Nucleophosmin Gene Mutations Identify a Favorable Risk Group in Childhood Acute Myeloid Leukemia with a Normal Karyotype.

Exon 12 gene mutations in nucleophosmin (NPM1) were recently discovered in approximately 30% of adult acute myeloid leukemia (AML) samples, and cluster in the normal karyotype subgroup (NK-AML). NPM1-mutated adult NK-AML has a favorable outcome (pOS in the 40-50% range), but in case a FLT3 internal tandem duplication (FLT3/ITD) is also present outcome is worse with 25–30% pOS. In pediatric AML, NPM1 mutations are less frequent (6–8%; Cazzaniga, Blood 2005 & Brown, Blood 2007). No studies have specifically addressed pediatric NK-AML, a subgroup lacking favorable prognostic cytogenetic aberrations and therefore mostly stratified in the intermediate risk arm of pediatric AML treatment protocols. We screened 292 newly diagnosed AML samples, and detected NPM1 mutations in 25 cases (8.6%). We also screened 46 initial diagnosis-relapse pairs, and no clonal instability was observed, which suggests that NPM1 mutations may be used for minimal residual disease detection. In contrast to adults, where type A mutations (TCTG-insertion) are most frequent (80%), in our cohort type B (CATG-insertion) mutations were found in 39% and type A in 23%. In the NK-AML cohort (n=98), 20% was NPM1-mutated, which was age dependent: <3 years, 0%; 3–10 years, 19%; >10 years, 29% (p=0.04). None of the 10 FAB M5 cases was NPM1 mutated (p=0.09). NPM1 mutations had an independent favorable prognostic impact on outcome in patients with NK-AML (5-year pEFS 77% vs. 41% for wild type patients; p=0.003), irrespective of FLT3 mutational status. In fact, NPM1-mutated patients with a FLT3/ITD did better than patients without an ITD, although this was not statistically significant (5-year pEFS 90% vs. 63%, respectively; p=0.48). In NK-AML without NPM1 mutations, patients with FLT3/ITD positive AML did significantly worse than wild type FLT3 AML patients (5-year pEFS 18% vs. 52%, p=0.002). The differential prognostic impact of FLT3/ITD between the NPM1-mutated vs. the wild type patients was not caused by differences in the FLT3/ITD allelic ratio or ITD length, nor was there a relationship with the type of NPM1 mutations. Multivariate analysis, including age, white blood cell count, NPM1 and FLT3 status and stem cell transplantation as time-dependent co-variable, showed that only NPM1 mutations had independent prognostic significance for pEFS (RR 0.34, p=0.02). We conclude that the incidence of NPM1 mutations increases with age, and that NPM1 mutations define a subgroup with favorable prognosis in pediatric NK-AML. Our data suggest that these molecular abnormalities allow stratification of children with NK-AML. However, different from adult NK-AML, we observed that all children with NPM1 mutations did well, irrespective of FLT3 status. Therefore, treatment in the ‘good risk’ arm should be considered for children with NPM1-mutated NK-AML.