Related Experiment Video
Updated: Jul 8, 2025
08:40
Quantitation of Protein Expression and Co-localization Using Multiplexed Immuno-histochemical Staining and Multispectral Imaging
Published on: April 8, 2016
12.9K
Atlas-based Semantic Segmentation of Prostate Zones
Jiazhen Zhang1, Rajesh Venkataraman2, Lawrence H Staib1,3,4
1Department of Radiology & Biomedical Imaging, Yale University, New Haven, CT, USA.
Medical Image Computing and Computer-Assisted Intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
|December 12, 2023
Summary
This study introduces a deep learning method for segmenting prostate zones in MRI scans. Integrating an anatomical atlas significantly improves the accuracy of identifying the central gland and peripheral zone.
Area of Science:
- Radiology
- Medical Imaging
- Artificial Intelligence
Background:
- Accurate segmentation of prostate anatomical zones in MRI is crucial for prostate cancer assessment.
- Distinguishing the central gland (CG) and peripheral zone (PZ) is of significant clinical interest.
Purpose of the Study:
- To develop and validate a deep learning framework for segmenting the CG and PZ in T2-weighted MRI.
- To integrate an anatomical prostate zone atlas into a semantic segmentation model.
Main Methods:
- Proposed a deep learning semantic segmentation framework incorporating a probabilistic prostate zone atlas.
- Utilized a dynamically controlled hyperparameter to combine atlas information with segmentation results.
- Enabled dynamic adjustment of the hyperparameter during inference for user-guided segmentation refinement.
Main Results:
- Achieved Dice similarity coefficients of 0.91±0.05 for the CG and 0.77±0.16 for the PZ on an external test dataset.
- Demonstrated significant improvement in segmentation performance compared to a baseline method without the atlas.
- The dynamic hyperparameter allowed for user-driven refinement of segmentation outcomes.
Conclusions:
- The proposed atlas-integrated deep learning approach enhances prostate zone segmentation accuracy in MRI.
- The dynamic hyperparameter offers a valuable tool for refining segmentation and clinical application.
- This method shows promise for improved radiological assessment of prostate cancer.
