing. All authors reviewed and approved the manuscript. The authors are indebted to all patients for their participation, as well as to all physicians and clinical research associates and GFM networks who contributed to patient care and recruitment for the study, particularly, Marie Sébert, Lorea Aguinaga, Guillaume Beziat, Teresa Botin, James Norwood, Houria Debarri, Claire Calmette, Bohrane Slama, Geoffroy Venton, Emmanuel Gyan, Lise Willems, Yann Leveneur. AI content: The authors declare that no AI content is included in this manuscript. The authors declare the following competing financial interests: Maud D'Aveni has participated in consulting or advisory boards for Abbvie, BMS and Sanofi and travel grants from Sanofi. Thibault Comont has participated in consulting or advisory boards for Abbvie, AMGEN, BMS, Grifols, Novartis, and has received travel grants from Abbvie, AMGEN, BMS, Grifols, Novartis, has received research funding from Novartis and has received payment for speaker bureau from Abbvie, AMGEN, BMS, Grifols, Novartis. Jose Miguel Torregrosa has participated in consulting advisory boards for BMS, Alexion, Novartis, Sobi, Sellas Biosciences, received travel grants from Sellas Biosciences, Novartis and Alexion. Aspasia Stamatoullas has received travel grants from BMS, Takeda, AbbVie. Lenaig Le Clech has participated in advisory boards for Abbvie. Etienne Daguindau has received consulting fees from Abbvie, BMS, Amgen, Sanofi, Novartis, Gilead-Kite, Takeda, Johnson & Johnson and Incyte, research funding from BMS, Novartis and travel accommodation from Abbvie. Clemence Santana has participated in advisory boards for Abbvie and BMS. Figures S1–S2. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.