Personalized White Matter Bundle Segmentation for Early Childhood.
Authors
Affiliations (8)
Affiliations (8)
- Dept. of Electrical and Computer Engineering, Vanderbilt University, Nashville, TN, USA.
- Dept. of Computer Science, Vanderbilt University, Nashville, TN, USA.
- Creighton University School of Medicine, Omaha, NE, USA.
- Dept. of Radiology, Cumming School of Medicine, Calgary, Canada.
- Dept. of Psychiatry, UMass Chan Medical School, Worcester, MA, USA.
- Dept. of Radiology and Radiological Sciences, Vanderbilt University Medical Center, Nashville, TN, USA.
- Vanderbilt Kennedy Center, Vanderbilt University, Nashville, TN, USA.
- Peabody College of Education, Vanderbilt University, Nashville, TN, USA.
Abstract
White matter segmentation methods from diffusion magnetic resonance imaging range from streamline clustering-based approaches to bundle mask delineation, but none have proposed a pediatric-specific approach. White matter tract segmentation tools optimized for adults have shown inconsistent performance in pediatric populations. We explore optimization of manually corrected masks integrated into a direct-set, deep learning framework for generalizable pediatric white matter tract segmentation. We hypothesize that a deep learning model with a similar approach to TractSeg will improve similarity between an algorithm-generated mask and an expert-labeled ground truth. Given a cohort of 56 manually labelled white matter bundles, we take inspiration from TractSeg's 2D UNet architecture, and we modify (a) inputs to match bundle definitions as determined by pediatric experts; (b) evaluation to use k-fold cross validation instead of a train/validation/test split; (c) the loss function to masked Dice loss. We evaluate Dice score, volume overlap, and volume overreach of 16 major regions of interest compared to the expert-labeled dataset. To test whether our approach offers statistically significant improvements over TractSeg, we compare Dice voxels, volume overlap, and adjacency voxels with a Wilcoxon signed-rank test followed by false discovery rate (FDR) correction. We find statistical significance (<i>p</i> < 0.05) across all bundles for all metrics with one exception in volume overlap. After we run TractSeg and our model, we combine their output masks into a 60-label atlas to evaluate if TractSeg and our model combined can generate a robust, individualized atlas, and observe smoothed, continuous masks in cases that TractSeg did not produce an anatomically plausible output. With the improvement of white matter pathway segmentation masks, we can further understand neurodevelopment on a population-level scale, and we can produce reliable estimates of individualized anatomy in pediatric white matter diseases and disorders.