Background : Tumor staging insufficiently discriminates between colon cancer patients with poor and better prognosis. We have evaluated, for the primary tumor, if the carcinoma‐percentage (CP), as a derivative from the carcinoma‐stromal ratio, can be applied as a candidate marker to identify patients for adjuvant therapy. Methods : In a retrospective study of 63 patients with colon cancer (stage I–III, 1990–2001) the carcinoma‐percentage of the primary tumor was estimated on routine H&E stained histological sections. Additionally these findings were validated in a second independent study of 59 patients (stage I–III, 1980–1992). (None of the patients had received preoperative chemo‐ or radiation therapy nor adjuvant chemotherapy.) Results : Of 122 analyzed patients 33 (27.0%) had a low CP and 89 (73.0%) a high CP. The analysis of mean survival revealed: overall‐survival (OS) 2.13 years, disease‐free‐ survival (DFS) 1.51 years for CP‐low and OS 7.36 years, DFS 6.89 years for CP‐high. Five‐year survival rates for CP‐low versus CP‐high were respectively for OS: 15.2% and 73.0% and for DFS: 12.1% and 67.4%. High levels of significance were found (OS p < 0.0001, DFS p < 0.0001) with hazard ratio’s of 3.73 and 4.18. In a multivariate Cox regression analysis, CP remained an independent variable when adjusted for either stage or for tumor status and lymph‐node status (OSp < 0.001, OSp < 0.001). Conclusions : The carcinoma‐percentage in primary colon cancer is a factor to discriminate between patients with a poor and a better outcome of disease. This parameter is already available upon routine histological investigation and can, in addition to the TNM classification, be a candidate marker to further stratify into more individual risk groups.