Improved detection of colorectal neoplasms with selective use of chromoendoscopy in 2005 consecutive patients

Background Colorectal cancer mortality is decreased by endoscopic polypectomy, but conventional colonoscopy may be inadequate for detecting subtle colonic lesions. Methods We selectively performed chromoendoscopy in all patients undergoing colonoscopy between January 1999 and December 2005 at the International Health Union of Rome. Patients with a history of colorectal polyps, inflammatory bowel disease, colorectal surgery or coagulopathy and those with poor bowel preparation were excluded from this analysis. Whenever colonoscopy revealed suspicious mucosal areas, dye-spraying with 0.2% indigo carmine solution was also performed. Findings from conventional and dyespraying views were classified morphologically, and specimens were analyzed histologically. Non-adenomatous lesions were classified as negative findings. Results A total of 2005 patients underwent conventional colonoscopy and in 305 cases (15%) chromoendoscopy was also performed. Conventional colonoscopy identified 508 neoplasms in 381 patients (19%). Selective chromoendoscopy found an additional 244 neoplasms in 212 patients (11%). Thus, chromoendoscopy was positive in 212 (70%) of 305 patients in whom the examination was performed. Overall, 56 large, ulcerated, advanced cancers and 696 non-advanced neoplasms were found. Of the 696 nonadvanced neoplasms, 448 (65%) were polypoid and 248 (35%) were non-polypoid. All but 4 non-polypoid lesions were only detected with chromoendoscopy. Of the 248 non-polypoid lesions, 12 (5%) were depressed and 236 (95%) were flat. Advanced histology was present in 39 non-polypoid lesions (15%) and was more common in depressed lesions than in flat ones (58% vs. 13%; p<0.001). Conclusions Our study confirms the existence of flat and depressed neoplasms in an Italian population. The vast majority of non-polypoid lesions were only detected by chromoendoscopy, and many lesions with advanced histology were missed by conventional colonoscopy. We therefore recommend selectively performing chromoendoscopy when conventional colonoscopy provides clues for non-polypoid lesions. Therefore, endoscopists should be trained in the detection of these subtle mucosal clues, as well as in the use of chromoendoscopy to enhance their detection. An erratum to this article is available at .