Next Generation Sequencing for Minimal Residual Disease Monitoring in AML Patients with FLT3-ITD,

Abstract 3548

Minimal Residual Disease (MRD) monitoring has become an important tool for risk and treatment stratification in hematological malignancies. MRD monitoring in FLT3 mutated patients has been difficult in the past as FLT3-ITDs vary from patient to patient and individual primer/probe sets would be required to assess MRD over time. In the present study we evaluated next-generation sequencing (NGS) as a new tool for MRD monitoring in patients with FLT3-ITD and NPM1 mutations.

Five pediatric and 5 adult AML patients with FLT3-ITD and 10 adult patients with NPM1 mutations were analyzed by NGS with a target coverage of 10,000 reads per amplicon. Pediatric samples were collected at diagnosis, day 22 and after consolidation chemotherapy while adult samples were collected at different time points (average 4 timepoints per patient). Samples were sequenced unidirectionally on eight-lane PicoTiterPlates on a GS FLX sequencing system. In total, 2,563,550 sequencing reads were generated, corresponding to a total of 1,176,171 high-quality sequencing reads. NPM1 mutations were analyzed by quantitative RT-PCR using the MutaQuant kit from Ipsogen (Ispogen, Marseille, France). Allelic ratios of FLT3-ITDs were determined by fragment analysis on a DNA sequencer using GeneMapper software 4.0.

First, the sensitivity of NGS to detect mutated alleles was evaluated by sequencing serial dilutions of a patient sample that had 46.3 percent mutated FLT3-ITD alleles at diagnosis. With a target coverage of 10,000 sequences and an allelic ratio of 46.3 percent the theoretical detection sensitivity was at most 1 in 4630 sequences. In fact, the allelic ratio in the sequenced samples linearly decreased in the tested dilutions down to the 5×10-4 dilution (Pearson correlation R2=.996). Samples from healthy volunteers were tested negative for both FLT3-ITD and NPM1 mutations (n=3). Allelic ratios from three diagnostic specimens of FLT3-ITD mutated patients were highly reproducible when determined in two independent NGS runs.

As proof of principle we analyzed NPM1 mutated patients by NGS and quantitative RT-PCR in parallel. The mean allelic ratio of NPM1 mutants at diagnosis was 0.37 (range 0.29–0.46). An allelic ratio of 0.37 and 0.4 was measured in peripheral blood of two patients, and thus was similar to ratios in bone marrow. Concordant results between NGS and qRT-PCR were found in 38 samples (95%), whereas in two samples one method did not detect the mutation while the other did (NGS and RT-PCR were negative once each). We analyzed relapse samples in four patients. The NPM1 mutation was detected consistently by both methods in three patients at allelic ratios of 0.013, 0.19, and 0.32, while one patient had lost the mutation at relapse. One patient had an atypical NPM1 mutation for which no RT-PCR kit was available. NGS allowed quantification of the allelic ratio in this patient, which was 0.37 at diagnosis, 0.06 after one cycle of induction therapy, and 0 after the second cycle of induction therapy.

In FLT3-ITD mutated patients we could determine insertion site, insertion length, number of individual clones, and allelic ratio from NGS data. The mean allelic ratio in diagnostic samples was 0.27 as measured by NGS and 0.4 as measured by fragment analysis. Three follow up samples were negative by fragment analysis, while a small clone could still be detected with NGS in these samples (allelic ratio 0.0004 to 0.001). All other samples were concordant between fragment analysis and NGS.

NGS was used to determine MRD status in 5 patients with childhood AML harboring mutated FLT3. A reduction of 2–3 orders of magnitude was achieved during induction chemotherapy. During consolidation a further decrease or disappearance of mutated alleles was achieved in 3 patients, who remained in remission. However, allelic burden increased in 2 patients after first consolidation treatment (HAM) by 9- and 735-fold compared to the allelic ratio after induction therapy, and they relapsed 74 and 303 days later. Thus, accurate determination of the FLT3-ITD allelic ratio by NGS may become useful to identify patients before overt relapse.

In summary, we show that NGS can be used for minimal residual disease assessment in FLT3-ITD mutated AML patients. The sensitivity of the method is scalable depending on the read depth, however, an adequate sensitivity level for efficient MRD detection still needs to be determined.