ASTX727 delivers superior response rates and associated survival benefit versus hydroxycarbamide/best supportive care in CMML and other MDS/MPN overlap syndromes: First results from the Phase 2 UK multicenter randomized ammo trial

Abstract

Abstract Background Chronic myelomonocytic leukemia (CMML) and the rarer MDS/MPN overlap syndromes share biological features, poor prognosis (median survival <2y) and few treatment options. Hypomethylating agents (HMA) are approved for CMML but less effective in proliferative disease, for which hydroxycarbamide (HC) remains standard-of-care (SoC). In Europe/UK HMA access is restricted to non-proliferative CMML-2. The phase 3 DACOTA randomized controlled trial (RCT) showed higher response for IV decitabine (DAC) v HC in proliferative CMML, but offset by toxicity for no survival benefit. ASTX727 is an oral fixed-dose combination of DAC 35mg/cedazuridine 100mg, delivering equivalent exposure to IV DAC 20mg/m2 and FDA-approved for MDS/CMML. Methods AMMO (ISRCTN30808508) is a UK phase 2 RCT comparing ASTX727 v HC/BSC in advanced MDS/MPN. Eligibility required Int-2/High risk by CPSS-Mol or other advanced/proliferative features requiring therapy; extending to non-CMML pts with IPSS-R higher-risk/advanced disease. Recruits were randomized (2:1) to ASTX727 (d1-5/28d) for 6 cycles, or BSC including HC (HMA not permitted). Responders could continue ASTX727 until progression. Primary outcome was best overall response rate (ORR) by C6, according to IWG/Savona criteria, analyzed by posterior probability distribution of difference using a Bayesian beta-binomial conjugate model. Secondary outcomes included overall, progression-free and transformation-free survival (OS; PFS; TFS), in a modified intention-to-treat population. Results Between Oct 2022-May 2024 77 pts from 11 centers were recruited, of whom 73 are evaluable for response. Median age was 73y (IQR 67-77). Disease distribution mirrored the epidemiology of MDS/MPN, with the majority (>60%) CMML. Most were higher-risk by CPSS-Mol/IPSS-R; 73% had proliferative disease mandating cytoreduction. Arms were well balanced other than more females (41% v 21%) randomized to ASTX727. Of 49 starting ASTX727, 12 stopped pre-C6 (8 for toxicity); 28 continued beyond C6. Median cycles received was 10 (IQR 5-18). Dose reductions were frequent, applied in (e.g.) 16% and 57% at C2 and C6, respectively. Of 24 on BSC, 75% received HC; 7 discontinued before end of C6 (4 due to progression/death). Best ORR for ASTX727 and HC/BSC was 53.0% (95% CrI 39.3-66.3%) and 30.3% (14.9-49.4%), respectively. Strict CR was achieved by 8.2% and 4.2%. Sampling posterior distribution models median ORR difference was 22.5% (95% HDI 0.1-44%) in favor of ASTX727. Total 227 CTCAE G3+ AEs were reported (in 48 pts): 208 (39 pts) on ASTX727 v 19 (9 pts) on HC/BSC. The only G3+ AEs occurring in ≥5% on either arm were neutropenia (38.8% ASTX727 v 8.3% HC/BSC), thrombocytopenia (30.6% v 8.3%), febrile neutropenia (24.5% v 0), anemia (20.4% v 16.7%) and lung infection (8.2% v 0). No unexpected safety signals were seen. Treatment-related toxicity (defined as any G3+ AE/G1+ SAE considered at least possibly attributable to drug) was experienced by 68% and 17% on ASTX727 and HC/BSC, respectively. The only treatment-attributed death occurred on HC/BSC. With 20mo median follow up, there were 32 deaths (18 on the ASTX727 arm) and 12 progressions (7 on ASTX727) / 4 to AML (1 on ASTX727). All survival measures showed significant advantage for ASTX727. Median PFS was 23.5mo (95% CI 15.8-NR) v 13.9mo (9.4-NR) [HR 0.44, 95% CI 0.22-0.86; p=0.017]. Median TFS was 23.5mo (18.4-NR) v 13.9mo (9.9-NR) [HR 0.40, 0.2-0.8; p=0.01]. At 12mo est survival probability was 0.831 (0.73-0.945) for ASXT727 v 0.724 (0.556-0.941) for HC/BSC. Median OS was 23.5mo (18.4-NR) v 16.9mo (13.9-NR) [HR 0.42, 0.21-0.86; p=0.017]. Survival benefits remained significant in CMML subgroup analyses. Multivariable Cox regression found no survival differences by other stratified variables. Conclusion ASTX727 displayed marked benefit compared with HC/BSC for advanced MDS/MPN. Rapid recruitment (6mo early) and high retention showcases feasibility/appetite for RCTs in this challenging group. Extending access beyond CMML AMMO also showcases HMA efficacy across related biological entities. The 53% ORR is in keeping with that expected for HMA monotherapy and almost double that for HC/BSC. Toxicity was higher, but acceptable and in line with prior ASTX727 trials. Crucially this did not negate its efficacy advantage, translating into significantly longer PFS, TFS and OS: the first randomized survival benefit demonstrated in MDS/MPN for ~30y and supporting ASTX727 as a new SoC for advanced MDS/MPN.