Authors: (including presenting author): :
SM Lam(1), YT Wong Anna(2), TF Tai(2), KF Wong(2), SK Leung(2), SY Wong(1), YY Lo(1), KM Yan(1), SK Tam(1), MF Wong(1), HL Chan Calvin(1)
Affiliation: :
1. Combined Endoscopy Unit, Tin Shui Wai Hospital
2. Department of Surgery, New Territories West Cluster
Introduction: :
The application of high technology in endoscopy affects the outcomes significantly. Artificial Intelligence (AI) in Colonoscopy may help increase the polyp detection rate.
Objectives: :
The aim of this study was to evaluate if the application of Artificial Intelligence (AI) Colonoscopy (ENDO-AID) could increase polyp detection rate.
Methodology: :
A single center, retrospective study was performed in Tin Shui Wai Hospital. Polyp detection rate in colonoscopy from 9/11/2020 to 23/12/2020 after the application of ENDO-AID/Artificial Intelligence (AI group) was compared to the cases from 2/6/2020 to 4/11/2020 before the application of the practice (non-AI group). Procedures were performed by a single endoscopist. Variables, such as patients’ demographic data, indications for colonoscopy, incidence of polyp detection rate, post colonoscopy complication rate between the 2 groups, were compared. Categorical and continuous variables were analyzed by using Chi-Square test (Fisher exact test if appropriate) and Mann-Whitney test respectively. Results were considered to be significant if p-value less than 0.05.
Result & Outcome: :
Total 119 patients with colonoscopy done were recruited. 72 patients (60.5%) were in the non-AI group while 47 patients (39.5%) were in the AI group. The mean age of the non-AI was higher than the AI group (65.3 vs 60.3, P-value 0.038), otherwise, there was no significant difference in sex, indications for colonoscopy (e.g. follow up colonoscopy for polyp/ cancer, per-rectal bleeding, altered bowel habit, history of colectomy done, etc.), completion colonoscopy rate (97.2% vs 97.9%) and complication rate (0% in both groups) between groups.
In the contrary, polyp detection rate was significantly higher in the AI group than the non-AI group. (74.5% vs 48.6%, P-value < 0.01)
AI colonoscopy is useful to detect polyp. This can be an effective and efficient option, acting as an extra-pair of eyes during the procedures to improve the polyp detection rate.
All the authors declared no conflict of interest.
Table 1: Comparisons between groups
Non AI group (72) AI group (47) P-value
Polyp No 37 12 < 0.01*
Yes 35 35
sex female 27 21 0.44
male 45 26
complication yes 0 0 1.0
no 72 47
Complete CLN Yes 70 46 1.0
no 2 (one poor bowel prep, one obstructive CA) 1 (one obstructive CA)
Age (mean) 65.3 60.3 0.038*
Indications FU polyp 10 10 0.29
FU CA 11 6 0.15
PRB 19 6 0.07
Altered bowel habit 12 12 0.24
Colectomy done 11 3 0.14
Abdominal pain 5 3 0.90
