The Prognostic Impact of Myelodysplasia-Related Gene Mutations in Acute Myeloid Leukemia with FLT3-ITD Mutations

Background

FLT3-ITD mutations are found in approximately one quarter of patients with acute myeloid leukemia (AML). One major change in the 2022 ELN risk classification is the re-classification of all FLT3-ITD mutated patients as intermediate risk independently of allelic burden or NPM1 co-mutation. Recommendations state that FLT3-ITD mutations without concurrent NPM1 mutations but with adverse-risk genetic aberrations should be classified as adverse risk. However, the prognostic impact of concurrent mutations in the newly-defined subgroup of myelodysplasia-related genes (MRG) in FLT3-ITD mutated patients remains unclear.

Methods

We retrospectively analysed 4078 intensively-treated AML patients within the AML HARMONY Consortium database with molecular genetic data from diagnosis including the mutational status of all nine MRG mutations (ASXL1, BCOR, EZH2, RUNX1, SF3B1, SRSF2, STAG2, U2AF1, ZRSR2), NPM1 and FLT3. All cases were annotated and organised using the OMOP common data model. Outcome was assessed by multivariate analysis (MVA) using a Cox model with backward elimination considering mutations in at least one MRG, hemoglobin levels, white blood cell counts (WBC), platelets, bone marrow blast count, age ≥60 years, TP53 mutation, NPM1 mutation, FLT3-ITD to wild type (wt) allelic ratio and complex karyotype. For multivariate analysis, data was imputed using logistic regression for binary variables and predictive mean matching for continuous variables with 5 imputations by chained equations.

Results

The prognostic impact of MRG mutations was analysed in 862 patients (21%) with a detectable FLT3-ITD mutation at diagnosis, and in subgroups defined by FLT3-ITD to wt allelic ratio and NPM1 mutational status. Median age was 51 years. Median follow-up was 6.2 years. 171 of these patients (20%) had at least one additional MRG mutation. Only 13 patients were treated with FLT3-inhibitors during induction therapy. The most common MRG mutations were RUNX1 (n=77), SRSF2 (n=37) and STAG2 (n=32). Patients with MRG mutations were older (median age 57 vs. 49.9 years), more likely to be male (61% vs. 42%) and had lower WBC at diagnosis (median WBC 26.6 vs. 47.8*10⁹/L). In MVA MRG mutations were neither associated with relapse free survival (RFS) (HR 1.10, 95%CI 0.85–1.41, p=0.5) nor overall survival (OS) (HR 1.21, 95%CI 0.97-1.50, p=0.085).

Grouping the patients according to low (n=231) or high (n=525) FLT3-ITD allelic ratio, MRG mutations were neither predictive for RFS (low: HR 1.38, 95%CI 0.86–2.19, p=0.2; high: HR 0.80, 95%CI 0.55–1.15, p=0.2) nor OS (low: HR 1.13, 95%CI 0.73–1.75, p=0.6, high: HR 1.12, 95%CI 0.83–1.51, p=0.5).

Among patients with NPM1 co-mutation (n=491), only 46 (9%) had an additional MRG mutation. In the small subgroup of FLT3/NPM1 co-mutated patients MRG mutations had a similar prognostic impact on RFS (HR 0.68, 95%CI 0.41–1.12, p=0.13) and OS (HR 0.74, 95%CI 0.43–1.26, p=0.3) compared to the remaining patients by MVA. Interestingly, in FLT3-ITD mutated patients without NPM1 mutation (n=371) multivariate analysis revealed MRG mutations as an independent prognostic marker for RFS (HR 1.44, 95%CI 1.06–1.95, p=0.021) and OS (HR 1.43, 95%CI 1.08–1.89, p=0.013) besides a high FLT3-ITD to wt ratio, age and high WBC count.

Both ASXL1 and RUNX1 mutations have been associated with an adverse outcome before the introduction of MRG mutations in the ELN 2022 classification. The presence of either RUNX1 or ASXL1 mutations (n=93) but not MRG mutations other than ASXL1 and RUNX1 was an independent predictor of worse RFS (HR 1.60, 95%CI 1.20–2.14, p=0.001) and OS (HR 1.32, 95%CI 1.00–1.73, p=0.046) by MVA in the full FLT3-ITD mutated cohort besides age ≥60 years, high WBC count, NPM1 wildtype status and a high FLT3-ITD to wild type ratio. Similarly, in FLT3-ITD mutated patients without NPM1 mutation, RUNX1 and/or ASXL1 mutations were associated with a worse RFS (HR 1.59, 95%CI 1.14–2.20, p=0.006) and OS (HR 1.40, 95%CI 1.04–1.89, p=0.026), while MRG mutations other than RUNX1 and ASXL1 were an independent predictor of worse OS (HR 1.49, 95%CI 1.00–2.22, p=0.0496), but not RFS (HR 1.23, 95%CI 0.72–2.10, p=0.4).

Conclusion

In the ELN 2022 subgroup of FLT3-ITD mutated patients MRG mutations were not associated with RFS and OS. However, in the subgroup of FLT3-ITD/NPM1 wildtype patients, MRG mutations were an independent risk factor for shorter RFS and OS.