Real-Time Feedback AI for Colonoscopy Quality Improvement

Recruiting at 1 trial location

Overseen By James Villar-Mead

Age: 18+

Sex: Any

Trial Phase: Academic

Sponsor: University of Minnesota

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

Compensation: Varies

Number of Visits: 1

Duration: 1 day

4150 Participants Needed

What You Need to Know Before You Apply

What is the purpose of this trial?

This trial tests if giving doctors timely feedback during endoscopic exams can improve the quality of these procedures for patients.

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

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

Is the AI system for colonoscopy generally safe for humans?

The research does not report any safety concerns related to the use of AI systems in colonoscopy, suggesting they are generally safe for human use. ^{1 2 3 4 5}

How does the AI treatment for colonoscopy differ from other treatments?

This AI treatment for colonoscopy is unique because it provides real-time feedback to improve the quality of the procedure by detecting and characterizing polyps more accurately than traditional methods. It uses advanced artificial intelligence techniques to match or exceed human expert performance, reducing the risk of missed lesions and potentially preventing colorectal cancer. ^{4 6 7 8 9}

What data supports the effectiveness of the treatment AI program for colonoscopy?

Research shows that AI systems in colonoscopy can match human experts in detecting and classifying polyps, which are small growths in the colon that can lead to cancer. AI can also provide real-time feedback to improve the quality of colonoscopy procedures, potentially reducing the number of missed polyps and improving overall detection rates. ^{3 8 10 11 12}

Who Is on the Research Team?

Piet de Groen, MD

Principal Investigator

UMN

Are You a Good Fit for This Trial?

This trial is for any endoscopist who is willing to participate and performs routine colonoscopy procedures. Specific details on exclusion criteria are not provided, but typically these would include factors that could interfere with the study or the safety of participants.

Inclusion Criteria

I am an endoscopist willing to participate and perform routine colonoscopies.

Timeline for a Trial Participant

Screening

Participants are screened for eligibility to participate in the trial

2-4 weeks

Treatment

Participants receive real-time feedback during colonoscopy to improve mucosal inspection and clearing of fecal debris

1 day

1 visit (in-person)

Follow-up

Participants are monitored for safety and effectiveness after treatment

5 months

What Are the Treatments Tested in This Trial?

Interventions

• AI program for colonoscopy

Trial Overview The trial is testing an AI program designed to provide real-time feedback during colonoscopies. The goal is to see if this technology can improve the quality of endoscopic examinations potentially leading to better outcomes in colorectal cancer screening.

How Is the Trial Designed?

Group I: Testing of degree of mucosal inspection Experimental Treatment 1 Intervention

AI provides real-time feedback related to circumferential views during endoscope removal

Group II: Testing of clearing of fecal debris Experimental Treatment 1 Intervention

AI provides real-time feedback related to removal of remaining fecal debris

Who Is Running the Clinical Trial?

University of Minnesota

Lead Sponsor

Trials

1,459

Recruited

1,623,000+

University of Washington

Collaborator

Trials

1,858

Recruited

2,023,000+

Johns Hopkins University

Collaborator

Trials

2,366

Recruited

15,160,000+

Published Research Related to This Trial

Computer-aided diagnosis in colonoscopy can significantly reduce the miss rates for polyps, which are currently as high as 22%, potentially decreasing the risk of interval colorectal cancers.

Recent advancements in artificial intelligence have led to algorithms that can match the performance of human experts in detecting and characterizing polyps, enhancing the reliability of optical biopsy techniques.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Artificial intelligence and computer-aided diagnosis in colonoscopy: current evidence and future directions. Ahmad, OF., Soares, AS., Mazomenos, E., et al.[2019]

The study evaluated the diagnostic accuracy of 67 endoscopists in identifying colorectal neoplasia, revealing a pooled sensitivity of 84.5% and specificity of 83% for detecting adenomas, indicating that while endoscopists perform well, there is still room for improvement.

Expert endoscopists demonstrated significantly higher sensitivity (90.5%) compared to non-experts (75.5%), suggesting that targeted training and competence evaluation, especially in the context of artificial intelligence validation, could enhance diagnostic performance.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Endoscopists performance in optical diagnosis of colorectal polyps in artificial intelligence studies. Pecere, S., Antonelli, G., Dinis-Ribeiro, M., et al.[2022]

Colonoscopy is the most effective method for preventing colorectal cancer, but its success relies heavily on the quality of the procedure, including how well it is planned and performed.

Improvements in detection rates of polyps and adenomas can be achieved through better patient preparation, advanced endoscopic technologies like wide-angle endoscopes, and the use of artificial intelligence tools for enhanced polyp detection.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Optimizing Screening Colonoscopy: Strategies and Alternatives. Allescher, HD., Weingart, V.[2020]

Artificial intelligence (AI) is increasingly being recognized as a valuable tool in gastroenterology, particularly for enhancing colonoscopy through applications like computer-aided polyp detection (CADe) and computer-assisted diagnosis (CADx).

While AI shows promise in improving colonoscopy quality metrics, such as assessing bowel cleansing and accurately locating lesions, there are concerns about cost-effectiveness and a lack of large-scale studies demonstrating its impact on critical outcomes like colorectal cancer incidence and mortality.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Artificial Intelligence Applied to Colonoscopy: Is It Time to Take a Step Forward? Gimeno-García, AZ., Hernández-Pérez, A., Nicolás-Pérez, D., et al.[2023]

This study aims to improve polyp detection rates among colonoscopists by using an automated feedback system that provides real-time performance data, specifically the optimal procedure-adjusted detection KPI (opadKPI), to endoscopists and their centers.

If successful, this intervention could enhance the quality of colonoscopy procedures across the UK by leveraging a learning health system that uses real-time data to inform and improve clinical practices, potentially reducing colorectal cancer incidence and mortality.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

The National Endoscopy Database (NED) Automated Performance Reports to Improve Quality Outcomes Trial (APRIQOT) randomized controlled trial design. Catlow, J., Sharp, L., Kasim, A., et al.[2021]

The AI-driven smartphone app significantly improved bowel preparation quality for colonoscopy, with 88.54% of patients achieving adequate preparation compared to 65.59% in the control group, based on a study of 524 outpatients.

Patients using the app also showed higher compliance with dietary restrictions (93.68% vs. 83.81%) and purgative instructions (96.05% vs. 84.62%), indicating that the app not only enhances preparation quality but also encourages better adherence to pre-colonoscopy guidelines.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Improving bowel preparation for colonoscopy with a smartphone application driven by artificial intelligence. Zhu, Y., Zhang, DF., Wu, HL., et al.[2023]

In a study involving 660 patients, the use of computer-aided detection (CADe) significantly increased the adenoma detection rate (ADR) among less experienced colonoscopists, achieving an ADR of 53.3% compared to 44.5% in the control group, indicating that CADe can enhance colonoscopy quality.

The findings suggest that the effectiveness of CADe in improving ADR is more impactful than the experience level of the endoscopist, as experience had a minor role in determining ADR outcomes.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Artificial intelligence and colonoscopy experience: lessons from two randomised trials. Repici, A., Spadaccini, M., Antonelli, G., et al.[2022]

An artificial intelligence (AI) system developed for real-time detection of colorectal cancer during colonoscopy demonstrated high sensitivity (97.3%) and specificity (99.0%), significantly reducing the risk of missed lesions.

The AI system is particularly effective in identifying non-polypoid lesions, which are often overlooked, and can predict cancerous regions in just 21.9 milliseconds per image, enhancing the early detection capabilities of endoscopists.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Development of a real-time endoscopic image diagnosis support system using deep learning technology in colonoscopy. Yamada, M., Saito, Y., Imaoka, H., et al.[2021]

In a study involving 45 subjects, trainee endoscopists (TEs) using artificial intelligence (AI) assistance achieved adenoma detection rates (ADR) and polyp detection rates (PDR) comparable to experienced endoscopists (EEs), indicating that AI can help bridge the performance gap between TEs and EEs.

The study found no significant differences in adenoma miss rate (AMR) or polyp miss rate (PMR) between AI-assisted TEs and EEs, suggesting that AI may enhance the quality of screening colonoscopy and potentially reduce missed cancers.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Real-time, computer-aided, detection-assisted colonoscopy eliminates differences in adenoma detection rate between trainee and experienced endoscopists. Biscaglia, G., Cocomazzi, F., Gentile, M., et al.[2022]

Providing individualized feedback on colonoscopy inspection quality (CIQ) through instructional videos led to a significant improvement in adenoma detection rate (ADR) and serrated detection rate (SDR) among lower-performing colonoscopists, with ADR increasing from 31.1% to 34.3% and SDR from 7.2% to 9.1%.

Higher-performing colonoscopists did not show any significant improvement in ADR or SDR after receiving feedback, indicating that the benefits of CIQ feedback are primarily for those with lower baseline performance.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Individualized feedback on colonoscopy skills improves group colonoscopy quality in providers with lower adenoma detection rates. Keswani, RN., Wood, M., Benson, M., et al.[2023]

Artificial intelligence (AI) is significantly enhancing the capabilities of gastrointestinal endoscopy, particularly in improving the detection of lesions during colonoscopy, which can help reduce the rate of missed neoplasia and interval cancers.

AI also aids in the characterization of lesions, potentially allowing for cost-saving strategies like 'leave-in-situ' for small hyperplastic lesions and 'resect and discard' for other small lesions, thus optimizing clinical decision-making.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Artificial intelligence-aided colonoscopy: Recent developments and future perspectives. Antonelli, G., Gkolfakis, P., Tziatzios, G., et al.[2021]

Automated image analysis methods for colonoscopy, particularly those using artificial intelligence, have shown promise in improving colonoscopy quality and detecting abnormalities, which may help prevent colorectal cancer.

Clinical trials, including multi-center studies, have demonstrated the effectiveness of these AI technologies, highlighting the need for further development to meet clinical requirements and enhance patient outcomes.

NCBI (Pubmed)

pubmed.ncbi.nlm.nih.gov

Artificial Intelligence for Colonoscopy: Past, Present, and Future. Tavanapong, W., Oh, J., Riegler, MA., et al.[2022]

Citations

1. Netherlands pubmed.ncbi.nlm.nih.gov

Artificial intelligence and computer-aided diagnosis in colonoscopy: current evidence and future directions. [2019]

2. England pubmed.ncbi.nlm.nih.gov

Endoscopists performance in optical diagnosis of colorectal polyps in artificial intelligence studies. [2022]

3. Switzerland pubmed.ncbi.nlm.nih.gov

Optimizing Screening Colonoscopy: Strategies and Alternatives. [2020]

4. Switzerland pubmed.ncbi.nlm.nih.gov

Artificial Intelligence Applied to Colonoscopy: Is It Time to Take a Step Forward? [2023]

5. Germany pubmed.ncbi.nlm.nih.gov

The National Endoscopy Database (NED) Automated Performance Reports to Improve Quality Outcomes Trial (APRIQOT) randomized controlled trial design. [2021]

6. England pubmed.ncbi.nlm.nih.gov

Improving bowel preparation for colonoscopy with a smartphone application driven by artificial intelligence. [2023]

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

Artificial intelligence and colonoscopy experience: lessons from two randomised trials. [2022]

8. England pubmed.ncbi.nlm.nih.gov

Development of a real-time endoscopic image diagnosis support system using deep learning technology in colonoscopy. [2021]

9. Germany pubmed.ncbi.nlm.nih.gov

Real-time, computer-aided, detection-assisted colonoscopy eliminates differences in adenoma detection rate between trainee and experienced endoscopists. [2022]

10. Germany pubmed.ncbi.nlm.nih.gov

Individualized feedback on colonoscopy skills improves group colonoscopy quality in providers with lower adenoma detection rates. [2023]

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

Artificial intelligence-aided colonoscopy: Recent developments and future perspectives. [2021]

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

Artificial Intelligence for Colonoscopy: Past, Present, and Future. [2022]