Intensive Induction Chemotherapy (IC) Vs Hypomethylating Agents + Venetoclax (HMA/VEN) in IDH1- or IDH2-Mutant Newly Diagnosed Acute Myeloid Leukemia (AML) - a Multicenter Cohort Study

Introduction:

Intensive induction chemotherapy (IC) remains the standard of care for younger and fit patients (pts) with acute myeloid leukemia (AML). Post-hoc analyses from clinical trials have associated IDH1 and IDH2 mutations with high rates of durable responses to venetoclax (VEN)-based therapies. However, these trials were conducted in pts ≥75 years of age or unfit for IC. As there are currently no trials comparing hypomethylating agents (HMA)+VEN and IC directly in IDH-mutant AML, we conducted a large multicenter retrospective cohort study of pts with newly diagnosed AML with IDH1 and IDH2 mutations to assess the impact of clinical and molecular characteristics on response and survival to IC versus HMA+VEN.

Methods:

We included pts with newly diagnosed IDH1- and IDH2-mutant AML who were treated at 4 large academic centers with IC (conventional 7+3 or liposomal cytarabine/daunorubicin) or HMA/VEN. Detection of IDH1 and IDH2 mutations was performed by next-generation sequencing per local standards at the participating sites. Responses were recorded according to European LeukemiaNet 2017 criteria. Composite complete response (cCR) was defined as complete response (CR) + CRi (CR with incomplete count recovery). Overall survival (OS) was compared between groups by log-rank test. Cox regression for OS was fitted to evaluate the effect of treatment, allogeneic stem cell transplantation (allo-SCT) as a time-varying covariate and additional covariates. All variables with p<0.1 in the univariate model were included in the multivariable stepwise backward selection model.

Results:

We reviewed 1132 pts with AML and included 98 and 131 pts with IDH1/2-mutant AML, respectively (9 pts had IDH1 and IDH2 co-mutations). As pts with IDH1- and IDH2-mutant AML had similar baseline demographic and disease characteristics, we combined both groups for further analyses.

Of the 220 pts with IDH1/2-mutant AML, 150 and 70 received IC and HMA+VEN, respectively (Table 1). Twenty pts in the IC group received concurrent treatment with the IDH1 inhibitor ivosidenib. Compared to pts treated with IC, pts who received HMA+VEN were older (median age 75 years with HMA+Ven [Range: 60 – 88 years] vs 63 years [R: 24 -76] with IC; p<0.001) and had higher rates of molecular defined secondary AML (57% vs 41%; p<0.001), CEBPA mutations (10.0% vs 2.7%; p=0.040), and TP53 mutations (13% vs 2.7%; p=0.005)

Despite these differences in baseline characteristics, rates of cCR were similar for IC and HMA+VEN-treated pts (68% vs 67%; p=0.90) with higher rates of CR seen with IC (64% vs 46%; p=0.010). The rate of allo-SCT was higher in pts treated with IC than with HMA+VEN (63% vs 23%; p<0.001).

The median OS for the entire cohort was 32 months (95% CI: 24 months [mo] – not reached [NR]) with a significant difference in OS by treatment type favoring IC over HMA+VEN (52 mo [95% CI: 30 mo – NR] vs 14 mo [95% CI: 11 – 39 mo]; p=0.002). However, in pts > 60 years (n=168 pts), median OS was 21 mo (95% CI: 16 - 32 mo) and not statistically significant (p=0.11) between IC (27 mo [95% CI: 20-57 mo]) and HMA+VEN (15 mo [95% CI: 11 mo – NR]). Comparable OS was seen among pts undergoing allo-SCT (n=111 pts; 24-mo OS 75% [95% CI: 66 - 84%] with IC vs 24-mo OS 79% [95% CI: 61 - 100% with HMA+VEN, p=0.71), as well as when censored at allo-SCT (IC 12 mo [95% CI: 7.7 – 25 mo]; HMA+VEN 12 mo [95% CI: 8.2 – 18 mo], p=0.58)

Although treatment with HMA+VEN vs IC was found to be associated with worse OS in univariable analysis (Table 2), it did not retain its prognostic value in the multivariable analysis (HR: 1.28; 95% CI: 0.74 – 2.19; p=0.37). Conversely, prior HMA, monosomal karyotype, JAK2 and FLT3-ITD mutations were associated with worse survival suggesting that these factors are more relevant for OS than the type of treatment.

Conclusion:

In pts with newly diagnosed IDH1/2-mutant AML, cCR rates were similar for pts treated with IC and HMA+VEN. While there was a difference in OS favoring treatment with IC over HMA+VEN in the unselected patient cohort, this difference was no longer present in subgroup analyses based on age >60 years and in pts who received an allo-SCT. Furthermore, treatment type did not impact OS in multivariable regression models suggesting that differences in OS between treatment groups are more likely to be explained by differences in patient and disease characteristics rather than treatment type. Due to the limitations of retrospective, real-world studies, a randomized clinical trial is needed.