Spectrum and implications of activating BRAF alterations in advanced prostate cancer (aPC).

220 Background: Activating genomic alterations (GA) in BRAF are rare in aPC and their impact on pathogenesis is poorly understood. However, emerging data suggest that these GA may represent clinically actionable targets (Fenor et al, Clin Transl Oncol, 2022). Using comprehensive genomic profiling (CGP) performed in a single commercial lab, we characterized the GA landscape of BRAF-activated tumors in aPC patients (pts). Methods: Tissue(N=15,864) and liquid (N=7,566) biopsies from aPC pts were profiled using FoundationOne CDx and FoundationOne Liquid CDx CGP assays, respectively. Each assay covers 324 cancer-related genes, including the full coding region of BRAF and introns 7-10, with additional sensitivity in exons 11-18 in the liquid assay. Activating GA were defined as hotspot missense mutations, in-frame indels in the kinase domain, or rearrangements (RE) that preserve the kinase domain. Tissue biopsies from non-aPC cancer types (N= 275,151) were used for comparison. Results: BRAF-activating GA were detected in 520/15,864 (3.3%) tissue biopsies: RE were the most common GA (243 samples, 1.5%), followed by K601E (101, 0.6%), and G469A (58, 0.4%). Median age of pts with BRAF-altered tumors was 69 (interquartile range 63-76), compared to 68 for BRAF wild-type (interquartile range 62-74). Rearrangement breakpoints occurred most frequently in intron 8 (37%), intron 9 (28%), intron 10 (20%), and intron 7 (13%). The most common RE were BRAF N-terminal truncations removing the auto-inhibitory domain (22%), SND1-BRAF fusions (13%), intragenic BRAF deletions of the auto-inhibitory domain (12%), and TMPRS22-BRAF fusions (5%). When studying cases with BRAF-activating GA compared to wild-type, we noted a larger proportion of CDK12 mutations (9.2% vs 5.2%, p=0.018), and a depletion of TMPRSS2-ERG fusions (11% vs 32%, p<0.0001), PTEN GA (17% vs 31%, p<0.0001), and APC GA (4.4% vs 8.9%, p = 0.018); alterations in AR occurred at similar rates (13% vs 13%, p=1.0). In liquid samples, overall BRAF GA were slightly less common (187/7566, 2.5%): 65 (0.9%) RE, 33 (0.4%) K601E, 12 (0.2%) G469A. Examining BRAF-altered samples across all cancer types, aPC had the highest proportions of RE (46%) and one of the lowest frequencies of V600E (0.1%). Conclusions: Activating BRAF alterations are detected in ~3% of aPC, with frequent BRAF-SND1 fusions. BRAF RE represent almost half of aPC BRAF GA, the highest fraction observed across a pan-tumor dataset. In addition, we detected a higher incidence of concurrent CDK12 GA and a lower relative frequency of concurrent PTEN, APC, and TMPRSS2-ERG GA. These findings suggest that genetic activation of BRAF in aPC pts may contribute to tumorigenesis and supports further clinical investigation of therapeutics targeting the MAPK pathway in this molecular subtype.