Comparative Analysis of AI Software for Enhanced Polyp Detection and Diagnosis

Comparative Analysis of AI Software and Conventional Colonoscopy for Enhanced Polyp Detection and Diagnosis: A Randomized Controlled Trial

Purpose & Research Questions

The purpose of this study is to evaluate whether artificial intelligence (AI) improves the detection of polyps and whether the system can classify the type and severity of detected changes. The investigators will also assess if there are any differences between the various AI systems and whether the polyps that may be missed are benign or malignant.

Study Overview

• Status: Recruiting
• Conditions: Colonic Polyp; Colon Lesion
• Intervention / Treatment: Device: AI

Detailed Description

Materials: Selection, Representativeness, and Grouping

The study is a randomized controlled clinical trial conducted at Västra Götaland region. Patients are randomized to receive either AI or conventional endoscopy, and each endoscopy unit has access to at least one AI device. Informed consent is obtained from patients before the examination.

Before the examination, patients are asked if they want to participate in the study. If the patient wishes to participate, informed consent is obtained. The endoscopist then draws a slip from an envelope, determining whether conventional colonoscopy or AI will be used. During conventional colonoscopy, the patient is examined according to routine practice. After the examination, a questionnaire is filled out, recording various findings from the examination. During AI examinations, the AI system is activated only during instrument withdrawal. The system identifies polyps and alerts the endoscopist. Additionally, two of the AI systems (GI-Genius, CADEYE) have the ability to characterize the detected polyps and provide an interpretation of whether the changes are benign or malignant. After the examination, a similar questionnaire is completed as in conventional colonoscopy.

Method: Data Collection

Basic data will be recorded, including the reason for the examination, other illnesses, medications, demographic data such as gender, age, family history of cancer, and the time of the examination (morning or afternoon). During the examination, the following will be recorded: the number of polyps, macroscopic classification of polyps, AI system's characterization of the polyps, bowel cleansing quality, laxative method, medication during the examination, complications occurring during or after the examination. Withdrawal time is recorded, as well as the number of false-positive polyps and supplementation with histological data of polyps after review by the pathologist. All materials and costs associated with the examination are recorded.

Three different AI systems will be used at the various hospital sites, namely EndoAid (Olympus), CADEYE (Fujinon), and GI-Genius (Medtronic). All units will randomize each examination to either AI or non-AI.

• Study Type: Interventional
• Enrollment (Estimated): 915
• Phase: Not Applicable

Contacts and Locations

• Study Contact: Name: Jonas Varkey, MD, PhD; Phone Number: 0046764030388; Email: jonas.varkey@vgregion.se

Study Locations

• Sweden
  • Gothenbburg
    • Gothenburg, Gothenbburg, Sweden, 41345
      • Recruiting
      • Sahlgrenska University Hospital
      • Contact: Jonas Varkey, MD, PhD; Phone Number: 0046764030388; Email: jonas.varkey@vgregion.se

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Adult; Older Adult
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• Age > 50 years
• Elective colonoscopy

Exclusion Criteria:

• Patient declines to participate in the study.
• Age < 50 years
• Unprepared bowel

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Diagnostic
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: None (Open Label)

Number of Arms

4

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
Active Comparator: CAD-EYE; AI system 1. Sealed envelopes in blocks of four were used for randomisation.	Device: AI; AI system
Active Comparator: GI-GENIUS; AI system 2. Sealed envelopes in blocks of four were used for randomisation	Device: AI; AI system
Active Comparator: Endo-AID; AI system 3. Sealed envelopes in blocks of four were used for randomisation	Device: AI; AI system
No Intervention: No AI; No AI. conventional examination. Sealed envelopes in blocks of four were used for randomisation

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Comparison of Adenoma Detection Rates (ADR) between colonoscopies assisted by Artificial Intelligence (AI) and conventional colonoscopies without AI	The ADR is typically expressed as a percentage and is calculated using the following formula:Number of colonoscopies with adenoma detection/ Total number of colonoscopies ×100.	During procedure (up to 40 minutes)
To compare the effectiveness of AI systems in characterizing polyps with conventional histology	The endoscopist's ability to characterize lesions with and without AI will be assessed, and this assessment will be compared to histology	During procedure (5minutes)
To assess and compare the costs associated with AI-assisted colonoscopy versus conventional colonoscopy.	Each cost associated with each examination, whether using AI-assisted or conventional colonoscopy, will be recorded prospectively.	During procedure (5minutes)

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
To compare the effectiveness of AI systems in characterizing polyps with conventional histology	Two of the AI systems include a characterization feature, and this assessment will be compared to conventional histology	During procedure (up to 40 minutes)

Collaborators and Investigators

• Sponsor: Sahlgrenska University Hospital
• Investigators: Study Director: Per Hedenstrom, MD. Ph D, Sahlgrenska University Hospital

Study record dates

Study Major Dates

• Study Start (Actual): March 1, 2023
• Primary Completion (Estimated): March 1, 2026
• Study Completion (Estimated): June 1, 2026

Study Registration Dates

• First Submitted: September 27, 2023
• First Submitted That Met QC Criteria: October 4, 2023
• First Posted (Actual): October 11, 2023

Study Record Updates

• Last Update Posted (Actual): November 24, 2025
• Last Update Submitted That Met QC Criteria: November 19, 2025
• Last Verified: November 1, 2025

More Information

Terms related to this study

• Additional Relevant MeSH Terms: Pathological Conditions, Anatomical; Polyps; Intestinal Polyps; Pathological Conditions, Signs and Symptoms; Colonic Polyps
• Other Study ID Numbers: 981840

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