Screening/early detection model for colorectal cancer. Why screen?

Publisher Website
Google Scholar
Abstract
Colon cancer is a leading cause of cancer death in the United States, causing approximately 60,000 deaths each year. Ideal screening would identify high risk patients and screen them with sensitive tests. Most cancers evolve from adenomatous polyps. There is now evidence that detection and removal of adenomas can prevent cancer. Unfortunately, our ability to identify high risk patients is limited. Screening of asymptomatic, average-risk individuals has been advocated, with the goal of reducing colon cancer mortality by detecting cancers at an early, curable stage or preventing cancer by detecting and removing adenomatous polyps. Recent data have suggested that screening populations older than age 50 with sigmoidoscopy and fecal occult blood tests can reduce colon cancer mortality. These reports are encouraging but also highlight significant limitations of this form of screening. Screening itself is designed merely to identify something that, once identified, needs further evaluation. Therefore, any discussion of colon screening must include consideration of how physicians will approach positive test results. The strategies for dealing with positive test results are costly and invariably lead to further surveillance. Current data suggest that patients with large polyps (> 1 cm) or villous adenomas have a high risk of colon cancer and are likely to benefit from full colonoscopy and subsequent surveillance. The benefits are far less clear for patients with small polyps. The subject of colon screening becomes even more complicated as consideration is given to when to start and stop screening, and how often to perform screening exams. There is clearly a need to improve upon the ability to identify patients most likely to develop colon cancer and design strategies to prevent cancer in this group. In the future, it may be possible to identify the high risk patient with some precision using genetic or biologic markers.