Introduction:FLT3 internal tandem duplications (ITD) are among the most frequent genetic alteration, found in approximately 30% AML, and are associated with a high risk of relapse and death. Measurable residual disease (MRD) NGS-based monitoring of FLT3-ITD mutations offers a new opportunity to detect leukemic cells during treatment. Here we aimed to retrospectively assess the prognostic value of FLT3-ITD MRD monitoring using a high sensitivity NGS-based approach in FLT3-ITD AML patients included in the AML-treated with ponatinib trial, a pan-BCR::ABL kinase inhibitor.

Methods: The AML-ponatinib trial (Clinicaltrial ID: NCT02428543) included AML patients positive for FLT3-ITD by fragment analysis in complete remission. Patients then received consolidation courses (2 to 3) with the combination of ponatinib 30 mg per day and Cytarabine (HDAC for the 18-55y cohort A and IDAC for the 55-70y cohort B). Ponatinib was used as a weak FLT3 inhibitor (FLT3 IC50: 12.6 nM).

In the first part of the work, FLT3-ITD MRD was monitored using 3 different techniques: 2 custom amplicon-based NGS, one according to Blätte et al (Leukemia, 2019) (AB technique), a second developed by the Toulouse team (Toulouse technique) and the last technique was a capture-based technique (capture technique) including the full sequence of FLT3 gene. For these 3 techniques, the detection and the quantification of the FLT3-ITD clones was performed using FiLT3R algorithm (4f569307). These 3 assays were compared to the reference FDA approved FLT3-ITD MRD NGS kit (IVS; Invivoscribe). Of note the 700 ng DNA input was similar for all the tests except for Toulouse (200ng).

Results: Using 72 samples, the comparison between the 3 in-house techniques and the IVS kit showed the following results: 99%, 97% and 93% concordance for the AB technique, capture technique and Toulouse technique, respectively. Regarding the strict similarity between the AB technique with IVS, the same sensitivity (10-5), the good median average reads obtained on the NextSeq 1000 (Illumina) and the advantage of FiLT3r algorithm (identification of the sequence and breakpoint of the different clones), we decided to apply the AB technique to the samples of the PONATINIB-AML trial.

Forty-seven patients were included, 27 in cohort A and 20 in cohort B. Median age was 57.8y (28-70), sex ratio was 0.4. NPM1 mutation was present in 36 patients (76.6%). FLT3 MRD was assessed in 40 out of 46 FLT3-ITD patients (87%). After induction (inclusion, timepoint 1 (TP1)), 25 (62.5%) and 20 (50%) of the patients achieved FLT3 MRD1 negativity at the threshold of 10-4 and 10-5 respectively. With a median follow-up of 4.1 years, FLT3 MRD negativity was associated with a strong benefit in overall survival (OS) (median survival not reached vs 1.2y in MRD1 positive patients, p=0.0003) with a 5y OS of 82.8% (95% CI:60.1-93.2). Sequential follow-up was performed in 31 patients with MRD post induction (TP1) and post consolidation 1 (TP2) with a threshold of 10-4. Median RFS was not reached, 0.8y and 0.35y for patients with neg/neg, pos/neg and pos/pos FLT3 MRD TP1/TP2 respectively (p=0.0001). NPM1 MRD was assessed at TP2 in 29 out of 36 NPM1 mutated patients (80%). Median RFS was not reached and 1y in NPM1 MRD negative and positive patients respectively (p=0.013). At TP2, MRD was assessed using NPM1 and FLT3 in 34 evaluable patients. No patients had NPM1 neg and FLT3 pos MRD. Median RFS was unreached, 1.2y and 0.2y for NPM1/FLT3 neg/neg, pos/neg and pos/pos MRD respectively (p=0.0001). Twenty-two patients relapses and 14 were evaluated for FLT3 MRD. Only 2 of them were negative, 9 had the same clone and 3 had a new FLT3-ITD clone emphasizing the major role of FLT3-ITD clones to trigger the relapse.

Conclusion: In conclusion, using an external quality control based on 72 samples, we are able to identify a very good FLT3-ITD MRD technique and to validate the methodology to assess FLT3-ITD MRD with a level of sensitivity of 10-5/10-4. In this post hoc study, FLT3-ITD was a major predictor of OS and RFS, particularly if negative at TP1. Patients with delayed later negativity had lower OS and RFS. The negativity of NPM1 and FLT3-ITD MRD at TP2 is predictive of a better RFS. Our data strongly support the use of FLT3-ITD MRD to monitor FLT3-ITD AML patients.

This content is only available as a PDF.
2025
Sign in via your Institution