Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Randomized Controlled Trial
. 2024 Oct 26;14(1):25453.
doi: 10.1038/s41598-024-77079-1.

A prospective multicenter randomized controlled trial on artificial intelligence assisted colonoscopy for enhanced polyp detection

Dong Kyun Park  1   2 , Eui Joo Kim  1 , Jong Pil Im  3 , Hyun Lim  4 , Yun Jeong Lim  5 , Jeong-Sik Byeon  6 , Kyoung Oh Kim  1 , Jun-Won Chung  1 , Yoon Jae Kim  7   8
Affiliations
Randomized Controlled Trial

A prospective multicenter randomized controlled trial on artificial intelligence assisted colonoscopy for enhanced polyp detection

Dong Kyun Park et al. Sci Rep. .

Abstract

Colon polyp detection and removal via colonoscopy are essential for colorectal cancer screening and prevention. This study aimed to develop a colon polyp detection program based on the RetinaNet algorithm and verify its clinical utility. To develop the AI-assisted program, the dataset was fully anonymized and divided into 10 folds for 10-fold cross-validation. Each fold consisted of 9,639 training images and 1,070 validation images. Video data from 56 patients were used for model training, and transfer learning was performed using the developed still image-based model. The final model was developed as a real-time polyp-detection program for endoscopy. To evaluate the model's performance, a prospective randomized controlled trial was conducted at six institutions to compare the polyp detection rates (PDR). A total of 805 patients were included. The group that utilized the AI model showed significantly higher PDR and adenoma detection rate (ADR) than the group that underwent colonoscopy without AI assistance. Multivariate analysis revealed an OR of 1.50 for cases where polyps were detected. The AI-assisted polyp-detection program is clinically beneficial for detecting polyps during colonoscopy. By utilizing this AI-assisted program, clinicians can improve adenoma detection rates, ultimately leading to enhanced cancer prevention.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Figure 1
Figure 1
Flow chart for the development of a polyp detection program using AI.
Figure 2
Figure 2
Actual polyp detection program image.
Figure 3
Figure 3
The actual operational screen of the polyp detection program transplanted into the PACS system.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Chen, C. D., Yen, M. F., Wang, W. M., Wong, J. M. & Chen, T. H. A case-cohort study for the disease natural history of adenoma-carcinoma and de novo carcinoma and surveillance of colon and rectum after polypectomy: implication for efficacy of colonoscopy. Br. J. Cancer. 88, 1866–1873. 10.1038/sj.bjc.6601007 (2003). - PMC - PubMed
    1. Dilly, C. K. & Kahi, C. J. Does increased Adenoma Detection reduce the risk of Colorectal Cancer, and how good do we need to be? Curr. Gastroenterol. Rep.21, 9. 10.1007/s11894-019-0678-5 (2019). - PubMed
    1. Vinsard, D. G. et al. Quality assurance of computer-aided detection and diagnosis in colonoscopy. Gastrointest. Endosc. 90, 55–63. 10.1016/j.gie.2019.03.019 (2019). - PubMed
    1. East, J. E. et al. Surface visualization at CT colonography simulated colonoscopy: effect of varying field of view and retrograde view. Am. J. Gastroenterol.102, 2529–2535. 10.1111/j.1572-0241.2007.01429.x (2007). - PubMed
    1. Lee, A. & Tutticci, N. Enhancing polyp detection: technological advances in colonoscopy imaging. Transl Gastroenterol. Hepatol.6, 61. 10.21037/tgh.2020.02.05 (2021). - PMC - PubMed

Publication types