S24 Transforming Endoscopic Documentation: Novel Artificial Intelligence Summarization Techniques for Pouchoscopy Videos

Background: Endoscopy is crucial in the evaluation and management of patients with Inflammatory Bowel Disease. Managing the patient clinically relies on endoscopic images, often performed by a different gastroenterologist. Ulcerative colitis is a chronic inflammatory bowel disease which may require surgical management including creation of an IPAA (ileal pouch anal anastomosis or J pouch). There are many potential surgical and inflammatory conditions which can involve the J pouch and require close management by IBD specialists. We postulate that having access to endoscopic videos will improve understanding of the complex anatomy and pathology encountered during pouchoscopy. Methods: Our GI division has a large repertoire of endoscopic videos. They are reviewed for patient care and serve as a valuable resource for medical education and training. However, the sheer volume and length of procedural videos presents challenges for information retrieval and consumption. Summarizing these videos by retaining only relevant information offers a concise overview, making the content more accessible and manageable. We developed tools for video summarization, including key frame extraction and video skimming using computer vision techniques. We employ existing endoscopic images to automatically summarize videos. Short clips are created around documented pictures and merged to form a comprehensive video, with anatomical locations titled for each segment. Chapters are added for easy navigation. Results: Following our methodology, 10 endoscopy videos were summarized. The mean length of original videos was 10.26 (5.22-16.21) minutes and summarized videos was 2.54 (1.34-4.02) minutes. The original and summarized videos were reviewed by 2 experienced IBD endoscopists to assess accuracy and veracity. Conclusions: This shift from text-based to video-based documentation in GI procedures represents a significant paradigm change. The summarized videos are available for physicians and medical care teams for documentation, follow-up care, or collaboration. Our method is generalizable to other procedures and departments. Effective video summarization enhances user experience by providing quick overviews, aiding in decision-making, and saving time.

Authors

Institutions

Publication Details

Journal
The American Journal of Gastroenterology
Published
2024-12-01
DOI
https://doi.org/10.14309/01.ajg.0001082628.88973.fe
Primary Topic
Colorectal Cancer Screening and Detection
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

S24 Transforming Endoscopic Documentation: Novel Artificial Intelligence Summarization Techniques for Pouchoscopy Videos

S. Shalini, Navreet Chowla
The American Journal of Gastroenterology
Colorectal Cancer Screening and Detection
article

S24 Transforming Endoscopic Documentation: Novel Artificial Intelligence Summarization Techniques for Pouchoscopy Videos

S. Shalini, Navreet Chowla
article en

Abstract

Background: Endoscopy is crucial in the evaluation and management of patients with Inflammatory Bowel Disease. Managing the patient clinically relies on endoscopic images, often performed by a different gastroenterologist. Ulcerative colitis is a chronic inflammatory bowel disease which may require surgical management including creation of an IPAA (ileal pouch anal anastomosis or J pouch). There are many potential surgical and inflammatory conditions which can involve the J pouch and require close management by IBD specialists. We postulate that having access to endoscopic videos will improve understanding of the complex anatomy and pathology encountered during pouchoscopy. Methods: Our GI division has a large repertoire of endoscopic videos. They are reviewed for patient care and serve as a valuable resource for medical education and training. However, the sheer volume and length of procedural videos presents challenges for information retrieval and consumption. Summarizing these videos by retaining only relevant information offers a concise overview, making the content more accessible and manageable. We developed tools for video summarization, including key frame extraction and video skimming using computer vision techniques. We employ existing endoscopic images to automatically summarize videos. Short clips are created around documented pictures and merged to form a comprehensive video, with anatomical locations titled for each segment. Chapters are added for easy navigation. Results: Following our methodology, 10 endoscopy videos were summarized. The mean length of original videos was 10.26 (5.22-16.21) minutes and summarized videos was 2.54 (1.34-4.02) minutes. The original and summarized videos were reviewed by 2 experienced IBD endoscopists to assess accuracy and veracity. Conclusions: This shift from text-based to video-based documentation in GI procedures represents a significant paradigm change. The summarized videos are available for physicians and medical care teams for documentation, follow-up care, or collaboration. Our method is generalizable to other procedures and departments. Effective video summarization enhances user experience by providing quick overviews, aiding in decision-making, and saving time.

The American Journal of GastroenterologyVol. 119(12S)
Mayo Clinic (US)
Openalex Percentile: Top 60%
Colorectal Cancer Screening and Detection
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.