The purpose of this study is to assess if artificial intelligence aid colonoscopy colonoscopy is superior to conventional colonoscopy for the detection of adenomas during surveillance colonoscopy in patients with Lynch syndrome.
To compare the performance of white-light colonoscopy (HD-WLE) versus white-light + CAD (CAD) for detection of adenomas (1:1 ratio). The CAD system will include an artificial intelligence-based medical device (GI Genius, Medtronic) trained to process colonoscopy images and superimpose them, in real time, on the endoscopy display as a green box over suspected lesion. An adequate bowel preparation and minimum withdrawal time of 6 min will be required. Colonoscopies will be performed by experienced endoscopists and with high-definition scopes. Lesions will be collected, and histopathology findings used as the reference standard. The primary outcome of this trial will be the mean number of adenomas per colonoscopy (APC).
Study Type
INTERVENTIONAL
Allocation
RANDOMIZED
Purpose
DIAGNOSTIC
Masking
DOUBLE
Enrollment
404
GI geniusTM (Medtronic) is an artificial intelligence device designed and validated on a dataset of white-light endoscopy videos from high-quality randomized controlled trials for the detection of colorectal lesions during colonoscopy. Based on a deep learning system using convolutional neural networks, GI genius works in real-time side-to-side with the gastroenterologist during colonoscopy highlighting the presence of precancerous lesions with a visual marker (green box) in real time and designed to be integrated with the majority of colonoscopies.
High definition white light endoscopy will be used during the scope withdrawal
María Pellisé. MD. PhD.
Barcelona, Spain
RECRUITINGOswaldo ortiz
Barcelona, Spain
RECRUITINGMean number of adenomas per colonoscopy
Compare mean number of adenomas in both arms calculated by dividing the total number of adenomas detected by the total number of colonoscopies in each arm
Time frame: 18 months
Withdrawal time
time spent on inspection, from the cecum to endoscope extraction through the anus, excluding the time needed for therapeutic measures (including polypectomy).
Time frame: 18 months
Mean number of polyps per colonoscopy
Compare mean number of polyps in both arms calculated by dividing the total number of polyps (including serrated lesions and adenomas) detected by the total number of colonoscopies in each arm
Time frame: 18 months
Adenoma detection rate
proportion of individuals undergoing a complete colonoscopy who had at least one adenoma detected between two arms.
Time frame: 18 months
Polyp detection rate
proportion of individuals undergoing a complete colonoscopy who had at least one polyp detected between two arms.
Time frame: 18 months
False positives in CAD arm
Combination of: * Lesions detected by CAD (\>2s green box) after an adequate inspection with a high confidence optical diagnosis negative for neoplasia made by the endoscopist. * Lesions detected by CAD (\>2s green box) after an adequate inspection resected by the endoscopist and with a clinically non-significant histology.
Time frame: 18 months
Mean number of non-clinically significant removed lesions
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.
Lesions removed and with a histology of normal or inflammatory mucosa or removed diminutive recto-sigmoid polyps with a histology of hyperplastic
Time frame: 18 months
Subjective endoscopist measures
Questionary addresing the easiness and satisfaction of the procedure, detection of lesions and performance of artificial intelligence will be filled in by each endoscopist
Time frame: 18 months