Improving Polyp Detection Rate by Artificial Intelligence in Colonoscopy

Haraldsplass Deaconess Hospital

Status: Enrolling

Conditions: Colonic Polyp

Treatments: Device: Standard white light colonoscope; Device: GI Genius

Study type: Interventional

Funder types: Other

Identifiers

NCT05322993

REK 275068

Details and patient eligibility

About

The aim of this study is to investigate if the use of artificial intelligence (AI) in colonoscopy improves the polyp detection rate, and if the use of AI has a learning effect.

Full description

The endoscopists will use GI Genius from Medtronic, a device that uses artificial intelligence (AI) based on machine learning to detect polyps in the colon in real time during colonoscopy. The device interprets the endoscopy pictures and superimposes possible polyps with frames.

The patients will be included in regular outpatient clinics in Western Norway. The endoscopists will be divided into groups depending on their experience. The endoscopists will perform colonoscopies in three phases; (1) before the use of AI, (2) during the use of AI and (3) after the use of AI. The investigators will then evaluate the polyp detection rate (PDR) in the three phases to see if AI increases PDR, and if there is a learning effect on PDR after the use of AI. The investigators will also evaluate if there is a difference in the learning-effect from AI-use depending on if the endoscopist is experienced or inexperienced.

The PDR's are registered as part of Norway's national quality register of colonoscopy, Gastronet. The data registered in Gastronet can also help the investigators evaluate other outcomes such as withdrawal time, bowel preparation, patient reported pain, patient satisfaction and complications.

Enrollment: 4,500 estimated patients

Sex: All

Volunteers: Accepts Healthy Volunteers

Inclusion criteria

• Patients coming to outpatient clinics to perform colonoscopies

Exclusion criteria

• Total colectomy
• Reservation against registration in Gastronet, the national quality register for colonoscopy in Norway

Trial design

Primary purpose: Diagnostic

Allocation: Non-Randomized

Interventional model: Crossover Assignment

Masking: None (Open label)

---

4,500 participants in 2 patient groups

With artificial intelligence (AI)

Active Comparator group

Description:

Use of GI Genius artificial intelligence device during colonoscopy.

Treatment:

Device: GI Genius

Without artificial intelligence

Active Comparator group

Description:

Use of standard colonoscopy equipment without GI Genius.

Treatment:

Device: Standard white light colonoscope