Detection of Colonic Polyps Via a Large Scale Artificial Intelligence (AI) System

NCT ID: NCT04693078

Last Updated: 2021-03-03

Basic Information

• Recruitment Status: COMPLETED
• Clinical Phase: NA
• Total Enrollment: 100 participants
• Study Classification: INTERVENTIONAL
• Study Start Date: 2020-05-18
• Study Completion Date: 2020-12-30

Brief Summary

Colonoscopy is the gold standard for detection and removal of precancerous lesions, and has been amply shown to reduce mortality. However, the miss rate for polyps during colonoscopies is 22-28%, while 20-24% of the missed lesions are histologically confirmed precancerous adenomas. To address this shortcoming, the investigators propose a new polyp detection system based on deep learning, which can alert the operator in real-time to the presence and location of polyps during a colonoscopy. The investigators dub the system DEEP: (DEEP) DEtection of Elusive Polyps. The DEEP system was trained on 3,611 hours of colonoscopy videos derived from two sources, and was validated on a set comprising 1,393 hours of video, coming from a third, unrelated source. For the validation set, the ground truth labelling was provided by offline gastroenterologist annotators, who were able to watch the video in slow-motion and pause/rewind as required; two or three specialist annotators examined each video.

This is a prospective, non-blinded, non-randomized pilot study of patients undergoing elective screening and surveillance colonoscopies using DEEP.

The aim of the study is to:

Assess the:

1. Number of additional polyps detected by the DEEP system in real time colonoscopy.

2. Safety by prospective assessment of the rate of adverse events during the study period attributed or not to the use of the DEEP system.

3. Stability of the DEEP system by measuring the rate of false positives (False Alarms) per colonoscopies 4 And to examine its feasibility and usefulness of in clinical practice by assessing the colonoscopist user experience while using the DEEP system in a 5 point scale.

Conditions

Colonic Polyp

Study Design

• Allocation Method: NA
• Intervention Model: SINGLE_GROUP
• Primary Study Purpose: SCREENING
• Blinding Strategy: NONE

Study Groups

Intervention Arm

Consecutive patients undergoing screening or surveillance colonoscopy in whom a new polyp detection system based on deep learning will be used during the procedure.

Group Type: EXPERIMENTAL

AI polyp detection system based on deep learning

Intervention Type: DEVICE

A Polyp detection system based on deep learning and artificial intelligence, which can alert the operator in real-time to the presence and location of polyps during a colonoscopy.

Interventions

AI polyp detection system based on deep learning

A Polyp detection system based on deep learning and artificial intelligence, which can alert the operator in real-time to the presence and location of polyps during a colonoscopy.

Intervention Type: DEVICE

Eligibility Criteria

Inclusion Criteria

• Healthy subjects undergoing routine screening or surveillance colonoscopy in an ambulatory non urgent setting.
• Able to understand the study protocol and sign inform consent.

Exclusion Criteria

• Previous surgery involving the colon or rectum
• Known diagnosis of colorectal cancer
• Known history of inflammatory bowel disease
• Known or suspected diagnosis of familial polyposis syndrome

• Minimum Eligible Age: 40 Years
• Maximum Eligible Age: 80 Years
• Eligible Sex: ALL
• Accepts Healthy Volunteers: Yes

Sponsors

Google LLC.

INDUSTRY

Sponsor Role: collaborator

Shaare Zedek Medical Center

OTHER

Sponsor Role: lead

Responsible Party

Responsibility Role: SPONSOR

Locations

Digestive Diseases Institute, Shaare Zedek Medical Center

Jerusalem, , Israel

Countries

Israel

References

Livovsky DM, Veikherman D, Golany T, Aides A, Dashinsky V, Rabani N, Ben Shimol D, Blau Y, Katzir L, Shimshoni I, Liu Y, Segol O, Goldin E, Corrado G, Lachter J, Matias Y, Rivlin E, Freedman D. Detection of elusive polyps using a large-scale artificial intelligence system (with videos). Gastrointest Endosc. 2021 Dec;94(6):1099-1109.e10. doi: 10.1016/j.gie.2021.06.021. Epub 2021 Jun 30.

Reference Type: DERIVED

PMID: 34216598 (View on PubMed)

Provided Documents

Document Type: Study Protocol and Statistical Analysis Plan

View Document

Other Identifiers

0309-19-SZMC

Identifier Type: -

Identifier Source: org_study_id