Performance of Artificial Intelligence in Colonoscopy for Right Colon Polyp Detection

NCT ID: NCT06216405

Last Updated: 2024-01-23

Basic Information

• Recruitment Status: COMPLETED
• Clinical Phase: NA
• Total Enrollment: 2000 participants
• Study Classification: INTERVENTIONAL
• Study Start Date: 2022-01-20
• Study Completion Date: 2024-01-03

Brief Summary

The aim of the study is to assess whether the use of artificial intelligence improves polyp detection in a segment of the colon (the right colon).

To achieve this objective, patients will be divided into two groups: one will undergo a standard colonoscopy, the other a colonoscopy with the artificial intelligence software connected to the machine.

This software does not modify the colonoscopy technique in any way, and does not require the administration of any product to the patient.

The study will compare the detection rate of right colon polyps between the group of patients who underwent standard colonoscopy and those who underwent colonoscopy with artificial intelligence. If this number does not differ between the two groups, the investigators can conclude that there is no point in using artificial intelligence.

Detailed Description

Colorectal cancer (CRC) is the third most common cancer and remains one of the leading causes of mortality among neoplastic diseases in the world. Screening colonoscopy with detection and resection of colorectal polyps have reduced CRC incidence and mortality.

Nevertheless, colonoscopy is an imperfect screening test and its effectiveness is influenced by a range of variables including the skill and expertise of the endoscopist. Indeed, a significant proportion of colorectal neoplasia is missed during colonoscopic examinations. In the majority of studies, the rate of interval CRC among all CRCs ranged from 2% to 9% and interval cancers occurred more likely in the right colon. Interval cancers can result from missed lesion, incomplete removal or newly developed cancer. Most missed polyps are smaller than 10 mm in diameter and are sessile or flat in appearance. Therefore, sessile serrated adenomas (SSAs), which predominantly occur in the right colon, are easily missed because they are small and sessile.

Colonoscopy is less effective in screening right sided CRCs, mainly because of the increased miss rate for polyps with sessile or flat appearance. In recent years, artificial intelligence (AI) is increasingly applied in gastrointestinal endoscopy, especially in the detection of colorectal polyps. In 2019, the first prospective randomized controlled trial including 1058 patients and comparing the polyp detection rate in colonoscopy with or without AI showed a significantly higher detection rate in the group with AI (29.1% vs 20.3%). Other previous prospective studies have showed that AI had great potential for improving colonic polyp detection.

To our knowledge, this is the first study to date to evaluate the performance of the AI in the detection of right colon polyps.

In this study, the investigators aim first of all to compare the rate of right colon polyp detection with AI-aided colonoscopy (AIC) to the rate obtained by the Standard (high-definition) colonoscopy (SC) in patient undergoing diagnostic colonoscopy. Then, the investigators would to evaluate the following endpoints:

1. Comparison of time required for AIC versus SC (withdrawal time)

2. Comparison of histological classification of all polyps detected in each group

3. Comparison of the number of right colon polyps detected in each group

4. A subgroup analysis comparing the rate of colon polyp detection between the two groups will be realized according to the procedure starting time (accounting for operator fatigability).

Conditions

Colon Cancer; Colon Polyp; Colon Lesion; Colon Adenocarcinoma

Study Design

• Allocation Method: RANDOMIZED
• Intervention Model: PARALLEL
• Primary Study Purpose: DIAGNOSTIC
• Blinding Strategy: SINGLE

Study Groups

AI-aided colonoscopy

Experimental arm : patients receive a colonoscopy using GI Genius™ intelligent endoscopy system (Medtronic Inc., Minneapolis, Minnesota, USA)

Group Type: EXPERIMENTAL

Colonoscopy

Intervention Type: PROCEDURE

Colonoscopy

Standard colonoscopy

Control arm :The artificial intelligence is not activated during the colonoscopy exam,the patient receive a standard high definition colonoscopy

Group Type: ACTIVE_COMPARATOR

Colonoscopy

Intervention Type: PROCEDURE

Colonoscopy

Interventions

Colonoscopy

Colonoscopy

Intervention Type: PROCEDURE

Eligibility Criteria

Inclusion Criteria

• Age>18 years
• Signed informed consent
• Affiliated to or beneficiary of a social security scheme

Exclusion Criteria

• Patients with history of inflammatory bowel diseases
• Patients with history of familial polyposis syndrome
• Patients with history of right colonic resection.
• Pregnant or breastfeeding women
• Patient benefiting from a legal protection measure : tutorship, curatorship, patient deprived of freedom
• Adults who are unable to give their consent
• People under psychiatric care who are not able to understand the aim if this research

• Minimum Eligible Age: 18 Years
• Eligible Sex: ALL
• Accepts Healthy Volunteers: No

Sponsors

Cosmo Artificial Intelligence-AI Ltd

INDUSTRY

Sponsor Role: collaborator

Groupe Hospitalier Diaconesses Croix Saint-Simon

OTHER

Sponsor Role: lead

Responsible Party

TANNOURY Jenny

Principal Investigator

Responsibility Role: PRINCIPAL_INVESTIGATOR

Locations

Groupe Hospitalier Diaconesses Croix Saint-Simon

Paris, Île-de-France Region, France

Countries

France

Other Identifiers

2021-A01243-38

Identifier Type: -

Identifier Source: org_study_id