ELN Risk Stratification Is Not Predictive of Outcomes for Treatment-Naïve Patients with Acute Myeloid Leukemia Treated with Venetoclax and Azacitidine

Abstract

Background: Venetoclax (Ven), a highly selective oral BCL-2 inhibitor, in combination with hypomethylating agents (HMAs) such as azacitidine (Aza), is approved for the treatment of patients (pts) with newly diagnosed acute myeloid leukemia (AML) who are ineligible for intensive chemotherapy (IC). An open-label, non-randomized phase 1b trial (NCT02203773) and the confirmatory VIALE-A phase 3 randomized trial (NCT02993523), both enrolling pts with intermediate and poor risk cytogenetics per NCCN Guidelines, demonstrated that the combination of Ven+Aza led to high rates of complete remission and longer overall survival (OS) than Aza alone (median OS of 14.7 months [mo] vs 9.6 mo), with similar safety profiles. In this exploratory post hoc analysis of the pooled treatment-naïve, IC ineligible pts treated with Ven+Aza in the phase 1b and phase 3 trials, patients were stratified according to the 2017 ELN risk classification system (which is based on data from patients receiving IC and utilizes cytogenetic and molecular determinants commonly associated with AML). We report clinical outcomes for these patients per ELN 2017 risk group, which will also be updated using the newly published ELN 2022 criteria. Methods: For the phase 1b trial, our analysis includes pts who were enrolled to evaluate safety and efficacy of Ven+Aza at a daily dose of 400 mg for VEN. For the phase 3 trial, pts were randomly assigned in a 2:1 ratio to the Ven+Aza group or placebo (Pbo)+Aza group. All pts received a standard dose of Aza (75 mg/m2 SC or IV on days 1-7 every 28-day cycle). Ven (400 mg) or Pbo was administered orally, once daily, in 28-day cycles. Pts with favorable risk cytogenetics such as t(8;21), inv(16), t(16;16) or t(15;17) as per the NCCN Guidelines Version 2 were excluded from the study. The pooled group of pts from the two trials having both site-reported cytogenetic data and bone marrow aspirate with centrally determined molecular data allowing for ELN genetic risk classification were included in this analysis. Pts were retrospectively classified per the ELN 2017 criteria as favorable-, intermediate-, or adverse-risk AML. Results: Inclusion of central molecular data allowed the reclassification of patients according to ELN 2017 recommendations (Figure 1). 63 of the 67 Ven 400mg + Aza pts enrolled in the phase 1b trial, and 329 of the 431 total pts enrolled in the VIALE-A trial were evaluable for inclusion in this analysis (279/353 in the pooled Ven+Aza group and 113/145 in the Pbo+Aza control group from the phase 3 study). In the Ven+Aza and Pbo+Aza groups, 46 (16.5%) and 20 (17.7%) pts had favorable-risk AML, 65 (23.3%) and 19 (16.8%) had intermediate-risk AML, and 168 (60.2%) and 74 (65.5%) had adverse-risk AML, respectively, by ELN. In favorable-risk AML pts, CR+CRi was achieved in 69.6% with VEN+AZA vs in 20% with Pbo+Aza; for intermediate-risk AML, CR+CRi was achieved in 75.4% vs 42.1%; for adverse-risk AML, CR+CRi was achieved in 61.3% vs 25.7%. Median OS was longer with Ven+Aza in all three risk categories when compared to Pbo+Aza: 21.09 vs. 12.96 mo, HR = 0.45 [95% CI: 0.24, 0.83] in the favorable-risk group; 23.26 vs. 13.14 mo, HR = 0.62 [95% CI: 0.35, 1.08] in the intermediate-risk group; 11.53 vs. 7.43 mo, HR = 0.59 [95% CI: 0.43, 0.82] in the adverse-risk group. The median OS was similar for favorable and intermediate risk groups who received Ven+Aza (21.09 mo and 23.26 mo, respectively; Figure 2). While the median OS of the Ven+Aza adverse risk group was lower (11.53 mo), two subgroups with distinct outcomes are apparent: adverse risk pts with a TP53 mutation (36.3% [61/168]) and adverse risk pts with a RUNX1 mutation (40.5% [68/168]) (one pt had both TP53 and RUNX1). The median OS for adverse risk pts with a RUNX1 mutation and TP53 WT was 22.9 mo while the median OS for adverse risk pts with a TP53 mutation and RUNX1 WT was 5.42 mo. Conclusions: ELN 2017 risk groups, based on younger patients who received IC regimens, are not prognostic in patients receiving Ven+Aza. This exploratory analysis suggests little OS distinction between favorable and intermediate risk groups. Furthermore, adverse risk ELN groups can be further refined to determine mutationally-defined cohorts with more favorable outcomes. A modified risk stratification system is needed for pts receiving newer therapies, such as Ven+Aza, that uses determinants that are prognostic for pts treated in the context of these regimens with distinct mechanisms of actions from IC. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal