A Study Comparing Standard and AI-Assisted Colonoscopies for Detecting and Characterizing Colorectal Lesions in Adults Aged 50-74 Undergoing Cancer Screening

Javier Santos Fernández

Status: Completed

Conditions: Adenoma Colon Polyp; Colorectal Neoplasms; Colorectal Cancer

Treatments: Diagnostic Test: Artificial Intelligence-Assisted Colonoscopy

Study type: Interventional

Funder types: Other

Identifiers

NCT07125300

2024/023

Details and patient eligibility

About

The goal of this clinical trial is to determine whether using artificial intelligence (AI) can improve the detection and characterization of abnormal growths (polyps) during colonoscopy in adults aged 50 to 74 years who are undergoing colorectal cancer screening after a positive stool test.

The main questions it aims to answer are:

• Does AI assistance increase the detection of adenomas or advanced colorectal neoplasia?
• Does AI provide more accurate optical diagnosis of polyps compared to standard assessment by endoscopists?

Researchers will compare colonoscopies performed with AI assistance (using the CAD EYE™ system) to standard colonoscopies without AI to see if AI improves detection rates or diagnostic accuracy.

Participants will:

• Undergo a screening colonoscopy after a positive fecal immunochemical test (FIT)
• Be randomly assigned to either an AI-assisted or standard colonoscopy group
• Have any detected polyps removed and analyzed
• Receive either AI-based or physician-based optical diagnosis of polyps during the procedure

This study helps evaluate whether AI can make colonoscopies more effective and reduce unnecessary polyp removals.

Enrollment: 368 patients

Sex: All

Ages: 50 to 74 years old

Volunteers: Accepts Healthy Volunteers

Inclusion criteria

• Adults aged 50 to 74 years
• Positive fecal immunochemical test (FIT) result (≥100 ng/mL)
• Scheduled for screening colonoscopy within a population-based colorectal cancer screening program
• Able and willing to provide written informed consent

Exclusion criteria

• Incomplete colonoscopy (e.g., failure to reach the cecum)
• Inadequate bowel preparation
• History of colorectal surgery
• Inability to provide informed consent

Trial design

Primary purpose: Diagnostic

Allocation: Randomized

Interventional model: Parallel Assignment

Masking: Triple Blind

---

368 participants in 2 patient groups

Conventional colonoscopy without AI assistance

No Intervention group

AI-assisted colonoscopy

Experimental group

Treatment:

Diagnostic Test: Artificial Intelligence-Assisted Colonoscopy