3520 Background: Monoclonal antibodies against the epidermal growth factor receptor (EGFR) are standard components of treatment algorithms in metastatic colorectal cancer (mCRC). It is already well established that only patients with wild-type KRAS tumors benefit from treatment with an anti-EGFR agent. Pyrosequencing is now used for a precise determination of KRAS mutation burden and a conservative cutoff of 10% was defined as the lower limit of quantification for the assignment of mutation status. Up to now, the impact of low-signal KRAS mutations below 10% on the response to anti-EGFR therapy in mCRC has not been evaluated. Methods: All consecutive patients treated by anti-EGFR for a mCRC between January 2006 and June 2011 have been retrospectively analyzed by pyrosequencing using the therascreen KRAS Pyro Kit (Qiagen). All patients were defined wild-type (WT) for KRAS status using direct sequencing. The PFS data were plotted as Kaplan–Meier curve and compared by the log-rank test. Results: A total of 141 patients treated by anti-EGFR for a mCRC were included in the study. Mean age was 64.1 ± 14.8 years and a majority of patients had synchronous metastases (68.6%). Patients benefited from anti-EGFR in first-line chemotherapy (30.7%), second-line (22.9%), third-line (35%) or later (11.4%). Majority of patients benefited from anti-EGFR combined with cytotoxic chemotherapy (91.4%), mostly irinotecan (78.6%). Using pyrosequencing, 117 tumors had a KRAS WT status and 24 tumors had low-signal mutation, between 2 and 10% (KRAS lowMT). Response rate according RECIST criteria were 13.6% versus 35.4% partial response, 13.6% versus 37.2% stabilization and 72.7% versus 27.1% progression (p<0.01), for KRAS lowMT versus KRAS WT respectively. The PFS was respectively 2.6 ± 0.2 months for KRAS lowMT versus 6.0 ± 0.5 months for KRAS WT (p=0.024). Overall survival was not different between the two groups (27.4 months versus 33.0 months, p=0.4). Conclusions: Patients with tumors harboring KRAS lowMT benefit less of anti-EGFR therapy than patients with tumor harboring KRAS WT. While these results invite to consider low-signal tumors as positives, generalization awaits a large prospective trial.