PF220 COMPARISON OF FOUR NGS PANELS AND ANALYSIS OF THEIR CLINICAL UTILITY IN MYELOID HEMATOLOGICAL MALIGNANCIES OF THE MYELOID LINAGE

Background:The diagnosis of myeloid neoplasms (MN) has significantly evolved through the last few decades. Nowadays, blood cell morphology, blasts count, cytogenetics and molecular testing are crucial for MN diagnosis, prognosis and therapy. The advent of personalized medicine has drastically influenced the way we currently diagnose and monitor MN, mainly thanks to Next Generation Sequencing (NGS), which is gradually becoming an essential tool to help clinicians with disease management. To this end, most specialized genetic laboratories have implemented NGS panels targeting a number of different genes relevant to MN.Aims:The aim of the present study is to evaluate the performance of four different targeted NGS gene panels based on their technical features and clinical utility.Methods:A total of 32 patient bone marrow samples were accrued: 17 acute myeloid leukemia (AML), 7 myeloproliferative neoplasms (MPN), 6 myelodysplastic syndromes (MDS), 2 chronic myelomonocytic leukemia (CMML). All of them had been tested with conventional molecular testing and harboured a variety of sequence variants, including single nucleotide variants (SNVs), insertions and deletions (indels). These samples were tested with four NGS panels, three were commercially available panels and one custom. The samples were tested as follows: 17 with Trusight Myeloid Panel (TSMP, 54 genes, amplicon based, Illumina), 16 with SureSeq (23 genes, capture‐based, Oxford Gene Technology), 15 with Myeloid Solutions (MYS, 30 genes, capture‐based, SOPHiA GENETICS), and all 32 with our custom Pan‐Myeloid Panel (PMP, 48 genes, capture‐based, SOPHiA GENETICS). Libraries were built following manufacturer's instructions and pair‐end sequenced on a Miseq sequencer (Illumina). Analysis of the sequencing data was carried out with a valid performance at the clinical variant allele frequency (VAF) cut‐off of 5%. Variants were classified by two geneticists with expertise in hematological malignancies; variants classified as “pathogenic” or “likely pathogenic” were kept as clinically relevant.Results:The average sequencing depth was over 5000X in PMP, MYS and TSMP; and over 1600X in SureSeq panel. Nineteen genes are included in all four panels (core myeloid gene set), being ASXL1, CALR, CEBPA, DNMT3A, ETV6, FLT3, IDH1, IDH2, JAK2, KIT, KRAS, MPL, NPM1, NRAS, RUNX1, TET2, TP53, U2AF1 and WT1. A total of 50 clinically relevant variants were detected by all four panels, and 37 of these fell in one gene of the core myeloid gene set. There are 11 discordant variants between panels, 3 in the core myeloid gene set missed only by SureSeq panel (2 FLT3‐internal tandem duplications (ITD) and one CALR mutation); and 8 mutations only called by the panels that include those genes in their design (Table 1). Of note, 2 additional FLT3‐ITD mutations of 36 bp of length that had been detected by conventional molecular analysis were not called by any of the NGS panels.Summary/Conclusion:After testing the four panels, our data show that there is a high risk of finding different mutations depending on the panel of choice, due both to the panel design and the data analysis method. Based on our data, ITD calling remains a challenge for NGS. This is a major issue, since AML management strongly depends on FLT3‐ITD detection. In addition, NGS testing times are hard to harmonise with turnaround time stablished in current European Leukemia Net guidelines. Therefore, conventional molecular testing might need to be kept in place for the correct diagnosis of MN for now.image