Epithelial growth factor receptor (EGFR) and KRAS mutation status have been reported as predictive markers of tumour response to EGFR inhibitors. High resolution melting (HRM) analysis is an attractive screening method for the detection of both known and unknown mutations as it is rapid to set up and inexpensive to operate. However, up to now it has not been fully validated for clinical samples when formalin-fixed paraffin-embedded (FFPE) sections are the only material available for analysis as is often the case.We developed HRM assays, optimised for the analysis of FFPE tissues, to detect somatic mutations in EGFR exons 18 to 21. We performed HRM analysis for EGFR and KRAS on DNA isolated from a panel of 200 non-small cell lung cancer (NSCLC) samples derived from FFPE tissues.All 73 samples that harboured EGFR mutations previously identified by sequencing were correctly identified by HRM, giving 100% sensitivity with 90% specificity. Twenty five samples were positive by HRM for KRAS exon 2 mutations. Sequencing of these 25 samples confirmed the presence of codon 12 or 13 mutations. EGFR and KRAS mutations were mutually exclusive.This is the first extensive validation of HRM on FFPE samples using the detection of EGFR exons 18 to 21 mutations and KRAS exon 2 mutations. Our results demonstrate the utility of HRM analysis for the detection of somatic EGFR and KRAS mutations in clinical samples and for screening of samples prior to sequencing. We estimate that by using HRM as a screening method, the number of sequencing reactions needed for EGFR and KRAS mutation detection can be reduced by up to 80% and thus result in substantial time and cost savings.