Related Experiment Video
Updated: Dec 27, 2025
06:53
3D Whole-heart Myocardial Tissue Analysis
Published on: April 12, 2017
9.2K
An Automated Method for Detecting the Scar Tissue in the Left Ventricular Endocardial Wall Using Deep Learning
Yashbir Singh1, Deepa Shakyawar1, Weichih Hu1
1Department of Biomedical Engineering, Chung Yuan Christian University, Zhongli, Taiwan.
Current Medical Imaging
|March 6, 2020
Summary
This study developed an automated deep learning method using Convolutional Neural Networks (CNNs) to detect left ventricular (LV) endocardial scar tissue in cardiac CT images, achieving high accuracy.
Area of Science:
- Cardiology
- Medical Imaging
- Artificial Intelligence
Background:
- Scar tissue evaluation is crucial for diagnosing cardiovascular diseases.
- Accurate segmentation of scar tissue is the initial step in morphological assessment.
- Deep learning on CT images aids in identifying left ventricular endocardial scar tissue.
Purpose of the Study:
- To create an automated deep learning method for detecting endocardial scar tissue in the left ventricle.
- To leverage Convolutional Neural Networks (CNNs) for scar tissue identification.
Main Methods:
- Extracted pixel values from the endocardial wall images.
- Applied morphological operations to define regions prone to scar tissue.
- Utilized CNNs to differentiate scar tissue from healthy LV tissue based on pixel values in cardiac CT images.
Main Results:
- Achieved 89.23% accuracy in detecting endocardial scars.
- Demonstrated 91.11% sensitivity and 87.75% specificity using the CNN-based approach.
- The CNN method proved effective in distinguishing scar tissue from healthy myocardium.
Conclusions:
- The CNN-based deep learning method offers robust accuracy for LV endocardial scar detection.
- This pixel-based deep learning approach provides a new direction for scar tissue assessment in cardiac imaging.
- The methodology shows significant translational potential for clinical applications in cardiovascular disease diagnosis.
