Clinical features and mutation analysis of class 1/2/3 BRAF mutation colorectal cancer.

BACKGROUND: BRAF (B-Raf proto-oncogene, serine/threonine kinase)-mutated colorectal cancer (CRC) still has poor prognostic. The efficacy of BRAF inhibitor is unpredictable just that intrinsic genetic complexity, immune microenvironment and partially unknown reason. Understanding the co-mutation mechanism can help improve treatment and follow-up strategies.

METHODS: We retrospectively analyzed 35 (BRAF-mutated/BRAF wild-type) Chinese CRC and 125 Western CRC who underwent next-generation sequencing (NGS). Co-occurrence mutation analysis, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was enabled in this study.

RESULTS: Thirty-five (10.32%) patients were BRAF-mutated, with 17 patients were BRAF V600E in Beijing Hospital. Patients with BRAF mutation had significant association with high tumor mutational burden (TMB-H) (P=0.0004) and high microsatellite instability (MSI-H) (P=0.0003) than those with BRAF wildtype. In 125 BRAF-mutated Western CRC patients, the frequency of age at diagnosis, gender, sample type, Tumor-Node-Metastasis (TNM), MSI, TMB, and BRAF mutation type was consistent with Chinese data. However, the primary tumor location showed significant statistical differences (P<0.0001). Class 1 were more likely to occur in elder and female. Western cohort was consistent with above in Chinese cohort. Other clinicopathological features were not significantly associated with mutation type. However, Western cohort showed class 1 exhibited primary sample type predominance in both class 1 vs. others (P<0.05) and class 1 vs. class 3 (P<0.05). Meanwhile, the data showed TMB-H (57.69% vs. 11.76%, P<0.001) and MSI-H (28.21% vs. 0%, P<0.05) of the class 1 BRAF mutation proportion were significantly higher, compared with class 3 BRAF mutation. In concurrent oncogenic mutations, compared with non-class 1 BRAF mutation, class 1 are more likely to co-occur with passenger mutation. Data from Western populations showed similar results. We also found that the class 1 mutation was mutually exclusive with co-KRAS (Kirsten rat sarcoma viral oncogene homologue) mutation in CRC, and co-APC (APC regulator of WNT signaling pathway) mutation appeared more frequently in non-class 1 BRAF mutation. KEGG pathway showed that fewer protocancer signaling pathways were enriched in the class 1, which further confirmed that this type had stronger tumorigenicity. GO enrichment also proved that class 1 had stronger tumorigenicity. Finally, prognostic analysis showed median overall survival (mOS) of 19.43 months in class 1 vs. 47.57 months in non-class 1 (P=0.0002). Further study showed that the mOS of class 1, class 2, class 3 and class NA (unknown) was 19.43, 28.50, 47.57 months and not reached (P=0.0001), respectively.

CONCLUSIONS: This study showed class 1/non-class 1 BRAF mutation in CRC had significantly differences in co-mutation features, genomic markers and prognostic. Understanding BRAF mutation types and comutation mechanism will contribute to accurately grasping treatment and follow-up strategies and promoting the development of precision therapy for CRC in the future.