Stool testing based on tumour-derived markers might offer a promising approach for non-invasive colorectal cancer (CRC) screening. The aim of this study was to estimate the potential of a new test for faecal tumour M2-PK to discriminate patients with CRC from a large sample of unselected older adults. Faecal tumour M2-PK concentrations were determined in 65 CRC patients and in a population-based sample of 917 older adults (median age: 65 and 62 years, respectively). Sensitivity and specificity of the test were calculated at different cutoff values, and receiver-operating characteristic curves (ROC) were constructed to visualise the discriminatory power of the test. The median (interquartile range) faecal tumour M2-PK concentration was 8.6 U ml À1 (2.8 -18.0) among CRC patients and o2 U ml À1 (o2 -3.2; Po0.0001) in the population sample. At a cutoff value of 4 U ml À1 , sensitivity (95% confidence interval) was 85% (65 -96%) for colon cancer and 56% (41 -74%) for rectum cancer. Specificity (95% confidence interval) was estimated to be 79% (76 -81%). Given the comparatively high sensitivity of the tumour M2-PK stool test (especially for colon cancer) and its simple analysis, the potential use of the test for early detection of CRC merits further investigation. Possibilities to enhance specificity of the test should be explored. Colorectal cancer (CRC) is one of the leading causes of cancerrelated morbidity and mortality worldwide (Ferlay et al, 2004). Even in countries where up-to-date therapeutic options are available, more than 40% of CRC patients still die from the disease within 5 years after diagnosis (Brenner, 2002;Brenner et al, 2005). As survival is considerably better for early, localised CRC than for later stages and CRC might be prevented altogether by detection and removal of precancerous lesions, enhanced screening will be of crucial importance for further progress in reducing the burden of the disease.Experience with faecal occult blood testing (FOBT) has shown the possibility of reducing both incidence and mortality owing to screening based on stool tests (Hardcastle et al, 1996;Mandel et al, 1999Mandel et al, , 2000 and the advantages of this screening modality regarding acceptability and practicality compared with invasive screening options. However, given the limitations of FOBT, mainly the inherently low sensitivity to detect precancerous or cancerous lesions, there is still need for improvement. Therefore, the development of stool tests with better performance characteristics has become a focus of current research, both in the field of DNAbased stool markers and in the field of protein-based stool markers. For some of the new tests, preliminary results regarding sensitivity and specificity are encouraging but evidence of the performance of the new methods is mostly restricted to small-scale investigations in the clinical setting (Haug and Brenner, 2005a).A novel approach based on the detection of proteins in stool derived from neoplastic colonocytes is the quantitative measurement of faecal tum...