Artificial intelligence system improves the quality of digestive endoscopy: A prospective pretest and post-test...

With the assistance of ENDOANGEL, a study was conducted at Hainan General Hospital to evaluate the effect of artificial intelligence (AI) system on the detection of gastrointestinal precancerous lesions.

The prospective, randomized, pretest and post-test, single-center clinical trial compared the detection rates of gastric precancerous lesions and intestinal adenomas between baseline and post-intervention phase among traditional digestive endoscopy (control groups i and ii, and experimental group i) and AI-assisted endoscopy (experimental group ii). Additionally, the effect of AI on the detection rate of different seniority physicians was analyzed.

AI assistance significantly increased the detection rates of intestinal metaplasia (experimental group ii vs control group ii: 14.23 % vs 9.15 %, P = 0.013), atrophy (experimental group ii vs control group ii: 22.76 % vs 17.28 %, P = 0.031) and intestinal adenomas (experimental group ii vs control group ii: 48.52 % vs 24.58 %, P < 0.001). The improvement was particularly notable among junior doctors, with significant enhancements in the detection rates of intestinal metaplasia (experimental group ii vs control group ii: 14.39 % vs 9.09 %, P = 0.008), atrophy (experimental group ii vs control group ii: 22.04 % vs 15.31 %, P = 0.004), and intestinal adenomas (experimental group ii vs control group ii: 45.18 % vs 29.27 %, P = 0.002).

AI systems have the potential to significantly improve the detection rates of precancerous conditions, particularly among less experienced endoscopists. This advancement can lead to more accurate and appropriate follow-up and review strategies for patients, ultimately reducing the risk of missed early cancer diagnoses.