Frequency and outcomes of BRAF alterations identified by liquid biopsy in metastatic non-colorectal gastrointestinal cancers.

808 Background: The impact of BRAF V600E, a known poor prognostic factor in metastatic colorectal cancer (CRC), has not been assessed broadly in non-CRC gastrointestinal (GI) cancers including pancreatic (PC), gastric/gastroesophageal (GC), hepatocellular carcinoma (HCC) and cholangiocarcinoma (CCA). In view of recent tumor-agnostic approval of a BRAF V600E inhibitor, it is important to determine the prevalence and impact of BRAF alterations on non-CRC GI cancer patients (pts). Methods: Genomic results from pts with PC, GC, HCC, or CCA tested with Guardant360 (G360, Guardant Health) as part of clinical care from June 2019 – June 2022 were retrospectively queried. Prevalence of characterized nonsynonymous BRAF genomic alterations (GA) was analyzed; differences between cancer types was assessed using the Chi-squared test. The GuardantINFORM database with aggregated commercial payer health claims and de-identified records from over 243,000 individuals with G360 results was used to derive real-world overall survival (rwOS), reported in months, calculated from diagnosis of metastatic cancer; differences between cancer types were compared using log-rank tests. Results: 23,069 patients were included: PC:49%; GC:24%; CCA:20%; HCC:6%. Pts were more likely to be male (p<0.01). PC/HCC pts were tested more often at new diagnosis (vs at disease progression) than GC/CCA (p=0.02, p<0.01, respectively). BRAF GAs were identified in 2.9%, comparable to tissue-based datasets (MSK-IMPACT: 2.4%); 21% of BRAF GAs were V600E (0.6% prevalence overall). PC/CCA pts were more likely to have BRAF GAs than the overall cohort (p<0.01); they also had more BRAF V600E GAs (p<0.01). rwOS differed by cancer type and BRAF status (Table). Notably, CCA pts with BRAF V600E had superior rwOS compared to CCA pts without BRAF findings (p=0.02). rwOS is not reported for HCC due to small n. Conclusions: Identification of BRAF GAs and V600E in the non-CRC GI population by liquid biopsy is similar to rates observed in tissue-based testing and may be reliably used to assess BRAF status. BRAF GAs and subtypes have mixed prognostic implications on survival for PC/CCA/GC patients that warrants further exploration. [Table: see text]