Computer-aided Detection During Screening Colonoscopy

Real-time Computer-aided Polyp/Adenoma Detection During Screening Colonoscopy: a Single-center Crossover Trial

Nowadays, colonoscopy is considered the gold standard for the detection of lesions in the colorectal mucosa. However, around 25% of polyps may be missed during the conventional colonoscopy. Based on this, new technological tools aimed to improve the quality of the procedures, diminishing the technical and operator-related factors associated with the missed lesions. These tools use artificial intelligence (AI), a computer system able to perform human tasks after a previous training process from a large dataset. The DiscoveryTM AI-assisted polyp detector (Pentax Medical, Hoya Group, Tokyo, Japan) is a newly developed detection system based on AI. It was designed to alert and direct the attention to potential mucosal lesions. According to its remarkable features, it may increase the polyp and adenoma detection rates (PDR and ADR, respectively) and decrease the adenoma miss rate (AMR). Based on the above, the investigators aim to assess the real-world effectiveness of the DiscoveryTM AI-assisted polyp detector system in clinical practice and compare the results between expert (seniors) and non-expert (juniors) endoscopists.

Colorectal cancer (CRC) is worldwide the second and third cancer-related cause of death in men and women, respectively. For the detection of lesions in the mucosa (premalignant and malignant), colonoscopy has been considered the gold standard. However, up to 25% of lesions can be missed during conventional colonoscopy. Some technical (i.e., bowel preparation) and operator-related (i.e., expertise, and fatigue) factors are related to these missing lesions. During the rapid-growing technological era, new tools were launched to improve the quality and performance of colonoscopies. Through the assistance of artificial intelligence (AI) an identification of a pattern can be achieved after a previous training from a large dataset of images. The DiscoveryTM AI-assisted polyp detector (Pentax Medical, Hoya Group, Tokyo, Japan), is a computer-assisted polyp/adenoma detection system based on AI. It detects classic adenomas and flat lesions, distinguished features like mucus cap or rim of debris with the advantage of a real-time and simultaneous multiple polyp detection. It was developed to minimize the missed lesions increasing as a result the polyp detection rate (PDR) and the adenoma detection rate (ADR). Lately, published data evaluating the AI-assisted polyp detectors has demonstrate high sensitivity, specificity, and interobserver agreement. Due to the importance of CRC diagnosis and prompt treatment, and taking advantage of the newly introduced DiscoveryTM AI system, the investigators aim to assess the real-world effectiveness of this AI-assisted polyp detector system in clinical practice and compare the results between expert (seniors) and non-expert (juniors) endoscopists.

This group is comprised by patients >45 years of age submitted for diagnostic colonoscopy. In the same session a HD-colonoscopy will be performed followed by an HD-colonoscopy with artificial intelligence assistance. The second procedure will be performed by an operator with the same-level-of -expertise in comparison to the initial procedure (expert or non-expert) and blinded to the results of the previous intervention.

This group is comprised by patients >45 years of age submitted for diagnostic colonoscopy. In the same session a HD-colonoscopy assisted by artificial intelligence will be performed followed by an HD-colonoscopy alone.The second procedure will be performed by an operator with the same-level-of -expertise in comparison to the initial procedure (expert or non-expert) and blinded to the results of the previous intervention.

HD-colonoscopy performed by an expert or non-expert endoscopist. All lesions will be recorded, assessed, and removed for histological analysis.

HD-colonoscopy with AI-assisted polyp detector. New polyps detected by AI will be recorded, removed, and studied.

The ADR will be determined by every new colonoscopy (second intervention) with at least one adenoma, histologically proven/NBI NICE classification. Results will be compared between experts and non-experts endoscopists.

The PDR will be determined by every new colonoscopy (second intervention) with at least one polyp. Results will be compared between experts and non-experts endoscopists.

The diagnostic performance of the AI-assisted system will be assessed by sensitivity, specificity, positive and negative predictive values (PPV and NPV) and observer agreement.

The AMR will be determined by the total number of missed adenomas on initial examination. The diagnosis of adenoma will be made by NBI NICE classification or biopsy.

Inclusion Criteria: Adults ≥45 years old Patients referred for screening colonoscopy Adequate bowel preparation, Boston Bowel Preparation Scale (BBPS) ≥8 Patients who authorized for endoscopic approach. Exclusion Criteria: Pregnancy Any clinical condition which makes endoscopy inviable. Patients with history of Colorectal Carcinoma. Patients with history of Inflammatory Bowel Disease (IBD) Inability to provide informed consent

Wang P, Berzin TM, Glissen Brown JR, Bharadwaj S, Becq A, Xiao X, Liu P, Li L, Song Y, Zhang D, Li Y, Xu G, Tu M, Liu X. Real-time automatic detection system increases colonoscopic polyp and adenoma detection rates: a prospective randomised controlled study. Gut. 2019 Oct;68(10):1813-1819. doi: 10.1136/gutjnl-2018-317500. Epub 2019 Feb 27.

Corley DA, Jensen CD, Marks AR, Zhao WK, Lee JK, Doubeni CA, Zauber AG, de Boer J, Fireman BH, Schottinger JE, Quinn VP, Ghai NR, Levin TR, Quesenberry CP. Adenoma detection rate and risk of colorectal cancer and death. N Engl J Med. 2014 Apr 3;370(14):1298-306. doi: 10.1056/NEJMoa1309086.

Kroner PT, Engels MM, Glicksberg BS, Johnson KW, Mzaik O, van Hooft JE, Wallace MB, El-Serag HB, Krittanawong C. Artificial intelligence in gastroenterology: A state-of-the-art review. World J Gastroenterol. 2021 Oct 28;27(40):6794-6824. doi: 10.3748/wjg.v27.i40.6794.

Parsa N, Byrne MF. Artificial intelligence for identification and characterization of colonic polyps. Ther Adv Gastrointest Endosc. 2021 Jun 29;14:26317745211014698. doi: 10.1177/26317745211014698. eCollection 2021 Jan-Dec.

Gong D, Wu L, Zhang J, Mu G, Shen L, Liu J, Wang Z, Zhou W, An P, Huang X, Jiang X, Li Y, Wan X, Hu S, Chen Y, Hu X, Xu Y, Zhu X, Li S, Yao L, He X, Chen D, Huang L, Wei X, Wang X, Yu H. Detection of colorectal adenomas with a real-time computer-aided system (ENDOANGEL): a randomised controlled study. Lancet Gastroenterol Hepatol. 2020 Apr;5(4):352-361. doi: 10.1016/S2468-1253(19)30413-3. Epub 2020 Jan 22. Erratum In: Lancet Gastroenterol Hepatol. 2020 Apr;5(4):e3.