Proliferating cells, especially tumour cells, express a special isoenzyme of pyruvate kinase, termed M2-PK, which can occur in a tetrameric form with a high affinity to its substrate, phosphoenolpyruvate (PEP), and in a dimeric form with a low PEP affinity. In tumour cells, the dimeric form is usually predominant and is therefore termed Tumour M2-PK. The levels of Tumour M2-PK within tumours and in EDTA-plasma correlate with staging and the ability of the tumour cells to metastasise. Since most colorectal tumours grow intraluminally, it appeared interesting to determine whether Tumour M2-PK is detectable in the faeces of tumour patients. Stool samples were tested by ELISA from controls without colorectal cancer and colorectal cancer patients. Whereas Tumour M2-PK levels were low in the control group (mean value7s.e.m.: 3.370.4, n ¼ 144), they were high in the case of colorectal cancer (56.1715.3, n ¼ 60). At a cutoff value of 4 U ml À1 , the sensitivity was 73%. TNM and Dukes' classification of the tumours revealed a strong correlation between faecal Tumour M2-PK levels and staging. The determination of Tumour M2-PK in faeces provides a new promising screening tool for colorectal tumours.

Compared to FOBT or genetic testing the M2-PK test seems to be superior for CRC screening. Concerning handling, effectiveness and analysis, M2-PK seems to be a good possibility for large scale-screening of colorectal carcinoma. It might even be used to detect larger adenomas. Elevated levels of M2-PK in patients with acute and/or chronic inflammatory bowel diseases are probably due to proliferation of epithelial cells and leucocytes in the inflammatory area.

Material and Methods The present study includes 338 patients who underwent complete colonoscopy after providing a sample for the determination of Tumor M2-PK. Stool samples of patients with CRC and patients without patholog ical findings were tested. Histology was obtained from the routine biopsies and/or from surgery. Tumor M2-PK in stool extracts was determined immunologically with a quantitative ELISA which is based on two monoclonal antibodies (ScheBo ® • Biotech AG, Germany). Results Data from 191 controls with no pathological finding upon colonoscopy and 147 patients with CRC have been evaluated to date (Table 1 and Figure 1). There is a highly significant difference (p < 0.001) between tumor patients and controls. At a cutoff point of 4 U/ml, the calculated sensitivity is 85% for colon cancer and 71% for rectal cancer and the specificity is 79%. Fecal Tumor M2-PK levels of 88 patients with CRC were correlated with tumor staging according to TNM (Figure 2) and Dukes classification (Figure 3). There is a significant difference between controls and tumor stages T1 and T2, and a highly significant difference (p < 0.00 1) between controls and tumor stages T3 and T4. Sensitivity for T1, T2, T3, and T4 is 56%, 61%, 82%, and 83%, respectively (Figure 2). Staging according to Dukes' classification revealed a significant difference between controls and Dukes A, and a highly significant difference (p < 0.001) between controls and Dukes B to Dukes D. Sensitivity is 52% for Dukes A, 76% for Dukes B, 81% for Dukes C, and 82% for Dukes D (Figure 3).