Evaluation of circulating tumor DNA as a prognostic and predictive biomarker in BRAF V600E mutated colorectal cancer-results from the FIRE-4.5 study

Abstract

The randomized FIRE-4.5 (AIO KRK0116) trial compared first-line therapy with FOLFOXIRI (folinic acid, fluorouracil, oxaliplatin, and irinotecan) plus either cetuximab or bevacizumab in B-Raf proto-oncogene, serine/threonine kinase (BRAF) V600E-mutant metastatic colorectal cancer (mCRC) patients. This study was accompanied by a prospective translational project analyzing cell-free circulating tumor DNA (ctDNA) in plasma to test whether ctDNA analysis may help to guide clinical treatment decision making. FIRE-4.5 included mCRC patients with BRAF V600E mutation detected by tissue-based analyses. Liquid biopsies (LBs) were collected at baseline (pre-treatment) and during therapy. Digital droplet PCR (ddPCR) technology was applied for determination of BRAF mutations and the in vitro diagnostics (IVD)-certified ONCOBEAM RAS procedure for analysis of RAS mutations. The BRAF V600E variants in ctDNA were analyzable in 66 patients at start of the therapy, at baseline. No BRAF V600E mutations were detected in 26% (17/66) of patients and was associated with a significantly longer progression-free survival (PFS: 13.2 vs 6.5 months; HR 0.47; P = 0.014) and overall survival (OS: 36.8 vs 13.2 months; HR 0.35; P = 0.02) as compared to ctDNA mutant patients. Patients with detectable BRAF mutations showed a clear superiority of FOLFOXIRI plus bevacizumab with regard to PFS (10.4 vs 5.7 months; HR 0.4; P = 0.009) and OS (16.6 vs 11.6 months; HR 0.5; P = 0.15), while this was not the case for BRAF wild-type patients. Follow-up LBs were obtained from 51 patients. Patients converting from BRAF V600E mutant to a BRAF V600 wild-type status (36%, N = 18) had a superior PFS (8.6 vs 2.3 months; P = 0.0002) and OS (17.4 vs 5.1 months; P < 0.0001) compared to patients with stable or increased mutational allele frequency (12%, N = 6). Those patients also achieved a significantly greater disease control rate (89% vs 20%; P = 0.008). In conclusion, LB evaluating ctDNA is informative and may help to guide treatment in patients with BRAF V600E-mutated mCRC.

Keywords: FOLFOXIRI; bevacizumab; cetuximab; circulating tumor DNA; liquid biopsy; mutational load.

Fig. 1

Consort diagram of patients with liquid biopsy subgrouped by mutational allele frequency of BRAF V600E in circulating tumor DNA (ctDNA). White boxes: Baseline liquid biopsy before treatment onset, gray boxes: Follow‐up sample collection during treatment, MAF‐mutational allele frequency (%), cut‐off value of the BRAF V600E MAF was set at 0.17 (see Fig. S2); BL baseline sample; * group of 7 patients with increased MAF in later follow‐up samples were included.

Fig. 2

BRAF V600E status in ctDNA at baseline and patient outcome. (A) Survival curves of patients with BL containing only BRAF V600 wild‐type molecules versus patients with BL containing BRAF V600E mutant (BL mut) variants reveal the better outcome of wild‐type BL patients (BL wt). (B) Disease control rate (DCR) and objective response rate (ORR) for patients grouped by ctDNA BRAF mutational status of baseline samples (BL). ctDNA, circulating tumor DNA; CET, FOLFOXIRI plus cetuximab; BEV, FOLFOXIRI plus bevacizumab; PFS, progression‐free survival; OS, overall survival; HR, Hazard ratio with 95% confidence interval (95%CI) was calculated with Kaplan–Meier method, P‐value by log‐rank test; OR, Odds ratio calculated by Fisher's exact test. Dotted line indicates 50% probability of survival.

Fig. 3

Comparison of treatment efficiency in patient groups classified by baseline liquid biopsy results. Survival curves of patients with liquid biopsy samples (A) with BL mut N = 49 (B) versus BL wt N = 17 (C). In the PFS time, a significant difference between the treatment arms could be measured: bevacizumab treatment seemed more beneficial for BL mutant group. In contrast, the efficiencies of both treatments were almost the same in the BL wild‐type patient group. BL, Baseline liquid biopsy before treatment onset; CET, FOLFOXIRI plus cetuximab; BEV, FOLFOXIRI plus bevacizumab; PFS, progression‐free survival; OS, overall survival; HR, Hazard ratio with 95% confidence interval (95%CI) and P values were calculated with Kaplan–Meier method and log‐rank test; Dotted line indicates 50% probability of survival.

Fig. 4

Grouping patients in respect to changes of mutational allele frequency (MAF %) in follow‐up liquid biopsies. The patients with initial BRAF V600E mutant ctDNA status and with increased MAF had the worst PFS and OS (A,B). On the other hand, the reduction in MAF is an indicator of good PFS and OS in response to treatment (C). OS, overall survival; PFS, progression‐free survival; P‐value were calculated by log‐rank test, except⁺. ⁺ P‐value was additionally calculated by Gehan‐Breslow‐Wilcoxon test, log‐rank P = 0.053 for this comparison. Dotted line indicates 50% probability of survival.