PulseTrendingJournal ClubResearchersJournalsExplore
Instagram
HomeTrendingJournal ClubExplore
Synapse
⌘+K
Synapse
January 23, 2026Journal of Crohn s and Colitis

P0492Deep learning for automated and objective in vivo assessment of intestinal epithelial barrier in IBD using probe-based confocal laser endomicroscopy

View Full Paper
Ask AI
Bookmark
Share

Authors

LPL ParisioRARaja AtreyaFCF Caprioli

Discussion

Loading...

Member takes

Overview

Automated deep learning models assess intestinal barrier function in IBD, suggesting enhanced clinical evaluation methods.

Key Points

  • This study aims to develop deep learning models for objective analysis of intestinal epithelial barrier integrity using pCLE.
  • Developed two deep learning models based on ResNet50 and ConvNeXtTiny architectures.
  • Model 1 filters pCLE image frames for informativeness.
  • Model 2 classifies frames for leakage or healing.
  • Models optimized for colonic and ileal datasets.
  • Video analysis produces a leakage score and classification.
  • Model 1 achieved 99% accuracy in classifying informative frames.
  • Model 2 showed test accuracies of 90% and 91% for colonic and ileal images, respectively.
  • Video pipeline distinguished leakage and healing cases with AUC values of 0.94 and 0.87.

Cite This Study

Parisio et al. (2026) studied this question.

synapsesocial.com/papers/69730f34c8125b09b0d1f0a2https://doi.org/10.1093/ecco-jcc/jjaf231.673
View Full Paper
Ask AI
Bookmark
Share