PulseExploreJournal ClubResearchersJournals
Instagram
HomeJournal ClubExplore
Synapse
⌘+K
Synapse
July 22, 2026Magnetic Resonance ImagingOpen Access

Comparison of semi-automatic myocardial scar segmentation methods on black-blood late gadolinium enhancement images

View Full Paper
Ask AI
Bookmark
Share

Why the study?

Manual myocardial scar segmentation is time-consuming and operator-dependent, and consensus is lacking on semi-automatic segmentation methods for black-blood LGE imaging.

Can conventional semi-automatic scar segmentation algorithms be accurately applied to black-blood late gadolinium enhancement images in patients with ischemic heart disease?

Comparison

Semi-automatic scar segmentation methods (n-SD, FWHM, and Otsu) on black-blood SPOT vs manual segmentation

Design

Observational imaging study

Key result

Conventional semi-automatic scar segmentation algorithms can be applied to black-blood imaging, with the 6-SD method achieving the highest Dice Similarity Coefficient (81.3%).

Authors

TGThaïs GénissonKNKalvin NarceauTRThéo Richard

Discussion

Loading...

Member takes

Overview

Supports feasibility of 6-SD segmentation on black-blood LGE; leaves open prospective validation before clinical use.

Study Design

Type

Observational (n=20)

Structured PICO

Can conventional semi-automatic scar segmentation algorithms be accurately applied to black-blood late gadolinium enhancement images in patients with ischemic heart disease?

P
Population
20 patients with ischemic heart disease evaluated for myocardial scar segmentation using bright- and black-blood LGE images.
E
Exposure
Semi-automatic scar segmentation algorithms (n-SD [n from 2 to 15], FWHM and Otsu thresholding) applied to black-blood SPOT images
C
Comparator
Manual segmentation by two experts on bright-blood (PSIR) and black-blood (SPOT) images
O
Outcome
Dice Similarity Coefficient (DSC), scar extent, volume and transmuralitysurrogate

Main Result

Absolute Event Rate: 81.3% vs 87.9%

Conventional semi-automatic scar segmentation algorithms developed for bright-blood imaging can be successfully applied to black-blood LGE imaging, with the 6-SD method showing the highest accuracy.

Limitations

  • Validation in larger populations is needed to determine and refine the most accurate approach.
  • Validation in larger populations is needed

Cite This Study

Génisson et al. (2026) conducted an observational in Ischemic heart disease (n=20). Semi-automatic scar segmentation algorithms (n-SD, FWHM, Otsu) vs. Manual segmentation was evaluated on Dice Similarity Coefficient (DSC). Conventional semi-automatic scar segmentation algorithms can be applied to black-blood imaging, with the 6-SD method achieving the highest Dice Similarity Coefficient (81.3%).

synapsesocial.com/papers/6a722a9d35aa2c282ce31e34https://doi.org/10.1016/j.mri.2026.110750
View Full Paper
Ask AI
Bookmark
Share