Tumor-agnostic genomic and clinical analysis of solid tumors with BRAF fusions.

Abstract

3146 Background: Approximately 15% of cancers harbor BRAF alterations, of which ~5% are BRAF fusions. BRAF fusions are class II alterations that have worse outcomes than class I V600E alterations. While prior investigational therapies for BRAF fusions had disappointing outcomes, novel therapies are in clinical trials, underscoring the need to better characterize these tumors. Methods: Data from adult and pediatric patients with BRAF fusion+ cancers identified between January 2014 and November 2022 utilizing a center-wide next generation sequencing (NGS) program of >99,000 sequenced samples were analyzed. DNA-based hybrid capture tumor NGS (MSK-IMPACT), ctDNA targeted NGS (MSK-ACCESS), and/or RNA anchored multiplex PCR tumor NGS (MSK-Fusion) were used. Fusions were manually reviewed and considered to have oncogenic potential if they had an in-frame protein fusion involving a non-BRAF partner gene and an intact BRAF kinase domain (exons 11-18). Fusions not previously reported in OncoKB, COSMIC, TCGA, NIH gene, Fusion GDB2, and PubMED were classified as novel. All cases underwent clinical data curation including baseline demographic, tumor characteristics, and treatment histories. Results: 212 patients (0.2%) with BRAF-fusion positive solid tumors were identified. 194 were identified by DNA-based tumor NGS (96 had sufficient tissue for confirmatory RNA-based NGS, all of which were positive). Six were identified by ctDNA NGS and 12 by RNA-based tumor NGS only (5 had insufficient tissue for DNA-based tumor NGS). 83 unique 5’ fusion partners were found, of which 42 were novel. The most frequent tumor types were pilocytic astrocytoma (n=30, 14%), prostate CA (n=28, 13%), melanoma (n=24, 11%), lung CA (n=21, 10%), and colon CA (n=15, 7%). 44% of patients with pilocytic astrocytoma had BRAF fusions, of which 90% were BRAF-KIAA1549. Concomitant alterations (≥10% frequency) included TP53 (24%), TERT (18%), CDKN2A deletions (13%), and CDKN2B deletions (12%). 20% (n=43) were treated with MAPK-pathway directed therapies spanning multiple histologies and lines of therapies. Of the 212 patients, 17 had acquired BRAF fusions after targeted therapy for another oncogene ( EGFR mutation (n=11), BRAF V600E (n=4), FGFR fusion (n=1), NTRK fusion (n=1)). The majority of patients with acquired BRAF fusions had EGFR-mutant lung adenocarcinoma (n=11, 76%); the median time from EGFR targeted therapy initiation to BRAF fusion detection was 25 months (range 16-38 months). Conclusions: A wide variety of adult and pediatric solid tumors harbored de novo BRAF fusions. Complementary RNA sequencing optimized fusion identification in many cases. Multiple novel fusion partners were found. Acquired BRAF fusions were identified after targeted therapy for a variety of distinct oncogenes, the majority of which were EGFR mutations.