Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 30, 2026Diagnostic PathologyOpen Access

Detection of collagen band–associated regions in H&E-stained colonic biopsies of collagenous colitis patients using superpixel-based feature extraction and neural network classification

View Full Paper
Ask AI
Bookmark
Share

Authors

VKVytautas KiudelisRPRobertas PetrolisRRRima Ramonaitė

Discussion

Loading...

Member takes

Overview

Machine learning detects collagen band regions in colonic biopsies, suggesting improved histopathological assessment.

Key Points

  • To develop a machine-learning algorithm for detecting collagen band regions in H&E-stained colonic biopsy images.
  • Collected H&E-stained colonic biopsy images from 36 patients with collagenous colitis.
  • Segmented images into superpixels using the Simple Linear Iterative Clustering (SLIC) algorithm.
  • Characterized superpixels with RGB histograms; trained a feed-forward neural network classifier.
  • Addressed class imbalance via data augmentation; applied size filtering for post-processing.
  • Achieved a superpixel-wise accuracy of 0.928 with sensitivity of 0.898 and specificity of 0.953.
  • Post-processing significantly reduced false-positive detections.
  • Final algorithm reached an image-level acceptability accuracy of 0.846 according to expert evaluations.

Cite This Study

Kiudelis et al. (2026) studied this question.

synapsesocial.com/papers/69ca134b883daed6ee09529chttps://doi.org/10.1186/s13000-026-01783-x
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Supervised Learning for Collagen Segmentation in Bright-Field Histology: A Comparative Evaluation of U-Net and MLP2026
  2. 2Abstract 4167: Virtual inference of collagen architecture from H&E to characterize stromal fiber morphology and organization in colorectal cancer2026
  3. 3Deep learning-based quantification of collagen and associated features from H&E-stained whole slide pathology images across cancer types2026
  4. 4Deep learning layer-specific collagen quantification correlates with activity and is associated with outcomes in Crohn’s disease2026
  5. 5Classification of colorectal tissue histopathological images using amalgamated hand-crafted features2026