Subgroup analyses from the randomized, Phase 3 VERONA study of venetoclax with azacitidine (Ven+Aza) versus placebo with azacitidine (Pbo+Aza) in patients with treatment-naïve, intermediate and higher-risk Myelodysplastic Syndromes (HR MDS)

Abstract

Abstract Background: Patients with HR MDS face poor prognosis and are often ineligible for hematopoietic cell transplantation (HCT). After encouraging safety and efficacy in a phase 1b study (Garcia Blood 2025), Ven+Aza was evaluated against Pbo+Aza in patients with treatment-naïve HR MDS in the randomized, phase 3 VERONA study (NCT04401748). Primary analysis at 41.2 months median follow-up showed no difference in overall survival (OS) with Ven+Aza (22.18 mo) vs Pbo+Aza (21.68 mo; HR=0.908 [95% CI, 0.733–1.126]; P=.38), but modified overall response (mOR) was higher with Ven+Aza vs Pbo+Aza (76.2% vs 57.7%; nominal P<.0001; Garcia-Manero SOHO 2025). Here, we present additional outcomes and pre-planned subgroup analyses from VERONA with the aim of identifying patient subsets that may have received clinical benefit from Ven+Aza. Methods: VERONA enrolled patients aged ≥18 years with a diagnosis of MDS (WHO 2016), Revised International Prognostic Scoring System (IPSS-R) score >3 (intermediate, high, very high), ECOG PS 0–2, not immediately SCT eligible, and had no prior MDS therapy and no therapy-related MDS. Patients were randomized 1:1 to receive oral Ven 400 mg or Pbo once daily on Days 1–14 combined with IV or SC Aza 75 mg/m2 for 7 days in each 28-day cycle. Primary endpoint was OS. Patients were stratified by IPSS-R risk (very high, high, intermediate), region (North America, Europe, Japan, China, rest of world), and HCT eligibility. Pre-planned subgroup analyses were conducted for the primary endpoint of OS by unstratified Cox proportional hazards model and for mOR (sum of complete remission [CR], partial remission, and marrow CR [mCR]) by risk difference. Data cutoff was 4/10/2025. Results: A total of 509 patients were randomized (Ven+Aza, n=256; Pbo+Aza, n=253) and included in efficacy analysis. Of these, 255 patients in the Ven+Aza arm and 246 in the Pbo+Aza arm received ≥1 dose of study drug and were evaluable for safety. Baseline patient and disease characteristics were well balanced between arms (Ven+Aza vs Pbo+Aza): median age 72 vs 72 years, 66% vs 59% aged 18 to <75 years, 93% vs 92% ECOG PS 0–1, 7% vs 8% ECOG PS 2, 22% vs 22% North America, 78% vs 78% outside of North America. IPSS-R risk group was very high in 36% vs 32%, high in 37% vs 39%, and intermediate in 27% vs 28%. There was a balanced distribution of poor prognostic and frequently mutated genes between study arms, including TP53 (25% vs 19%), ASXL1 (30% vs 33%), RUNX1 (15% vs 24%), and EZH2 (3% vs 6%). Reasons for treatment discontinuation (Ven+Aza, n=231 [91%]; Pbo+Aza, n=234 [95%]): HCT (16% vs 8%), progressive disease (30% vs 45%), adverse event (20% vs 15%), and patient withdrawal (14% vs 17%). Overall, Ven was used as post-study therapy in 16% of patients (Ven+Aza, n=27 [11%]; Pbo+Aza, n=53 [21%]). Rate of AML transformation was 15% with Ven+Aza vs 20% with Pbo+Aza. There were no differences in OS across subgroups, though trends favoring Ven+Aza were observed in younger patients (18 to <75 years, HR=0.835 [95% CI, 0.630–1.107]; ≥75 years, HR=1.160 [95% CI, 0.834–1.613]) and those with excess blasts (≥5% to <20% blasts, HR=0.858 [95% CI, 0.676–1.090]; <5% blasts, HR=1.313 [95% CI, 0.795–2.169]). Subgroup analysis showed patients with TP53 mutation had an HR of 1.064 (95% CI, 0.670–1.688). Post-study HCT was received by 17% of patients (43/256) in the Ven+Aza arm and 13% (33/253) in the Pbo+Aza arm at a median of 5.6 months (range, 2.9–18.1) for those treated with Ven+Aza and 6.7 months (range, 2.2–33.9) for those treated with Pbo+Aza. The best response on study treatment prior to HCT were CR 25.6%, mCR 60.5%, and stable disease (SD) 14.0% in the Ven+Aza arm and CR 27.3%, mCR 33.3%, and SD 39.4% in the Pbo+Aza arm. Additional post-study therapy was given prior to HCT in 4/43 (9%) in the Ven+Aza arm and 10/33 (30%) in the Pbo+Aza arm. Subgroup analyses showed trend toward higher mOR with Ven+Aza vs Pbo+Aza in patients <75 years and those with very high IPSS-R, >5% blasts, and mutations in ASXL1, TP53, or RUNX1. mCR with hematologic improvement (mCR+HI) was higher with Ven+Aza vs Pbo+Aza (27.3% vs 18.2%). Conclusions: VERONA did not meet the primary endpoint of OS. A higher proportion of patients treated with Ven+Aza achieved mOR, mCR+HI, and numerically lower risk of AML transformation vs those treated with Pbo+Aza. Subgroup analyses showed favorable trends in response for Ven+Aza vs Pbo+Aza in younger patients, excess blasts, and select mutations.