Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
May 6, 2026Cardiology PlusOpen Access

Histological validation of artificial intelligence–driven automatic plaque characterization in coronary OCT: a head-to-head comparison with clinicians

View Full Paper
Ask AI
Bookmark
Share

Authors

MCMiao ChuFRFrancesca RazziGMGiovanni Luigi De Maria

Discussion

Loading...

Member takes

Overview

Observational study shows AI correlates well with histology in plaque characterization of atherosclerotic arteries, implying reduced subjectivity.

Key Points

  • This research evaluates the performance of an AI model in characterizing coronary plaque against histological standards.
  • Analyzed OCT pullbacks and histological sections from 25 plaques in 11 swine arteries.
  • Used a coarse-to-fine approach for OCT-histology co-registration with anatomical references.
  • Three independent readers annotated OCT images for comparison against AI model outputs.
  • The AI model showed high correlation with histology for fibrous and lipidic components (Spearman's ρ = 0.907 and ρ = 0.900, respectively).
  • Median percentages of fibrous, lipidic, and calcific components were reported with IQRs.
  • Agreement with histology was superior for the AI model compared to human experts, particularly for fibrous components.

Cite This Study

Chu et al. (2026) studied this question.

synapsesocial.com/papers/69faa1eb04f884e66b532a24https://doi.org/10.1097/cp9.0000000000000158
View Full Paper
Ask AI
Bookmark
Share