Analytical performance of a PCR assay for the detection of KRAS mutations (codons 12/13 and 61) in formalin-fixed paraffin-embedded tissue samples of colorectal carcinoma.

KRAS mutation testing is mandatory before prescribing anti-epidermal growth factor monoclonal antibodies in the treatment of advanced colorectal cancer. We describe the performance of a TaqMelt polymerase chain reaction (PCR) assay-the cobas® KRAS Mutation Test-designed to detect 19 mutations in codons 12, 13, and 61. The limit of detection was determined using DNA blends from cell lines, plasmids, and formalin-fixed paraffin-embedded tissue specimens. Assay performance was compared to Sanger sequencing using a panel of 188 specimens. Discordant specimens were subjected to next generation pyrosequencing (454). Assay repeatability was assessed using a panel of six specimens. A >95% correct mutation call rate was obtained in all specimen types with ~5% mutant alleles at DNA inputs of 0.8-6.3 ng per PCR reaction; 100% detection rate was observed at the recommended DNA input of 50 ng. The positive percent agreement with Sanger was 97.5% (79/81) for codons 12/13 and 85.7% (6/7) for codon 61. Negative percent agreement was 94.4% (101/107) for codon 12/13 and 99.4% (180/181) for codon 61. Nine of 10 discordant specimens yielded 454 results consistent with the cobas® results. With repeated testing, the assay showed a correct call rate of 100% (192/192) for all operators, instruments, reagent lots, and days tested. The cobas® test detects KRAS mutations in codons 12, 13, and 61 at a limit of detection of <5%. The PCR assay was more sensitive and specific than Sanger sequencing, and performance was highly reproducible. Test performance was not influenced by various endogenous interfering substances or common gut microbes.