Evaluation of Minimal Residual Disease Based on NPM1 Mutations in AML with Intermediate Risk Cytogenetics: A Prospective Study of 36 Patients.

Mutations in exon 12 of the nucleophosmin (NPM1) gene occur in approximately 50% of adult acute myeloid leukemia (AML) with normal karyotype. More than 40 mutant variants have been identified. Most of these mutations consist of a 4-bp insertion, which can be used as a target for minimal residual disease (MRD) monitoring. We previously checked the stability of NPM1 mutations at relapse in 21 NPM1-mutated patients at initial diagnosis. In this prospective study, we evaluated MRD by real-time quantitative PCR (RQ-PCR) in 36 NPM1-mutated AML patients corresponding to 33 adult and 3 pediatric cases, treated according to the French ALFA9801 or ALFA9802 and ELAM02 protocols, respectively. Out of these patients, 31/34 (91%) had normal karyotype, 13/33 (39%) had a high initial white blood cell count, and 10/36 (28%) were FLT3-Intern Tandem Duplication (FLT3-ITD) positive. 28 (78%) patients carry NPM mutation A, 3 (8%) mutation B and 5 (14%) other rare variants. RQ-PCR assays using a mutation-specific primer were performed on cDNA for mutation A and B and on genomic DNA for other NPM1 mutants. In our experiments, the maximal reproductible sensitivity of NPM1-based MRD detection is about 10⁻⁴ on genomic DNA and 10⁻⁵ to 10⁻⁶ on cDNA. The median follow-up was 260 days [40–791]. 2 to 9 follow-up samples from bone marrow and/or peripheral blood were analysed per patient. No correlation was found between leukocytosis at diagnosis and initial expression ratio of NPM1 mutation. The study of MRD log reduction after induction therapy shows that molecular responses are very heterogeneous (from 4.10⁻² to more than 1.10⁻⁵), but 50% of patients reach at least a 4 log reduction in NPM1 levels. Patients with FLT3-ITD tend to have lower log reduction after induction than patients without FLT3-ITD, although not statistically significant (P=0.07). The analysis of NPM1-MRD in bone marrow and in peripheral blood at the same follow-up time-points shows that NPM1 levels can be until 1 log higher in bone marrow. This indicates that the evaluation of NPM1-MRD in bone marrow is more informative than in peripheral blood. We found all relapses had NPM1-MRD levels comparable to those observed at diagnosis. Among the 5 patients who relapsed so far, 2 were predictable by increasing MRD levels 1 to 4 months before hematological relapse. In 29 out of 36 cases, we could monitor MRD by both NPM1 mutation and WT1 gene expression. The comparison of the MRD profiles obtained by these two approaches reveals some discordant results, which can be, at least in part, explained by difference in the sensitivity of the RQ-PCR techniques, since the sensitivity of WT1 expression as MRD target is generally not higher than 10⁻³. In conclusion, NPM1 mutations are very specific and sensitive markers for MRD monitoring in AML. Further studies are required to determine if NPM1-MRD provides an independent prognostic factor and may be useful for therapeutic stratification in AML patients with intermediate risk cytogenetics.