Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Contrast-Free Myocardial Scar Segmentation in Cine MRI using Motion and Texture Fusion

Guang Yang, Jingkun Chen, Xicheng Sheng, Shan Yang, Xiahai Zhuang, Betty Raman, Lei Li, Vicente Grau

TL;DR

This work tackles the challenge of detecting myocardial scars without contrast agents by introducing MTI-MyoScarSeg, a two-stage framework that first extracts cardiac motion from cine MRI relative to the end-diastolic frame and then fuses motion with texture cues to segment LV myocardium and scars. The motion extraction uses a U-Net with a spatial transformer to obtain framewise displacements and warps to a fixed reference, optimized by motion and smoothness losses. The segmentation stage combines the full cine sequence with motion fields to produce Myo and scar masks, trained with Dice and BCE losses. Results on a paired cine–LGE dataset show that the method achieves competitive performance to LGE-based models, with ablation and motion studies confirming the value of joint motion–texture fusion, offering a promising contrast-free alternative with potential clinical impact despite higher computational cost.

Abstract

Late gadolinium enhancement MRI (LGE MRI) is the gold standard for the detection of myocardial scars for post myocardial infarction (MI). LGE MRI requires the injection of a contrast agent, which carries potential side effects and increases scanning time and patient discomfort. To address these issues, we propose a novel framework that combines cardiac motion observed in cine MRI with image texture information to segment the myocardium and scar tissue in the left ventricle. Cardiac motion tracking can be formulated as a full cardiac image cycle registration problem, which can be solved via deep neural networks. Experimental results prove that the proposed method can achieve scar segmentation based on non-contrasted cine images with comparable accuracy to LGE MRI. This demonstrates its potential as an alternative to contrast-enhanced techniques for scar detection.

Contrast-Free Myocardial Scar Segmentation in Cine MRI using Motion and Texture Fusion

TL;DR

This work tackles the challenge of detecting myocardial scars without contrast agents by introducing MTI-MyoScarSeg, a two-stage framework that first extracts cardiac motion from cine MRI relative to the end-diastolic frame and then fuses motion with texture cues to segment LV myocardium and scars. The motion extraction uses a U-Net with a spatial transformer to obtain framewise displacements and warps to a fixed reference, optimized by motion and smoothness losses. The segmentation stage combines the full cine sequence with motion fields to produce Myo and scar masks, trained with Dice and BCE losses. Results on a paired cine–LGE dataset show that the method achieves competitive performance to LGE-based models, with ablation and motion studies confirming the value of joint motion–texture fusion, offering a promising contrast-free alternative with potential clinical impact despite higher computational cost.

Abstract

Late gadolinium enhancement MRI (LGE MRI) is the gold standard for the detection of myocardial scars for post myocardial infarction (MI). LGE MRI requires the injection of a contrast agent, which carries potential side effects and increases scanning time and patient discomfort. To address these issues, we propose a novel framework that combines cardiac motion observed in cine MRI with image texture information to segment the myocardium and scar tissue in the left ventricle. Cardiac motion tracking can be formulated as a full cardiac image cycle registration problem, which can be solved via deep neural networks. Experimental results prove that the proposed method can achieve scar segmentation based on non-contrasted cine images with comparable accuracy to LGE MRI. This demonstrates its potential as an alternative to contrast-enhanced techniques for scar detection.
Paper Structure (11 sections, 6 equations, 2 figures, 2 tables)

This paper contains 11 sections, 6 equations, 2 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Cardiac Motion Extraction
  4. Motion and Texture Fusion based Scar Segmentation
  5. Experiments and Results
  6. Dataset and Evaluation
  7. Implementation Details
  8. Ablation Study and Comparison Methods
  9. Motion Effectiveness Study
  10. Discussion and Conclusion
  11. Acknowledgments

Figures (2)

  • Figure 1: The architecture of the proposed cardiac motion and texture informed model for contrast-free myocardial scar localization from cine MRI. Note that the input of Seg-Net is the whole cardiac cycle (including $I_0$) and estimated cardiac motion.
  • Figure 2: Visualisation of the LV scar segmentation results of different methods from LGE and cine MRIs, respectively.