Artificial Intelligence Increased Adenoma and Advanced Adenoma Detection Rates in Multicenter Randomized Controlled Trial
Douglas K. Rex, MD, MASGE, reviewing Xu H, et al. Clin Gastroenterol Hepatol 2022 Jul 18.
Artificial intelligence (AI) has been shown to increase adenoma detection rates (ADRs) by an average of about 10% in randomized controlled trials (RCTs). Meta-analyses of RCTs indicate that advanced adenoma detection is also increased, but this has not been shown in an individual RCT. For this study, colonoscopies were performed by 12 nonexpert endoscopists (<5000 colonoscopies) and 12 experts, with nonexperts and experts balanced across 6 centers.
In a large multicenter RCT involving 3059 asymptomatic patients who initially underwent direct colonoscopy screening or fecal immunochemical test (FIT)-based screening, AI-assisted colonoscopy increased conventional colonoscopy ADR from 32.4% to 39.9%, advanced ADR from 4.9% to 6.6%, expert ADR from 32.8% to 42.3%, and nonexpert ADR from 32.1% to 37.5%. Adenomas per colonoscopy increased from 0.45 to 0.59. Longer cecal intubation times (0.3 minutes) and withdrawal times (0.5 minutes) were present in the AI-assisted arm. Improvements in detection were even present after multivariate regression analysis controlling for multiple factors, including withdrawal time. Sessile serrated lesion (SSL) detection was not improved by AI.
COMMENTThis computer-aided detection (CADe) program was not extensively trained to detect SSLs and is another indication that CADe programs are limited by the types and range of lesions they are trained to detect. Nevertheless, the demonstrated improvement in advanced lesion detection in a single randomized controlled trial is an important achievement.
Note to readers: At the time we reviewed this paper, its publisher noted that it was not in final form and that subsequent changes might be made.
CITATION(S)
Xu H, Tang RSY, Lam TYT, et al. Artificial intelligence-assisted colonoscopy for colorectal cancer screening: a multicenter randomized controlled trial. Clin Gastroenterol Hepatol 2022 Jul 18. (Epub ahead of print) (https://doi.org/10.1016/j.cgh.2022.07.006)
