Assisted documentation as new focus for artificial intelligence in endoscopy: The precedent of reliable withdrawal time and image reporting.

Researchers

Journal

Modalities

Models

Abstract

Reliable documentation is essential for maintaining quality standards in endoscopy. However, in clinical practice the report quality varies. We developed an artificial intelligence based prototype for the measurement of withdrawal and intervention time, as well as automatic photo-documentation.A multi class deep learning algorithm distinguishing different endoscopic image content was trained with 10,557 images (1,300 examinations, 9 centers, 4 processors). Consecutively, the algorithm was used to calculate withdrawal time and extract relevant images. Validation was performed on 100 colonoscopy videos (5 centers). Reported and predicted withdrawal time were compared to video-based measurement and photo-documentation were compared regarding documented polypectomies.Video-based measurement in 100 colonoscopies revealed a median absolute difference of 2.0min between measured and reported withdrawal time compared to 0.4min for predictions. The original photo-documentation represented the cecum in 88 compared to 98 of 100 examinations for the generated documentation. For 39 of 104 polypectomies, the examiners’ photos included the instrument compared to 68 for the artificial intelligence’s images. Lastly, we demonstrated real-time capability (10 colonoscopies).Our system calculates withdrawal time, provides an image report, and is real-time-ready. After further validation, the system may improve standardized reporting while decreasing workload caused by routine documentation.The Author(s). This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. (https://creativecommons.org/licenses/by/4.0/).