Atypical non-V600E BRAF (aBRAF) mutations as a prognostic and predictive factor in real-life metastatic colorectal cancer patients from the Nordic countries

Approximately 2% of metastatic colorectal cancers (mCRC) have atypical BRAF mutations (aBRAFmt), i.e. non-V600E. Mutations in KRAS/NRAS (RASmt) and BRAF (BRAF V600E mt) are negative prognostic and predictive biomarkers for EGFR monoclonal antibodies (mAbs), restricting their use to RAS/BRAF wildtypes (wt). Little is known about aBRAFmt in real-life mCRC patients. A selected patient series reported median overall survival (mOS) >60 months (Jones et al JCO 2017). The prospective real-life Finnish RAXO and Scandinavian population-based PRCRC studies were combined. RAS and BRAF analyses were performed in 1409/1912 tumours. Forty-two had aBRAFmt, 182 BRAF V600E mt, 712 RASmt and 473 were RAS/BRAFwt. Progression-free survival (PFS) and OS were estimated with Kaplan-Meier and compared with log-rank and COX-regression analysis. Twenty different aBRAF mutations were seen (most common D594G/N, G466E, G496R). They were class II (signaling as dimers) in 10 (23%), III (kinase activity low/none) in 28 (67%) and unclassified in 4 (10%). aBRAFmt was not concomitant with MSI-H (0% of 27 tested) but was with BRAF V600E mt (3%), KRASmt (28%) and NRASmt (8%). Median age was 67 years (range 23-97) without statistical differences. aBRAFmt was male predominant (aBRAFmt 67% vs BRAF V600E mt 39%, RASmt 58%, RAS/BRAFwt 63%). Baseline ECOG was worse (PS 2-3 in 33% vs 35/19/23%). Primary tumour location was rectal predominant in aBRAFmt (50%), right colon in BRAF V600E mt (70%), left colon in RASmt (34%) and RAS/BRAFwt (45%). Presentation was predominantly synchronous (aBRAFmt 67% vs BRAF V600E mt 59%, RASmt 64%, RAS/BRAFwt 57%). Metastatic sites differed: liver 60% vs 55/72/75%, lung 40% vs 23/34/25%, lymph nodes 38% vs 40/24/29%, peritoneal10% vs 27/16/14% and local relapse 2% vs 9/4/6%. OS was 14.5 months vs 11.2/24.0/30.1 months (p 95% 1.12-2.35] vs 2.72 [2.25-3.28]/1.18 [1.03-1.35]/reference). First-line mPFS was 9.7 months vs 6.0/10.3/12.0 months (p V600E mt [12.5 vs 19.1], RASmt [23.4 vs 29.1] and RAS/BRAFwt [27.4 vs. 35.5]). Five in the aBRAFmt group received later line EGFRmAb without any responses: SD in 2 (class II, II) and PD in 3 (class II, III, III). Metastasectomy was performed in 19% vs 9/30/32%, resulting in prolonged survival (aBRAFmt [mOS 14.1 vs 62.0 months], BRAF V600E mt [10.7 vs 28.5], RASmt [18.3 vs 68.3] and RAS/BRAFwt [19.9 vs. 82.8]). In multivariable analysis the strongest factors predicting shorter OS were ECOG 2-3 (HR 2.55;CI 95% 2.22-2.94), BRAFV600Emt (2.12;1.73-2.61), right-sided primary (1.36;1.19-1.56), aBRAFmt (1.36;0.94-1.97), age >70 (1.32;1.16-1.49); whereas synchronous presentation (1.06;0.94-1.20) and female sex (0.99;0.88-1.12) were not predictive. aBRAFmt represents a distinct subgroup regarding gender, primary location, metastatic sites, and MSI-H. Performance status was worse and less oncologic therapy was given than in BRAFwt. aBRAFmt shows shorter OS and PFS compared with BRAFwt, contrary to Jones et al, but slightly better than BRAF V600E mt. The addition of bevacizumab may prolong survival in aBRAFmt, but aBRAFmt may be a negative predictive marker for EGFRmAb efficacy. aBRAF should be tested.