Real-world data study of BRAF mutant metastatic colorectal cancer patients across the Greater Manchester region prior to BEACON trial results

Abstract

Background BRAF mutations are known to impact both prognosis and anti-EGFR response in metastatic colorectal cancer (mCRC). Methods All consecutive pts (n=80) with BRAF mutation from 2016 to 2018 were included in this real-world data (RWD) study. Data were obtained from electronic patient records. Survival univariate analysis (UVA) was performed with Kaplan-Meier curves and log-rank test. Multivariable survival analysis (MVA) was performed by Cox regression. Results Median age for the 80 pts was 68y (range 32 to 82). Median follow up was 10.2 months (ms). Females 36 pts (45%) and males 44 (55%). Anatomically, 42.5% were ascending colon, 13.7% transverse, 21.9% sigmoid and 17.8% rectum; in summary, right colon tumours accounted for 51.5%. Seventy (87.2%) pts had BRAF V600E mutation. KRAS mutation was present in 7 cases (8.8%), 4 (5%) were co-expressed with BRAF V600E mutations. Sixteen (20%) had a PIK3CA mutation. The most common first-line chemotherapy backbone was FOLFOX (41.5%), followed by FOLFIRI (40%). Only 12 pts (15%) received anti-EGFR antibodies, 9 of them with BRAF V600E mutations. The median variant allele frequency for BRAF variants was 22%, ranging from 2 to 56%. Responses by RECIST criteria to first line were, complete response (CR) in 4 (6.5%) out of 61 evaluable patients, partial response (PR) in 12 (19.7%), stable disease (SD) in 16 (26.2%) and progressive disease (PD) in 29 (47.6%). Overall survival (OS) for the full cohort was 11.4 ms 95%CI (9.2-13.7). First-line progression-free survival (PFS) for the full cohort was 5.3 ms 95%CI (3.7-6.9). In the UVA for OS, sidedness was statistically significant (p=0.048) with right-sided tumours having a median OS of 9.7 ms 95%CI (5.6-13.7) vs left-sided 13.6 ms 95%CI (5.4-21.8). However, when the analysis was stratified by BRAF mutation, the difference was not statistically significant (p=0.204); a trend was observed for the non-V600E (p=0.058). No differences in OS were found regarding the chemotherapy backbone or anti-EGFR antibodies. In the first line, responders had a better OS of 14.8 ms (95%CI 0.0-30.2) vs the OS of non-responders which was 9.2 ms (95%CI 6.9-11.4, p=0.005). There were no differences regarding baseline ECOG. In the UVA for PFS, responders had a better PFS 9.7 ms (95%CI 5.0-14.3) vs non-responders PFS 3.2 (95%CI 2.5-3.9, p=0.001) and it was the only significant variable. In the MVA analyses for OS, response by RECIST remained as an independent prognostic factor when adjusted for BRAF variant, gender, EGOG and sidedness (HR: 0.6 95%CI 0.4-0.8 with p=0.003). In the MVA analyses for PFS, response by RECIST remained significant when corrected for BRAF variant, gender ECOG and sidedness (HR: 0.6 95%CI 0.4-0.7, p< 0.001). Conclusion In our RWD study, response to treatment was the main independent factor associated with PFS and OS for the first line of treatment. This is in keeping with the utility of chemotherapy triplets in this subgroup.