Results from paradigm - a phase 2 randomized multi-center study comparing azacitidine and venetoclax to conventional induction chemotherapy for newly diagnosed fit adults with acute myeloid leukemia

Abstract

Abstract Introduction: For decades, standard upfront treatment for acute myeloid leukemia (AML) among fit patients (pts) has been intensive induction chemotherapy (IC), typically with cytarabine and anthracyclines. However, long-term outcomes remain poor, and IC continues to pose substantial short- and long-term morbidities, with significant burden to pts and hospitals. VIALE-A, a phase 3 trial of IC-ineligible pts, established superiority of azacitidine and venetoclax (aza-ven) vs aza alone, with a median OS of 15 months and composite remission rate (CCR) of 66%, which compare favorably to that expected for IC among older pts. We prospectively tested whether aza-ven was superior to IC in fit pts and could challenge the current treatment standard. Methods: This open-label, multicenter, phase II randomized clinical trial compared the therapeutic activity of conventional IC (7+3 regimen or liposomal daunorubicin and cytarabine [CPX351]) to aza-ven among IC-eligible pts aged ≥ 18. Pts with core binding factor fusions, FLT3 mutations, or NPM1 mutations (unless aged ≥ 60) were excluded. The IC arm incorporated cycles of chemotherapy consolidation following induction, for use as clinically appropriate per current standard of care. Pts were allowed to proceed to transplant (HCT) on both arms following response on protocol-directed therapy. The primary endpoint was event free survival (EFS), with events defined as progressive disease, persistent disease prompting therapy change, relapse, hospice, or death. Among secondary endpoints were response rates, OS, toxicity, measurable residual disease (MRD), hospitalization metrics, and quality of life (QOL). Pts were stratified by age (≥ 65 vs < 65) and pre-randomization selection in the IC arm to 7+3 or CPX351. EFS is estimated by Kaplan Meier method and assessed by log rank test and Cox PH regression. Clinical responses are assessed using an exact binomial test. Significance is declared at the two-sided 0.1 type I error. Results: As of 7/25/2025, 172 pts at 9 US centers were randomized to aza-ven (n=86) or IC (n=86), and accrual is complete. Median age was 64 for aza-ven and 65 for IC pts. 55% and 60% were male in the aza-ven and IC arms, respectively, and 65% and 69% were White. The majority of pts were ELN 2022 adverse risk (72%), 15% were intermediate-, and 12% favorable-risk. Distribution of risk categories did not differ across arms (p=0.44), nor did the proportions of TP53, NPM1, or IDH1/2 mutations. Median number of cycles completed were 4 for aza-ven and 2 for IC pts. Among IC pts, 54% received 7+3, and the rest, CPX351. By intent to treat analysis, the rate of overall response (OR=CR+CRh+CRi+PR+MLFS) was significantly higher in the aza-ven arm (88% vs 62%; P<0.001), as was CCR (CR+CRh+CRi) (81% vs 55%; P<0.001). CR rates for aza-ven and IC were not significantly different (59% vs 50%; P=.066). Procession to HCT following response from protocol therapy differed across arms (P=0.009) - 52 (61%) pts on aza-ven and 34 (40%) on IC arms. With a median follow-up of 16 months, 1-year EFS was 53% for aza-ven and 39% for the IC arm. EFS overall was significantly superior in the aza-ven arm (HR 0.61; P=0.017). There was no significant age effect on EFS when included with treatment in the Cox PH model, and HR for treatment remained significant (0.61; P=0.018). Grade (G)3/4 therapy-related adverse events in ≥10% were mainly hematologic and at similar rates across arms. G3/4 lung infections and sepsis occurred in 12% and 7% of aza-ven, and 15% and 11% of IC pts. 30- and 60-day mortality on the aza-ven arm were both 0%, vs 3.5% and 4.7% in the IC arm, respectively. At 2 weeks, aza-ven pts reported significantly better QOL (P=0.001), symptom burden (P=0.019), and depression symptom (P=0.007) scores, vs IC. Aza-ven pts were less likely to need ICU care (0 vs 9.8%; P=0.003), and experienced fewer inpatient days for the index hospitalization (15 vs 36; P<0.001) and during the first 6 months (41 vs 58; P<0.001). Conclusion: The study has met its primary endpoint. Aza-ven was associated with significantly improved EFS, as well as higher rates of OR and CCR, vs IC in younger, IC-eligible pts. Overall survival data continue to mature. A greater proportion of pts successfully went to HCT following response to aza-ven than those to IC. Aza-ven led to numerically fewer serious infectious complications, significantly improved QOL and symptom burden during initial therapy, with less time in the hospital and the ICU.