LBA-3 Integrated analysis of cell-free DNA (cfDNA) BRAF mutant allele fraction (MAF) and whole exome sequencing in BRAFV600E metastatic colorectal cancer (mCRC) treated with BRAF-antiEGFR +/- MEK inhibitors

Abstract

The combination of encorafenib and cetuximab has led to a paradigm change in BRAFV600E mCRC with significant improved outcomes over the standard of care. Our group previously reported that BRAFV600E mCRC patients treated with BRAF/EGFR+/-MEK inhibitors could be split into three different prognostic subgroups according to tumor load surrogates. Here, we integrate the analysis of BRAFV600E MAF in cell-free DNA (cfDNA) and whole exome sequencing data to unveil predictive/prognostic profiles. An international, prospective, multicentric cohort included 59 BRAFV600E mutated mCRC patients from Vall d'Hebron and Università della Campania L. Vanvitelli, Naples that received encorafenib-cetuximab +/- binimetinib. Basal plasma samples were extracted from 40 patients and BRAF MAF by ddPCR was performed baseline. Moreover, samples for plasma Whole Exome Sequence (plWES) were also extracted baseline from 23 patients. All patients received BRAF/EGFR+/-MEK inhibitors. Responses were evaluated every 8 weeks. The primary endpoint is to correlate BRAF MAF with overall survival and overall response rate. Secondary endpoints included the identification of a WES driving signature that could predict clinical outcomes. Event rates over time will be estimated within each treatment group using the Kaplan-Meier method. In total, 70 patients were included in the analysis. Median age was 61y (33-83), 56% female, 34% right-sided tumors, and 56% received 2 or more prior chemotherapy lines. Overall, median PFS was 5.2 months (m) (CI 95% 4.1-8.5) and median OS (mOS) was 10.3 m (CI 95% 6.5-20.2). BRAF MAF was >5% in 14 pts (35%), <5% in 10 pts (25%), and undetectable in 16 pts (40%) with a mOS of 4.2 m, 17.1 m (HR 0.21, p<0.001), and 17.5 m (HR 0.15, p<0.001), respectively. In the multivariate analysis the most parsimonious model included five factors: ECOG, presence of liver metastases, CEA, NLR levels, and BRAF MAF. We stratified patients into three risk prognostic groups based on their number of presenting risk factors. These three prognostic groups showed differentiated OS outcomes, with a median OS of 5.1 m, 8.4 m (HR 0.26, p<0.001), and 21.8 m (HR 0.05, p<0.005), respectively. Added to this, WES analysis showed a clinical benefit of targeted therapy in those tumors with a BRAF/EGFR gain as well as an enrichment of response in RNF43 mutated tumors. Our preliminary data suggest that cfDNA BRAF MAF may constitute a significant tumor load surrogate with correlation with overall survival. cfDNA BRAF MAF and WES could help to identify three subgroups of patients with different prognostic and predictive features that could potentially have therapeutic implications.