Background: The impact of KRAS mutation status in mCRC treated with anti-VEGF therapy is controversial. The randomized phase 3 CAIRO3 study showed benefit of first-line CAP-B maintenance treatment compared to observation in mCRC patients (pts) with stable disease or better after 6 cycles CAPOX-B. We investigated the predictive value of KRAS mutation status within the CAIRO3 study. Methods: We extracted tumor DNA from paraffin blocks and used Next Generation Sequencing (NGS) to detect mutations in KRAS exon 2-4. We supplemented this with KRAS data collected from local laboratories. We used stratified Cox regression models adjusted for baseline covariables to estimate hazard ratios (HR) and to analyze whether treatment effect was modified by KRAS mutation status. Time to first progression (PFS1), time to second progression after CAPOX-B reintroduction (PFS2, primary endpoint), time to second progression on any treatment (TT2PD) and overall survival (OS) were analyzed. Results:KRAS mutation status was available from 448/557 pts (80%), determined by NGS (n = 317) or already available data (n = 131). Tumors were KRAS wild-type (wt) in 240 (54%) and mutant (mt) in 208 (46%) pts. Pts with KRAS wt tumors had significantly more benefit from CAP-B maintenance treatment compared to observation than pts with KRAS mt tumors for all endpoints, except for PFS1. Pts with KRASmt tumors showed significant benefit from maintenance treatment only in PFS1 and PFS2 (Table). Conclusions: Our data suggest that in the first-line treatment of mCRC, pts with KRAS wt tumors benefit most from CAP-B maintenance treatment compared to observation after 6 cycles CAPOX-B. We are currently analyzing the predictive value of RAS, BRAF and mismatch repair status. Clinical trial information: NCT00442637