AI-assisted Detection of Missed Colonic Polyps

Artificial Intelligence-Assisted Real-time Detection of Missed Lesions During Colonoscopy: A Prospective Study

A prospective validation of real time deep learning artificial intelligence model for detection of missed colonic polyps

Study Overview

• Status: Completed
• Conditions: Colonic Polyp; Colon Cancer; Colon Adenoma
• Intervention / Treatment: Device: Artificial intelligence-Assisted real time colonoscopy

Detailed Description

Consecutive adult patients, age 40 or above, who were scheduled to have outpatient colonoscopy in the Queen Mary Hospital were invited to participate. Patients were excluded if they were unable to provide informed consent, considered to be unsafe for taking biopsy or polypectomy including patients with bleeding tendency and those with severe comorbid illnesses. Also, patients with history of inflammatory bowel disease, familial adenomatous polyposis, Peutz-Jeghers syndrome or other polyposis syndromes were excluded.

The primary endoscopist conducted the colonoscopic examination in the usual manner. All colonoscopy procedures were performed with high-definition colonoscopes (EVIS-EXERA 290 video system, Olympus Optical, Tokyo, Japan). The colonoscopy was first advanced to the cecum in all patients as confirmed by identification of the appendiceal orifice and ileocecal valve or by intubation of the ileum. After cecal intubation, the colonoscopy was slowly withdrawn to the rectum by the primary endoscopist. The AI real time detection was then activated with the output displayed in a different monitor and was only viewed by an independent investigator, who was an experienced endoscopist. The primary endoscopist was blinded to the AI real time detection result al.

The colon was divided into three segments during the examination: right side, transverse and left side colon, using hepatic flexure and splenic flexure as dividing landmark. All polyps were marked for size (measured with biopsy forceps), location and morphology according to the Paris classification, and then removed or biopsied for histological examination. After examination of each segment, segmental unblinding of the AI results were provided by the independent viewer. If additional polyps were detected by AI but not by the endoscopist, that segment were reexamined to look for the missed polyp. If no additional polyp was detected by the AI, the next colonic segment was examined. Missed lesions were defined as lesions identified by AI and then confirmed on reexamination by the endoscopist.

The first withdrawal time (minus the polypectomy site) was measured. The Boston Bowel Preparation Scale score (BPPS) was used for evaluation of bowel cleanliness.

• Study Type: Interventional
• Enrollment (Actual): 52
• Phase: Not Applicable

Contacts and Locations

Study Locations

• Hong Kong
  • Hong Kong, Hong Kong
    • Queen Mary Hospital

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 40 years to 90 years (Adult, Older Adult)
• Accepts Healthy Volunteers: No
• Genders Eligible for Study: All

Description

Inclusion Criteria:

• consecutive adult patients, age 40 or above, who were scheduled to have outpatient colonoscopy in the Queen Mary Hospital were invited to participate

Exclusion Criteria:

• Patients were excluded if they were unable to provide informed consent, considered to be unsafe for taking biopsy or polypectomy including patients with bleeding tendency and those with severe comorbid illnesses.
• Also, patients with history of inflammatory bowel disease, familial adenomatous polyposis, Peutz-Jeghers syndrome or other polyposis syndromes were excluded.

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Diagnostic
• Allocation: N/A
• Interventional Model: Single Group Assignment
• Masking: None (Open Label)

Number of Arms

1

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Experimental: Artificial intelligence-Assisted real time colonoscopy; AI assisted real-time detection of colonic lesions	Device: Artificial intelligence-Assisted real time colonoscopy; The colonoscopy was performed under artificial intelligence assistance

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Adenoma miss rate	The number of patient had at least one missed adenoma	During the colonoscopy procedure

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Total number of adenoma missed	The total number of missed polyps for all subjects	During the colonoscopy procedure
Colonic polyp miss rate	The number of patient had at least one missed adenoma	During the colonoscopy procedure
Total number of missed polyps	The total number of missed polyps for all subjects	During the colonoscopy procedure

Collaborators and Investigators

• Sponsor: The University of Hong Kong
• Investigators: Principal Investigator: Ka Luen, Thomas Lui, Queen Mary Hospital, the University of Hong Kong

Publications and helpful links

General Publications

• Lui TKL, Hui CKY, Tsui VWM, Cheung KS, Ko MKL, Foo DCC, Mak LY, Yeung CK, Lui TH, Wong SY, Leung WK. New insights on missed colonic lesions during colonoscopy through artificial intelligence-assisted real-time detection (with video). Gastrointest Endosc. 2021 Jan;93(1):193-200.e1. doi: 10.1016/j.gie.2020.04.066. Epub 2020 May 4.

Study record dates

Study Major Dates

• Study Start (Actual): January 1, 2020
• Primary Completion (Actual): February 1, 2020
• Study Completion (Actual): March 1, 2020

Study Registration Dates

• First Submitted: January 10, 2020
• First Submitted That Met QC Criteria: January 13, 2020
• First Posted (Actual): January 14, 2020

Study Record Updates

• Last Update Posted (Actual): March 4, 2020
• Last Update Submitted That Met QC Criteria: March 2, 2020
• Last Verified: March 1, 2020

More Information

Terms related to this study

• Keywords: Detection; Colonoscopy; Artificial intelligence; Deep learning; Computer-aided
• Additional Relevant MeSH Terms: Neoplasms by Histologic Type; Neoplasms; Neoplasms, Glandular and Epithelial; Pathological Conditions, Anatomical; Intestinal Polyps; Adenoma; Polyps; Colonic Polyps
• Other Study ID Numbers: UW 19-309

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No