AI boosts adenoma detection in colonoscopy but increases removal of non-neoplastic polyps
Click Here to Manage Email Alerts
Key takeaways:
- Artificial intelligence-assisted colonoscopy resulted in a higher adenoma detection rate vs. standard colonoscopy (44% vs. 35.9%).
- There also was an increase in removal of non-neoplastic polyps with AI.
Although artificial intelligence-assisted colonoscopy significantly increased detection of adenomas, it also resulted in higher removal rates of non-neoplastic polyps, according to researchers.
“Recently, a series of AI-based tools have been developed to guide the colonoscopist in detection of polyps (computer-aided detection [CADe]),” Cesare Hassan, MD, PhD, of the department of biomedical sciences at Humanitas University, and colleagues wrote in Annals of Internal Medicine. “Increased polyp detection with CADe may lead to better cancer prevention by reducing the 15% to 30% miss rate for clinically relevant lesions.
“However, CADe adoption may also lead to more overdiagnosis and overtreatment of nonneoplastic polyps and thus may increase potential patient harm and burden for individuals and society.”
To quantify the benefits and harms of CADe, Hassan and colleagues conducted a systematic review and meta-analysis of 21 randomized controlled trials published between 2019 and 2023. Trials included 18,232 participants, of whom 9,142 underwent standard colonoscopy and 9,090 underwent colonoscopy with CADe.
Benefit outcomes included adenoma detection rate (proportion of patients with ≥ 1 adenoma), number of adenomas detected per colonoscopy (APC), advanced APC (≥ 10 mm with high-grade dysplasia and villous histology), the number of serrated lesions per colonoscopy and adenoma miss rate. Harm outcomes included the number of polypectomies for nonneoplastic lesions and inspection time.
Results demonstrated a significantly higher ADR in the CADe group vs. the standard colonoscopy group (44% vs. 35.9%; RR = 1.24; 95% CI, 1.16-1.33) which corresponded to a 55% (risk ratio = 0.45; 95% CI, 0.35-0.58) relative reduction in adenoma miss rate. There was no significant difference in mean advanced APC between groups.
However, more non-neoplastic polyps were removed in the CADe group (0.52 vs. 0.34 per colonoscopy), and there was a slight increase in mean inspection time with CADe vs. standard colonoscopy (9.22 vs. 8.73 minutes; mean difference = 0.47 minutes; 95% CI, 0.23-0.72).
“Our meta-analysis showed that CADe may increase the detection of colorectal neoplasia, probably by reducing the miss rate, but that this benefit may be offset by an increase in the removal of non-neoplastic polyps,” Hassan and colleagues concluded.
Collapse
Click Here to Manage Email Alerts