KRAS mutation has been unambiguously identified as a marker of resistance to cetuximab-based treatment in metastatic colorectal cancer (mCRC) patients. However, most studies of KRAS mutation analysis have been performed using homogenously archived CRC specimens, and studies that compare freshly frozen specimens and formalin-fixed paraffin-embedded (FFPE) specimens of CRC are lacking. The aim of the present study was to evaluate the impact of tissue preservation on the determination of KRAS mutational status. A series of 131 mCRC fresh-frozen tissues were first analyzed using both high-resolution melting (HRM) and direct sequencing. KRAS mutations were found in 47/131 (35.8%) using both approaches. Out of the 47 samples that were positive for KRAS mutations, 33 had available matched FFPE specimens. Using HRM, 2/33 (6%) demonstrated suboptimal template amplification, and 2/33 (6%) expressed an erroneous wild-type KRAS profile. Using direct sequencing, 6/33 (18.1%) displayed a wild-type KRAS status, and 3/33 (9.1%) showed discordant mutations. Finally, the detection of KRAS mutations was lower among the FFPE samples compared with the freshly frozen samples, demonstrating that tissue processing clearly impacts the accuracy of KRAS genotyping.