Endoscopy 2025; 57(S 02): S158
DOI: 10.1055/s-0045-1805404
Abstracts | ESGE Days 2025
Oral presentation
Colorectal lesion detection: AI and beyond! 05/04/2025, 10:30 – 11:30 Room 122+123

A novel switching of artificial intelligence to generate simultaneously multimodal images to assess inflammation and predict outcomes in Ulcerative Colitis

Authors

  • M Iacucci

    1   APC Microbiome Ireland, College of Medicine and Health, University College of Cork, Cork, Ireland
    2   University of Birmingham, Institute of Immunology and Immunotherapy, Birmingha, United Kingdom

  • I Zammarchi

    1   APC Microbiome Ireland, College of Medicine and Health, University College of Cork, Cork, Ireland

  • G Santacroce

    1   APC Microbiome Ireland, College of Medicine and Health, University College of Cork, Cork, Ireland

  • B B Kolawole

    3   School of Engineering Computer Science and Informatics, London South Bank University, London, United Kingdom

  • U Chaudhari

    3   School of Engineering Computer Science and Informatics, London South Bank University, London, United Kingdom

  • R Del Amor

    4   Instituto de Investigación e Innovación en Bioingeniería, Universitat Politècnica de València, Valencia, Spain

  • P Meseguer

    4   Instituto de Investigación e Innovación en Bioingeniería, Universitat Politècnica de València, Valencia, Spain
    5   valgrAI – Valencian Graduate School and Research Network of Artificial Intelligence, Valencia, Spain

  • V Naranjo

    4   Instituto de Investigación e Innovación en Bioingeniería, Universitat Politècnica de València, Valencia, Spain

  • M Puga-Tejada

    1   APC Microbiome Ireland, College of Medicine and Health, University College of Cork, Cork, Ireland

  • I Capobianco

    1   APC Microbiome Ireland, College of Medicine and Health, University College of Cork, Cork, Ireland

  • I Ditonno

    1   APC Microbiome Ireland, College of Medicine and Health, University College of Cork, Cork, Ireland

  • A Buda

    6   Department of Gastrointestinal Oncological Surgery, Santa Maria del Prato Hospital, Feltre, Italy

  • B Hayes

    7   Department of Histopathology, Cork University Hospital, Cork, Ireland

  • R Crotty

    7   Department of Histopathology, Cork University Hospital, Cork, Ireland

  • R Bisschops

    8   UZ Leuven Gasthuisberg Campus, Leuven, Belgium

  • S Ghosh

    1   APC Microbiome Ireland, College of Medicine and Health, University College of Cork, Cork, Ireland

  • E Grisan

    3   School of Engineering Computer Science and Informatics, London South Bank University, London, United Kingdom
Further Information
Also available at
 

Aims Virtual Chromoendoscopy (VCE) is pivotal for assessing activity and predicting outcomes in Ulcerative Colitis (UC), though inter- and intra-observer variability and the need for expertise persist. Artificial intelligence (AI) has the potential to offer standardised VCE-based assessment. This study introduces a novel AI model to detect, generate and transition between various endoscopic modalities, enhancing AI-driven inflammation assessment and outcome prediction in UC.

Methods Endoscopic videos in high-definition white-light (HD-WLE), iScan2, iScan3 and NBI modalities from UC patients of the international PICaSSO iScan and Narrow-Band Imaging (NBI) cohort (302 and 54 patients, respectively) were used to develop a neural network (NN) able to identify the acquisition modality of each frame and for inter-modality image switching. 2535 frames were switched to different endoscopic modalities and used to train a deep-learning model for inflammation assessment using single and multimodal inputs on 169 videos of the iScan cohort. Subsequently, the model was tested on a subset of the iScan and NBI cohort (72 and 51 videos, 1080 and 765 frames, respectively). The model performance in predicting endoscopic and histological activity and outcomes and the agreement with experts were evaluated.

Results The model efficiently classified and converted images across modalities (92% NN classifier accuracy). The model excellently predicted endoscopic and histological remission in the iScan cohort. It exhibited an accuracy of 91.67% (95% CI 82.74-96.88) and 88.89% (79.28-95.08) and AUROC of 0.96 and 0.90 in predicting endoscopic remission by Ulcerative Colitis Endoscopic Index of Severity (UCEIS) and Paddington International Virtual Chromoendoscopy (PICaSSO) scores, respectively. Similarly, the AI multimodal model achieved an accuracy of 90.28% (80.99-96), 88.89% (79.28-95.08) and 90.28% (80.99-96) and AUROC of 0.94, 0.92 and 0.95 in predicting histologic remission by Robarts Histopathology Index (RHI), Nancy Histopathology Index (NHI) and Picasso Histologic Remission Index (PHRI), respectively. Excellent performance was also confirmed when testing in the NBI cohort. Moreover, it showed a remarkable ability to predict clinical outcomes in the iScan and NBI cohort (HR 3.18 [0.98-10.35] and 1.7 [0.7-4.11] by endoscopy; 5.75 [1.77-18-71] and 3.9 [1.15-13.28] by histology, respectively). Finally, the agreement with the assessment performed by endoscopists and pathologists was good [1] [2].

Conclusions Our multimodal “AI-switching” model innovatively detects, generates and transitions between different endoscopic enhancement modalities and platforms, refining inflammation assessment, outcome prediction and precise UC management by integrating model-derived images.



Publication History

Article published online:
27 March 2025

© 2025. European Society of Gastrointestinal Endoscopy. All rights reserved.

Georg Thieme Verlag KG
Oswald-Hesse-Straße 50, 70469 Stuttgart, Germany

 

  • References

  • 1 Chaudhari U.. et al. 'Simultaneous Detection and Conversion Among Endoscopic Enhancement Modalities,'. 2024 IEEE International Symposium on Biomedical Imaging (ISBI); Athens, Greece: 2024. pp 1-4
    CrossrefSearch in Google Scholar
  • 2 Kolawole B.B.. et al. 'Inflammation Detection Using Ensemble Endoscopic Multimodal Assessment in Inflammatory Bowel Disease,'. 2024 IEEE International Symposium on Biomedical Imaging (ISBI); Athens, Greece: 2024. pp 1-4
    CrossrefSearch in Google Scholar