The objective of this study was to determine the prognostic information provided by flow cytometric DNA analysis in non-small-cell lung cancer. Lung samples of 132 consecutive patients submitted to surgery were prospectively processed. When no aneuploid populations were detected in fresh frozen samples, the process continued as a second step in paraffin-embedded tissue, consuming all the tumor available. The influence of ploidy on the postoperative outcome was studied by both a univariate and a multivariate analysis. Aneuploidy was found in 81 patients (61.4%). Fourteen patients showed no aneuploidy in fresh frozen samples; and only after further analysis in paraffin-embedded tissue was abnormal DNA detected. Overall, the 36-month survival was 69% for the diploid group and 24% for the aneuploid group (p = 0.0006). Including subjects submitted to complete tumor removal (stages I, II, and IIIA) in a multivariate analysis adjusted for TNM stage and histologic type, bearers of aneuploid tumors exhibited a higher risk of relapse (hazard ratio 2.65; CI 95% 1.5-4.66;p = 0.004) or death (hazard ratio 2.17; CI 95% 1.08-4.39;p = 0.032) than patients with diploid tumors. DNA ploidy resulted an independent prognostic factor of survival and tumor relapse in completely resected non-small-cell lung cancer. Sequential analysis of fresh and paraffin-embedded samples can help avoid the bias due to intratumoral DNA content heterogeneity. DNA ploidy could be an useful parameter in any future multifactorial analysis of outcome in such tumors.