Related Experiment Video
Updated: Nov 30, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.2K
Context aware deep learning for brain tumor segmentation, subtype classification, and survival prediction using
Linmin Pei1, Lasitha Vidyaratne1, Md Monibor Rahman1
1Vision Lab, Electrical and Computer Engineering, Old Dominion University, Norfolk, VA, 23529, USA.
Scientific Reports
|November 13, 2020
Summary
This study introduces deep learning for brain tumor segmentation, classification, and survival prediction using MRI scans. The method shows robust performance in segmenting tumors and predicting survival, achieving second place in a classification challenge.
Area of Science:
- Neuroimaging
- Artificial Intelligence
- Oncology
Background:
- Accurate brain tumor segmentation and classification are crucial for patient prognosis and treatment.
- Multimodal magnetic resonance imaging (mMRI) provides rich data for analyzing brain tumors.
Purpose of the Study:
- To develop and evaluate a context-aware deep learning framework for brain tumor segmentation, subtype classification, and overall survival prediction.
- To leverage structural mMRI data for improved diagnostic and prognostic capabilities.
Main Methods:
- A 3D context-aware deep learning model was proposed for brain tumor segmentation, incorporating location uncertainty.
- A 3D convolutional neural network (CNN) was applied to segmented tumors for subtype classification.
- A hybrid deep learning and machine learning approach was used for overall survival prediction.
Main Results:
- The proposed method demonstrated robust performance in brain tumor segmentation and overall survival prediction on the BraTS 2019 dataset.
- Tumor classification achieved a second-place ranking in the 2019 CPM-RadPath Challenge.
- Performance was validated using metrics like Dice score, Hausdorff distance, classification accuracy, and mean square error.
Conclusions:
- The developed context-aware deep learning framework offers a robust solution for brain tumor analysis.
- The findings highlight the potential of deep learning in improving brain tumor segmentation, classification, and survival prediction.
- This approach contributes to advancing precision medicine in neuro-oncology.
