Related Experiment Video
Updated: Jul 17, 2026
09:11
Use of MRI-ultrasound Fusion to Achieve Targeted Prostate Biopsy
Published on: April 9, 2019
22.6K
ProSeg: multi-scale context fusion for high-precision prostate segmentation in MRI.
1Guangdong Pharmaceutical University, Guangzhou, China. qinjw@gdpu.edu.cn.
Scientific Reports
|March 17, 2026
Summary
ProSeg, a new deep learning method, improves prostate MRI segmentation by precisely mapping peripheral zones and central glands. This advanced technique enhances diagnostic accuracy and treatment planning for prostate cancer.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Radiology
Background:
- Prostate MRI segmentation is vital for diagnosis and treatment planning.
- Challenges include irregular peripheral zone boundaries, homogeneous central gland textures, and imaging protocol variability.
Purpose of the Study:
- To introduce ProSeg, a novel deep learning framework for accurate prostate zonal segmentation.
- To address limitations in current segmentation methods using specialized deep learning techniques.
Main Methods:
- Developed ProSeg, a deep learning framework with a specialized ProSeg block.
- Integrated anisotropic convolutions for peripheral zone boundary delineation.
- Utilized cross-slice attention mechanisms for central gland texture modeling.
Main Results:
- ProSeg achieved state-of-the-art performance on Promise12 and Promise158 datasets.
- Achieved Dice scores of 84.31% (peripheral zone) and 57.92% (central gland) on Promise12.
- Achieved Dice scores of 83.15% (peripheral zone) and 56.38% (central gland) on Promise158, outperforming existing methods.
Conclusions:
- ProSeg demonstrates superior accuracy in prostate zonal segmentation compared to existing methods.
- The framework shows consistent performance across diverse imaging protocols.
- ProSeg holds significant clinical potential for reliable prostate segmentation in real-world applications.
