Background: Activation of EGFR pathway in RAS-BRAF wild-type models can be modulated by scaffold proteins kinase suppressor of ras (KSRs) that binds RAS promoting ERK activation. We previously showed in patients with KRAS-BRAF wild-type tumor treated with FOLFIRI-cetuximab that KSR2 rs11068551 TT was associated to significantly longer PFS. On the other hand, among female patients treated with FOLFIRI-bevacizumab the KSR1 rs2241906 CC genotype was associated to significantly shorter PFS. We aimed at validate our finding by investigating the correlation between KSR variants and clinical outcomes in patients enrolled in the FIRE 3 trial. Methods: Genomic DNA was isolated from tissue samples of 592 patients with KRAS exon 2 wild-type enrolled in the randomized phase III, FIRE 3 trial that was designed to compared the efficacy of two targeted agents when associated to FOLFIRI regimen. Patients were randomly assigned to first-line FOLFIRI associated with either bevacizumab (n = 295) or cetuximab (n = 297) (NCT00433927). In both arms, DNA was amplified by PCR and KSR variants were detected by direct sequencing. Results: We confirmed prospectively that KSR2 rs11068551 could predict outcomes in patients receiving cetuximab. 1/Patients with the TT genotype (n = 40) demonstrated a significant lower response rate than patients carrying C (n = 133) in the cetuximab arm (55% vs. 77%, p= 0.008). 2/ Patients with left-sided tumors and TT genotype (n = 32) showed a significant longer PFS (12.9 vs. 10.2 months, log rank p = 0.049) than patients harboring any of the C allele (n = 128) in the cetuximab arm. 3/ No association between KSR2 rs11068551 and outcomes was seen in the bevacizumab arm. We did not confirm the relation between KSR1 rs2241906 and outcomes (response rate, PFS or OS) even when adjusted with gender and location. Conclusions: We prospectively validated the predictive value of KSR2 rs11068551 mCRC pts treated with first-line cetuximab-containing therapy. This predictive value may dependent on tumor location. We could not confirm the predictive value of KSR1 rs2241906 in patients receiving bevacizumab-containing treatment.