Objective To analyze the impact of artificial intelligence (AI) on colorectal adenoma detection rate (ADR) in fatigue state of endoscopists.
Methods A total of 784 patients undergoing colonoscopy at the Endoscopy Center of Hengshui People's Hospital in Hebei Province were enrolled. Patients were divided into group A (n=405, time from 14: 00 to 15: 59) and group B (n=379, time from 16: 00 to 17: 30) based on the operation time. Patients in both groups who underwent AI-assisted examination were included in AI subgroup, while those without AI assistance were included in non-AI subgroup. Baseline data, including age, gender and bowel preparation quality assessed by the Boston Bowel Preparation Scale (BBPS) score were collected. The ADR and polyp detection rate (PDR) for colorectal lesions were calculated for both groups, and the ADR was compared between groups with and without AI assistance.
Results The overall ADR was 32.10% in the group A, which was higher than 26.91% in the group B, but the difference was not statistically significant (P>0.05). Without AI assistance, the ADR in the group A was 25.85%, which was significantly higher than 17.30% in the group B (P < 0.05). In the group B with AI assistance, the ADR was significantly higher than that in the group A without AI assistance (P < 0.05). Without AI assistance, the PDR in the group A was 33.17%, which was significantly higher than 22.70% in the group B (P < 0.05). After using AI assistance, there was no significant difference in PDR between the two groups (P>0.05). In the group A, the ADR was 38.50% in the AI subgroup (n=200) and 25.80% in the non-AI subgroup (n=205), with a statistically significant difference (P=0.006). In the group B, the ADR was 36.08% in the AI subgroup (n=194) and 17.30% in the non-AI subgroup (n=185), with a statistically significant difference (P < 0.001).
Conclusion AI-assisted colonoscopy can effectively improve the ADR of endoscopists under fatigue conditions.
DownLoad: