AML-MR Mutations Drive the Benefit of CPX-351 over 7+3 in the Pivotal Phase 3 AML Trial

Introduction

Gene mutations that define distinct biologic subsets of acute myeloid leukemia (AML) were integrated into the recently revised diagnostic (WHO) and prognostic (ELN 2022) models. Mutations in SRSF2, SF3B1, U2AF1, ZRSR2, ASXL1, EZH2, BCOR, or STAG2 now define the adverse risk disease subtype termed AML, myelodysplasia related (AML-MR) independent of clinical history of MDS; TP53 mutations are linked with poor overall outcomes and mediate the adverse prognosis of therapy-related AML (t-AML), and germline DDX41 mutations cause a common AML predisposition with favorable treatment outcomes. CPX-351 is approved for initial induction therapy in patients with secondary AML (s-AML) defined by cytogenetics, morphology, or clinical history and t-AML based on a phase 3 trial that showed improved overall survival (OS) compared with daunorubicin and cytarabine (7+3). However, the role of CPX-351 in the AML treatment landscape is unclear as it is not known how the results of the pivotal phase 3 trial apply to patients grouped according to current classifications. To address this gap, we analyzed outcomes according to treatment arm and current AML classification at diagnosis.

Methods

We performed targeted mutational analysis on all available pre-treatment samples (184 of 309, 60%) from patients aged 60-75 years with newly diagnosed s-AML or t-AML who were randomized to CPX-351 (N=93) or 7+3 (N=91) in the phase 3 CPX-351-301 trial (NCT01696084). We defined four molecular groups in hierarchical order based on the presence of a TP53 mutation (N=62, 34%), a germline DDX41 mutation without concurrent TP53 mutation (N=10, 5%), at least 1 AML-MR defining mutation (N=88, 48%), or all remaining patients (de novo; N=24, 13%). We investigated the association between these genetic groups and (1) efficacy outcomes derived from the 5-year follow-up analysis (OS; complete remission [CR] or CR with incomplete count recovery [CRi]) and (2) myelotoxicity outcomes (reflected by time to absolute neutrophil count [ANC] > 1 x109/L and time to platelet recovery > 100 x109/L).

Results

In the overall cohort, median OS (95% CI) and CR/CRi rates were significantly different across molecular groups (p<0.001): TP53 5.0 (3.3-7.3) months and 35%; DDX41: 50.1 (17.1-60.7) months and 100%; AML-MR 8.0 (5.7-10.6) months and 38%; de novo 12.4 (5.7-29.3) months and 63%. The time to platelet recovery (median [IQR]) was longer in the AML-MR group (49 [38-73] days) than in those in the TP53 (35 [29-49] days, p=0.022), DDX41 (37.5 [31-43] days, p=0.047) and de novo (36 [31-42] days, p=0.024) groups. Time to neutrophil recovery was similar across groups.

To determine the effect of CPX-351 compared with 7+3 in each molecular group, we analyzed outcomes by treatment arm separately in each group. The median (95% CI) and 2-year OS were significantly better with CPX-351 than with 7+3 only in the AML-MR group (9.7 [6.2-13.7] vs 6.8 [3.6-9.6] months, 2yr 27% vs 8%, p=0.037). OS was similar with CPX-351 and 7+3 in the TP53 (4.5 [2.9-7.6] vs 5.1 [2.9-7.3] months, 2yr 4% vs 8%, p=0.70), and de novo (median 11.3 [5.6-NA] vs 16.9 [4.0-NA] months, 2yr 38% vs 36%, p=0.77) groups. In patients with a DDX41 mutation, median OS was 56.4 [36.6-NA] vs 17.1 [4.6-NA] months, with 2 year OS 100% vs 40%, p=0.11. Transplantation and CR/CRi rates were similar between treatment arms in each group.

Treatment with CPX-351 was associated with prolonged platelet recovery time (median [IQR]) in the de novo (41 [36-56] vs 30.5 [29-32] days, p=0.018) and DDX41 groups (43 [41-49] vs 30 [29-34] days, p=0.056). Platelet recovery time by treatment was comparable in the AML-MR (52.5 [42-77] vs 44 [34-50] days, p=0.2) and TP53 (39.5 [34-47] vs 33 [28-49] days, p=0.3) groups. There was no treatment-specific difference in time to ANC recovery within any molecular group.

Conclusion

In a post hoc analysis of the pivotal phase 3 study of CPX-351 vs 7+3, patients with TP53 or AML-MR defining mutations had poor overall outcomes. CPX-351 improved survival without pronounced myelotoxicity in patients with AML-MR mutations, had no benefit over 7+3 in those with TP53 mutations and may add myelotoxicity without survival benefit in those with de novo mutations. In this high-risk AML cohort, 5% had a germline DDX41 mutation with 100% CR and prolonged survival. Our results indicate that the benefit of CPX-351 over 7+3 is driven by the presence of AML-MR defining mutations.