Overall Survival and Progression-Free Survival of Patients Following Luspatercept Treatment in the MEDALIST Trial

Introduction: Overall survival (OS) is rarely assessed in patients (pts) with lower-risk myelodysplastic syndromes (LR-MDS) despite a previous analysis highlighting the impact of treatment on factors influencing OS (Santini V, et al. Clin Lymphoma Myeloma Leuk 2022;doi.org/10.1016/j.clml.2022.05.001). Luspatercept is US FDA-approved to treat anemia in adults with Revised International Prognostic Scoring System (IPSS-R) Very low- to Intermediate (Int)-risk MDS with ring sideroblasts (RS) following erythropoiesis-stimulating agent (ESA) failure. Luspatercept promotes late-stage erythropoiesis and significantly reduced transfusion burden (TB) in pts with LR-MDS vs placebo (PBO) in the phase 3 MEDALIST trial (Fenaux P, et al. N Engl J Med 2020;382:140-151). However, the baseline (BL) pt characteristics predictive of the greatest survival benefit from luspatercept treatment have yet to be investigated. This study aimed to report the probability of greater OS and progression-free survival (PFS) benefit from luspatercept vs PBO in the MEDALIST trial. Methods: Pts were ≥ 18 years of age with IPSS-R Very low-, Low-, or Int-risk MDS-RS who received ≥ 2 red blood cell (RBC) units/8 weeks (wk) in the 16 wk before randomization and were unlikely to respond/refractory to ESAs. Pts in the intent-to-treat (ITT) population were randomly assigned to receive luspatercept (n = 153) at 1.0 mg/kg (titration up to 1.75 mg/kg) or PBO (n = 76) subcutaneously every 3 wk; no crossover was allowed at any point of the study. Responders were defined as pts who achieved transfusion independence ≥ 8 wk during the first 24 wk of treatment. OS was defined as time from randomization to death from any cause. PFS was defined as time from MDS diagnosis to acute myeloid leukemia progression. Kaplan-Meier method was used to evaluate ITT and responder OS (Jan 15, 2021 data cut), and ITT PFS (Jan 15, 2022 data cut). Analysis of BL characteristics associated with PFS and OS included pts who remained on treatment or discontinued before 36 months (mo; Jan 15, 2022 data cut). Results: Median OS (95% confidence interval [CI]) of the ITT population was not significantly different between luspatercept (not applicable [NA]; 46.1-NA mo) and PBO (NA; 43.1-NA mo; P < 0.9604). Luspatercept responders had statistically significantly longer OS than luspatercept non-responders at wk 25 (hazard ratio 0.319; P = 0.0003; Figure 1A). Probability of 36-mo OS was similar in luspatercept vs PBO pts when stratified by the BL RBC TB categories < 6 units/8 wk (50/87 [57.5%] vs 23/43 [53.5%] pts, respectively; odds ratio [OR] 1.18; P = 0.6680) and ≥ 6 units/8 wk (35/66 [53.0%] vs 20/33 [60.6%], respectively; OR 0.73; P = 0.4768). Probability of 36-mo OS was greater for pts diagnosed with IPSS-R Very-low risk MDS receiving luspatercept vs PBO (14/18 [77.8%] vs 1/6 [16.7%], respectively; OR 17.5; P = 0.0088). Of pts with a BL platelet count > 400 × 109/L, those receiving luspatercept also had a significantly higher 36-mo OS probability than those receiving PBO (13/17 [76.5%] vs 3/9 [33.3%], respectively; OR 6.5; P = 0.0349). BL serum erythropoietin (EPO) levels and time from MDS diagnosis did not impact 36-mo OS probability in the luspatercept vs PBO arms. Median PFS (95% CI) of the ITT population was not significantly different between the luspatercept (NA; 223.57-NA mo) and PBO (NA; 95% CI NA-NA) arms (P = 0.3514; Figure 1B). No significant PFS advantage was observed when comparing luspatercept vs PBO pts by BL RBC TB categories < 6 units/8 wk (81/87 [93.1%] vs 36/43 [83.7%], respectively; OR 2.62; P = 0.0947) and ≥ 6 units/8 wk (62/66 [93.9%] vs 32/33 [97.0%], respectively; OR 0.48; P = 0.5184). Probability of 36-mo PFS of pts with a BL serum EPO level of 100 to ≤ 200 U/L who received luspatercept was significantly higher than that of pts who received PBO (36/37 [97.3%] vs 15/19 [78.9%], respectively; OR 9.6; P = 0.0238). Conclusions: Although the MEDALIST trial was not specifically powered to assess OS or PFS, these data show that achieving response with luspatercept treatment increased OS probability. Luspatercept was associated with increased 36-mo OS probability for pts with IPSS-R Very low-risk MDS and 36-mo PFS probability in pts with a BL serum EPO level of 100 to ≤ 200 U/L. Therefore, pts with LR-MDS with these BL characteristics may derive greater survival benefit from luspatercept. Future studies with longer follow-up times will help to clarify the long-term impact of luspatercept therapy. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal