AI-Assisted Colonoscopy for Polyp Detection

Overseen By Zack Jacob

Age: 18+

Sex: Any

Trial Phase: Academic

Sponsor: Northwestern University

Disqualifiers: Polypectomy, Right colon surgery, others

No Placebo Group All trial participants will receive the active study treatment (no placebo)

Approved in 2 Jurisdictions This treatment is already approved in other countries

Compensation: Varies

Number of Visits: 2

180 Participants Needed

Trial Summary

What is the purpose of this trial?

Based on prior studies, trainee and practicing gastroenterologists miss pre-cancerous polyps (adenomas and serrated polyps) during colonoscopy. The use of computer-aided detection (CADe) systems, a form of artificial intelligence (AI) has been shown to help identify colorectal lesions for practicing gastroenterologists. However, less is known how AI impacts polyp detection for trainees. The investigators are conducting a tandem colonoscopy study wherein a portion of the colon is examined first by the trainee and then the attending physician. For each procedure, randomization will occur which will determine whether or not the trainee will utilize AI for their examination of the colon. At the end of the study, the investigators will determine whether AI helps trainees miss fewer polyps during colonoscopy. The investigators will also conduct interviews with trainees to understand how AI impacts colonoscopy training.

Do I need to stop my current medications for the trial?

The trial information does not specify whether you need to stop taking your current medications.

What data supports the effectiveness of the treatment AI-Assisted Colonoscopy for Polyp Detection?

Research shows that using AI-assisted systems during colonoscopy can improve the detection of adenomas (a type of polyp that can turn into cancer) by highlighting their location, which helps doctors find them more easily. This approach has been shown to increase the adenoma detection rate (ADR) and reduce the number of missed adenomas, making the procedure more effective. ^{1 2 3 4 5}

Is AI-assisted colonoscopy safe for humans?

There is no specific safety data available for AI-assisted colonoscopy in the provided research articles. ^{6 7 8 9 10}

How is AI-assisted colonoscopy different from other treatments for polyp detection?

AI-assisted colonoscopy uses computer-aided detection (CADe) systems to help doctors find polyps more effectively during the procedure by highlighting their locations in real-time, which can reduce the number of missed polyps compared to traditional colonoscopy methods. ^{1 2 3 4 11}

Research Team

Rajesh Keswani, MD

Principal Investigator

Northwestern Medicine

Eligibility Criteria

This trial is for trainee gastroenterologists performing colonoscopies. It aims to see if using AI can help them spot pre-cancerous polyps more effectively. Participants must be in training for gastroenterology and involved in conducting colonoscopies.

Inclusion Criteria

I am an adult scheduled for a colonoscopy.

Exclusion Criteria

Poor bowel preparation (Boston Bowel Preparation Score less than or equal to 6)

I am referred for a procedure to remove polyps or for a colonoscopy.

I have had surgery on the right side of my colon.

Timeline

Screening

Participants are screened for eligibility to participate in the trial

2-4 weeks

Training and Consent

Trainees complete an hour-long meeting with the study team for protocol explanation and consent

1 week

1 visit (in-person)

Colonoscopy Procedure

Trainees perform colonoscopy with or without AI assistance, followed by attending's examination

1 day

1 visit (in-person)

Follow-up

Participants are monitored for safety and effectiveness after the colonoscopy procedure

1 week

Treatment Details

Interventions

• Colonoscopy With Computer-Aided Detection

Trial Overview The study tests whether a computer-aided detection (CADe) system, which is a type of AI, helps trainees find more polyps during a colonoscopy compared to not using the technology. Trainees will either use AI or not by random choice and results are compared.

Participant Groups

Group I: Colonoscopy with AI Active Control 1 Intervention

Trainee using AI during colonoscopy inspection

Group II: Colonoscopy without AI Active Control 1 Intervention

Trainee not using AI during colonoscopy inspection

Colonoscopy With Computer-Aided Detection is already approved in United States, European Union for the following indications:

Approved in United States as Colonoscopy with CADe for:

• Colorectal cancer screening
• Detection of adenomas and serrated polyps

Approved in European Union as AI-Assisted Colonoscopy for:

• Colorectal cancer screening
• Detection of adenomas and serrated polyps

Who Is Running the Clinical Trial?

Northwestern University

Lead Sponsor

Trials

1,674

Recruited

989,000+

Findings from Research

The EndoMind computer-aided polyp detection system successfully identified all adenomas in a study of 41 colonoscopies, achieving an adenoma detection rate (ADR) of 41.5% and a median time to first detection of just 130 milliseconds.

With a low false positive rate of 2.2% and a high usability score of 96.3, EndoMind shows promise for enhancing polyp detection in clinical settings, although further research is needed to confirm its clinical benefits in randomized controlled trials.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Pilot study of a new freely available computer-aided polyp detection system in clinical practice. Lux, TJ., Banck, M., Saßmannshausen, Z., et al.[2022]

In a study involving 2015 participants, real-time computer-aided detection (CADe) during colonoscopy improved the adenoma detection rate (ADR) to 37.5% compared to 33.7% in standard colonoscopy, suggesting a potential benefit in polyp detection.

The study indicated that while CADe-assisted colonoscopy showed a slight increase in ADR, the difference was not statistically significant (p=0.051), and safety was generally good with only one minor bleeding event reported.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Effect of real-time computer-aided detection of colorectal adenoma in routine colonoscopy (COLO-GENIUS): a single-centre randomised controlled trial. Karsenti, D., Tharsis, G., Perrot, B., et al.[2023]

In a study of 790 patients, the use of a computer-aided detection (CADe) system during colonoscopy significantly improved the adenoma detection rate (ADR) from 20.91% to 29.01%, indicating enhanced polyp detection capabilities.

The integration of the CADe system did not increase physician fatigue levels, suggesting it can be a beneficial tool for improving colonoscopy outcomes without compromising the comfort of the endoscopists.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

The single-monitor trial: an embedded CADe system increased adenoma detection during colonoscopy: a prospective randomized study. Liu, P., Wang, P., Glissen Brown, JR., et al.[2022]

In a study involving 223 patients undergoing tandem colonoscopy, the use of an artificial intelligence-based computer-aided polyp detection (CADe) system significantly reduced the adenoma miss rate (AMR) compared to traditional high-definition white light (HDWL) colonoscopy, with an AMR of 20.12% in the CADe group versus 31.25% in the HDWL group.

The CADe system also demonstrated a lower miss rate for sessile serrated lesions (SSL) and a higher adenomas per colonoscopy (APC) rate, indicating that CADe can enhance polyp detection and improve the effectiveness of colonoscopy screenings.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Deep Learning Computer-aided Polyp Detection Reduces Adenoma Miss Rate: A United States Multi-center Randomized Tandem Colonoscopy Study (CADeT-CS Trial). Glissen Brown, JR., Mansour, NM., Wang, P., et al.[2022]

The use of artificial intelligence computer-aided detection (CADe) during colonoscopy significantly increases adenoma detection rates (ADRs), with a rate of 44.0% in the CADe group compared to 35.9% in the standard group, indicating a 55% reduction in adenoma miss rates.

However, CADe also leads to a higher number of unnecessary removals of nonneoplastic polyps (0.52 vs. 0.34 per colonoscopy), suggesting a potential risk of overdiagnosis and overtreatment.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Real-Time Computer-Aided Detection of Colorectal Neoplasia During Colonoscopy : A Systematic Review and Meta-analysis. Hassan, C., Spadaccini, M., Mori, Y., et al.[2023]

Colon capsule endoscopy (CCE) is a reliable and safe method for screening colorectal cancer (CRC), with a success rate of 68.9% in a study of 161 patients across six German centers from 2010 to 2015.

The procedure detected significant pathological findings, including polyps in 32.3% of patients and five confirmed carcinomas, indicating its efficacy in identifying potential cancerous conditions.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Colon Capsule Endoscopy: Indications, Findings, and Complications - Data from a Prospective German Colon Capsule Registry Trial (DEKOR). Hausmann, J., Tal, A., Gomer, A., et al.[2021]

Information technology methods, such as event monitoring and natural language processing, can effectively and cost-efficiently detect adverse events like adverse drug events and nosocomial infections in clinical databases.

These techniques are already in use in some hospitals and have the potential to be adapted for broader detection of various adverse events as more medical data becomes computerized.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Detecting adverse events using information technology. Bates, DW., Evans, RS., Murff, H., et al.[2022]

A new computerized system for detecting adverse drug events (ADEs) was developed in China, featuring five functional modules that enhance the detection and assessment of ADEs in hospitalized patients.

The system demonstrated high positive predictive values (44.4% to 95.8%) for monitoring specific ADEs like drug-related thrombocytopenia and liver injury, showing improved performance over previous methods that relied solely on laboratory values.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Development of a Computer-Assisted Adverse Drug Events Alarm and Assessment System for Hospital Inpatients in China. Chen, C., Jia, W., Guo, D., et al.[2021]

In a 6-month study at a 140-bed community hospital, a computer-based alerting system generated 3547 alerts, with 528 classified as high or critical priority, indicating its potential to identify serious medication-related issues.

The review of high-priority alerts revealed that approximately 37 adverse drug events (ADEs) could be prevented annually, demonstrating the system's effectiveness in mitigating medication toxicity through timely clinical intervention.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Adverse drug event detection in a community hospital utilising computerised medication and laboratory data. Seger, AC., Jha, AK., Bates, DW.[2018]

The web-based adverse event tracking system (eAETS) has effectively supported 175 clinical protocols over four years, capturing 2,440 adverse event reports, including minor symptoms that could indicate serious issues.

Out of the reported AEs, 1,053 did not align with the initial risk profiles, leading to corrective recommendations in 13% of the protocols, highlighting the system's role in enhancing subject safety by identifying unanticipated patterns.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

The impact of minor adverse event tracking on subject safety: a web-based system. Shenvi, NV., Gebhart, SS.[2009]

In a multicenter randomized controlled trial involving 3059 asymptomatic participants, AI-assisted colonoscopy significantly improved the overall adenoma detection rate (ADR) to 39.9% compared to 32.4% with conventional colonoscopy, indicating enhanced efficacy in detecting polyps.

The AI-assisted method also increased the detection of advanced adenomas and showed better performance for both expert and non-expert endoscopists, while slightly increasing the median withdrawal time during the procedure.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Artificial Intelligence-Assisted Colonoscopy for Colorectal Cancer Screening: A Multicenter Randomized Controlled Trial. Xu, H., Tang, RSY., Lam, TYT., et al.[2023]

References

1. Germany pubmed.ncbi.nlm.nih.gov

Pilot study of a new freely available computer-aided polyp detection system in clinical practice. [2022]

2. Netherlands pubmed.ncbi.nlm.nih.gov

Effect of real-time computer-aided detection of colorectal adenoma in routine colonoscopy (COLO-GENIUS): a single-centre randomised controlled trial. [2023]

3. England pubmed.ncbi.nlm.nih.gov

The single-monitor trial: an embedded CADe system increased adenoma detection during colonoscopy: a prospective randomized study. [2022]

4. United States pubmed.ncbi.nlm.nih.gov

Deep Learning Computer-aided Polyp Detection Reduces Adenoma Miss Rate: A United States Multi-center Randomized Tandem Colonoscopy Study (CADeT-CS Trial). [2022]

5. United States pubmed.ncbi.nlm.nih.gov

Real-Time Computer-Aided Detection of Colorectal Neoplasia During Colonoscopy : A Systematic Review and Meta-analysis. [2023]

6. Korea (South) pubmed.ncbi.nlm.nih.gov

Colon Capsule Endoscopy: Indications, Findings, and Complications - Data from a Prospective German Colon Capsule Registry Trial (DEKOR). [2021]

7. England pubmed.ncbi.nlm.nih.gov

Detecting adverse events using information technology. [2022]

8. Switzerland pubmed.ncbi.nlm.nih.gov

Development of a Computer-Assisted Adverse Drug Events Alarm and Assessment System for Hospital Inpatients in China. [2021]

9. New Zealand pubmed.ncbi.nlm.nih.gov

Adverse drug event detection in a community hospital utilising computerised medication and laboratory data. [2018]

10. United States pubmed.ncbi.nlm.nih.gov

The impact of minor adverse event tracking on subject safety: a web-based system. [2009]

11. United States pubmed.ncbi.nlm.nih.gov

Artificial Intelligence-Assisted Colonoscopy for Colorectal Cancer Screening: A Multicenter Randomized Controlled Trial. [2023]