Authors (including presenting author) :
Gabriel YC Sit(1), KF Wong(1), SK Leung(1)
SM Lam(2), SY Wong(2), YY Lo(2), KM Yan(2), SK Tam(2), MF Wong(2), HL Chan Calvin(2)
Affiliation :
(1) Department of Surgery, New Territories West Cluster
(2) Centre of Endoscopy Unit, Tin Shui Wai Hospital
Introduction :
Cerebro is an artificial intelligence (AI) system that uses 29 views of 12 landmarks to enhance endoscopic coverage and increase pathology detection of the upper gastrointestinal tract. The endoscopist is encouraged to ensure complete coverage as the anatomical check-list is progressively completed. The inspection time for each site is recorded and shown in real-time to improve examination quality. The system can automatically capture photos of the landmarks and provide recommendations to improve image quality.
Objectives :
The objective of this report is to investigate the effect of Cerebro on the improvement of quality, and in training of young endoscopists of oesophagogastroduodenoscopy (OGD).
Methodology :
This is a descriptive report of OGDs performed with the application of Cerebro in Tin Shui Wai Hospital in a two-month period. All procedures were performed by trained endoscopists. Commonly missed landmarks were summarised and impacts of the system were concluded.
Result & Outcome :
A total of 69 OGDs with Cerebro were done in our centre. The three most commonly missed views are lower gastric body (26.7%), gastric fundus-cardia (21.0%) and middle-upper gastric body (17.9%). There was an increase in the averaged views seen as time progresses with the use of Cerebro from 84.5% in the beginning to 97.3% at the end.
The number of missed areas from a single endoscopy is affected by the endoscopist's experience, reflected by the accuracy of the examination and total scope time. Other possible factors include system malfunction, and actions done during the procedure. Endoscopists reflected increased awareness of anatomy, more dedicated endoscopic coverage and examination of areas under the presence of Cerebro. This provides improved quality assurance by setting benchmark in anatomical coverage, quality of images, and as a tool for training young endoscopists.
