Brain Tumor Segmentation and Tractographic Feature Extraction from Structural MR Images for Overall Survival Prediction

Po-Yu Kao, Thuyen Ngo, Angela Zhang, Jefferson Chen, B. S. Manjunath

Published 2018-07-20Version 1

This paper introduces a novel methodology to integrate human brain connectomics and parcellation for tumor segmentation and survival prediction. For segmentation, we utilize an existing brain parcellation atlas in the MNI152 1mm space and map the parcellation to each individual subject data. We leverage current state-of-the-art deep networks together with hard negative mining to achieve the final voxel level classification. This method achieves mean Dice scores of 0.904, 0.785 and 0.805 on segmentation of the whole tumor, tumor core and enhancing tumor, respectively. For survival prediction, we present a new method for combining features from connectomics data, brain parcellation, and the segmented brain tumor. Since the BraTS dataset does not include diffusion tensor imaging data for computing tractography, we use the average connectome information from the Human Connectome Project and map each subject brain volume onto this common connectome space. From this, we compute tractographic features that describe potential disruptions due to the brain tumor. These features are then used to predict the overall survival time. The proposed tractographic features achieve an average accuracy of 69% on training data and 50% on validation data.