Accuracy of colonoscopy for the detection of colorectal polyps.

The findings at colonoscopy were compared with the pathologic findings of the surgical specimen in 235 patients who underwent a colon resection for a primary colorectal neoplasm from January 1980 to December 1987 at Roswell Park Cancer Institute. Seven patients (3 percent) were found to have synchronous primary colon carcinomas, and 100 patients (43 percent) were found to have synchronous adenomatous polyps identified by colonoscopy and/or pathology. In patients with polyps 10 mm or greater in diameter, the findings on colonoscopy agreed with the pathology report 96 percent of the time. When polyps of all sizes were included, with many less than 5 mm in diameter, colonoscopy agreed with the pathology in 89 percent of patients. When only the area of the colon resected was used to determine the ability of colonoscopy to locate polyps, 58 percent of polyps of all sizes were located. The majority of the missed polyps were adjacent to a carcinoma. One cecal carcinoma was not seen by colonoscopy because of technical inabilities to reach the cecum. A second carcinoma (20 mm x 17 mm) was not seen at the splenic flexure.