Clonal dynamics of FLT3-ITD positive acute myeloid leukemia following relapse after allogeneic stem cell transplantation

Abstract

Abstract Introduction: Allogeneic stem cell transplantation (alloSCT) offers a potential cure for patients with FLT3-internal tandem duplication (ITD)-mutated acute myeloid leukemia (AML), particularly when performed in first complete remission (CR1). However, relapse occurs in up to 30% of patients post-SCT. The clonal dynamics of FLT3-ITD and co- mutations in post-transplant relapse remain poorly characterized. We investigated the impact of maintenance therapy and the molecular patterns of relapse post alloSCT in FLT3-ITDAML Materials and methods: We retrospectively analyzed the cytomolecular landscape of paired pre- and post-transplant relapsed AML in patients with FLT3-ITD AML, who underwent first alloSCT between 2017 and 2023. The length of ITD was used to compare FLT3 ITD clones in pre- and post alloSCT samples.We assessed the length of ITD by capillary electrophoresis. Results: Among 153 patients, 67 relapsed, and 40 had paired samples available for analysis. The median age at transplant was 48 years (range, 19-72). According to ELN 2017 risk classification, 18 patients (41.9%) had adverse, 13 (30.2%) intermediate and 12 (27.9%) favorable risk disease. Only 12 underwent alloSCT in CR1; the remainder were in CR2 (n=1), marrow CR (n=19) or had active disease (n=8) at transplant. Among the 40 patients with evaluable paired samples, 24 (70%) received post-transplant maintenance therapy: 22 with FLT3 inhibitors (FLT3i) alone, 1 with hypomethylating agent (HMA) monotherapy, 3 with HMA plus FLT3i, and 2 with venetoclax plus FLT3i. Notably, 33 of the 40 relapses (82.5%) occurred within the first year post-transplant, and 24 (60%) within the first 6 months. These findings may also be influenced by disease status at the time of alloSCT, as a substantial proportion of patients proceeded in marrow CR or active disease. Median overall survival following relapse was 6.6 months (IQR, 2.6-11.9 months) with estimated one-year post alloSCT survival of 31.1% (95% CI, 17.5%-45.8%). 40% (16/40) of patients had >1 FLT3-ITD at diagnosis (median VAF 22%, range 1-80), compared to only 12.5% (5/40) at relapse. In 27 patients (61.9%), at least one FLT3-ITD clone persisted at relapse. FLT3-ITDbecame undetectable at relapse in 9 of 40 (22.5%) patients, nearly all of whom (but one) had received FLT3i maintenance. Four patients exhibited a new ITD or changes in ITD insertion site/length between diagnosis and relapse. Four patients had alternate FLT3 clones (D385Y, F691) emerge at relapse, albeit at very low VAFs. We observed persistence in mutations related to epigenetic modifications at relapse, including DNTM3A (retained in 20/22 patients), TET2 (8/8) and IDH2 (4/4). All 21 patients with NPM1 mutations retained these at relapse. Conversely, mutations in other functional gene groups showed more variability between diagnosis and relapse. The most frequent emergent mutations at post-transplant relapse involved tumor suppressors (WT1, TP53; n=4, 10%), followed by transcription factors (GATA2, ETV6; n=3; 7.5%), spliceosome genes (U2AF1, SF3B1, ZRSR2; n=3, 7.5%), chromatin modifiers (ASXL1, ASXL2, PH6; n=3, 7.5%) and the cohesin complex (RAD21, STAG2, n=2; 5%). Interestingly, RAS/MAPK pathway alterations were rare, with a single event in KRAS, unlike the frequent RAS-driven escape seen in non-transplant cohorts. Conclusions: This study provides novel insight into clonal dynamics of FLT3-ITD AML relapse following alloSCT, contrasting with previously reported patterns in the non-transplant setting. Most relapses after alloSCT are driven by clonal persistence of FLT3-ITD, while selective loss of FLT3-ITD among patients receiving FLT3i maintenance therapy suggests clonal suppression and escape. Our results provide biologic rationale for MRD-directed therapeutic approaches and highlight the need for broader post-transplant maintenance strategies that can address both FLT3-dependent and -independent relapse pathways.