The treatment of non-small cell lung cancer (NSCLC) remains unsatisfactory, with most patients succumbing to metastatic disease within 5 years of diagnosis. Improved selection of patients for aggressive adjuvant therapy may contribute to improved survival. Mutation of the k-ras oncogene is considered a potentially clinically useful prognostic biomarker for this purpose. This report presents the results of a meta-analysis performed to determine whether the existing data support such a role for k-ras mutations in NSCLC. Two year survival data derived from 881 NSCLC patients in eight published studies were analyzed using a general variance-based meta-analytical method employing confidence intervals. The outcome of interest was a summary relative risk (RR(s)) reflecting the risk of death at 2 years associated with k-ras mutation-positive versus k-ras mutation-negative disease. Prior to calculation of RR(s), analysis for heterogeneity showed Q to equal 15.52 (7 df, P = 0.03). This indicated heterogeneity across the analyzed studies in terms of their estimate of effect. Possible sources of heterogeneity were identified and included selection bias and various other sources of uncontrolled confounding. Although a RR(s) of 2.35 (95% CI = 1.61-3. 22) was found when all eight studies were combined (favoring a negative prognostic role for k-ras mutation), it is unclear whether the magnitude of the RR(s) would persist after adjusting for other well-established prognostic indicators (e.g. stage). The existing data suggest that k-ras mutation may be associated with shortened survival in NSCLC, although this finding awaits confirmation in well-designed multivariate analyses which adjust for the effect of known prognostic indicators.