Clinical Validation of Polydeep: an Artificial Intelligence-based Computer-aided Polyp Detection (CADe) and Characterization (CADx) System

Clinical Validation of Polydeep: an Artificial Intelligence-based Computer-aided Polyp Detection (CADe) and Characterization (CADx) System

This study is a clinical validation of Polydeep, a computer-aided polyp detection (CADe) and characterization (CADx) system. Polydeep Advance 1 is an unicentric prospective diagnostic tests trial with a paired study design. The hypothesis of the study is that Polydeep, a CAD system, is more sensitive than a blinded endoscopists for the detection of colorectal polyps in a high definition colonoscopy.

Colorectal cancer (CRC) is the most frequently cancer in western world. A fundamental tool for detection and prevention is the colonoscopy. The detection and endoscopic resection of colorectal polyps, the precursor lesion of CRC can reduce CRC incidence and mortality. Adenoma detection rate is the most used endoscopic quality indicator. The improvement of this indicator is related to the reduction of postcolonoscopy CRC incidence and mortality. Colorectal polyp diagnosis is based on endoscopic resection and histological analysis. An accurate optical diagnosis could avoid histological lesion of smaller lesions, reducing the costs associated with histological diagnosis. The NICE international classification has proposed the use of high definition endoscopes that have Narrow Band Imaging. However, NICE must be used by endoscopists who are sufficiently prepared and who have overcome the learning curve. Therefore, optical histology diagnosis with high accuracy independently of the center and the endoscopist is necessary. Computer Aid Diagnosis (CAD) systems based on Artificial Intelligence are experiencing exponential development in the field of medical image analysis. The development of the CAD system is based on the creation of large databases of endoscopic images and/or videos, on the training, development and validation of diagnostic algorithms in such databases and, finally, on prospective clinical validation in patients undergoing colonoscopy. The goal of CAD systems in colonoscopy is double. First, it aims to increase the detection of polyps (CADe) in general, and of adenomas and serrated lesions in particular. The second objective is to characterize (CADx) the histology of detected lesion. Polydeep CAD is a functional prototype. It is capable of detecting, locating and classifying colorectal polyps. In vivo validation data shows that Polydeep has high diagnostic accuracy for polyp identification and that this accuracy can be accommodated. The aim of Polydeep advance 1 is to perform the clinical validation within a diagnostic test trial with a paired study design. We will compare the sensitivity of Polydeep to endoscopists blinded to Polydeep in high definition colonoscopy.

A high definition colonoscopy will be performed by a high experienced endoscopists. The endoscopists will be blinded to a second monitor with the Polydeep evaluation. A second observer will annotate the lesions detected during colonoscopy withdrawal and will inform the endoscopists if Polydeep detects a lesion that is not detected by the endoscopists.

Both diagnostic interventions will be performed in all patients: High definition colonoscopy and Polydeep system.

Both diagnostic interventions will be performed in all patients High definition colonoscopy performed by high experienced endoscopists blinded to Polydeep. Polydeep: a CADe and CADx system. The gold standard will be the histological diagnosis of the lesion.

To compare the sensitivity of Polydeep to a high experienced endoscopist for colorectal polyp detection (adenoma or serrated lesion histologically confirmed)

To compare the sensitivity of Polydeep to a high experienced endoscopist for serrated lesions detection

To compare the sensitivity of Polydeep to a high experienced endoscopist for advanced colonic lesions (serrated lesions ≥10mm and/or dysplasia, adenoma ≥10mm and/or villous histology and/or high grade dysplasia) detection

To compare the sensitivity of Polydeep to a high experienced endoscopist for diminute lesions (≤5mm) detection

Inclusion Criteria: First diagnostic colonoscopy performed after a positive fecal immunochemical test performed within the CRC screening program. Surveillance after resection of colorectal adenomas. Acceptance after reading the information sheet and signing the informed consent Exclusion Criteria: Colonoscopies with insufficient intestinal cleansing (Boston Bowel Preparation Scale <6 or <2 in any of the evaluated segments). Detected lesions without histologic diagnosis. Previous CRC Previous colonic resection Hereditary CRC syndromes Serrated polyposis syndrome Incomplete colonoscopy without cecal intubation.

Cubiella J, Marzo-Castillejo M, Mascort-Roca JJ, Amador-Romero FJ, Bellas-Beceiro B, Clofent-Vilaplana J, Carballal S, Ferrandiz-Santos J, Gimeno-Garcia AZ, Jover R, Mangas-Sanjuan C, Moreira L, Pellise M, Quintero E, Rodriguez-Camacho E, Vega-Villaamil P; Sociedad Espanola de Medicina de Familia y Comunitaria y Asociacion Espanola de Gastroenterologia. Clinical practice guideline. Diagnosis and prevention of colorectal cancer. 2018 Update. Gastroenterol Hepatol. 2018 Nov;41(9):585-596. doi: 10.1016/j.gastrohep.2018.07.012. Epub 2018 Sep 20. English, Spanish.

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.

Zhao S, Wang S, Pan P, Xia T, Chang X, Yang X, Guo L, Meng Q, Yang F, Qian W, Xu Z, Wang Y, Wang Z, Gu L, Wang R, Jia F, Yao J, Li Z, Bai Y. Magnitude, Risk Factors, and Factors Associated With Adenoma Miss Rate of Tandem Colonoscopy: A Systematic Review and Meta-analysis. Gastroenterology. 2019 May;156(6):1661-1674.e11. doi: 10.1053/j.gastro.2019.01.260. Epub 2019 Feb 6.

ASGE Technology Committee; Abu Dayyeh BK, Thosani N, Konda V, Wallace MB, Rex DK, Chauhan SS, Hwang JH, Komanduri S, Manfredi M, Maple JT, Murad FM, Siddiqui UD, Banerjee S. ASGE Technology Committee systematic review and meta-analysis assessing the ASGE PIVI thresholds for adopting real-time endoscopic assessment of the histology of diminutive colorectal polyps. Gastrointest Endosc. 2015 Mar;81(3):502.e1-502.e16. doi: 10.1016/j.gie.2014.12.022. Epub 2015 Jan 16.

Puig I, Lopez-Ceron M, Arnau A, Rosinol O, Cuatrecasas M, Herreros-de-Tejada A, Ferrandez A, Serra-Burriel M, Nogales O, Vida F, de Castro L, Lopez-Vicente J, Vega P, Alvarez-Gonzalez MA, Gonzalez-Santiago J, Hernandez-Conde M, Diez-Redondo P, Rivero-Sanchez L, Gimeno-Garcia AZ, Burgos A, Garcia-Alonso FJ, Bustamante-Balen M, Martinez-Bauer E, Penas B, Pellise M; EndoCAR group, Spanish Gastroenterological Association and the Spanish Digestive Endoscopy Society. Accuracy of the Narrow-Band Imaging International Colorectal Endoscopic Classification System in Identification of Deep Invasion in Colorectal Polyps. Gastroenterology. 2019 Jan;156(1):75-87. doi: 10.1053/j.gastro.2018.10.004. Epub 2018 Oct 6.

Jin EH, Lee D, Bae JH, Kang HY, Kwak MS, Seo JY, Yang JI, Yang SY, Lim SH, Yim JY, Lim JH, Chung GE, Chung SJ, Choi JM, Han YM, Kang SJ, Lee J, Chan Kim H, Kim JS. Improved Accuracy in Optical Diagnosis of Colorectal Polyps Using Convolutional Neural Networks with Visual Explanations. Gastroenterology. 2020 Jun;158(8):2169-2179.e8. doi: 10.1053/j.gastro.2020.02.036. Epub 2020 Feb 29.

Hassan C, Spadaccini M, Iannone A, Maselli R, Jovani M, Chandrasekar VT, Antonelli G, Yu H, Areia M, Dinis-Ribeiro M, Bhandari P, Sharma P, Rex DK, Rosch T, Wallace M, Repici A. Performance of artificial intelligence in colonoscopy for adenoma and polyp detection: a systematic review and meta-analysis. Gastrointest Endosc. 2021 Jan;93(1):77-85.e6. doi: 10.1016/j.gie.2020.06.059. Epub 2020 Jun 26.

Parmar R, Martel M, Rostom A, Barkun AN. Validated Scales for Colon Cleansing: A Systematic Review. Am J Gastroenterol. 2016 Feb;111(2):197-204; quiz 205. doi: 10.1038/ajg.2015.417. Epub 2016 Jan 19.

Parsa N, Rex DK, Byrne MF. Colorectal polyp characterization with standard endoscopy: Will Artificial Intelligence succeed where human eyes failed? Best Pract Res Clin Gastroenterol. 2021 Jun-Aug;52-53:101736. doi: 10.1016/j.bpg.2021.101736. Epub 2021 Feb 22.

Wani S, Rastogi A. Narrow-band imaging in the prediction of submucosal invasive colon cancer: how "NICE" is it? Gastrointest Endosc. 2013 Oct;78(4):633-6. doi: 10.1016/j.gie.2013.06.015. No abstract available.

Mangas-Sanjuan C, Santana E, Cubiella J, Rodriguez-Camacho E, Seoane A, Alvarez-Gonzalez MA, Suarez A, Alvarez-Garcia V, Gonzalez N, Lue A, Cid-Gomez L, Ponce M, Bujanda L, Portillo I, Pellise M, Diez-Redondo P, Herraiz M, Ono A, Pizarro A, Zapater P, Jover R; QUALISCOPIA Study Investigators. Variation in Colonoscopy Performance Measures According to Procedure Indication. Clin Gastroenterol Hepatol. 2020 May;18(5):1216-1223.e2. doi: 10.1016/j.cgh.2019.08.035. Epub 2019 Aug 22.

https://intranet.pacifico-meetings.com/amsysweb/faces/publicacionOnlineDOI.xhtml?id=711&idComunicacion=160479

Effect of a deep-learning computer-aided detection system on adenoma detection during colonoscopy (CADe-DB trial): a double-blind randomised study.