Computer-aided detection of colonic neoplasia yields modest improvements

08 Mar 2024
Computer-aided detection of colonic neoplasia yields modest improvements

The use of real-time, computer-aided detection of colonic neoplasia leads to modest increases in adenoma detection rate and adenomas detected per colonoscopy in average-detector community endoscopists, as shown in the results of a pragmatic randomized controlled trial.

The trial included 1,100 patients (median age 55.5 years, 61 percent female, 72.7 percent of Hispanic ethnicity) undergoing colonoscopy for colorectal cancer screening, surveillance, positive faecal-immunohistochemical tests, and diagnostic indications at a single community-based centre. These patients were randomly assigned to undergo traditional colonoscopy or real-time computer-aided detection. Histopathologic findings were analysed by pathologists who were blinded to the intervention.

Of the patients, 9.1 percent had inadequate bowel preparation. The primary outcome of adenoma detection rate (the percentage of patients with at least one histologically proven adenoma or carcinoma) was significantly higher after computer-aided detection than after traditional colonoscopy (42.5 percent vs 34.4 percent; p=0.005). The main reason for the better adenoma detection rate was the successful identification of smaller adenomas (<5 mm).

The mean adenomas detected per colonoscopy was likewise higher among patients who underwent computer-aided detection (0.89 vs 0.60; p<0.001).

Computer-aided detection was also associated with a higher sessile-serrated lesion detection rate compared with traditional colonoscopy (4.7 percent vs 2.0 percent; p=0.01).

Clin Gastroenterol Hepatol 2024;doi:10.1016/j.cgh.2024.02.021