AIM: To review the risk of proximal colon cancer in patients undergoing colonoscopy. METHODS: We estimated the risk of advanced proximal adenomas and cancers in 6 196 consecutive patients that underwent colonoscopy (mean age 60 years, 65% males, without prior history of colorectal examination). Neoplasms were classified as diminutive adenoma (5 mm or less), small adenoma (6-9 mm), advanced adenoma (10 mm or more, with villous component or high-grade dysplasia) and cancer (invasive adenocarcinoma). The sites of neoplasms were defined as rectosigmoid (rectum and sigmoid colon) and proximal colon (from cecum to descending colon). RESULTS: The trend of the prevalence of advanced proximal adenoma was to increase with severe rectosigmoid findings, while the prevalence of proximal colon cancer did not increase with severe rectosigmoid findings. Among the 157 patients with proximal colon cancer, 74% had no neoplasm in the rectosigmoid colon. Multivariate logistic-regression analysis revealed that age was the main predictor of proximal colon cancer and existence of rectosigmoid adenoma was not a predictor of proximal colon cancer. CONCLUSION: Sigmoidoscopy is inadequate for colorectal cancer screening, especially in older populations. METHODS: We estimated the risk of advanced proximal adenomas and cancers in 6 196 consecutive patients that underwent colonoscopy (mean age 60 years, 65% males, without prior history of colorectal Neoplasms were classified as diminutive adenoma (5 mm or less), small adenoma (6-9 mm), advanced adenoma (10 mm or more, with villous component or high-grade dysplasia) and cancer (invasive adenocarcinoma). The sites of neoplasms were defined as rectosigmoid (rectum and sigmoid colon) and proximal colon (from cecum to descending colon). RESULTS: The trend of the prevalence of advanced proximal adenoma was to increase with severe rectosigmoid findings, while the prevalence of proximal colon cancer did the increase with severe rectosigmoid findings. Among the 157 patients with proximal colon cancer, 74% had no neoplasm in the rectosigmoid colon. Multivariate logistic-regression analysis revealed t hat age was the main predictor of proximal colon cancer and existence of rectosigmoid adenoma was not a predictor of proximal colon cancer. CONCLUSION: Sigmoidoscopy is inadequate for colorectal cancer screening, especially in older populations.