Related Experiment Video
Updated: May 28, 2025
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
Deformation registration based on reconstruction of brain MRI images with pathologies
1School of Information Science and Engineering, East China University of Science and Technology, Shanghai, China.
Abstract:
Deformable registration between brain tumor images and brain atlas has been an important tool to facilitate pathological analysis. However, registration of images with tumors is challenging due to absent correspondences induced by the tumor. Furthermore, the tumor growth may displace the tissue, causing larger deformations than what is observed in healthy brains. Therefore, we propose a new reconstruction-driven cascade feature warping (RCFW) network for brain tumor images. We first introduce the symmetric-constrained feature reasoning (SFR) module which reconstructs the missed normal appearance within tumor regions, allowing a dense spatial correspondence between the reconstructed quasi-normal appearance and the atlas. The dilated multi-receptive feature fusion module is further introduced, which collects long-range features from different dimensions to facilitate tumor region reconstruction, especially for large tumor cases. Then, the reconstructed tumor images and atlas are jointly fed into the multi-stage feature warping module (MFW) to progressively predict spatial transformations. The method was performed on the Multimodal Brain Tumor Segmentation (BraTS) 2021 challenge database and compared with six existing methods. Experimental results showed that the proposed method effectively handles the problem of brain tumor image registration, which can maintain the smooth deformation of the tumor region while maximizing the image similarity of normal regions.
Insights
This study introduces a novel network for brain tumor image registration, reconstructing tumor regions to improve accuracy. The method enhances pathological analysis by enabling precise spatial transformations in deformed brain images.
Area of Science:
- Medical Imaging
- Neuroscience
- Artificial Intelligence
Background:
- Deformable registration of brain tumor images to atlases is crucial for pathological analysis but challenging due to tumor-induced absent correspondences and tissue displacement.
- Existing methods struggle with the significant deformations caused by tumor growth, limiting their effectiveness in clinical applications.
Purpose of the Study:
- To develop a novel reconstruction-driven cascade feature warping (RCFW) network for accurate deformable registration of brain tumor images.
- To address the challenges of absent correspondences and large deformations in brain tumor imaging.
Main Methods:
- Proposed a reconstruction-driven cascade feature warping (RCFW) network incorporating a symmetric-constrained feature reasoning (SFR) module to reconstruct normal appearance in tumor regions.
- Introduced a dilated multi-receptive feature fusion module to capture long-range features for improved tumor region reconstruction, especially in large tumor cases.
- Utilized a multi-stage feature warping module (MFW) to progressively predict spatial transformations using reconstructed tumor images and an atlas.
Main Results:
- The RCFW network was evaluated on the BraTS 2021 challenge database, outperforming six existing methods.
- The proposed method demonstrated effective handling of brain tumor image registration challenges.
- It successfully maintained smooth deformation in tumor regions while maximizing image similarity in normal brain areas.
Conclusions:
- The RCFW network offers a significant advancement in brain tumor image registration.
- This approach improves pathological analysis by providing accurate spatial transformations for tumor-affected brains.
- The method shows promise for enhanced diagnostic and treatment planning in neuro-oncology.
