RAS testing of circulating cell-free DNA (cfDNA) from tissue RAS/BRAF wild-type metastatic colorectal carcinoma (mCRC) patients: Preliminary data from the Liquid Biopsy Monoclonal Antibodies in mCRC (LIBImAb) study

3046 Background: In mCRC patients RAS/BRAF testing is usually performed on tissue biopsy in order to select patients for treatment with anti-EGFR monoclonal antibodies. Several studies described discordant RAS status between tissue and cfDNA testing from the same patients. In addition, up to 30% of patients with secondary resistance to anti-EGFR therapies become RAS mutant at progression. The LIBImAb study is a phase III, randomized, open-label, comparative, multi-centre trial to assess the superiority in terms of efficacy of bevacizumab versus cetuximab in combination with FOLFIRI chemotherapy in patients with mCRC, RAS/BRAF wild-type on tumor tissue and RAS mutant (RAS mut ) at liquid biopsy. This study is supported by the Italian Drug Agency (AIFA). We report preliminary data of liquid biopsy testing. Methods: Plasma samples from enrolled patients are analyzed for variants in exons 2, 3 and 4 of KRAS and NRAS genes and for codon 600 BRAF mutations using the fully automated Idylla system (Biocartis). Patients RAS mut on liquid biopsy at baseline are randomized to receive FOLFIRI plus Cetuximab or FOLFIRI plus Bevacizumab. Patients RAS wild-type on cfDNA at baseline, are treated with FOLFIRI plus Cetuximab up to 8 cycles and, if not progressed, they are retested for RAS mutations on cfDNA. RAS mut patients at re-screening are randomized to continue Cetuximab or to switch to Bevacizumab. Results: Plasma samples from 169 tissue RAS/BRAF wild-type mCRC patients at baseline and 75 patients at 8 cycles of treatment have been tested as of January 31, 2023. Analysis of baseline plasma samples detected 16 RAS mutations in 16/169 patients (9.5%). In particular, 12 KRAS variants (7.1%, 8 in exon 2 and 4 in exon 3) and 4 NRAS variants (2.3%, all in exon 2) were found. The Idylla test also detected 3 BRAF V600 mutations (1.8%) in 3 additional patients. The overall concordance between tissue and cfDNA testing for RAS/BRAF variants was 88.8%. RAS mutations were also detected in 6/75 (8%) plasma samples obtained at 8 cycles of FOLFIRI/cetuximab treatment. The cfDNA test revealed the presence of 5 KRAS (6.7%, 3 in exon 2 and 1 in exon 3) and 1 NRAS mutation (1.3%, exon 2). A BRAF V600 variant was detected in 1 additional patient (1.3%). The rate of RAS/BRAF mutant cases at 8 cycles of treatment was 9.3%. Conclusions: These preliminary data suggest that liquid biopsy might better recapitulate the heterogeneity of mCRC and might be useful in clinical practice to complement tissue testing. The clinical relevance of RAS variants detected in cfDNA will be determined in the LIBImAb trial.