Last update:
02-Oct-2011

OBJECTIVE The aim of this study was to carry out a multivariate meta-analysis of studies which investigated the predictive value of KRAS mutations in the efficacy of cetuximab in patients suffering from colorectal cancer.

METHOD A systematic literature search was made in the PubMed, Medline and Cochrane databases using a combination of the key words "cetuximab", "anti-egfr" "KRAS or K-RAS" and "colorectal cancer". A 2-dimensional random-effects model was used for the analysis of sensitivity and specificity. Data from each study was entered in 2×2 tables from which sensitivities, specificities and predictive values (negative and positive) were estimated. The statistical program Stata 10 (StataCorp) was used for all the statistical analyses and significant results were considered those with p-value <0.05.

RESULTS A total of 26 relevant reports were found from the search of the databases, of which 13 were appropriate for this specific meta-analysis. Initially, it was found that in these studies the specificities were much higher than the sensitivities. On combining the data from the 13 studies, it was found that mutations in the KRAS gene constitute a negative predictive biomarker for the response to cetuximab, with very high specificity, 0.96 (0.84−0.99), but low sensitivity, 0.47 (0.43−0.50), and that publication bias may often occur.

CONCLUSIONS The clinical importance of these findings is that cetuximab should be administrated only to patients who have the wild type KRASw oncogene. Mutations in the KRAS gene are a negative predictive factor for the response to cetuximab, with very high specificity and low sensitivity. The low value of sensitivity is probably due to the presence of additional mechanisms of resistance to anti-EGFR therapies, such as mutations in BRAF.

Key words: Anti-EGFR, Cetuximab, Colorectal cancer, Efficacy, KRAS or K-RAS.