Quantum-First: Deep-Dive Analysis of FMS-like Tyrosine Kinase 3–Internal Tandem Duplication (FLT3-ITD) Detection Methods Utilized for Enrollment and Longitudinal MRD Detection in Newly Diagnosed (nd) Patients (pts) with FLT3-ITD–Positive (FLT3-ITD+) Acute Myeloid Leukemia (AML)

Introduction

The phase 3 QuANTUM-First study (NCT02668653) showed that in nd FLT3-ITD+ AML pts, adding the oral, highly potent, selective, type 2 FLT3 inhibitor quizartinib to standard chemotherapy ± transplantation, followed by quizartinib or placebo monotherapy for up to 3 y, decreased relative risk of death by 22.4% vs placebo, with a generally manageable safety profile (PMID: 37116523). In QuANTUM-First, FLT3-ITD mutation status was determined using a FLT3-ITD mutation detection clinical trial assay (CTA). We describe concordance between the CTA and next-generation sequencing (NGS) measurable residual disease (MRD) assay in detecting the same FLT3-ITD sequences at screening and longitudinally. We provide FLT3-ITD detection results using the MRD assay in paired bone marrow (BM) aspirates and peripheral blood (PB) samples where available.

Methods

The CTA, developed by Navigate BioPharma Services, Inc., utilizes PCR and capillary electrophoresis (PCR + CE). FLT3-ITD mutations were detected in genomic DNA (gDNA) from nd FLT3-ITD+ AML pts by PCR with primers flanking exons 14 & 15 of the FLT3 gene. PCR products underwent fragment analysis on the ABI 3130xL Genetic Analyzer using ABI 3130xL Data Collection Software to resolve and generate different-sized PCR products. A DNA sample was considered CTA+ if the variant allele frequency (VAF; FLT3-ITD/total FLT3) was ≥3%. For the PCR + NGS MRD assay, gDNA from pts with composite complete remission (CRc = complete remission [CR] + CR with incomplete neutrophil or platelet recovery) was amplified by PCR with primers flanking exons 14 & 15 of FLT3. PCR products were sequenced after library preparation using the Illumina MiSeq System, with a LLOQ of 10⁻⁴ and estimated LLOD of ~2×10⁻⁶ leukemia cells. FLT3-ITD mutations detected after induction (IND) were cross validated against FLT3-ITD mutations detected at enrollment for each pt. Using a custom bioinformatics program, FLT3-ITD mutations were detected, and VAFs calculated with a sensitivity of 10⁻⁵ leukemia cells. Estimates of FLT3-ITD VAF at study entry were derived from peak heights in the PCR + CE assay. In both technologies utilized, gDNA from BM was primarily used. If a BM sample was not available, a PB sample was used. Paired BM and PB samples were compared by the NGS MRD assay from a small pt subset.

Results

Among 342 pts (with 417 corresponding ITD sequences) with CRc by the end of IND, tested by the PCR + CE and the PCR + NGS MRD assays at screening, concordance in FLT3-ITD detection of exact ITD length was revealed in 329 (96.2%) pts (with 393 corresponding ITD sequences). Median difference in ITD length was 6 base pairs in the remaining ITDs that did not match in length with 11 ITDs in pts with >1 ITD detected and 13 ITDs in pts with only 1 ITD detected. Persistent FLT3-ITD sequences originally detected at screening were assessed at longitudinal time points in 265 CRc pts who were still MRD+ (cutoff of 0 leukemia cells) at the end of IND (125 [65.1%] in the quizartinib arm and 140 [79.5%] in the placebo arm), to determine if major ITD presence changed over time. Additional analyses of paired sampling at various time points revealed that among the overall MRD data from the study, there were 27 paired BM and PB samples (corresponding to 18 pts), in which MRD was analyzed from both sample types at the same time point, and in which a FLT3-ITD clone was detected in 1 or both paired samples. The median VAFs of clones detected in BM samples were 6.6-fold higher, compared with corresponding VAFs derived from paired PB samples (P=.0011). In 18/27 (66.7%) cases, a clone was detected in both samples. In 8/27 cases, a clone was detected in the BM sample but not in the paired PB sample, while in 1/27 cases a clone was detected in PB but not in BM.

Conclusions

In QuANTUM-First, FLT3-ITD detection at screening by the PCR + CE and PCR + NGS MRD assays was concordant to the exact ITD length detected in 96.2% of 342 pts analyzed. Longitudinal analyses assess persistence/variability of the major ITD initially present at screening to gain awareness of ITD changes over time during therapy. Paired sampling analyses based on review of 27 paired BM and PB samples, confirm that the PCR + NGS MRD assay is more sensitive when BM is used as the DNA source vs PB, as anticipated, but the ability to detect the same FLT3-ITD from PB in 66.7% of paired samples speaks to the versatility of the NGS MRD assay and the ability to still obtain meaningful data when only PB is available for specimen collection.