Artificial Intelligence (AI) Validation Study for Polyp Detection

Not Applicable

Completed

• Conditions: Polyp of Colon
• Interventions: Diagnostic Test: Artificial intelligence for polyp detection

• Registration Number: NCT04378660
• Lead Sponsor: Universitaire Ziekenhuizen KU Leuven

Brief Summary

Interventional prospective multicenter study: Polyp detection by an automated endoscopic tool as second observer during routine diagnostic colonoscopy

Detailed Description

This is an investigator-initiated multicentre non-randomized prospective interventional trial to validate the performance of a novel state-of-the-art computer-aided diagnostic (CAD) tool in polyp detection implemented as second observer during routine diagnostic colonoscopy and to evaluate its feasibility in daily endoscopy. Patients referred for a screening surveillance or therapeutic colonoscopy will undergo a colonoscopy performed by an endoscopist with moderate to high adenoma detection rate (ADR \> 20% and \< 50%) while a second observer will follow the procedure on a bedside AI-tool to count the number of detections made by the AI system and categorize the results into positive or negative results as follows (1) true positive, (2) false negative, (3) other positive and (4) obvious false positive. When a doubtful detection is made by the AI-system, the second observer will ask to re-evaluate the indicated region. When the detection is clear, the endoscopist and second observer do not communicate. The entire procedure is recorded.

Study Timeline

• Study Start: 2019-10-30
• Status Verified: 2020-04
• Primary Completion: 2020-04-18
• Study Completion: 2020-04-18
• Study First Submitted: 2020-04-20
• Study First Submitted QC: 2020-05-05
• Last Update Submitted: 2020-05-05
• Study First Posted: 2020-05-07
• Last Update Posted: 2020-05-07

Recruitment & Eligibility

• Status: COMPLETED
• Sex: All
• Target Recruitment: 357

Inclusion Criteria

• > 18 years
• Diagnostic or screening colonoscopy
• Therapeutic colonoscopy

Exclusion Criteria

• Inability to give informed consent by the patient or legal representative
• < 18 years old
• Any contraindication for colonoscopy or biopsies of the colon
• Uncontrolled coagulopathy
• Confirmed diagnosis of inflammatory bowel disease
• Short bowel or ileostomy
• Pregnancy

Study & Design

• Study Type: INTERVENTIONAL
• Study Design: SINGLE_GROUP

Arm && Interventions

Group	Intervention	Description
Intervention arm	Artificial intelligence for polyp detection	Colonoscopy with AI

Primary Outcome Measures

Name	Time	Method
Diagnostic accuracy of the novel AI system compared to endoscopic diagnosis as gold standard	4 months	Determination of the diagnostic accuracy of the novel AI system as second observer

Secondary Outcome Measures

Name	Time	Method
Endoscopist's polyp miss rate as number of extra AI detections	4 months	Determination of the AI system's precision and extra value as second observer

Trial Locations

Locations (1)

University Hospitals Leuven; 🇧🇪 Leuven, Vlaams-Brabant, Belgium