BRAF Copy Number Alterations and Ultralow BRAF mRNA Expression Are Prognostic for Poor Overall Survival in Prostate Cancer

Abstract

Background: Prostate cancer (PC) is the second most common cancer among men in the US, with adenocarcinomas being the most prevalent histologic subtype (≥99%). Emerging evidence suggests that BRAF alterations, mostly annotated by protein altering mutations, may be drivers of prostate cancer progression in up to 6% of patients. Here we characterized, to our knowledge, the largest cohort of whole-exome and -transcriptome sequenced BRAF-altered PCs. We aimed to interrogate molecular features and clinical outcomes in BRAF-altered PCs.

Methods: The Tempus Lens Platform (Tempus AI, Inc., Chicago, IL) was used to query the Tempus multimodal de-identified database and establish a cohort of 485 patients with prostate adenocarcinoma with xT (DNA) and xR (RNA) testing. Of those, 137 had BRAF alterations while the remaining 348 were considered BRAF wildtype (wt). BRAF alterations were defined as including a copy number alteration (CNA) or a single nucleotide variant (SNV) resulting in a short variant mutation. Of the BRAF-wt group, 94 were primary-tumor biopsies while 43 were metastatic biopsies. Within the BRAF-wt group, there were 210 primary tumors and 138 metastatic biopsies. RNA expression data was normalized and quantified as transcripts per million (TPM). BRAF mRNA expression from metastatic tissue were grouped as ultra-low (10th percentile), low (25th percentile), mid (50th percentile), and high (75th percentile). Overall survival (OS) was defined as the time from biopsy collection to death or loss to follow up and conducted using risk set adjustment through Cox proportional hazard analysis and reported by hazard ratios (HR).

Results: We confirmed prior studies in that Class 2 BRAF mutations (p.K601) were the predominant form of alteration in PC (46%, 63 of 137). CN gains were the second most common BRAF alteration (15%, 21 of 137). Patients with BRAF Class 1, 2, 3 mutations demonstrated no significant difference in OS. However, metastatic PC patients with BRAF CN gain had worse OS relative to BRAF-wt (HR: 2.4, CI 95% 1.18-4.85, p=0.01). Among BRAF-wt PCs from metastatic tissues, those with ultra-low BRAF mRNA expression (bottom 10% within metastatic patients) exhibited worse OS compared to mid BRAF expression (HR:2.43, CI 95% 1.22-4.84, p=0.01). The median OS for PC patients with CN gain and ultralow BRAF gene expression were at 9.4 and 6.6 months, as compared to the wt group of metastatic patients at 15.6 months.

Conclusions: The current paradigm for reporting BRAF status in PC patients requires revision. In addition to SNVs (class 1,2,3), we recommend reporting BRAF CN gain and BRAF mRNA ultralow expression – both may be indicative of a subset of prostatic adenocarcinomas with very poor OS. Future studies should examine the functional consequences of these BRAF perturbations in larger cohorts of PC patients as well as their therapeutic implications.