The evaluation of nutritional status in cancer patients is often neglected in spite of the fact that poor nutritional status may adversely affect prognosis and treatment tolerance. In day-to-day oncology practice, a sensitive but simply applied nutritional assessment tool is needed to identify at-risk patients. Several tools exist; however, none has been universally accepted. The aim of this study was to compare two potential tools, the Mini-Nutritional Assessment (MNA) and the scored Patient Generated Subjective Global Assessment (PGSGA). The MNA is more simply applied and does not require a trained dietitian. The PGSGA has been previously validated in cancer patients. One hundred fifty-seven newly diagnosed cancer patients were assessed using both tools. Of these, 126 were reassessed at 4-6 wk, and 104 were reassessed at Weeks 8-12 after initial assessment. A significant negative correlation was found between the tools at all three time periods (at baseline r = -0.76; P < 0.001). Taking the PGSGA as the most accepted nutritional assessment tool, at baseline the MNA demonstrated a sensitivity of 97% and specificity of 54%. At 4-6 wk MNA sensitivity was 79% and specificity was 69%. At 8-12 wk MNA sensitivity was 93% and specificity was 82%. When comparing the tools in elderly patients alone (>65 yr), similar results were obtained. Both tools were able to correctly classify patients as malnourished, although the MNA lacks specificity. Therefore, the PGSGA should be the tool of choice for nutritional assessment in cancer patients.