NPM1 mutant variant allele frequency correlates with leukemia burden but does not provide prognostic information in NPM1‐mutated acute myeloid leukemia

A mutation in nucleophosmin 1 (NPM1) occurs in approximately 30% of patients with acute myeloid leukemia (AML) and constitutes a distinct category with a more favorable clinical outcome.1 FLT3-internal tandem duplication (ITD) and/or DNMT3A mutations commonly co-occur in NPM1-mutated AML and confer a worse prognosis.2 Several studies have shown that higher FLT3-ITD allelic ratio is correlated with worse survival.3 One single-center retrospective study suggested that higher variant allelic frequency (VAF) of mutant NPM1 was associated with worse overall survival (OS) and event-free survival (EFS) in newly diagnosed AML, particularly in patients with a DNMT3A mutation or who underwent hematopoietic stem cell transplantation (HSCT) in first remission.4 In light of these data, we sought to determine the impact of mutant NPM1 VAF on outcomes in patients with newly diagnosed AML treated at our institution with idarubicin and cytarabine (IA)-based regimens. We identified 147 patients with NPM1-mutated AML who received frontline therapy with an IA-based regimen,5 with or without addition of a nucleoside analogue, at The University of Texas MD Anderson Cancer Center between 2008 and 2018 and were evaluable for response. Baseline characteristics are shown in Supporting Information Table 1. NPM1 VAF was measured by capillary electrophoresis in 63 patients (43%) and by next-generation sequencing in 84 patients (57%). The median mutant NPM1 VAF was 43% (range, 4%-62%; interquartile range, 37%-48%). Seventy-eight percent (54/69) of patients with FLT3-ITD mutation received FLT3 inhibitors (92% with sorafenib). Twenty-two patients (15%) received an investigational agent in addition to the IA-based regimen. A total of 142 (96.5%) patients achieved complete remission or complete remission with incomplete platelet recovery. Fifty-nine patients (43%) underwent HSCT in first remission. With a median follow-up of 26.8 months, the median OS and EFS for the entire cohort were 69.0 months and 41.1 months, respectively. Patients with FLT3-ITD mutations had significantly higher mutant NPM1 VAF compared to FLT3-ITD wild type patients (mean ± SE: 42.7% ± 1.1% vs 39.1% ± 1.2%, P = 0.03). There were also significant positive correlations between NPM1 VAF and white blood cell count (ρ = 0.26, P = 0.001), peripheral blood blast percentage (ρ = 0.36, P <0.001), and bone marrow blast percentage (ρ = 0.2, P = 0.012), and a negative correlation with platelet count (ρ = −0.16, P = 0.04). However, there was no difference in mutant NPM1 VAF with respect to other characteristics, including gender, age, DNMT3A mutation, cytogenetics, method of VAF assessment, or whether or not HSCT was performed in first remission. To assess the impact of mutant NPM1 VAF on outcomes, we then conducted survival analyses for OS and EFS across different NPM1 VAF levels (divided by either quartiles or the median). There was no cutoff across different quantiles that predicted OS or EFS. Since the median VAF in our population (43%) was very similar to the previous cutoff reported by others to predict survival (44%),4 we used a cutoff of 43% to correlate NPM1 VAF with outcomes. For the entire population, OS and EFS were not different between patients with VAF <43% and ≥43% (Figure 1A,B). We also did not find a statistically significant difference in OS and EFS for subgroup analyses by gender, age (<60 years vs ≥60 years), cytogenetics, DNMT3A or FLT3-ITD mutation status, and whether HSCT was performed in first remission, when patients were stratified by mutant NPM1 VAF. Among the 54 patients with a FLT3-ITD who received a FLT3 inhibitor, patients with NPM1 VAF ≥43% had shorter median OS (29 months) compared to those with VAF <43% (42 months) but it was not statistically significant (P = 0.24). Among patients who underwent HSCT in first remission, there was no difference outcomes based on mutant NPM1 VAF. We then performed Cox proportional-hazards model univariate analysis for predictors of survival in NPM1-mutated AML. Age less than 60 years (P = 0.03) and higher hemoglobin levels (P < 0.01) were associated with better OS. There was no significant correlation with respect to gender, HSCT, FLT3-ITD mutation, FLT3 inhibitor use, DNMT3A mutation, white blood cell or platelet counts, or blast percentage. In multivariate analysis, age and hemoglobin maintained significance with respect to OS (hazard ratio [HR], 0.40; confidence interval [CI], 0.21-0.82, P = 0.007; and HR, 0.67; CI, 0.52-0.86; P = 0.001). Our data demonstrate no significant correlations between NPM1 VAF and either OS or EFS in the entire cohort of patients nor any subgroup. These findings are consistent with a report by the German AML Cooperative Group that also found no independent association between NPM1 VAF and outcomes.6 Overall, the range of mutant NPM1 VAF in our study was relatively small (interquartile range, 37%-48%); furthermore, the variation in VAFs was mostly related to peripheral white blood cell count and the percentage of blasts in the periphery and bone marrow. Together, these findings suggest that differences in NPM1 VAF are due to slight differences in disease burden but do not represent a different disease biology that influences long-term outcomes. This is in contrast to FLT3-ITD-mutated AML in which there is a wide range of FLT3-ITD allelic ratios that represent the relative role of the FLT3-ITD mutation in driving disease phenotype.3 These findings are contrary to a previous report that found that higher mutant NPM1 VAF was associated with worse survival.4 The differences in these two studies could be attributed to several treatment-related factors. We used IA-based regimens with higher doses of cytarabine during induction phases, whereas the patients in the study by Patel et al. received the "7 + 3" regimen with lower doses of cytarabine during induction but higher dose cytarabine during consolidation. Second, our HSCT rate was lower (43% vs 57%), which may have altered outcomes, although notably we did not detect an impact of NPM1 VAF on outcomes when according to HSCT status. Among FLT3-mutated patients, the rates of FLT3 inhibitor use was higher in our study (78% vs 57%), which may have also led to differential outcomes and abrogated the prognostic impact of NPM1 VAF. However, ultimately, because there is no clear biological mechanism to believe higher NPM1 VAF would be associated with worse outcomes, our findings strongly suggest that mutant NPM1 VAF should not be used in prognostication and clinical decision-making. The authors have no conflict of interest to report. Supplementary Figure 1 NPM1 VAF distribution across all patients Supplementary Table 1. Demographic and clinical characteristics of the entire population (N = 147) Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.