Abstract
Introduction: Receptor tyrosine kinases (RTKs) initiate signaling cascades, including growth and differentiation. Activation can occur through chromosomal rearrangements that lead to gene fusions. RTK fusions are potential targets for small molecule inhibitors to treat advanced cancers. The original Afirma Xpression Atlas (XA) reported 761 selected variants and 130 fusion pairs in Bethesda III/IV Afirma Genomic Sequencing Classifier (GSC) suspicious or Bethesda V/VI nodules. The landscape of additional potentially actionable gene fusions has not been explored in treatment-naïve patients.
Methods: Anonymized RNA-seq data from >37,000 Bethesda III-VI samples were examined with STAR-fusion to determine gene/gene fusions. All samples were examined for NTRK1, NTRK3, RET, ALK, and BRAF fusions, regardless of fusion partner. Fusions were evaluated for being in-frame, with an intact kinase domain at the 3’ end of the fusion pair. Fusion pairs not currently reported by XA and not reported in thyroid TCGA fusion data are denoted “additional”. All fusion pairs were searched for in the literature and public fusion databases.
Results: Examining the Veracyte clinical database revealed 7 additional NTRK1/3 fusions, with 3 NTRK fusions observed more than once - SQSTM1/NTRK3, VIM/NTRK3, and EML4/NTRK3. One of the 7 NTRK fusions had not been previously reported. Eight additional ALK fusions were identified, with 4 observed more than once- ITSN2/ALK, PPP1R21/ALK, PDE8B/ALK, NPAT/ALK. Five of these 8 ALK fusions had not been previously described. Seventeen additional RET fusions were identified, with 5 observed recurrently - KIAA1217/RET, AFAP1L2/RET, ACBD5/RET, SQSTM1/RET, and TFG/RET. Six of the 17 RET fusions had not been previously reported. Seventy-two additional BRAF fusions were identified, and 58 of them have not been previously reported. Eight of the 72 BRAF fusions were observed more than once. Examining >50,000 Afirma samples, NTRK1, NTRK3, RET, ALK, or BRAF fusions were not identified among the Afirma GSC Benign, and were present in 3.2% of 16,594 Bethesda III/IV Afirma GSC Suspicious samples, and 8.0% of 1,692 Bethesda V/VI samples. Correlation with surgical histology is unknown.
Conclusions: By examining a large cohort of patients with an unbiased, whole-transcriptome RNA-seq assay, we identified potentially actionable kinase fusions in thyroid nodules beyond those described in TCGA. All fusions described here are either novel and not previously reported, rarely reported in one or two case studies, or not described in thyroid cancers. Additional NTRK, ALK, RET and BRAF fusions were found, all of which may be targeted with specific kinase inhibitors currently available. Future studies may determine genotype-phenotype correlations regarding the natural history of these neoplasms. Because of the potential clinical implications of these genomic markers for patient management, all 104 fusions described here are now included among the 235 gene pairs reported by the expanded Afirma XA.
Fusion Genes and RNAs in Cancer Development
