ifferent paraffin embedded carcinoma tissues with immunohistochemistry. PCR was applied to amplify the mycoplasma DNA from the positive samples for confirming immunohistochemistry. RESULTS Fifty of 90 cases (56%) of gastric carcinoma were positive for mycoplasma hyorhinis. In other gastric diseases, the mycoplasma infection ratio was 28% (18/49) in chronic superficial gastritis, 30% (14/46) in gastric ulcer and 37% (18/49) in intestinal metaplasia. The difference is significant with gastric cancer (χ2=12.06, P<0.05). In colon carcinoma, the mycoplasma infection ratio was 55.1% (32/58),but it was 20.9% (10/49) in adenomarous polyp (χ2 = 13.46, P<0.005). Gastric and colon cancers with high differentiation had a higher mycoplasma infection ratio than those with low differentiation (P<0.05). Mycoplasma infection in esophageal cancer, lung cancer, breast cancer and glioma was 50.9% (27/53), 52.6% (31/59), 39.7% (25/63)and 41% (38/91), respectively. The mycoplasma DNA was successfully amplified with the DNA extracted from the cancer tissues that were positive for mycoplasma infection (detected with antibody PD4). CONCLUSION There was high correlation between mycoplasma infection and different cancers, which suggests the possibility of an association between the two. The mechanism involved in oncogenesis by mycoplasma remains unknown.