Frequency and impact of FLT3-ITD microclones in younger AML patients

Abstract

Abstract Introduction: FLT3 internal tandem duplications (ITD) are found in approximately 25% AML and defined by an allelic ratio (AR)≥0.05, hereinafter referred to as macroclones (Mclones). We previously described the prevalence of FLT3-ITD microclones (µclones), defined by an AR<0.05, in a cohort of patients with FLT3-ITD Mclones (Joudinaud, Blood Adv 2025). Here we aimed at assessing the prevalence and prognostic value of FLT3-ITD µclones in the French multicenter BIG-1 trial. Methods: The BIG-1 trial (NCT02416388) included patients (pts) aged 18-60 years with newly diagnosed AML treated by intensive chemotherapy (CBF-AML and APL excluded). This post-hoc analysis required DNA fragment analysis (FA) to detect FLT3-ITD Mclones (AR≥0.05). Samples with negative or AR<0.05 FA were used for NGS on a NovaSeq 6000. Regions of interest were sequenced with an average depth of 3000x then analyzed with FiLT3r algorithm for quantification of FLT3-ITD (Boudry, BMC Bioinformatics 2022) allowing an AR threshold detection at 4x10e-4. Patients with an AR 0.0004-0.05 were referred to µclones, other patients were considered without FLT3-ITD. Midostaurin (MIDO) has been introduced in 07/2018 according to its label. Results: 1,733 pts with available DNA were included in this analysis from 01/2015 to 02/2022, 353 (20.4%) had FLT3-ITD Mclones (±µclones, MACRO group) including 175 (49.6%) pts with high AR (≥0.5), 240 (13.8%) had FLT3-ITD µclones only (MICRO group) and 1,140 (65.8%) were FLT3-ITD negative. Overall, median age was 50y and 848 pts were female without differences between the three groups. As expected, according to ELN-2022, risk groups were favorable for 0.6%/40.9%/35.2%, intermediate for 82.4%/28.3%/15.8% and adverse for 17.0%/30.8%/49.0% of pts in MACRO, MICRO and negative groups, respectively. Among the 668 pts with NPM1 mutation, 237 (35.5%), 130 (19.4%) and 301 (45.1%) pts were in MACRO, MICRO and negative groups, respectively, with significant differences for each pairwise comparison. MIDO was administered to 134 patients (38%) in the MACRO group, 18 (7.5%) in the MICRO group, and 18 (1.6%) in the negative group, the latter two based on FLT3-TKD mutations. In multivariate analyses, compared to absence of FLT3-ITD, µclones were significantly and independently associated with an increased risk of relapse (sHR 1.46 [1.15-1.85]; P=0.002), as well as presence of FLT3-ITD Mclones (sHR 2.09 [1.65-2.66]; P<0.001), as a whole, or divided in two groups with low AR<0.5 (sHR 1.93 [1.46-2.56]; P<0.001) and high AR≥0.5 (sHR 2.36 [1.73-3.22]; P<0.001), adjusted for confounding factors including allo-SCT in CR1 as time-dependent variable. The other independent predictive factors in multivariate analyses for CIR were: age, cytogenetic risk, NPM1, CEBPA bZIP and TP53 mutations, MIDO and allo-SCT. 5y-CIR estimations were 47.5%, 47.9% and 37.5%, 2y-CIR estimations were 42.5%, 45.1% and 29.4% in MACRO, MICRO and negative groups, respectively. Presence of FLT3-ITD µclones was also significantly and independently associated with an increased risk of death or relapse (aHR: 1.39 [1.13-1.71]; P=0.002), as were Mclones with low and high AR. 5y-RFS were 41.9%, 40.6% and 47.7% in MACRO, MICRO or negative groups, respectively. Given the impact of FLT3-ITD mutations in the ELN 2022 risk definition of patients with NPM1 gene mutation, we conducted an analysis in this subgroup. In multivariate analyses, FLT3-ITD µclones were significantly and independently associated with increased CIR (sHR 1.66 [1.24-2.24]; P=0.001), shorter RFS (aHR 1.71 [1.29-2.27]; P<0.001) and shorter OS (aHR 1.48 [1.02-2.14]; P=0.04) compared to negative group. After two courses of chemotherapy, NPM1 BM MRD was ≥1% in 24.1%, 15.7%, 16.3% and 6.7% of pts in MACRO group with high and low AR, MICRO group and negative group respectively. Conclusion: In younger AML pts, FLT3-ITD microclones increase the risk of relapse and shorten RFS. In the NPM1 mutated subgroup, such FLT3-ITD microclones also shorten OS. Our results advocate for a change of practice from fragment analysis to high-sensitivity molecular techniques for FLT3-ITD detection, consider the mutation in the ELN classification regardless of the diagnostic threshold of 0.05 and investigate FLT3 inhibitors in this population.