Pb1779: flt3-itd mutation characterization with classical pcr methodology versus capture- and amplicon-based ngs platforms: a pethema ngs-aml project

Topic: 3. Acute myeloid leukemia - Biology & Translational Research Background: The identification of FLT3 mutation in acute myeloid leukemia (AML) has become a key issue since the incorporation of FLT3 inhibitors in front-line therapy. Aims: To compare the identification of FLT3 mutations by the classical PCR technique vs. NGS with the two main platforms (capture-based Illumina and amplicon-based ThermoFisher). Methods: Data was collected from the 7 NGS reference laboratories of PETHEMA AML diagnostic network. A total of 354 FLT3-ITD positive AML samples at diagnosis or relapse were analyzed by PCR (followed by capillary electrophoresis) and NGS. A ratio ≥0.03 was established for a positive PCR result. ThermoFisher (amplicon-based technology) and Illumina (capture-based technology) platforms were used to characterize FLT3-ITD mutations in 196 and 158 samples respectively. NGS mutation characterization was performed with the commercial and custom myeloid panels and software used in routine diagnosis at each reference laboratory. Results: Using the ThermoFisher NGS platform, 14/196 samples (7.1%) were detected by PCR but not NGS. Of them, 6 samples had a PCR ratio between 0.03–0.19 and 7 samples presented an ITD of 129–200bp length. Meanwhile, all NGS FLT3-ITD positive samples were also identified by PCR. With Illumina platform, 12/158 (7.6%) were not characterized with PCR (but all of them were NGS positive). Ten out of the 12 samples presented a VAF between 1-2.8%, and 1 sample had an ITD mutation at position c.2503, thus, outside the PCR amplicon. We also compared concordance of PCR-ratio and VAF-ratio (mutated VAF/wild-type VAF) for each NGS platform. Twenty six samples with high ITD length (> 110bp) were excluded from the analyses as they showed highly dissimilar results. As shown in Figure 1, Illumina platform showed higher concordance (mean PCR-ratio: 0.902, mean VAF-ratio: 0.937, R of lineal regression=0.875, N=145) compared to ThermoFisher (mean PCR-ratio: 0.856, mean VAF-ratio: 0.399, R of lineal regression=0.655, N=182). Considering the whole series, we identified 12 (12/158, 7.6%) highly discordant cases with the Illumina platform, with a difference between PCR-ratio and VAF-ratio of at least two units. Half of these cases showed an ITD insert between 114–237bp which corresponded with overestimated VAF values between 74–99.4%. We also found 11 highly discordant cases (11/196, 5.6%) with the ThermoFisher platform, although only 2 cases had a high length insert (112 and 180bp) which conversely corresponded to infraestimated or negative NGS VAF values. Summary/Conclusion: Most of the samples were characterized by both PCR and NGS methods and the Illumina capture-based technique showed higher sensitivity and concordance. High-length ITD mutations are problematic as they are either undetected by amplicon-ThermoFisher technology or overestimated with the capture-Illumina workflow. These may be relevant considerations in the accurate identification of FLT3-ITD mutations as well as for NGS-MRD monitoring. Figure1. Lineal regression of PCR-ratio vs. VAF-ratio in each NGS platformKeywords: Acute myeloid leukemia, Flt3-ITD, PCR