Automatic Segmentation and Overall Survival Prediction in Gliomas using Fully Convolutional Neural Network and Texture Analysis

TL;DR

This study presents a fully automated method combining a deep convolutional neural network for glioma segmentation with texture and shape features for survival prediction, achieving high segmentation accuracy and moderate survival prediction accuracy on MRI data.

Contribution

The paper introduces a novel integrated approach using a 23-layer FCNN for glioma segmentation and texture-based features with XGBoost for survival prediction, demonstrating improved performance on BraTS 2017 data.

Findings

Dice scores of 0.83, 0.69, and 0.69 for tumor segmentation.

Survival prediction accuracy of 52%.

Effective combination of deep learning and texture analysis.

Abstract

In this paper, we use a fully convolutional neural network (FCNN) for the segmentation of gliomas from Magnetic Resonance Images (MRI). A fully automatic, voxel based classification was achieved by training a 23 layer deep FCNN on 2-D slices extracted from patient volumes. The network was trained on slices extracted from 130 patients and validated on 50 patients. For the task of survival prediction, texture and shape based features were extracted from T1 post contrast volume to train an XGBoost regressor. On BraTS 2017 validation set, the proposed scheme achieved a mean whole tumor, tumor core and active dice score of 0.83, 0.69 and 0.69 respectively and an accuracy of 52% for the overall survival prediction.