DNMT3A, NPM1, and FLT3-ITD triple-mutated AML has a favorable prognosis in the era of FLT3 inhibitors

Abstract

Abstract Background: FLT3-ITD mutations are common in AML, occur in a diverse co-mutation landscape, and are generally associated with a poor prognosis. DNMT3A and NPM1 are commonly co-mutated with FLT3. Historically, DNMT3A, NPM1, and FLT3-ITD triple-mutated AML (TM-AML) has been reported to have a higher risk of relapse and worse overall survival in response to treatment with chemotherapy alone. However, recent approval and widespread use of multiple FLT3 inhibitors have significantly improved disease outlook. We evaluated the outcomes of DMNT3A/NPM1/FLT3-ITD triple-mutated AML from a cohort of newly diagnosed AML patients treated with modern therapeutic approaches, including intensive chemotherapy, azacitidine and venetoclax, allogeneic transplant, and FLT3 inhibition, guided by a high-sensitivity FLT3-ITD MRD assay. Methods: The study cohort consists of 431 adult patients with newly diagnosed AML who were admitted and treated in consecutive fashion on the leukemia service at the Johns Hopkins Hospital between January 2020-July 2024. Targeted sequencing of a panel of 60 leukemia-associated genes was carried out using genomic DNA derived from peripheral blood or bone marrow specimens collected at diagnosis. Clinical records were reviewed to determine patients' baseline characteristics, MRD status (using the Invivoscribe FLT3-ITD MRD assay), FLT3 inhibitor use, event free survival, and overall survival. Statistical testing, survival analysis, and Cox regression were performed using R. Results: FLT3 mutations were found in 101/431 (23.4%), 72 with FLT3-ITD (17%), 25 (6%) with FLT3-TKD, and 8 (2%) with non-canonical mutations. Of the 72 FLT3-ITD patients, 17 patients had DNMT3A, NPM1, and FLT3-ITD triple mutations (24% of FLT3-ITD; 4% of all AML patients) and 10/17 (59%) of these received allogeneic transplant. The median white count for the TM subgroup at diagnosis was 78, and the median age was 61 with a wide range (27-85). Karyotype was intermediate risk in 100% of TM subgroup (82% diploid). 15/17 (88%) TM-AML patients received FLT3 inhibitors during their treatment course. Of the 2 patients who did not receive FLT3 inhibitors, both were over the age of 70 and were induced with azacitidine and venetoclax, and one transitioned to hospice during induction. Kaplan-Meier survival analysis revealed patients with TM-AML (n=17) had significantly better overall survival (OS) and event-free survival (EFS) compared to non-TM FLT3-ITD AML (n= 55) (2-year OS: 82% vs. 50%; 2-year EFS: 84% vs. 43%; log-rank, p < 0.05). Patients with TM-AML had a comparable overall survival to patients with APL (n=25) and CBF AML (n=24) treated during the same time period (log-rank, p = 0.94). Subgroup analysis of FLT3-ITD AML demonstrated TM-AML had the longest OS and EFS compared to any other combination of those 3 mutations (log-rank, p < 0.01).Multivariate Cox regression confirmed a significant 3-fold ([1.03-8.41], p<0.05) increased hazard ratio with DNMT3A alone and yet paradoxically improved survival with both DNMT3A and NPM1 (HR 0.08 [0.01-0.55], p < 0.05). FLT3 inhibitor use is additionally associated with decreased mortality (HR 0.35 [0.13-0.97], p < 0.05). We also evaluated FLT3-ITD MRD negativity rates following 2 cycles of treatment among FLT3-ITD patients who received a FLT3 inhibitor during treatment. We found higher MRD negative rates in 8/10 (80%) TM-AML patients vs. 10/25 (40%) non-TM-AML while accounting for FLT3 inhibitor exposure, suggesting that TM-AML is more responsive to FLT3 inhibition (Fisher's exact, p < 0.05). In the non-TM-AML subgroup, 24/55 (44%) harbored MDS-defining molecular abnormalities. The median FLT3-ITD VAF was 27% in the TM-AML group versus 10% in the non-TM-AML group arising from antecedent MDS (Wilcoxon rank-sum, p < 0.05), highlighting dominance of the FLT3-ITD clone at diagnosis in TM-AML versus a sub-clonal event in myelodysplasia-related AML. Conclusions: FLT3 mutations by themselves do not define an AML subtype, but in this new era of FLT3 inhibitors, DNMT3A/NPM1/FLT3-ITD triple-mutated AML appears to represent a unique subgroup of AML with especially favorable outcomes and a prognosis similar to CBF and APL. This subgroup may be uniquely dependent on FLT3 signaling and therefore distinctly responsive to FLT3 inhibition. High sensitivity MRD assays can be used to guide therapeutic decisions for this subset, which may contribute to the overall favorable outcome.