NGS-Based Monitoring of Measurable Residual Disease in FLT3-ITD Positive Acute Myeloid Leukemia

Background: In acute myeloid leukemia (AML), most patients respond to chemotherapy initially, but the risk of relapse remains high. Continued monitoring of measurable residual disease (MRD) following therapy onset enables early assessment of treatment response and clinical intervention. However, while appropriate assays exist for several genetic markers, some of the most common aberrations in AML, FLT3 internal tandem duplications (FLT3-ITDs), have remained a challenging target due to insertion site and length heterogeneity. The established diagnostic assay based on PCR and fragment analysis (FA), on the other hand, lacks sufficient sensitivity. In consequence, extensive monitoring of FLT3-ITD MRD in AML patients over time and treatment is not yet available. At the same time, with the approval of the FLT3-targeting tyrosine kinase inhibitor (TKI) midostaurin, the need for detailed monitoring of FLT3-ITD clones has increased.

Aims: We have developed a new next-generation sequencing (NGS) based assay and analysis program to detect FLT3-ITDs with high sensitivity and specificity. We aimed to demonstrate our method's applicability to MRD monitoring and evaluated the progression of FLT3-ITD MRD in patients treated with midostaurin within the AMLSG 16-10 trial.

Methods: FLT3 exons 14 and 15 were PCR-amplified and sequenced on the Illumina MiSeq to generate at least 1 million 250 bp paired-end reads. ITDs were detected using our novel analysis program that performs all steps of the analysis, does not require any manual filtering and reports ITDs extensively annotated to the user. To assess our assay's concordance with established methods, FLT3-ITDs in all diagnosis and relapse samples were characterized by FA and Sanger sequencing (SSeq).

Results: Assay specificity was confirmed with 0 ITDs reported in four FLT3-ITD⁻ control samples (healthy volunteers n=2, AML patient n=1, AML cell line HL60 n=1). To assess assay sensitivity, we serially diluted DNA of the ITD⁺ AML cell line MOLM14 in DNA of the ITD⁻ AML cell line HL60. The expected 21 bp ITD was detected in all four dilutions, down to a variant allele frequency (VAF) of 6.7x10⁻⁵ (0.0067 %). We further processed matched diagnosis and relapse samples from 9 ITD⁺ AML patients at low coverage (mean: 430000 paired-end reads) and confirmed concordance of ITD site, length (range 21-198 bp) and VAF with results from FA/SSeq. For 4/9 patients, we independently sequenced a total of 12 matched diagnosis, remission and relapse samples to our target coverage of 1.5 million paired-end reads. In all of these, ITDs detected by FA were confirmed and ITD site, length and VAF were again concordant between FA/SSeq and our NGS-based approach; results from the diagnosis and relapse samples sequenced to both low and high coverage were highly reproducible.

In 11/12 samples sequenced to high coverage, our assay identified additional ITD clones that were not identified by FA: The mean number of ITDs detected was 4.8 at diagnosis (1.5 by FA), 2.0 at clinical remission (not measured by FA) and 1.5 at relapse (0.5 by FA). VAFs of these ITDs ranged from 0.007-38 % (mean 5.3 %); ITDs that were solely detected by our NGS-assay were present at lower VAFs than those also identified by FA/SSeq (NGS-assay only: mean 0.21 %, range 0.007-2 %; also detected by FA/SSeq: mean 21.4 %, range 2-38 %). About half of the ITDs detected by NGS at diagnosis (11/19 ITD clones) were not detected at any of the later time points and thus presumably eradicated by treatment. Conversely, 8/19 ITD clones persisted at remission (6/19) or reoccurred at relapse (2/19). In fact, nearly all of the ITD clones present at relapse were already identified at an earlier time point.

Conclusion: In sum, we could demonstrate that our NGS-based assay and analysis program achieve high accuracy and precision and enable MRD monitoring in FLT3-ITD positive AML patients. Its high sensitivity enables the detection of additional ITD clones occurring at low allelic frequencies not only at remission but also at diagnosis and relapse. Further evaluation of MRD monitoring in FLT3-ITD positive AML patients treated with midostaurin within the AMLSG 16-10 trial is currently ongoing.