Patients
Eight NTRK fusion sarcomas were retrieved from the archives of pediatric sarcomas under 18-year-old from the archives of the Department of Pathology, Seoul National University Children's Hospital from 2002 to 2019. Among them, authors reviewed clinicopathological and genetic findings of 8 NTRK fusion-positive sarcomas, detected by either fluorescence in situ hybridization (FISH) or next-generation sequencings (NGS), such as RNA sequencing or customized gene panel study. They were one case of ETV6-NTRK3 fusion-positive IMT, one case of TPR-NTRK1 fusion-, one case of LMNA-NTRK1 fusion- and six cases of ETV6-NTRK3 fusion-sarcomas.
Pathology, immunohistochemistry (IHC), and fluorescence in situ hybridization-study (FISH)
All tumors were reviewed by two pathologists (JWK and SHP). Immunohistochemical (IHC) stain performed on an immunostaining system (BenchMark ULTRA system, Ventana-Roche, Mannheim, Germany) using primary antibodies including Trk (1: 50, Cell signaling, Boston, USA), nestin (1: 200, Millipore, Temecula, USA), vimentin (1: 500, DAKO, Grostrup, Denmark), S100 protein (1: 3000, DAKO), CD34 (1: 200, Dako), CD10 (RTU, Novocastra, Newcastle, UK), Ki67 (1: 100, MAb MIB-1; Dako), and phosphohistone-H3 (1: 5000, Cell Marque, Rocklin, USA), TLE1 (1: 20, Cell Marque, Rocklin, US), Fli1 (1: 300, Becton and Dickinson, Flanklin Lakes, US), p53 (1: 100, DAKO), ERG (rtu, Ventana, Export, US), CD99 (1: 200, Novocastra (Leica), Muchen, Germany), Smooth muscle actin (SMA, 1: 500, DAKO), Desmin (1:200, DAKO), Myogenin (1: 500, DAKO), cytokeratin (1: 300, DAKO), epithelial membrane antigen (EMA, 1: 300, DAKO), Integrase interactor 1 (INI-1, 1: 100, Cell signaling,), STAT6 (1: 100, ABCAM, Cambridge, UK) (Suppelementary file: table 1). Appropriate positive controls were included, and for the negative control, primary antibodies were omitted. Mitotic activity was assessed with pHH3 immunostain on 4µm thick formalin-fixed, paraffin-embedded (FFPE) slides by counting mitotic figures in 10 high power fields (HPF; area, 2.38 mm2). The immunohistochemical antibodies that were used are summarized in Supplementary file, Table 1.
For ETV6 break-apart FISH study, locus-specific identifier (LSI) Vysis ETV6 fluorescence dual-color break apart DNA probes, ETV6 (CEN) SpectrumGreen and Vysis LSI ETV6 (TEL) SpectrumOrange (Abbott Molecular, Abbott Park, US), was used.
DNA extraction and customized brain tumor gene panel study
On hematoxylin and eosin-stained FFPE sections, representative areas of tumors with at least 90 % tumor cell purity were outlined for microdissection. DNA-extraction from the serial sections of the microdissected tumor tissue using the Maxwell® RSC DNA FFPE Kit (Promega, USA) was carried out according to the manufacturer's instructions.
The customized targeted gene panel (FIRST brain tumor panel and FIRST pan-cancer panel), which was customized and verified by the Department of Pathology of Seoul National University Hospital (SNUH), containing 172 genes and ten fusion genes, 1.7 Mb/run by NextSeq550Dx in Hi-Output. The produced sequencing data was analyzed using the pipeline of SNUH First Brain Tumor Panel Analysis. First, we performed the quality control of the Fastq file and analyzed only the data that passed the criteria. Paired-end alignment to HG19 reference genome was performed using BWA-men and the GATK Best Practice.[18] After finishing the alignment step, an "analysis-ready BAM" was produced, and second quality control was performed to determine if further variant calling is appropriate. In the pipeline, single nucleotide variation (SNV), insertion and deletion (InDel), copy number variation (CNV), and translocation, were analyzed using at least more than two analysis tools, including in-house and open-source software. The open-source tools used were GATK UnifiedGenotyper, SNVer, and LoFreq for SNV/InDel detection[19], Delly and Manta for Translocation discovery[20], THetA2 for purity estimation, and CNVKit for CNV calling[21], respectively. SnpEff annotated detected variants with various databases such as RefSeq, COSMIC, dbSNP, ClinVar, and gnomAD. Then germline variant was filtered using the population frequency of these databases (> 1% population frequency). Finally, the variants were confirmed throughout a comprehensive review of a multidisciplinary molecular tumor board.
RNA extraction, RNA sequencing, and fusion analysis
For RNA sequencing, the tumor RNA was extracted from the paraffin block (tumor fraction: >90%) with Maxwell® RSC RNA FFPE Kit (Promega, USA). The library was generated with SureSelectXT RNA Direct Kit (Agilent, Santa Clara, USA) and sequenced on an Illumina NovaSeq 6000 at Macrogen (Seoul, Republic of Korea). Raw sequencing reads were analyzed with three kinds of algorithms of DIFFUSE, Fusion catcher, and Arriba (https://github.com/suhrig/arriba/) to detect gene fusions, and compared the results.
Briefly, Fastq files were aligned by the STAR aligner on the hg19 reference genome for Arriba analysis. The chimeric alignments file and the read-through alignments file were produced, and fusion candidates were generated with a set of filters that detect artifacts based on various characteristic features.