Targeting rare variants of a pan-cancer target: The landscape of BRAF non-v600 mutations and BRAF fusions from 172,005 adult patients with cancer.

person Niamh Coleman Trinity St. James's Cancer Institute, Dublin, Ireland info_outline Niamh Coleman, Vivek Subbiah

Authors person Niamh Coleman Trinity St. James's Cancer Institute, Dublin, Ireland info_outline Niamh Coleman, Vivek Subbiah Organizations Trinity St. James's Cancer Institute, Dublin, Ireland, Sarah Cannon Research Institute, Nashville, TN Abstract Disclosures Research Funding No funding sources reported Background: BRAF is a recognized pan-cancer target, with dabrafenib in combination with trametinib receiving FDA accelerated approval for unresectable metastatic solid tumors with BRAF V600E mutation. BRAF fusions and mutations at codons aside from V600 (non-V600) activate the MAP kinase/ERK-signaling pathway, however the spectrum of these alterations is not yet well-defined. Preclinical data suggests that BRAF non-v600 mutations are diverse, with evidence of alternate signaling, requiring dimerization to function, while BRAF V600E tumors activate the pathway as monomers. We include a comprehensive analysis of non- BRAF V600 mutations and BRAF fusions in pan-cancer adult malignancies. Methods: 198,041 samples from 172,005 patients available from AACR Project GENIE v.15 (1) database were analyzed for the prevalence of non-V600 BRAF mutations, fusions and copy number alterations in a range of cancer types. Results: A total of 6374 separate non-V600 BRAF alterations were identified in 5876 samples (2.97%), including 1176 fusions (18.4%), 4435 missense mutations (69.6%), 299 truncating mutations (4.7%), and 245 in-frame mutations (3.8%), splice variants (3.4%). Most frequent tumors included lung (19%), melanoma (14%), colorectal (9.7%) glioma (5%), thyroid (1.4%). BRAF fusions were observed in 0.7% of tumor samples, most identified in glioma, prostate, lung, thyroid, and colorectal cancer (20%, 7.3%, 6.1%; 4%, 4% of identified BRAF fusions, respectively). Most fusions were considered driver events (1123, 95.5%); frequent fusion gene partners included KIAA1549 , intragenic fusions, SND1 , AGK , MKRN1 (39.9%, 15%, 4%, 3.3%, 2.4% of 1176 samples respectively). Of 4435 missense mutations identified, most were considered driver events (2734, 62%); 1701 missense mutations were variants of uncertain significance (VUS) using OncoKB database (38%). Missense mutations occurred across codons, most frequently involving codon 469 (n=584, 13.2%, G469A/V/R/E/S/L/K/I), 466 (n=310, 10.5%, G466V/E/A/R), 594 (n=506, 11.4%, D594G/N/A/E/H/Y/V/F), 601 (n=279, 6.3%, K601E/N/I/T/_S602delinsNT), 581 (n=231, 5.2%, N581S/I/Y/H/T/D/K) 597 (n=118; 2.7%, L597R/Q/S/V/P/H/I). BRAF amplification occurred in 0.14% of samples. Conclusions: While BRAF fusions are rare events across cancers, non-v600 BRAF alterations occur in 3% of malignancies. Atypical non-V600 BRAF alterations and BRAF fusions represent a distinct molecular cohort across cancers. Most BRAF fusion events and non-V600 missense mutations are characterized as oncogenic, underlining the spectrum of BRAF inhibition, and the urgent need to expand agents beyond the current BRAF V600 indications. Functional characterization of atypical BRAF variants is therefore crucial, together with enrolling patients with these rare alterations on rationally designed clinical trials. 1. Cancer Discov. 2017.