Heterogeneity Of Patients With Cytogenetically Normal Acute Myeloid Leukemia (CN-AML) and Nucleophosmin-1 (NPM1) Gene Mutation: An Acute Leukemia French Association (ALFA) Study

CN-AML with NPM1 mutation accounts for 25% of AML cases and is individualized as a provisional entity in the WHO classification. In this large AML subset, additional gene mutations could influence patient outcome, like it has been shown for bad-prognosis FLT3 internal tandem duplication (FLT3-ITD). We thus reviewed the characteristics and outcome of 409 patients with de novo NPM1-mutated CN-AML treated in sequential ALFA studies.

Patients were treated between 1990 and 2012 (median follow-up, 31 months) according to ALFA-9000 (#16), 9801 (#91), 9802 (#89), 9803 (#19) and 0701 (#84) published protocols or in the ongoing 0702 study (#110). In addition to NPM1, the following mutations were centrally assessed using standard-PCR methods and Sanger sequencing: FLT3-ITD, mutations of FLT3 tyrosine kinase domain (FLT3-TKD, i.e. FLT3D835/I836), IDH1R132, IDH2R140, IDH2R172, TET2 (exons 3-11), and DNMT3A (exons 8-9, 11-23). Primary endpoint was cumulative incidence of relapse (CIR) after censoring the 73 patients who received allogeneic stem cell transplantation (SCT) in first CR at SCT time. Analyses were performed with multiple imputations of missing values, followed by multivariable logistic regression or Cox model.

Median age was 52 years (17-79). Median white blood cell count (WBC) was 23.5 x 10⁹/L (0.8-377). Gene mutations are indicated in Table 1. Multivariable analyses showed that: 1) FLT3-ITD was associated with higher median WBC (60 versus 18 x 10⁹/L, P<0.001) and less frequent FLT3-TKD, IDH1R132 and IDH2R140 mutations (P=0.003, 0.037 and <0.001, respectively); 2) IDH2R140 mutations were associated with higher median WBC (42 versus 24 x 10⁹/L, P<0.001) and less frequent IDH1R132 and DNMT3AR882 mutations (P<0.001 and 0.001, respectively); 3) TET2 mutations were associated with higher median WBC (37 versus 22 x 10⁹/L, P=0.003) and less frequent FLT3-ITD, FLT3-TKD, and IDH1R132 mutations (P=0.007, 0.028 and 0.022, respectively); and 4) non-R882 DNMT3A mutations, but not R882 mutations, were markedly associated with advanced age (22% in patients ≥ 50 years versus 9% in younger patients; P<0.001). Overall, 372 patients (91%) reached complete remission (CR). In multivariable models, advanced age and higher WBC negatively influenced CR achievement (P=0.044 and 0.037, respectively), but none of the additional mutations. In CR patients, 3-year CIR was 51% (45-58). In a first multivariable model not considering TET2 mutational status (analyzed in 232 patients only), advanced age was associated with a higher specific hazard of relapse (SHR) (P<0.001). At 3 years, CIR was 63% in patients ≥ 50 years versus 35% in younger patients. The other factors associated with a higher CIR were FLT3-ITD (SHR, 2.04 [1.35-3.09]; P=0.001) and non-R882 DNMT3A mutations (SHR, 1.94 [1.10-3.42]; P=0.023), while FLT3-TKD was associated with a lower CIR (SHR, 0.38 [0.16-0.94]; P=0.036). Conversely, WBC, IDH1/2 mutations and R882 DNMT3A mutations did not influence CIR. Advanced age, FLT3-ITD and non-R882 DNMT3A mutations were also identified as independent risk factors in a second multivariable model performed in the subset of patients tested for TET2 mutations. Here, TET2 mutations had no impact on CIR. Overall 3-year CIR estimates were 41%, 61%, 68% and 94% in the FLT3-ITD-/non-R882 DNMT3A-, FLT3-ITD+/non-R882 DNMT3A-, FLT3-ITD-/non-R882 DNMT3A+, and FLT3-ITD+/non-R882 DNMT3A+ subsets, respectively.

This study confirms the heterogeneity of patients with NPM1-mutated CN-AML, independently of the known poor prognosis of FLT3-ITD. Higher WBC was associated with FLT3-ITD, IDH2R140 and TET2 mutations, but did not represent per se an independent risk factor for relapse. Higher incidence of unfavorable non-R882 DNMT3A mutations in older patients may explain, at least in part, why older patients do worse than younger patients in this peculiar AML subset. A potential favorable prognostic role of FLT3-TKD mutations in these patients needs to be confirmed.

No relevant conflicts of interest to declare.