Monoclonal antibodies directed against the EGF-receptor (EGFR) have recently been approved for the treatment of metastatic colorectal cancer (CRC) patients with EGFR-positive tumors at immunohistochemistry (IHC). Surprisingly, data demonstrate a lack of correlation between the tumor's EGFR expression at IHC and outcome. Indeed, as pointed out from small experiences, patients with EGFR-IHC-negative metastatic CRC have the same chance as EGFR-IHC-positive patients to benefit from an anti-EGFR monoclonal antibody, underlying the importance of different, more reliable, selection criteria. In particular, the identification of such predictive factors is important as these agents are expensive, have side effects and are really only effective in a minority of patients. Several potential clinical and biological predictive markers of activity and/or efficacy for such agents have been evaluated in retrospective series with promising results. Moving from clinical data suggesting that there could be a subpopulation of CRC patients that are more liable to benefit from anti-EGFR monoclonal antibodies, here we review major studies on determinants of outcome in this field.