Quantum-First Trial: FMS-like Tyrosine Kinase 3-Internal Tandem Duplication (FLT3-ITD)–Specific Measurable Residual Disease (MRD) Clearance Assessed through Induction (IND) and Consolidation (CONS) Is Associated with Improved Overall Survival (OS) in Newly Diagnosed (nd) FLT3-ITD+ AML Patients (pts)

Background

Detection of MRD by flow cytometry or gene fusion transcript quantitation increasingly is used to guide treatment decisions in AML. Molecular monitoring of recurrent gene mutations for MRD detection is an alternative strategy but, other than for nucleophosmin 1 (NPM1) gene quantitation, remains controversial. FLT3-ITD is among the most common recurrent mutations in AML and confers a worse prognosis vs FLT3–wild-type and FLT3-TKD+ AML. Routine use of FLT3-ITD detection in remission as a predictor of relapse risk or OS has been limited by the low sensitivity of both conventional PCR-based detection methods and broad NGS platforms in clinical practice settings. More recently, FLT3-ITD–specific PCR-NGS assays such as getITD (PMID: 31089248) show greater sensitivity and ease of interpretation that point to eventual routine clinical application. However, the clinical value of these measurements has not been evaluated prospectively in large scale randomized controlled trials of FLT3 inhibitors. The phase 3 QuANTUM-First study (NCT02668653) evaluated the novel, potent, and highly selective type II FLT3 inhibitor quizartinib (Quiz) in nd FLT3-ITD+ AML pts and demonstrated that Quiz added to intensive IND and CONS, ± transplant, followed by single-agent continuation (CONT) therapy (Tx) resulted in a significant improvement in OS (PMID: 37116523). We analyzed if FLT3-ITD–specific MRD in QuANTUM-First pts impacted the clinical outcome or the benefits provided by Quiz in nd FLT3-ITD+ AML pts.

Methods

Genomic DNA, isolated from bone marrow aspirates or peripheral blood from pts after achievement of remission after 1 or 2 courses of IND and at end of CONS (prior to transplant or CONT cycle 1 day 1 [C1D1] for transplant pts and prior to CONT C1D1 for non-transplant pts), was analyzed with a FLT3-ITD PCR-NGS assay specifically developed for this trial (PMID: 31722002). ITD mutations detected after IND were cross-validated against the ITD detected at enrollment for each patient. Using a custom bioinformatics program, ITD mutations were identified, and variant allele frequencies (VAFs) were calculated with a sensitivity of 10⁻⁵. MRD was classified as undetectable (using the 0 cutoff) or detectable above or below a 10⁻⁴ predefined cutoff (based on lower limit of quantification for the assay). Comparisons of complete response (CR), composite complete response (CRc=CR+CRi), and the rate of pts achieving CRc during IND with no MRD between arms were made using a stratified Cochran-Mantel-Haenszel test. Comparison of FLT3-ITD VAF during IND and during CONS between arms was made using Wilcoxon rank sum test. All P-values were not adjusted for multiplicity.

Results

In QuANTUM-First, 539 nd FLT3-ITD+ AML pts were randomized to Quiz (n=268) or placebo (PBO; n=271). Of the 539 randomized pts, 368 (68.3%) achieved CRc after 1 or 2 courses of IND, and MRD analysis was performed on 321 (87.2%) of these pts (162 pts in Quiz and 159 pts in PBO) using samples collected at the time of response assessment during IND, before further Tx. MRD was also assessed in 337 pts (172 pts in Quiz and 165 pts in PBO) by end of up to 4 cycles of CONS prior to CONT/maintenance Tx. Of these pts, 166 additionally received transplant (87 in Quiz, 79 in PBO). The % of pts in CRc at end of IND with FLT3-ITD MRD of <10⁻⁴ was similar between study arms (25.4% Quiz vs 21.8% PBO, nominal P = 0.3430), however, a greater % of pts in CRc had undetectable MRD with Quiz than PBO (12.3% vs 7.0%, nominal P = 0.0403). Among pts with CRc at end of IND, the median best FLT3-ITD VAF by end of CONS was lower (0% versus 0.0017%; nominal P = 0.0006) with Quiz vs PBO (Fig 1) and, using the 0 cutoff at end of IND, a longer OS was observed with Quiz vs PBO, regardless of MRD status, with a greater effect with Quiz (HR, 0.789 in MRD− pts; HR, 0.749 in MRD+ pts; Fig 2). In MRD+ pts, the median OS was not reached with Quiz and 35.4 months with PBO (Fig 2). Performing this analysis with an MRD cutoff of 10^‒4 yielded similar findings (HR, 0.696 in MRD− pts; HR, 0.800 in MRD+ pts). Additional data about MRD impact on pts ± transplant in this trial will be presented.

Conclusions

These findings demonstrate the potential prognostic utility of FLT3-ITD–specific MRD measurements in the clinical management of pts with FLT3-ITD+ AML. Our data suggest that long-term OS benefits conferred by Quiz in part derive from a deep and sustained reduction of the FLT3-ITD+ leukemia burden.