Background:An everyday clinical practice dilemma in the 20–30% of metastatic colorectal cancer (CRC) patients that have not been operated on their primary tumour, is, under which specific histopathology and molecular circumstances, an endoscopic biopsy could be considered adequate to provide a representative RAS/BRAF molecular status to guide treatment.Methods:A consecutive series of 193 paired biopsy and primary CRC tumour samples between August 2008 and 2010 available in the Department of Pathology archives, University Hospitals, KU Leuven were retrieved. For a pair to be included, in the endoscopic biopsy, 20% of invasive adenocarcinoma cells should be present and enough slides to yield an extracted DNA concentration of ⩾5 ng μl−1, and no <2 ng μl−1 should be available for cutting. Exons 2–4 KRAS/NRAS, BRAF, PIK3CA molecular evaluation was performed with RT–PCR and Sequenom.Results:From 165 deemed adequate by the pathologist pairs, 85 (51.5%) were concordantly mutated in at least one of the tested genes, 70 (42.5%) were wt and 10 (6%) were discordant, harbouring a mutation in the primary and not in the endoscopic biopsy. In the re-evaluation, when more slides were cut per discordant pair, mutational status changed in two of the six discordantly KRAS-mutated pairs. A strong strength of agreement for both runs was observed (Cohen's kappa, k=0.877, P<0.001 and k=0.901, P<0.001, respectively) between the surgically acquired and the endoscopic biopsy specimens' evaluation.Conclusions:Based on our results, an endoscopic biopsy could provide an accurate mutational profile and become a justified alternative to a surgically removed primary tumour specimen, as long as specific histopathology criteria are met.