The Mutational Landscape of Acute Myeloid Leukaemia Predicts Responses and Outcomes in Elderly Patients from the PETHEMA-FLUGAZA Phase 3 Clinical Trial

Abstract

We sought to predict treatment responses and outcomes in older patients with newly diagnosed acute myeloid leukemia (AML) from our FLUGAZA phase III clinical trial (PETHEMA group) based on mutational status, comparing azacytidine (AZA) with fludarabine plus low-dose cytarabine (FLUGA). Mutational profiling using a custom 43-gene next-generation sequencing panel revealed differences in profiles between older and younger patients, and several prognostic markers that were useful in young patients were ineffective in older patients. We examined the associations between variables and overall responses at the end of the third cycle. Patients with mutated DNMT3A or EZH2 were shown to benefit from azacytidine in the treatment-adjusted subgroup analysis. An analysis of the associations with tumor burden using variant allele frequency (VAF) quantification showed that a higher overall response was associated with an increase in TET2 VAF (odds ratio (OR), 1.014; p = 0.030) and lower TP53 VAF (OR, 0.981; p = 0.003). In the treatment-adjusted multivariate survival analyses, only the NRAS (hazard ratio (HR), 1.9, p = 0.005) and TP53 (HR, 2.6, p = 9.8 × 10-7) variants were associated with shorter overall survival (OS), whereas only mutated BCOR (HR, 3.6, p = 0.0003) was associated with a shorter relapse-free survival (RFS). Subgroup analyses of OS according to biological and genomic characteristics showed that patients with low-intermediate cytogenetic risk (HR, 1.51, p = 0.045) and mutated NRAS (HR, 3.66, p = 0.047) benefited from azacytidine therapy. In the subgroup analyses, patients with mutated TP53 (HR, 4.71, p = 0.009) showed a better RFS in the azacytidine arm. In conclusion, differential mutational profiling might anticipate the outcomes of first-line treatment choices (AZA or FLUGA) in older patients with AML. The study is registered at ClinicalTrials.gov as NCT02319135.

Keywords: NGS; acute; azacytidine; clinical trials and observations; complete remission; cytarabine; genetic risk; leukemia; leukemic cells; myelocytic; myeloid neoplasia; older adults; prognostic factors; variants.

Conflict of interest statement

M.T. declares honoraria for lectures from Celgene, Pfizer, Novartis, Janssen, Merck Sharp. Dohme (MSD), Daiichi, and Servier SL, and membership on advisory boards with Celgene, Novartis, Roche, and Astellas. J.S. declares honoraria for lectures, and membership on advisory boards with Daiichi Sankyo, Pfizer, Celgene, Novartis, Roche, and Amgen. F.R. declares travel grants from Celgene, Novartis, Amgen, AbbVie, Janssen, Roche, MSD, and Daiichi Sankyo; consulting fees from Celgene, Novartis, Amgen, and AbbVie; and advisory board membership with, as well as research grants from, Celgene. E.L. declares honoraria for lectures from Janssen and Novartis, and advisory board membership with Janssen, Celgene, Astellas, and Amgen. M.B.V. declares honoraria for lectures from, and membership on advisory boards with, Janssen, BMS, Novartis, Roche, Astellas Pharma, and Jazz Pharmaceuticals. C.G. declares honoraria for lectures from Celgene, Amgen, Janssen, and Pfizer, and advisory board membership with Celgene. B.P. declares honoraria for lectures from, and membership on advisory boards with, Amgen, Bristol Myers Squibb (BMS), Celgene, Janssen, Merck, Novartis, Roche, and Sanofi; unrestricted grants from Celgene, EngMab, Sanofi, and Takeda; and consultancy for Celgene, Janssen, Sanofi, and Takeda. M.A.S. declares a consulting or advisory role for Teva, Daiichi Sankyo, Orsenix, AbbVie, Novartis, and Pfizer. J.M.-L.. declares honoraria for lectures from, and membership on advisory boards with, Janssen, BMS, Sanofi, Novartis, Incyte, Roche, and Amgen; and membership on the boards of directors of Hosea and Altum Sequencing. P.M. declares advisory board and speaker’s bureau membership with, as well as research support from, AbbVie, Janssen, Novartis, Pfizer, and Teva; research support, being a consultant for, and speaker’s bureau and advisory board membership with Astellas, Celgene, and Daiichi Sankyo; being a consultant for Agios, Tolero Pharmaceutical, Glycomimetics, and Forma Therapeutics; speaker’s bureau and advisory board membership with Incyte; and research support from, and advisory board membership with, Karyopharm. The remaining authors declare no competing financial interests.

Figures

Figure 1

Subgroup analysis of responders to treatment via biological and genomic characteristics. OR: odds ratio, cytogenetic risk: low–intermediate vs. high risk as per ELN 2017 classification; high risk score was defined by the presence of mutated NRAS or TP53. A score predicting an AZA response was defined by the presence of mutated EZH2, U2AF1, DNMT3A, or TET2 genes. Patients with baseline mutations in DNMT3A (odds ratio (OR) 0.20, p = 0.023) or a score predicting AZA response (OR 0.448, p = 0.046) could benefit from azacytidine.

Figure 2

Mutation status has prognostic significance for overall survival and relapse-free survival. In the global series, the AZA arm’s and FLUGA arm’s overall survival and disease free-survival are represented by Kaplan–Meier plots. Wild-type status is indicated in blue and mutated status is indicated in red. NRAS mutations (a) and TP53 mutations (b) are adverse factors affecting overall survival in the FLUGA arm. BCOR mutations are adverse factors affecting relapse-free survival in the AZA and FLUGA arms (c). AZA: azacytidine, FLUGA: fludarabine plus low-dose cytarabine (LDAC).

Figure 3

Subgroup analysis of overall survival and progression-free survival via biological and genomic characteristics. HR: hazard ratio, cytogenetic risk: low–intermediate vs. high risk as per ELN 2017 classification; a high-risk score was defined by the presence of mutated NRAS or TP53. (a) in the subgroup analyses of overall survival via biological and genomic characteristics, we observed that patients with low–intermediate cytogenetic risk (hazard ratio (HR0 1.51, p = 0.045) and mutated NRAS (HR 3.66, p = 0.047) could benefit from azacytidine. (b) in the subgroup analyses of relapse-free survival, we observed that patients with mutated TP53 (HR 4.71, p = 0.009) and scores indicating a high risk for AML (HR 2.69, p = 0.013) showed a higher chance of relapse-free survival under the AZA arm.

Figure 4

Kaplan–Meier overall survival curves in acute myeloid leukemia (AML) patients classified by the presence or absence of high molecular risk score (HMR). HMR pattern defined by presence of NRAS or TP53 mutations which is associated with unfavorable outcomes and shorter survival after azacytidine (left) or FLUGA (right) schemes. HMR absent is indicated in blue and HMR present is indicated in red. Number of censored patients with respect to the stratified groups and the number at risk is indicated. p values are considered significant (p < 0.05). OS: overall survival.